Comprehensive Chiroptical Spectroscopy, Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules (Volume 2)

COMPREHENSIVE CHIROPTICAL SPECTROSCOPY Volume 2 COMPREHENSIVE CHIROPTICAL SPECTROSCOPY Volume 2 Applications in Stere...

Author: Nina Berova | Prasad L. Polavarapu | Koji Nakanishi | Robert W. Woody

24 downloads 385 Views 16MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

COMPREHENSIVE CHIROPTICAL SPECTROSCOPY Volume 2

COMPREHENSIVE CHIROPTICAL SPECTROSCOPY Volume 2

Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules

Edited by Nina Berova Prasad L. Polavarapu Koji Nakanishi Robert W. Woody

A JOHN WILEY & SONS, INC., PUBLICATION

Copyright © 2012 by John Wiley & Sons, Inc. All rights reserved Published by John Wiley & Sons, Inc., Hoboken, New Jersey Published simultaneously in Canada No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750–8400, fax (978) 750–4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748–6011, fax (201) 748–6008, or online at http://www.wiley.com/go/permission. Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and speciﬁcally disclaim any implied warranties of merchantability or ﬁtness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of proﬁt or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762–2974, outside the United States at (317) 572–3993 or fax (317) 572–4002. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com. Library of Congress Cataloging-in-Publication Data: Advances in chiroptical methods/edited by Nina Berova . . . [et al.]. p. cm. Includes index. ISBN 978-0-470-64135-4 (hardback : set)—ISBN 978-1-118-01293-2 (v. 1)—ISBN 978-1-118-01292-5 (v. 2) 1. Chirality. 2. Spectrum analysis. 3. Circular dichroism. I. Berova, Nina. QP517.C57A384 2012 541.7–dc23 2011021418 Printed in the United States of America 10 9 8 7 6 5 4 3 2 1

IN MEMORY OF CARLO ROSINI (1948–2010)

Carlo Rosini obtained his degree in Chemistry (1973) at the University of Pisa, where he completed his thesis on the stereochemistry of Ni(II) complexes. He entered the Italian CNR by joining the group of Professor Piero Salvadori and the research on determination of absolute conﬁguration by Circular Dichroism. Later on, Carlo Rosini spent two years (1977–1979) at the King’s College in London, under the supervision of Professor Stephen F. Mason. During this period he studied polarized-light-based spectroscopy and its application to structural determinations. He was appointed as associate professor (1992) at the University of Pisa and then as a full professor (1997) at the University of Basilicata, Potenza. The ﬁeld of chirality was fundamental to the scientiﬁc activity of Carlo Rosini. His broad scientiﬁc interests included many aspects of organic stereochemistry, like asymmetric organic synthesis, chiral discrimination mechanisms, chiral stationary phases for enantioselective chromatography, and structural characterization of organic molecules by Circular Dichroism. The last research projects of Carlo Rosini were oriented toward chemical/computational approaches for the determination of absolute conﬁguration by linking experimental and theoretical studies. We miss his enthusiasm and his charisma, but we will remember his life and his contributions to the science and the chemical community. Carlo Rosini was one of the ﬁrst scientists who accepted to contribute a chapter to this volume. Although his premature and tragic death prevented his submission, his spirit never died and is now, not only in the chapter contributed by his co-workers and former students, but also in the minds of all of us who had the privilege to know him and collaborate with him.

CONTENTS

PREFACE CONTRIBUTORS

PART I A HISTORICAL OVERVIEW

1

THE FIRST DECADES AFTER THE DISCOVERY OF CD AND ORD BY AIME´ COTTON IN 1895 Peter Laur

PART II ORGANIC STEREOCHEMISTRY

2

SOME INHERENTLY CHIRAL CHROMOPHORES—EMPIRICAL RULES AND QUANTUM CHEMICAL CALCULATIONS

xi xiii

1 3

37 39

Marcin Kwit, Pawel Skowronek, Jacek Gawronski, Jadwiga Frelek, Magdalena Woznica, and Aleksandra Butkiewicz

3

ELECTRONIC CD OF BENZENE AND OTHER AROMATIC CHROMOPHORES FOR DETERMINATION OF ABSOLUTE CONFIGURATION

73

Tibor Kurt´an, S´andor Antus, and Gennaro Pescitelli

4

ELECTRONIC CD EXCITON CHIRALITY METHOD: PRINCIPLES AND APPLICATIONS

115

Nobuyuki Harada, Koji Nakanishi, and Nina Berova

5

CD SPECTRA OF CHIRAL EXTENDED π -ELECTRON COMPOUNDS: THEORETICAL DETERMINATION OF THE ABSOLUTE STEREOCHEMISTRY AND EXPERIMENTAL VERIFICATION

167

Nobuyuki Harada and Shunsuke Kuwahara

vii

viii

CONTENTS

6

7 8

ASSIGNMENT OF THE ABSOLUTE CONFIGURATIONS OF NATURAL PRODUCTS BY MEANS OF SOLID-STATE ELECTRONIC CIRCULAR DICHROISM AND QUANTUM MECHANICAL CALCULATIONS Gennaro Pescitelli, Tibor Kurt´an, and Karsten Krohn DYNAMIC STEREOCHEMISTRY AND CHIROPTICAL SPECTROSCOPY OF METALLO-ORGANIC COMPOUNDS James W. Canary and Zhaohua Dai CIRCULAR DICHROISM OF DYNAMIC SYSTEMS: SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

217

251

289

Angela Mammana, Gregory T. Carroll, and Ben L. Feringa

9 10

11

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS Cheng Yang and Yoshihisa Inoue

317

THE ONLINE STEREOCHEMICAL ANALYSIS OF CHIRAL COMPOUNDS BY HPLC-ECD COUPLING IN COMBINATION WITH QUANTUM-CHEMICAL CALCULATIONS Gerhard Bringmann, Daniel G¨otz, and Torsten Bruhn

355

DETERMINATION OF THE STRUCTURES OF CHIRAL NATURAL PRODUCTS USING VIBRATIONAL CIRCULAR DICHROISM

387

Prasad L. Polavarapu

12

DETERMINATION OF MOLECULAR ABSOLUTE CONFIGURATION: GUIDELINES FOR SELECTING A SUITABLE CHIROPTICAL APPROACH

421

Stefano Superchi, Carlo Rosini, Giuseppe Mazzeo, and Egidio Giorgio

PART III INORGANIC STEREOCHEMISTRY

13

APPLICATIONS OF ELECTRONIC CIRCULAR DICHROISM TO INORGANIC STEREOCHEMISTRY

449 451

Sumio Kaizaki

PART IV BIOMOLECULES

14

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS Robert W. Woody

473 475

ix

CONTENTS

15

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES Claudio Toniolo, Fernando Formaggio, and Robert W. Woody

499

16

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

545

Claudio Toniolo and Fernando Formaggio

17

CIRCULAR DICHROISM SPECTROSCOPY OF NUCLEIC ACIDS Jaroslav Kypr, Iva Kejnovsk´a, Kl´ara Bedn´arˇ ov´a, and Michaela Vorl´ıcˇ kov´a

18

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

575

587

Roberto Corradini, Tullia Tedeschi, Stefano Sforza, and Rosangela Marchelli

19 20 21 22

23 24

CIRCULAR DICHROISM OF PROTEIN–NUCLEIC ACID INTERACTIONS Donald M. Gray

615

DRUG AND NATURAL PRODUCT BINDING TO NUCLEIC ACIDS ANALYZED BY ELECTRONIC CIRCULAR DICHROISM George A. Ellestad

635

PROBING HSA AND AGP DRUG-BINDING SITES BY ELECTRONIC CIRCULAR DICHROISM Mikl´os Simonyi

665

CONFORMATIONAL STUDIES OF BIOPOLYMERS, PEPTIDES, PROTEINS, AND NUCLEIC ACIDS. A ROLE FOR VIBRATIONAL CIRCULAR DICHROISM Timothy A. Keiderling and Ahmed Lakhani

707

STRUCTURE AND BEHAVIOR OF BIOMOLECULES FROM RAMAN OPTICAL ACTIVITY Laurence D. Barron and Lutz Hecht

759

OPTICAL ROTATION, ELECTRONIC CIRCULAR DICHROISM, AND VIBRATIONAL CIRCULAR DICHROISM OF CARBOHYDRATES AND GLYCOCONJUGATES

795

Tohru Taniguchi and Kenji Monde

25

ELECTRONIC CIRCULAR DICHROISM IN DRUG DISCOVERY Carlo Bertucci and Marco Pistolozzi

INDEX

819

843

PREFACE

Chirality is a phenomenon that is manifested throughout the natural world, ranging from fundamental particles through the realm of molecules and biological organisms to spiral galaxies. Thus, chirality is of interest to physicists, chemists, biologists, and astronomers. Chiroptical spectroscopy utilizes the differential response of chiral objects to circularly polarized electromagnetic radiation. Applications of chiroptical spectroscopy are widespread in chemistry, biochemistry, biology, and physics. It is indispensable for stereochemical elucidation of organic and inorganic molecules. Nearly all biomolecules and natural products are chiral, as are the majority of drugs. This has led to crucial applications of chiroptical spectroscopy ranging from the study of protein folding to characterization of small molecules, pharmaceuticals, and nucleic acids. The ﬁrst chiroptical phenomenon to be observed was optical rotation (OR) and its wavelength dependence, namely, optical rotatory dispersion (ORD), in the early nineteenth century. Circular dichroism associated with electronic transitions (ECD), currently the most widely used chiroptical method, was discovered in the mid-nineteenth century, and its relationship to ORD and absorption was elucidated at the end of the nineteenth century. Circularly polarized luminescence (CPL) from chiral crystals was observed in the 1940s. The introduction of commercial instrumentation for measuring ORD in the 1950s and ECD in the 1960s led to a rapid expansion of applications of these forms of chiroptical spectroscopy to various branches of science, and especially to organic and inorganic chemistry and to biochemistry. Until the 1970s, chiroptical spectroscopy was conﬁned to the study of electronic transitions, but vibrational transitions became accessible with the development of vibrational circular dichroism (VCD) and Raman optical activity (ROA). Other major extensions of chiroptical spectroscopy include differential ionization of chiral molecules by circularly polarized light (photoelectron CD), measurement of optical activity in the X-ray region, magnetochiral dichroism, and nonlinear forms of chiroptical spectroscopy. The theory of chiroptical spectroscopy also goes back many years, but has recently made spectacular advances. Classical theories of optical activity were formulated in the early twentieth century, and the quantum mechanical theory of optical rotation was described in 1929. Approximate formulations of the quantum mechanical models were developed in the 1930s and more extensively with the growth of experimental ORD and ECD studies, starting in the late 1950s. The quantum mechanical methods for calculations of chiroptical spectroscopic properties reached a mature stage in the 1980s and 1990s. Ab initio calculations of VCD, ECD, ORD, and ROA have proven highly successful and are now widely used for small and medium-sized molecules. Many books have been published on ORD, ECD, and VCD/ROA. The present two volumes are the ﬁrst comprehensive treatise covering the whole ﬁeld of chiroptical spectroscopy. Volume 1 covers the instrumentation, methodologies, and theoretical xi

xii

P R E FA C E

simulations for different chiroptical spectroscopic methods. In addition to an extensive treatment of ECD, VCD, and ROA, this volume includes chapters on ORD, CPL, photoelectron CD, X-ray-detected CD, magnetochiral dichroism, and nonlinear chiroptical spectroscopy. Chapters on the related techniques of linear dichroism, chiroptical imaging of crystals and electro-optic absorption, which sometimes supplement chiroptical interpretations, are also included. The coverage of theoretical methods is also extensive, including simulation of ECD, ORD, VCD, and ROA spectra of molecules ranging from small molecules to macromolecules. Volume 2 describes applications of ECD, VCD, and ROA in the stereochemical analysis of organic and inorganic compounds and to biomolecules such as natural products, proteins, and nucleic acids. The roles of chiroptical methods in the study of drug mechanisms and drug discovery are described. Thus, this work is unique in presenting an extensive coverage of the instrumentation and techniques of chiroptical spectroscopy, theoretical methods and simulation of chiroptical spectra, and applications of chiroptical spectroscopy in inorganic and organic chemistry, biochemistry, and drug discovery. In each of these areas, leading experts have provided the background needed for beginners, such as undergraduates and graduate students, and a state-of-the-art treatment for active researchers in academia and industry. We are grateful to the contributors to these two volumes who kindly accepted our invitations to contribute and who have met the challenges of presenting accessible, up-to-date treatments of their assigned topics in a timely fashion. Nina Berova Prasad L. Polavarapu Koji Nakanishi Robert W. Woody

CONTRIBUTORS

S´andor Antus, University of Debrecen, Research Group for Carbohydrates of the Hungarian Academy of Sciences, Debrecen, Hungary Laurence D. Barron, Department of Chemistry, University of Glasgow, Glasgow, United Kingdom Kl´ara Bedn´arˇ ov´a, Institute of Biophysics, Academy of Sciences of the Czech Republic, v.v.i., Brno, Czech Republic Nina Berova, Department Chemistry, Columbia University, New York, New York, USA Carlo Bertucci, Department of Pharmaceutical Sciences, University of Bologna, Bologna, Italy Gerhard Bringmann, Institute of Organic Chemistry, University of W¨urzburg, W¨urzburg, Germany Torsten Bruhn, Institute of Organic Chemistry, University of W¨urzburg, W¨urzburg, Germany Aleksandra Butkiewicz, Polish Academy of Sciences, Institute of Organic Chemistry Warsaw, Poland James W. Canary, Department of Chemistry, New York University, New York, New York, USA Gregory T. Carroll, Chemical Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California, USA Roberto Corradini, Department of Organic and Industrial Chemistry, University of Parma, Parma, Italy Zhaohua Dai, Department of Chemistry and Physical Sciences, Pace University, New York, New York, USA George A. Ellestad, Department of Chemistry, Columbia University, New York, New York, USA Ben L. Feringa, Stratingh Institute for Chemistry, University of Groningen, Groningen, The Netherlands Fernando Formaggio, Department of Chemistry, University of Padova, Padova, Italy Jadwiga Frelek, Polish Academy of Sciences, Institute of Organic Chemistry, Warsaw, Poland Jacek Gawronski, Department of Chemistry, A. Mickiewicz University, Poznan, Poland Egidio Giorgio, Department of Chemistry, University of Basilicata, Potenza, Italy xiii

xiv

CONTRIBUTORS

Daniel G¨otz, Institute of Organic Chemistry, University of W¨urzburg, W¨urzburg, Germany Donald M. Gray, Department of Molecular and Cell Biology, The University of Texas at Dallas, Richardson, Texas, USA Nobuyuki Harada, Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Sendai, Japan Lutz Hecht, Department of Chemistry, University of Glasgow, Glasgow, United Kingdom Yoshihisa Inoue, Department of Applied Chemistry, Osaka University, Suita, Japan Sumio Kaizaki, Department of Chemistry, Graduate School of Science, Osaka University, Osaka, Japan Timothy A. Keiderling, Department of Chemistry, University of Illinois at Chicago, Chicago, Illinois, USA Iva Kejnovsk´a, Institute of Biophysics, Academy of Sciences of the Czech Republic, v.v.i., Brno, Czech Republic Karsten Krohn, Department of Chemistry, University of Paderborn, Paderborn, Germany Tibor Kurt´an, Department of Organic Chemistry, University of Debrecen, Debrecen, Hungary Shunsuke Kuwahara, Department of Chemistry, Toho University, Funabashi, Japan Marcin Kwit, Department of Chemistry, A. Mickiewicz University, Poznan, Poland Jaroslav Kypr, Institute of Biophysics, Academy of Sciences of the Czech Republic, v.v.i., Brno, Czech Republic Ahmed Lakhani, Department of Chemistry, University of Illinois at Chicago, Chicago, Illinois, USA Peter Laur, Institute of Inorganic Chemistry, RWTH Aachen University, Aachen, Germany Angela Mammana, Department of Chemistry, University of Dayton, Dayton, Ohio, USA Rosangela Marchelli, Department of Organic and Industrial Chemistry, University of Parma, Parma, Italy Giuseppe Mazzeo, Department of Chemistry, University of Basilicata, Potenza, Italy Kenji Monde, Faculty of Advanced Life Science, Frontier Research Center for Postgenome Science and Technology, Hokkaido University, Sapporo, Japan Koji Nakanishi, Department of Chemistry, Columbia University, New York, New York, USA Gennaro Pescitelli, Department of Chemistry and Industrial Chemistry, University of Pisa, Pisa, Italy Marco Pistolozzi, Department of Pharmaceutical Sciences, University of Bologna, Bologna, Italy

CONTRIBUTORS

Prasad L. Polavarapu, Department of Chemistry, Vanderbilt University, Nashville, Tennessee, USA Carlo Rosini, (deceased) Department of Chemistry, University of Basilicata, Potenza, Italy Stefano Sforza, Department of Organic and Industrial Chemistry, University of Parma, Parma, Italy Mikl´os Simonyi, Chemical Research Center, Department of Molecular Pharmacology, Hungarian Academy of Sciences, Budapest, Hungary Pawel Skowronek, Department of Chemistry, A. Mickiewicz University, Poznan, Poland Stefano Superchi, Department of Chemistry, University of Basilicata, Potenza, Italy Tohru Taniguchi, Faculty of Advanced Life Science, Frontier Research Center for Postgenome Science and Technology, Hokkaido University, Sapporo, Japan Tullia Tedeschi, Department of Organic and Industrial Chemistry, University of Parma, Parma, Italy Claudio Toniolo, Department of Chemistry, University of Padova, Padova, Italy Michaela Vorl´ıcˇ kov´a, Institute of Biophysics, Academy of Sciences of the Czech Republic, v.v.i., Brno, Czech Republic Robert W. Woody, Department of Biochemistry and Molecular Biology, Colorado State University, Fort Collins, Colorado, USA Magdalena Woznica, Polish Academy of Sciences, Institute of Organic Chemistry, Warsaw, Poland Cheng Yang, Department of Applied Chemistry, Osaka University, Suita, Japan

xv

a

c

o

Projection along b-axis

Figure 5.38. Absolute stereostructure of the C60 fullerene cis-3 bisadduct (R,R,f,s A)-[CD(+)280]-32 (top) and projection along b-axis (bottom). (Redrawn from reference 54, with permission.)

(a)

(b)

(c)

Figure 8.1. Dynamic chirality at the molecular and supramolecular level detected by CD spectroscopy. (a) A chiral molecule can direct achiral molecules to self-assemble into chiral supramolecular structures. (b) A chiral molecular switch or motor undergoes conformational changes that include inversion of molecular helicity. (c) Chiral molecules can self-assemble into chiral supramolecular structures, the chirality of which is determined by the enantiomer in excess.

S

Rotor Axle Stator O

Legs

O O

O

n

n S

S

Au Surface

2 CD (mdeg)

hν

0 –2

hν

200

240

280

320

λ (nm)

Figure 8.6. Assembly of thiol-terminated light-driven rotary molecular motors on a semitransparent gold ﬁlm provides a monolayer of chiroptical material that can be analyzed with CD spectroscopy. The CD signals invert between positive and negative bands, corresponding to changes in the helicity of the molecules comprising the monolayer upon the application of photons and thermal energy. The initial spectrum (solid black) inverts (dotted black) after irradiation with UV light (λmax = 365 nm) at room temperature. After heating the surface (70◦ C, 2 h) the spectrum inverts again to restore the original (solid gray). A second dosage of photons inverts the signal (dotted gray). Heating brings the rotors back to the original orientation relative to the substrate [30].

UV O NH N2N

S

S

O

DET-4o

O NH

Vis

HN NH2

H2N

+

S

S

O HN NH2

DET-4c

DET-4o

–

+ –

dG

dC

Figure 8.18. A photoswitchable chiroptical DNA complex. At the top is shown the photoequilibrium between the open (DET-4o) and closed (DET- 4c) forms of a dithienylethene molecular switch that contains pendant ammonium groups to confer water solubility and allow the switches to bind electrostatically to the polyanionic backbone of DNA when the amine is protonated. Photochemical ring closure was accomplished through the use of a UV lamp (λmax = 365 nm) using a 340-nm cutoff ﬁlter. Photochemical ring opening was performed with visible light using ® a 520-nm cutoff ﬁlter. Molecular models (created using Hyperchem ) show that the distance between the terminal ammonium functionalities closely resembles the distance between the anionic phosphate groups of a guanosine (G)–cytosine (c) base pair [58].

Interface

Interface

Compression

M-Chiral

Compression

Achiral

P-Chiral

Figure 8.24. A supramolecular chiroptical switch composed of achiral amphiphiles. Space-ﬁlling structures of achiral amphiphile (TARC18), which forms a Langmuir–Schaefer ﬁlm at the air–water interface, and chiral supramolecular structures formed upon interface compression. (Reprinted by permission of John Wiley & Sons, Inc. [66].)

OR

complexation N

I

N

N

N

N

N

N

N

NOH

N OH

N mutarotation of glucose

N

N

42a: R = (C2H4O)8CH3 42b: R = n-C4H9

complexation

N

N

n

42

N

N

N

N

N OH

N N N

OH

left-handed helical complex

right-handed helical complex

Figure 9.24. Chiral self-aggregation of achiral polymer induced by a saccharide. (Reprinted with permission from reference 80. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.)

3 MeOH / water 5:1 6:1 7:1 CD/(mdeg)

CD/(mdeg)

2

1

0

–1 300

8

8:1 9:1 10:1 310

320

330

340

350

360

6 β-glucose +7 +2.4 mdeg 4 (337 nm) 2

time 0h 1h 3h 8h 15 h 24, 48 h 15, 24, 48 h 8h

0

5h 3h α-glucose –2 –3 +2.4 mdeg 1h 0h (337 nm) –4 300 310 320 330 340 350 360

λ (nm)

λ (nm)

(a)

(b)

Figure 9.25. (a) Induced CD spectra of a mixture of 42a (1mM in monomer unit) and D-glucose (0.3M) in 5:1–10:1 MeOH/H2 O at 25◦ C. (b) Time-dependent CD spectra of a mixture of 42a (1mM in monomer unit) and α- or β-D-glucose (0.3M) in 5:1 MeOH/H2 O at 25◦ C. (Reprinted with permission from reference 80. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.)

(a)

(b)

Figure 12.6. Exciton chirality deﬁned by the allowed B naphthalene and S = O transitions in 1

(a) E-conformer of (S)-8 and (b) Z-conformer of (S)-9.

0.05 0

Δε

−0.05 Nd

−0.15

ν −0.25 14.00

(103 cm−1) 16.00

18.00

20.00

0.3 MI

Δε

0 −0.3 −0.6 −0.9 20.00

Ln Ho -SAPR-8-C4(llll)-M[Ln(+)-(hfbc)4] with an encapsulated alkali metal ion

ν (103 cm−1) 21.00

22.00

23.00

24.00

25.00

0.04

Δε

0 −0.04 −0.08 −0.12

ν (103 cm−1)

Er

−0.16 12.00 14.00 16.00 18.00 20.00 22.00 24.00

Figure 13.6. CD spectra in the hypersensitive 4f –4f transitions of Cs[Ln((+)-hfbc)4] in CHCl3 (left) and the proposed structure in solution (right).

200

Δε

100 0 −100 −200 250

300 λ (nm)

350

Figure 13.9. Exciton CD spectra of M[La((+)-hfbc)4 ] in CHCl3 . M: Cs (red), Rb (green), K (blue), Na (black).

Internuclear (net effect) Intranuclear (Ln end) Intranuclear (Cr end)

Δε (M−1cm−1)

50

CrIII

LnIII 0

N

N

−50

N

N

N

N

L2 275

300

375 325 350 Wavelength/nm

400

425

O N

Figure 13.10. Right: Structure of the ligand L2(below) and -[LnIII CrIII (L2)3 ]6+ (above). Left: Schematic vertical lines summering the dominant coupling effects in the CD spectra of -LnIII CrIII (L2)3 ]6+ . The black line corresponds to the CD spectrum of -[GdCr(III)(L2)3 ]6+ in CH3 CN.

10 8 6 4 Ellipticity (mdeg)

2 0 –2 –4 B-DNA hZαADAR1 yabZαE3L IsZαE3L orfZαE3L spZαE3L vZαE3L

–6 –8 –10 –12 –14 –16 230

240

250

260

270

280

290

300

310

320

2800

3200

3600

Wavelength (nm) (a) 6

Ellipticity (mdeg) at 255 nm

4 2 0 hZαADAR1 yabZαE3L IsZαE3L orfZαE3L spZαE3L vZαE3L

–2 –4 –6 –8 –10 0

400

800

1200

1600

2000

2400

Time (sec) (b)

Figure 19.1. (a) CD spectra of poly[d(G–C)] in the B form and in the presence of Zα domains from human ADAR1 editing enzyme (hZα), yaba-like disease virus (yabZα), lumpy skin disease virus (lsZα), orf virus (orfZα), swinepox virus (spZα), and vaccinia virus (vZα), listed in decreasing order of their abilities to convert B-DNA to Z-DNA. CD spectra of the added proteins contributed the negative signals at wavelengths shorter than about 250 nm. (b) Kinetics of the B to Z conversion in the presence of the same domains. Proteins were added to poly[d(G–C)] at a protein to base-pair ratio of 0.4 (with the ﬁnal protein concentration being 90 μM), in a buffer of 10 mM HEPES, pH 7.4, 10 mM NaCl, and 0.1 mM EDTA, except for yabZα where the buffer included 100 mM NaCl. Spectra were taken at 25◦ C using a 2-mm-pathlength cell. CD values are in mdeg ellipticity. (Reproduced from Quyen et al. [6] by permission of Oxford University Press, copyright 2007.)

θ (mdeg)

5

0

–5

Sp1ZF6 (ER)4 + [2GC (10)] Sp1ZF6 (KE)4 + [2GC (10)] Sp1ZF6 (G4S)4 + [2GC (10)] Free [2GC (10)]

–10 200

220

240

260

280

300

320

Wavelength (nm)

Figure 19.8. CD spectra of a DNA containing two GC-box sequences separated by a 10-bp spacer, 2GC(10), complexed with each of three peptides containing six zinc ﬁngers but with different linkers between zinc ﬁngers 3 and 4: Sp1ZF6(ER)4 with linker (Glu–Ala–Ala–Ala–Arg)4 , Sp1ZF6(KE)4 with linker (Lys–Ala–Ala–Glu–Ala)4 , and Sp1ZF6(G4 S)4 with linker (Gly–Gly–Gly–Gly–Ser)4 . Spectra were taken at 20◦ C using a 1-mm-pathlength cell. Samples contained 4.5 μM peptide–DNA complex in 10 mM Tris–HCl (pH 8.0), 50 mM NaCl, 0.005% Nonidet P-40, and 1 mM dithiothreitol. CD values are mdeg ellipticity. (Reprinted with permission from Yan et al. [42], ©2005, American Chemical Society)

B-DNA

A-DNA

Z-DNA

Binding modes minor groove binding

major groove minor groove

major groove

intercalation

minor groove

Figure 20.1. Representation of the three principal secondary structures of DNA. The right¨ handed A and B form are obtained from standard parameters within the Schrodinger–Maestro graphical interface. The thinner and more elongated Z form is obtained from X-ray parameters of a hexamer as imported from the protein data bank (PDB). In this representation, three units of the hexamer are stacked in order to display the overall left-handed zig-zag helicity. The structure on the far right depicts drug–DNA double-helix interactions with the drug colored black: minor groove binding (top) and intercalation between base pairs (bottom).

Figure 21.1. Binding sites are indicated by speciﬁc ligands in white, warfarin (Site I, right-hand side) and diazepam (Site II, left-hand side). (Reprinted with permission from reference 3, copyright 1996, Elsevier.)

Cleft Thyroxine 5 2°: lodipamide

IIIB FA 5 Thyroxine 2,3 2° : Oxyphenbutazone 2° : Propofol

IIIA: Drug Site 2 FA 3, 4 Thyroxine 4 Diflunisal Diazeapam Halothane Ibuprofen Indoxyl sulphate Propofol 2° : CMPF

IB FA 1 Hemin 2° : Azapropazone 2° : Indomethacin 2° : TIB

FA 2 IIA: Drug Site 1 FA 7 Thyroxine 1 Azapropazone CMPF DIS Indomethacin Iodipamide Oxyphenbutazone IIA-IIB Phenylbutazone FA 6 2° : Diflunisal TIB 2° : Halothane Warfarin 2° : Ibuprofen 2° Indoxyl sulphate 3° Diflunisal

Figure 21.2. Ligand-binding capacity of HSA deﬁned by crystallographic studies. (Reprinted with permission from reference 9, copyright 2005, Elsevier.)

Figure 21.16. Mutual positions of quercetin (molecular modeling 48) and warfarin (X ray [8]) in the cavity of Site I subdomain IIA. (Reprinted with permission from reference 48, copyright 2003, Elsevier.)

IB

IA

IIIB

site II IIA site I

IIIA

IIB

Figure 21.21. X-ray crystallographic structure of HSA [8] with curcumin molecules localized by docking. Subdomains are indicated. (Reprinted with permission from reference 55, copyright 2003, Elsevier.)

1 5

7 2 4 3 6 C16:0

Figure 21.37. Structure of HSA complexed with seven palmitic acid molecules. (Reprinted with permission from reference 105, copyright 2000, Elsevier.)

(a)

(b)

Figure 21.38. Conformational changes in warfarin binding (Site I) as a result of fatty acid binding. (a) Helices h2 and h3 are shown by light shades for defatted and by dark shades for myristate bound HSA. (Reprinted with permission from reference 8, copyright 2001, American Society for Biochemistry and Molecular Biology.) (b) The volume of Site I in defatted HSA is depicted by a light brown semitransparent surface that becomes expanded upon myristate binding (blue semitransparent surface), the red arrows point to structural changes associated with fatty acid binding. (Reprinted with permission from reference 9, copyright 2005, Elsevier.)

Figure 21.40. Two crocetin molecules ﬁtted to FA3 and FA4 sites; the negative exciton dictates the horizontal crocetin molecule to be behind the slanting one. (Reprinted with permission from reference 108, copyright 2001, Elsevier.)

(a)

(b)

Figure 21.41. Electrostatic potentials of genetic variants. (a) lysophospholipid ligand binding at the surface of variant F1-S. (b) variant a. (Reprinted with permission from reference 124, copyright 2006, ACS.)

Figure 21.55. Preferred conformer of diazepam docked into the crystal structure of AGP F1. Hydrogen bonds of the carbonyl oxygen to Glu64 and Gln66, as well as contacts of the ring nitrogens with Arg90 and Tyr127, are indicated. (Reprinted with permission from reference 125, copyright 2008, Elsevier.)

HR1-25

40

Millidegrees

20 10

Millidegrees

30

HR2-25

10

0% TFE 2.5% TFE 5% TFE 7.5% TFE 10% TFE 12.5% TFE 15% TFE

0

0 0% TFE 2.5% TFE 5% TFE 7.5% TFE 10% TFE 12.5% TFE 15% TFE

–10

–10 –20

–20

–30 200

210

220 230 240 Wavelength (nm)

250

200

260

210

220

230

240

1:1 Mixture 10

0% TFE 2.5% TFE 5% TFE 7.5% TFE 10% TFE 12.5% TFE 15% TFE

20 5

0 0% TFE 2.5% TFE 5% TFE 7.5% TFE 10% TFE 12.5% TFE 15% TFE

–10 –20 –30 –40 220

230

240

250

260

Millidegrees

Millidegrees

10

210

260

Difference Spectra

30

200

250

Wavelength (nm)

0

–5

–10 200

210

Wavelength (nm)

220 230 240 Wavelength (nm)

250

260

Figure 25.9. CD spectra of HR1–C25, HR2–C25, and their 1:1 mixture: [peptide] 50 μM, PBS, pH7.4, TFE as the co-solvent (0%, 2.5%, 5%, 7.5%, 10%, 12.5%, 15%). Difference CD spectra as a function of TFE concentration (0%, 2.5%, 5%, 7.5%, 10%, 12.5%, 15%) is also depicted. The difference CD spectra are calculated subtracting the two individual peptide spectra from those of the mixture. (Reproduced with permission from reference 69.)

PART I A HISTORICAL OVERVIEW

1 THE FIRST DECADES AFTER THE DISCOVERY OF CD AND ORD BY AIME´ COTTON IN 1895 Peter Laur

1.1. SCOPE: SUBJECTS AND TIME FRAME TO BE REVIEWED The story of the Cotton effect begins with its discovery in 1895. Although the news was hailed by leading physicists and chemists, studies to extend, exploit, and apply Cotton’s ﬁndings developed at a slower pace than one might have anticipated. One of the reasons for this delay was simply the necessity of the researchers to construct their own optical apparatus. Gradual technical improvements eventually allowed one, in the 1920s, to take chiroptical measurements in the ultraviolet as well as the visible, thus making accessible in principle a great many Cotton effects in colorless (mostly organic) compounds. Despite the paramount importance of such developments, neither the technical details nor the physics involved will be discussed in the following. Rather, a chemist’s view will prevail, paying attention chieﬂy to experimental results and the application of chiroptics to chemical problems. Since much of the work during the ﬁrst 20 or so years after Cotton’s discovery was done by physicists and physicochemists, it is not surprising that many investigations were interconnected with or even motivated by the concomitant progress of the theory of optical activity. But also the discussion of this part of (theoretical) physics will be curtailed in the following. The exclusion in this chapter appears justiﬁed, because various comprehensive reviews are readily available, as they are for the ﬁeld of optical instrumentation. By about 1935, Cotton effect measurements were possible with most organic and inorganic chromophores. It is rather surprising that not much use was made of the chiroptical techniques, especially by organic chemists. On the other hand, physical chemists had demonstrated the feasibility of Cotton effect studies in various classes of chemical compounds, but seemed satisﬁed with this result. Likewise, the advancement of optical Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

3

4

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

instruments for chiroptical measurements slowed down. All this led to a certain climax of chiroptical studies in the early 1930s, to be followed by a near standstill. It is not unreasonable to symbolically connect this phenomenon with the death in 1936 of T. M. Lowry, one of the most active scholars in the ﬁeld. Arguably, the death of T. M. Lowry ended the ﬁrst, pioneering period of chiroptical studies. The present chapter will concentrate on reviewing these ﬁrst “historical” decades. Some work on the experimental study of the Cotton effect continued after 1936 until World War II on a minor scale, on, for example, organic compounds (S. Mitchell) or platinum complexes (I. Lifschitz). But at exactly the same time, new developments took place in the theory of optical activity and its application to chemical problems: Werner Kuhn’s calculation of the absolute conﬁguration of lactic acid in 1935 rang in a new era. The waning interest of the experimentalists contrasts with the increased activity of theoretical chemists like J. G. Kirkwood, E. U. Condon, H. Eyring, or W. Kauzmann, who in the late 1930s advanced different models of optical activity. Still, chemistry had to wait for the period of 1950–1960 for a revitalization of chiroptics. Some reasons for the animation are: (1) the development of X-ray scattering methods for the determination of the absolute conﬁguration, thus anchoring the stereochemistry unambiguously, following J. M. Bijvoet’s seminal publication of 1951; (2) the advent of new, commercially available measuring devices of ORD and CD; and (3) growing interest in natural products chemistry and, generally, optically active systems. But to discuss these topics would need another chapter.

1.2. EARLY CHIROPTICAL STUDIES The discovery of optical activity is credited to the two distinguished French mathematicians, physicists, astronomers, and geodesists (and more) Dominique-Franc¸ois Jean Arago (1786–1853, of Catalan origin) and Jean-Baptiste Biot (1774–1862) [1]. Arago and Biot had been closely associated at least since 1806 in the pursuit of other scientiﬁc subjects, and they sometimes published together. Both investigated the optical activity of quartz, and apparently they also shared their equipment to some extent. If, on the one hand, Arago was the ﬁrst to go into print, Biot, on the other hand, soon became more active in this ﬁeld and extended the studies. He undoubtedly observed optical activity for the ﬁrst time in organic compounds such as natural oils and terpenes, or solutions of camphor [2] and cane sugar [3]. Biot continued his research on optical activity throughout his life, later concentrating particularly on tartaric acid. He noticed the wavelength dependence of the optical rotation even at the very beginning of his studies, albeit in a rather qualitative way. Whereas eventually the rotatory dispersion of quartz could be elucidated satisfactorily (which led to Biot’s law, stating that the rotation is inversely proportional to the square of the wavelength), similar solution studies were seriously impeded by experimental deﬁciencies, particularly the lack of suitable monochromatic light sources. Genuine chiroptical studies were, therefore, rather infrequent until the end of the nineteenth century. One of the most important papers here is a report by the Norwegian physicist Adam Arndtsen, who discussed his studies of aqueous solutions of (+)-tartaric acid [4]. Using sunlight, he was able to visually determine the angle of rotation at some of the principal Fraunhofer lines, that is, C (656), D (589), E (527), b (517), F (486), and e (438 nm). He could conﬁrm and extend Biot’s earlier ﬁnding that the rotation exhibits a maximum in the spectral region studied, with its wavelength shifting from the

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

blue to the red on increasing concentration. This unexpected and intriguing result led the Swiss chemist Hans Landolt (an important pioneer of the investigation and application of optical activity, as well as one of the “fathers” of Physical Chemistry) in 1877 to introduce the expression “anomale Rotationsdispersion” (anomalous rotation dispersion) [5], which since has become established for the description of such rotatory dispersion curves that run through a maximum or minimum, or show a reversal of sign. It had thus become apparent that spectropolarimetry promised to develop into an interesting ﬁeld in the future. In his last and comprehensive paper on optical activity, Biot [6] suggested, therefore (translation from the French by the present author): I should like to draw the attention of experimentalists to a class of phenomena which, hitherto, has been little studied but which, nevertheless, for both theoretical and practical purposes, ranks in importance with that of the optical rotatory power itself of which it is a constituent element. I refer to the speciﬁc mode of dispersion that each optically active substance or compound imparts to plane polarized light of different wavelengths [literally: refrangibility].

Despite this exhortation, reports on rotatory dispersion remained scarce until the end of the century. This is also evident from the very ﬁrst book on optical activity, where all data known at that time are summarized, which was published in Germany in 1879 by Landolt [7]. Here, he also describes in detail the optical equipment used by himself and his predecessors. Therefore, it is not necessary to dwell at this point on the measuring devices and optical methods. Although most of the rotations listed (many of which had been determined or redetermined by Landolt himself) refer to the sodium D line only, his book also has short sections on normal and anomalous rotatory dispersion. It is important to realize that so far all reported optically active liquids or solutions were based on organic compounds without absorption bands in the visible. In fact, Landolt emphasized that there is not a single inorganic substance known which shows optical activity in solution (or in the gas phase), from which he tentatively—but incorrectly—concluded that optical activity might be restricted to carbon compounds, except for the solid phase. Surprisingly, he gave no reference to any optically active transition metal complex, although at least Fehling’s solution (a mixture of several Cu(II) tartrate complexes) had been around since 1848 [8]. One might speculate whether such coordination compounds (of a still unknown nature) were ignored as a result of theoretical considerations. It should also be mentioned that measurements in general were limited to practically colorless samples and to merely certain frequencies of the visual solar spectrum. The only other reasonably monochromatic light sources available were based on lithium, sodium, and thallium salts heated in a Bunsen burner (invented in 1866), giving access to the wavelengths 671 nm (Li), 589 nm (Na), and 535 nm (Tl), respectively. It is worthwhile to brieﬂy turn to the “anomalous” refractive dispersion using unpolarized light—that is, the characteristic sigmoidal variation of the index of refraction in the absorption region, running through a maximum and minimum, instead of steadily increasing as the wavelength decreases, as in normal dispersion. This behavior had been discovered in iodine vapor in 1862 by the French physicist F.-P. Leroux [9], and around 1870 it attracted the attention of several investigators, who published independently on anomalous dispersion in the visible, using solutions of organic dyes like fuchsine [10]. There was some dispute as to priority among the Danish physicist Christian Christiansen, the Swiss chemist and physicist Jacques-Louis Soret, and the German physicist August Kundt. While it is clear that Christiansen was the ﬁrst to publish, the most extensive studies were carried out by Kundt. The relevance of these ﬁndings to the present subject

5

6

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

lies in the fact that here was proven the possibility of successfully studying the index of refraction even near absorption bands in the visible. Consequently, a similar anomalous dispersion could be expected to exist for the optical rotation, keeping in mind the relation between the velocity of light, the index of refraction, and the optical rotation. This anomalous dispersion feature of the rotation should have been accessible by existing techniques, if only suitable colored optically active samples had been available. It took more than two decades, however, before this problem was addressed.

1.3. THE DISCOVERY OF THE COTTON EFFECT In 1895, two short papers (“notes”) appeared in the fortnightly journal of the French Academy of Sciences, entitled “Unequal absorption of right and left circularly polarized light by certain optically active substances” [11] and “Anomalous rotatory dispersion of absorbing substances” [12]. The author was the 26-year-old physicist Aim´e Auguste ´ Cotton (Bourg-en-Bresse 1869–S`evres 1951), a PhD student at the prestigious Ecole Normale Sup´erieure in Paris. The ﬁrst of these papers describes and names the property of “dichro¨ısme circulaire” (what we now call “CD”) associated with an absorption band of an optically active compound in solution, and the second one introduces the corresponding effect in the dispersion mode (now called “ORD”). The full paper of 85 pages, also incorporating studies on magnetic optical activity, was published in 1896 under the heading “Investigations of the absorption and the dispersion of light by optically active media” [13]. It summarizes A. Cotton’s Th`ese de Doctorat, which he prepared from ´ November 1893 to July 1896 at the Physics Laboratory of the Ecole Normale with Professors Marcel Brillouin and Jules Violle as advisors. Based on his important discoveries, Cotton was accorded the degree of Docteur e` s Sciences in 1896. In his thesis, Cotton for the ﬁrst time reports data of (a) optical rotations close to both sides of an absorption band in the visible, using solutions of Cu(II) and Cr(III) coordination compounds with tartrate or malate ligands, and (b) the associated circular dichroism. It is quite obvious that Cotton was successful to a large degree owing to both the quality of his optical components and the skillful and precise construction of the measuring devices, especially for the determination of very small values of the ellipticity, but also to his power of observation, and—last but not least—to a fortunate choice of optically active samples. In this chapter, however, his technical equipment and the underlying physical principles shall not be discussed in detail, because Cotton himself gives a full description in his major paper, and there are also comprehensive reviews elsewhere as, for example, in the books by Mitchell and Lowry (see below). While Cotton’s expression “dispersion rotatoire anomale” (anomalous rotatory dispersion) is self-explanatory, a comment concerning his novel term “dichro¨ısme circulaire” (circular dichroism) may be appropriate. Cotton did not always measure directly or indirectly the difference in absorption of left- and right-circularly polarized light by his sample [i.e., (εL –εR )], but rather the ellipticity of the emerging elliptically polarized light. In this case, his measuring device included, apart from a Nicol prism to provide plane-polarized light, a “double circular polarizer” consisting of two quarter-waveplates placed side by side in the plane-polarized light beam in such a way that their principal axes are at 90◦ to one another and at 45◦ to the plane of the incident light. This arrangement allowed the observation of left- and right-circularly polarized light beams next to each other. On the introduction of a sample showing circular dichroism, the beams would be differently absorbed, which could be detected visually or photometrically. The ﬁeld of vision was thus divided into two halves by these λ/4 plates. When

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

he used white instead of monochromatic light for the examination of his optically active sample solutions, these two halves showed different colors. This reminded him of the dichroism observed with certain doubly refracting crystals, where the ordinary and the extraordinary ray are absorbed unequally, as found by Biot in tourmaline and later by the Austrian mineralogist Haidinger in many other cases [14]. Cotton therefore chose the term “circular dichroism.” In fact, Haidinger had already discovered this phenomenon in amethyst quartz in 1847 [15]. Nowadays, the expression “circular dichroism” probably just awakens vague memories of the original visual observations; this context has been largely forgotten nowadays, with the advent of automated electronic spectropolarimeters. Actually, Cotton himself had already performed some photometric measurements, but had found them inferior to his visual results.

1.4. THE FIRST CD AND ORD CURVES Cotton’s measurements obviously were not only restricted to the visible, but also quite limited as to the wavelengths available. Even under favorable conditions, at most eight spectral lines were at his disposal, namely, 657 (red, near C), 589 (yellow, sodium D line), 581 (orange, near D), 562 (greenish yellow), 522 (green, between E and b), 475 (blue, near F), 459 (blue-violet), and 437 nm (violet, near e) [the letters C, D, E, b, F, and e refer to the Fraunhofer lines so designated]. A comparison with Arndtsen’s paper of 1858 [4] shows that hardly any improvement of the spectral availability had taken place until the end of the nineteenth century. However, on the positive side it can be seen that these lines are spread rather evenly across the whole visual region. Nevertheless, the generation of continuous absorption and rotation curves, as often published, on the basis of observations at some of these individual wavelengths, leaves much to the whim of the draftsman, especially concerning the position and magnitude of any maxima and minima. Such “data” should not be overinterpreted. This situation would prevail in the decades to come. It is not unexpected that, at the onset of his investigations, Cotton chose Fehling’s solution (“liqueur de Fehling”) for his studies. It is, after all, in the direct line of Biot’s research to look at derivatives of active tartaric acid. Secondly, the only area of importance where the application of optical activity had become established was saccharimetry; and thirdly, Fehling’s solution was a proven and powerful reagent in carbohydrate chemistry [16]. It seems that Cotton systematically progressed from the complex and notoriously unstable Fehling’s solution to simpler alkali copper(II) tartrates, the preparation of which he describes in detail. By the way, it is amusing to note that in one case he reports the precipitation of copper tartrate from a copper sulfate solution by adding the aqueous solution of a crystal of Seignette salt (potassium sodium tartrate); this crystal had been prepared by Pasteur himself. Unfortunately, these copper complexes proved to be very unstable; they changed or simply decomposed with time or at elevated temperature and were also light-sensitive. Furthermore, the chemical composition of these aqueous solutions was unknown (and, to some extent, still is), and attempts at isolating any well-deﬁned compound failed. A solution of crystalline copper malate, perhaps more stable, did not show any observable circular dichroism. Despite these drawbacks, Cotton did obtain many ORD and some CD data, but obviously the reproducibility of the experiments remains questionable, and the curves shown in print [13] should be interpreted with caution.

7

8

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The most convincing chiroptical effects, however, were observed with aqueous solutions of potassium chromium(III) tartrate, prepared in situ. They shall be discussed here in some detail. Figure 1.1 is Cotton’s Figure 18 on page 408 of reference 13, and it shows a complete “Cotton effect” in the ORD and the CD near 570 nm. Because of its seminal importance, this ﬁgure has been later reprinted by others a number of times. The actually measured data are given as follows: 657 nm, rotation ρ + 1◦ 26 , ellipse [sic] φ + 32 ; similarly: 589, +2◦ 30 , (−3◦ 40 ); 581, +1◦ 46 , −4◦ 54 ; 562, −1◦ 21 , −4◦ 16 ; 522, −2◦ 50 , −1◦ 25 ; and 475, 1◦ 52 [no sign given in the paper; from the curve it is evident that ρ must be negative], +28 . Data were thus collected at six wavelengths only, because the onset of a second strong absorption band made observations at shorter wavelengths impossible. The parentheses around the ellipticity value at the sodium D line are Cotton’s and indicate that this number results from photometric measurements. Despite its beautiful appearance, there are unfortunately some ﬂaws in this ﬁgure and the data as printed. A comparison of the ﬁgure with the data listed above makes evident two discrepancies at 562 nm: In the ﬁgure, the angle φ is given as −4◦ 46 (not −4◦ 16 ), and the corresponding angle ρ is given as −0◦ 21 (not −1◦ 21 ). On reexamination, the true values were veriﬁed to be φ − 4◦ 46 and ρ − 1◦ 21 . The ﬁgure should be redrawn, therefore, using this value of ρ. Such a modiﬁcation would necessarily modify the shape of the ORD curve, while not basically changing it. Cotton gives these corrections in

+3°

+2°

ρ) ons (

Rotati

+1°

0

Elli s pse

–1°

(ϕ)

–2°

–3°

–4°

–5°

Figure 1.1. CD and ORD of potassium chromium(III) 657

589 581562

522

475 459 437

tartrate (solvent H2 O). (From A. Cotton, Ann. Chim. Physique 1896, [7] 8, 347; Figure 18, p. 408.)

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

a letter to Professor Ladislas Natanson in Cracow, Poland, quoted on pages 33/34 of reference 17. He explains the ﬁrst error as a misprint, and he states that the second one is a mistake by the “dessinateur” (draftsman). However, the really important disagreement between the ORD and CD, as we can see immediately from the curves—with hindsight—lies in the incompatibility of their signs. If we accept the rotation values as correct, as seems reasonable, the sign of the CD is in error. And so it is! Cotton himself redressed this ﬂaw two years later [18] in a paper, the ﬁrst sentence of which runs as follows (translated from the French): ‘It is easy to be mistaken as to the sense of a circular vibration.’ Admitting his mistake in the assignment of the direction of the rays circularly polarized by a Fresnel rhomb, he imputed it to his misinterpretation of some of Billet’s tenets in the latter’s “Trait´e ” [19]. Apparently, Billet had used the expression “principal section” of a mica crystal in an unorthodox way and had also treated this crystal as positive, contrary to the common practice. Therefore, all of Cotton’s CD curves, and the sign of all ellipticities published before 1898, ought to be inverted. But not everyone read or responded to this correction; others did so, but without indicating it. The confusion that might have been generated was fortunately curtailed by the fact that very few scientists, apart from Frenchmen, studied the circular dichroism in the following decades. But as late as in 1923, (Ms.) N. Wedeneewa in Moscow (for example) still used the earlier “wrong” sign of the CD, when she reported the ORD and CD of camphor quinone [20]. Similarly, T. M. Lowry just reprinted Cotton’s Figure 18 in his famous classic of 1935 [21] without any comment, whereas S. Mitchell in his treatise on the Cotton effect [22] of 1933 simply shows an inverted CD curve in ostensibly the same ﬁgure (see Figure 1.2), also without any further comment. Another point of criticism could be raised because of the all-too-vague identity of the samples investigated. Although Cotton carefully describes the preparation of his samples, as mentioned earlier, their inherent instability cannot preclude changes with time, perhaps also as the result of shifting equilibria between the several complexes present. Indeed, small changes even in the synthesis of the tartrate complexes can lead to the total inversion of the anomalous rotatory dispersion, as has been observed by Wedeneewa [20]. All this calls for caution with respect to the early ORD and CD publications. However, concerning the key compound discussed at length, potassium chromium(III) tartrate, all doubts were ﬁnally set to rest by W. Kuhn [23], who much later very carefully repeated Cotton’s work and found it fully correct (Figure 1.3).

1.5. THE REACTION OF THE LEARNED WORLD TO COTTON’S DISCOVERIES Cotton’s papers raised the immediate attention of Wilhelm Ostwald (Nobel Prize 1909), who, one year after the publication of the original notes in the Comptes Rendues [11, 12], wrote two abstracts thereof himself for his journal Zeitschrift f¨ur Physikalische Chemie [24]. This was followed by his six-page review of Cotton’s full paper [13] in the same year [25], with several CD and ORD curves reprinted, including Cotton’s original Figure 18, discussed above. It should be pointed out that the lack of correspondence of the sign of the ORD and the CD could not have been noticed by Ostwald at that time, since the necessary theoretical background had not yet been provided. With these reviews, Ostwald acquainted the chemical world with Cotton’s results, and his name carried much weight. It is certainly unusual that preliminary notes by a foreign physics student and extracts of his

9

10

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

5°

4

3

2

1

0

–1

–2

–3 6600

6200

5800

5400

5000

Wave-length Å.U. Rotations Ellipticities

4600

Figure 1.2. CD and ORD of potassium chromium(III) tartrate (solvent H2 O). (From S. Mitchell, The Cotton Effect, Bell, London, 1933; Figure 12, p. 23; reproduced with permission.)

thesis should induce an already famous physical chemist to such a presentation. Incidentally, already the “sponsoring” of Cotton’s notes by the renowned physicist Gabriel Lippmann from Luxembourg (Nobel Prize 1908)—such notes had to be presented by an academician—attests to the importance attributed to them. One might well say that chiroptics had a splendid start. The speed with which the news was reported and hailed is altogether breathtaking. For example, the physical chemist Landolt referred to Cotton’s studies already in the second edition of his book, published in 1898 [7]. Mention should also be made of the German physicist Paul Drude, who included a treatment of Cotton’s “(anomalous) rotary dispersion” in his famous Lehrbuch der Optik of 1900 [26]. So, by the beginning of the twentieth century, the international world of physics and physical chemistry was well aware of Cotton’s results. It took only a few additional years before a thorough theoretical treatment was provided by L. Natanson, Professor of Theoretical Physics at the Jagiellonian University Krak´ow (Cracow, Poland). The title of his important paper, “On the elliptic polarization of light transmitted through an absorbing naturally-active medium” [27], with a supplementary note [17], needs no further comment. Here, Natanson treated the interdependence of absorption, optical rotation, and circular dichroism. Probably in order to spread his results further, also an amalgamated and shortened French translation of both papers by the Count of Ballehache was published very shortly thereafter [28]. The relations presented here between the sign of the rotation and the circular dichroism have become known as the “R`egle de Natanson” or “Natanson’s Rule” [29]. This ﬁnally allowed the

11

xz–xT

log x

[A]–10–3

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

2,75 7,5 3,5

2,50 5,0 3,0

Z

3

1

2,25 2,5 2,5 + 0 2,0

2,00 −

2

1,75 –2,5 1,5 1 1,50 –5,0 1,0 g –0,10

1,25 –7,5 0,5 4

–0,05

g 1,00 –10,0 0

–0,05 –12,5 –0,5

7000 Å

6000

5000

4000

3000

λ

Figure 1.3. UV, CD, and ORD of potassium chromium(III) tartrate (solvent H2 O). (From W. Kuhn, ¨ A. Szabo, Z. Phys. Chem. 1931, B15, 59; Figure 1, p. 62; Oldenbourg Wissenschaftsverlag Munchen, reproduced with permission.)

prediction of the sign of the circular dichroism associated with a speciﬁc absorption band, based on just the anomalous rotation curve, which should not be too difﬁcult to obtain. Natanson’s papers included the following sentences on the ﬁrst page: “Effects of this kind have been observed and investigated by Monsieur A. Cotton” [27] and “Des ph´enom`enes de ce genre ont e´ t´e observ´e et analys´es par M. Cotton” [28]. Here we ﬁnd the seed that has developed into the important technical terms “Cotton’s Phenomenon” and “Cotton Effect,” which have been used ever since, with the ﬁrst one preferred in the early decades of the twentieth century. At this point it may be timely to more formally give a deﬁnition of the Cotton effect as we understand it today. It may be interesting to compare the deﬁnition given in 1933 by Stotherd Mitchell on page 24 of his book on the Cotton effect (incidentally the ﬁrst monograph of this kind) [22] with the deﬁnition by Werner Kuhn from 1960 [30]. Mitchell wrote: “A maximum ellipticity and zero rotation are found in this region [of the absorption band]. The rotation reaches a maximum value on one side of the band and

12

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

a minimum on the other. This variation of rotation and ellipticity in the neighbourhood of an absorption band has been called the Cotton effect.” [Mitchell’s italics] Kuhn stated: “[Cotton] found that optical rotatory power as a function of the wavelength often shows, in the region where the substances show ordinary absorption, a characteristic anomaly which is associated with a circular dichroism in the absorption region and which after the name of its discoverer is called a Cotton effect.” [Kuhn’s italics] It is satisfactory that both deﬁnitions, published some 30 years apart, fully agree with one another; furthermore, we still can subscribe to both of them, even 50 years later. Many similar deﬁnitions can be found over the last 80 years, all of them stressing the point that the ensemble of rotatory dispersion and circular dichroism in the absorption region collectively constitute the Cotton effect. Nevertheless, quite commonly the term Cotton effect has loosely been used to characterize merely the “anomalous” rotation features, since in the decades following Cotton’s discoveries the available data were mostly limited to the optical rotation. In fact, in many cases it has been considered sufﬁcient to have reached the ﬁrst maximum of the rotatory dispersion curve, still outside the absorption band, to apply the term Cotton effect. In recent decades, when ORD effectively disappeared in favor of CD, the term usually means the CD curve only.

1.6. MORE TARTRATES: THE PHYSICIST’S PLAYGROUND Cotton’s discovery of circular dichroism raised so much interest in Brace’s Physics Laboratory at the University of Nebraska that it was decided to construct an improved and more sensitive apparatus for measuring both elliptical polarization and rotation, in order to repeat and extend the French ﬁndings. The American physicist DeWitt Bristol Brace was himself active in the ﬁeld of optical activity and had in 1904 described an elliptical polarizer and compensator that was incorporated not only in the optical system used in Nebraska, but also later in Europe. Brace died in 1905 and had, therefore, no part in the further development. The ﬁrst results on, for example, complex chromium, copper, cobalt, and nickel tartrates and copper malate were presented by M. F. McDowell in 1905 [31]. The ellipticity had been measured in “all parts of the spectrum,” which means at some 10 different wavelengths of the visual solar spectrum. Unfortunately, the calculation of the ellipticity was found to be incorrect, and some compounds were irreproducible. This was carefully rectiﬁed at the same laboratory in 1912 by L. B. Olmstead, who studied tartrates, malates, and lactates of chromium, copper, cobalt, and manganese [32]. Also here, the so-called “monochromatic” light, with a spectral band width of perhaps 20 nm, was obtained from sunlight. Although the optical part of the investigation seems to be impeccable (except that Cotton’s ﬁrst—incorrect—sign protocol of the circular dichroism was still used), the identity of the compounds studied is uncertain. Olmstead himself points out: “No chemical analyses of the compounds were made; the names assigned being merely for convenience, and not indicating that the chemical formulæ are known.” [Olmstead’s italics]. He observed that Cotton’s results for potassium chromium tartrate could be repeated quantitatively when the sample was prepared from potassium dichromate and potassium tartrate, but an oppositely signed Cotton effect developed when the potassium dichromate was replaced by chromium acetate. Undoubtedly, the samples consisted of a mixture of complexes, as was also indicated by color changes of the solutions, depending on variations of the concentration and with time. As a result, even these carefully collected data are of a qualitative nature only.

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

The same qualiﬁcations pertain to a large number of papers of the early twentieth century on the rotatory dispersion of transition metal complexes with optically active ligands. In addition, the rotatory dispersion data were often collected at only four or ﬁve rather ill-deﬁned spectral bands in the visible. Examples are found in the report by H. Grossmann and A. Loeb on copper tartrate and malate coordination compounds [33], as well as in the paper by H. Volk on copper, nickel, and cobalt complexes with lactate ligands [34]. Also these investigations were initiated, by the way, in order to verify and extend Cotton’s ﬁndings. Cotton himself did not continue his work on the optically active tartrates, but motivated his student Georges Bruhat to address the problem again [35]. Bruhat tried to synthesize and isolate individual, well-deﬁned compounds, but succeeded with most tartrates and malates in part only, because of the easy decomposition of the respective solutions. He did isolate and investigate uranyl tartrate that seemed to be stable and showed a Cotton effect near 500 nm. His optical equipment limited the quality of his measurements rather severely, however. After a disruption of the research by the First World War, he resumed his studies again in 1919 with a much more advanced apparatus. This allowed him to reduce the spectral band width from 30 nm to 10 nm, which was essential to avoid “ﬂattened-out” dispersion curves. In this way, he obtained splendid CD and ORD data for uranyl tartrate and ammoniacal cobalt tartrate [36], for example. But regrettably, not even the high quality of the physical data allows any better analysis of the compounds responsible. The interest in this topic was not yet put to rest in Cotton’s laboratory. In the earlier work, the copper complexes had been particularly unsatisfactory. Therefore, the study of alkaline copper tartrate solutions was taken up again in order to enhance the quality of the samples [37]. Somewhat later, complex chromium [38] and cobalt tartrates [39] were reinvestigated. Good CD data could be collected, but the chemical identity of the species in solution remained uncertain, despite Mathieu’s extensive experiments. Such tartrate studies were not wholly limited to Paris. Also W. Pﬂeiderer in Basel, Switzerland, had returned to measuring the optical rotation of aqueous alkaline solutions of copper tartrate, and he found his data to qualitatively agree with Cotton’s of 1895 [40]. The chiroptical instability that Nina Wedeneewa in Moscow, Russia, had encountered with alkaline chromium tartrates in the absorption region has been mentioned already [20]. Last but not least, attention is drawn to W. Kuhn’s reevaluation of the same problem, as outlined earlier [23]. The overview presented here is not exhaustive. Because of the similarity of the problems, the preceding discussion pertains also to, for example, optically active lactates, malates, and “sucrates” of transition metals. The respective chiroptical results are not basically different from those with tartrate ligands. It remains to report that even many years later the chemical identity of these complicated coordination compounds has not been fully understood, with several questions still unsettled even today [41]. While some of the variability observed is certainly caused by the gradual replacement of coordinated water by the organic ligands, condensation processes leading to multinuclear species also seem to be involved. It is intriguing that the chiroptical properties of tartrate complexes dominate the study of circular dichroism for three decades. In fact, during these years very few CD measurements have been carried out outside of this area (see later). One might speculate whether this conservatism would perhaps result from the fact that practically all researchers were physicists, who might have had limited awareness of colored optically

13

14

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

active compounds in other ﬁelds of chemistry. After all, with, for example, organic xanthates, such compounds did exist, but their investigation was largely conﬁned to only one chemical research group in Russia, as will be seen. The question might well be asked, What motivated these tenacious tartrate investigations? It is obvious that genuine chemical topics were not addressed, such as the stereochemical correlation and the application of optical activity to the study of reaction mechanisms, or as a tool in the elucidation of the chemical constitution. This contrasts with the aims in connection with the Werner complexes, to be discussed in the following section of this report. But as to tartrates, the investigators were primarily interested in the circular dichroism in its own right. They tried to effect improvements in the optical instrumentation in order to enhance the sensitivity and precision of their measuring devices. For testing the various theories of optical activity and to compare calculated and experimental data, the latter should be measured with a maximum of accuracy and reliability. It might well have been felt that the continuation with “well-known” samples like the complex tartrates would be advantageous, with a host of data already existing for comparison.

1.7. WERNER COMPLEXES: INORGANIC CHEMISTS LEARN TO MAKE USE OF THE COTTON EFFECT According to common practice, the tartrate systems discussed in the previous section can be considered to belong to the realm of inorganic chemistry. But they were chosen for chiroptical research without paying much attention to their chemical nature. Chemists have performed hardly any systematic studies of these compounds and have instead tended to neglect them. The situation was quite different with regard to the chemically and structurally welldeﬁned octahedral transition metal complexes, following the Alsatian Alfred Werner’s (1866–1919) introduction in 1893 of his geometric model for centers with the coordination number six [42]. At that time, Werner worked in the laboratory of his doctoral advisor, Professor Arthur Rudolf Hantzsch, at the University of Zurich, Switzerland. He had obtained his doctoral degree only in 1890, but was quickly promoted to a chair of chemistry at this University in 1895. Although his revolutionary concept eventually secured him the Nobel Prize in 1913, it met much resistance among his chemical colleagues. The opposition diminished, however, after he had achieved the resolution of some of his complexes into enantiomers [43], since the occurrence of optical activity was hard to reconcile with other than the octahedral geometry. The optically active compounds, mostly Co(III) complexes, were of greatly varying chemical and optical stability, often racemizing at room temperature within a few hours. It was found that chelating ligands like oxalate ions (O, O -donor ligands) or 1,2-diamines (N , N -donor ligands) led to increased stability. In the beginning, the optical rotation at only one wavelength was considered sufﬁcient to characterize a speciﬁc complex. But it was soon realized that the stereochemical correlation should not be based on such an individual value, since it varied too much in magnitude and even in sign from one complex to the next, notwithstanding a close chemical relationship. Furthermore, as the enantiomeric purity of the compounds was often uncertain, a particular, selected rotation value could be quite misleading. On the other hand, these complexes were well-suited to measurements of the Cotton effect due to their color, which often brought them within the range of visual observation. The sign of this Cotton effect, associated with electron transitions at the coordinating metal

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

center, was found to be a reliable, characteristic feature. Consequently, measuring the rotatory dispersion including, hopefully, a more or less complete Cotton effect became a common goal. The ﬁrst Werner complex for which both ORD and CD data through an absorption band were obtained, namely, potassium (trisoxalato)iridate(III) dihydrate, K3 [Ir(C2 O4 )3 ]·2 H2 O, was resolved by M. Del´epine [44], and its chiroptical properties were determined by G. Bruhat [45]. A Cotton effect near 450 nm was found. This is, incidentally, one of the rare instances where the circular dichroism of a Werner complex has been determined in the early part of the twentieth century. In practically all other cases, the term Cotton effect just refers to the rotatory dispersion near or at an absorption band in the visible. Werner himself had already reported similar dissymmetric, optically active complexes with bidentate ligands like [Co(en)3 ]3+ or [Cr(ox)3 ]3 – (en = 1,2-diaminoethane; ox = oxalate ion C2 O4 2− ), but without any further spectropolarimetric data [46]. It is amusing to see that in some other cases the Cotton effect—or, rather, the anomalous rotatory dispersion—rests on measurements at only three different wavelengths as for [Rh(ox)3 ]3− , for example [47]. Werner’s concept achieved its ﬁnal breakthrough when he published the resolution of the “completely inorganic” complex [Co{Co(NH3 )4 (OH)2 }3 ]Br6 , an octahedral Co(III) complex with bidentate O, O ligands. This complex, without any carbon atom, is sterically related to the simpler [Co(en)3 ]3+ system. It showed a Cotton effect near 600 nm [48]. This ﬁnding ﬁnally put to rest the long-lived but obsolete theory that the presence of carbon atoms was essential for the unfolding of optical activity. Werner showed no particular interest in the Cotton effect in its own right, however, but rather made use of it for the stereochemical correlation and the characterization of his compounds. A further example which may be mentioned is [Co(NO2 )2 (en)(pn)], with pn = 1,2-diaminopropane, with either rac-pn or l -pn, that had Cotton effects in the 530- to 540-nm range [49]. In general, Werner’s interest in spectropolarimetry remained limited, and usually he left further chiroptical studies to others. Even when he reported Cotton effects, it is not always clear how, where, and by whom the data were obtained. Meanwhile, a new “center of gravity” for the examination of Werner complexes was developing in Groningen in the Netherlands. Here, the stereochemist Franciscus Mauritius Jaeger (1877–1945), Professor of Inorganic and Physical Chemistry at the Rijksuniversiteit Groningen (RUG) from 1908 to 1945, had by 1915 embarked on a program of the comprehensive investigation of these systems [50]. A great many new, and some already known, Werner complexes were synthesized and studied. Jaeger’s interests lay largely in their crystallographic description, but he also included optical activity in his research program. Most of the spectropolarimetric data generated “plain” ORD curves only, because the anticipated Cotton effect was often beyond the wavelength limit of the optical devices or inaccessible because of too strong an absorption. In Jaeger’s extensive paper of 1919 [51], many such plain curves are reported, but only two cases of a bona ﬁde Cotton effect, namely, in K3 [Cr(ox)3 ] (Cotton effect near 565 nm) and K3 [Co(ox)3 ]·H2 O (near 620 nm). The chromium compound had already been synthesized and resolved by Werner in 1912, but had not been investigated by spectropolarimetry, whereas the (trisoxalato)cobaltate(III) was new. It was very advantageous that Jaeger could induce Israel Lifschitz (1888–1953), Private Docent at the University of Zurich, to join his laboratory in 1921. Lifschitz, whose special area of research had been the absorption spectroscopy and photochemistry of organic compounds, now became Private Docent of Electrochemistry and Photochemistry

15

16

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

and tenured staff member at RUG. He turned to the spectroscopic study of transition metal complexes, with particular attention to spectropolarimetry. Largely owing to his efforts, the laboratory developed into a center of chiroptical spectroscopy, of international repute. Lifschitz had taken his doctorate in 1911 with A. Hantzsch in Germany (Hantzsch had changed from Zurich to the University of Leipzig) and had moved to Switzerland in 1914. It is interesting to note that he there worked close to Werner, who also had been a student of Hantzsch. Perhaps the shift of Lifschitz from organic photochemistry to coordination stereochemistry has thereby been inﬂuenced, but presently nothing is known of a personal interaction with Werner. It is difﬁcult to assay in detail Lifschitz’s contribution to the Groningen laboratory, except for his own publications, since Jaeger as laboratory head regularly included the research results of his local colleagues in his own publications, without giving any individual credit. This procedure was rather common in those days. Only sometimes is Lifschitz mentioned in a vague way as a “collaborateur” (coworker). In 1923 Lifschitz started a series of papers called “Investigations of Rotatory Dispersion” [translated from the original German]. In the ﬁrst paper, he presented and discussed ORD data of complexes of Cr(III), Co(III), Ni(II), and UO2 2− ions with optically active camphor derivatives, including nitrocamphor. These compounds exhibited Cotton effects in the visible. The paper is also noteworthy, because here the technical term “Cotton effect” was introduced into the chemical literature; the earlier term had been “Cotton phenomenon” [52]. In the second paper of this series, Lifschitz reported the ORD Cotton effects of Co and Cu complexes with amino acid ligands (alanine, asparagine), and also of the complex [Cu-(l -pn)2 ]SO4 (effect at 510 nm) [53]. Slightly later, Jaeger extended the chiroptical studies to cobalt complexes with 1,2-diamino ligands, and he reported the Cotton effects in [Co(rac-trans-1,2-diaminocyclopentane)3 ]Cl3 · 4H2 O at 470 nm and in [Co(rac-trans-1,2-diaminocyclopentane)(en)2 ]Br3 · 2H2 O at 500 nm [54]. A few years later, some of these Werner complexes were reinvestigated by Werner Kuhn, making use of advanced instrumentation. Measurements had now become possible down to 280 nm. With potassium (trisoxalato)cobaltate(III), for example, the Cotton effect at ∼600 nm was measured both in rotatory dispersion and in circular dichroism [55]. But now the aim had shifted from using the spectropolarimetric data for chemical and stereochemical correlation, as had been the purpose in Zurich and Groningen, to probing the stereochemistry in depth, with the elucidation of the absolute conﬁguration in mind, and to testing new theoretical models of the optical activity. But it would take a few additional decades before eventually another experimental reinvestigation of the circular dichroism of some of these complexes, in connection with an analysis based on an improved theory, led to the desired knowledge of both the structure in solution and the absolute conﬁguration. To this end, once again the circular dichroism of the (trisethylenediamine)cobalt(III) cation [56] and of the (trisoxalato)cobaltate(III) anion [57] was studied. But to trace this development would far exceed the scope of the present overview.

1.8. THE COTTON EFFECT IN ORGANIC CHEMISTRY, A RUSSIAN DOMAIN It would not be correct to claim that organic chemists neglected optical activity in the early twentieth century, except to characterize compounds by their D-line rotation. As an example to the contrary can be cited a series of papers by the Swiss chemist Hans Rupe (1866–1951) on the inﬂuence of the constitution on the rotatory power of optically active

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

compounds, starting in 1903 [58]. However, usually colorless solutions were examined in the visible, as also done by Rupe, and could provide data on “plain” rotatory dispersion only. These investigations are, therefore, outside of the scope of the present overview. However, it must be admitted that, in general, organic chemists seem to have been less interested in chiroptical effects than their inorganic or physicochemical colleagues. Thus the study of organic molecules, as much as there is, had to rest on the “good will” of people from the latter fold. Here, credit is to be given ﬁrst to the distinguished English physical chemist, Thomas Martin Lowry (1874–1936). Lowry’s interest in optical activity dates back to 1898, when he noted the change of optical rotation on nitrocamphor with time and introduced the term “mutarotation” to characterize this phenomenon [59]. He greatly improved the mathematical treatment and the theoretical understanding of rotatory dispersion and circular dichroism, based in part on the experimental data collected in his own research group in London, and later in Cambridge. It is interesting to see that there was a certain lack of understanding of the theory of optical activity on the part of some organic chemists. Even Rupe himself, for example, maintained as late as in 1921 that there had not been established with certainty any connection between the Cotton effect and the absorption of light [60]. In the earlier period of his career, Lowry dealt mostly with features outside of absorption bands; that is, he did not penetrate by experiment into the Cotton effect region itself. The state of the research on optical activity by the year 1914 has been summarized in the report “Optical Rotatory Power. A general Discussion” [61], and later in Lowry’s classic book [21], and it need not be described further in this paper. Lowry’s work in the early 1930s on the Cotton effect of organic molecules will be discussed later in this section. The most active pioneer in the study of the rotatory dispersion of organic molecules, and the only one who obtained data for the Cotton effect before World War I, is undoubtedly the Russian Leo [Lev] Alexandrovitsch Tschugaev (1873–1922) [62]. Tschugaev was a proliﬁc research worker, who from the beginning of his career engaged in the chemistry of compounds like terpenes and camphor and, secondly, that of transition metal complexes. It was probably the study of optically active natural products that aroused his interest in optical activity generally. Eventually, he turned to a third research topic, after he had become Professor of Inorganic Chemistry at the Imperial University of St. Petersburg, and began in 1909 a series of papers on (anomalous) rotatory dispersion [63]. Tschugaev was fully aware of Cotton’s ground-breaking discoveries, and he was also aware of the problems inherently connected with the samples chosen for this early work. Therefore, he proudly, and correctly, pointed out in his initial papers that he now for the ﬁrst time employed well-deﬁned compounds for the study of the rotatory dispersion (he himself had no instrument that would allow him to measure the circular dichroism, in addition). But also he was still limited to visual observations at certain spectral lines. He used samples from two different families of sulfur-containing colored derivatives of optically active terpene alcohols like borneol, menthol, or fenchol, namely, xanthates RO–C(S)–SR and related compounds, along with “dithiourethanes” RO–C(S)–NPh–C(S)Ph (with Ph = C6 H5 ) and similar compounds. In all cases he found an anomalous dispersion of the rotation, but for different reasons. The xanthates give colorless or yellowish solutions, because there are no absorption bands in the visible. The anomalous rotatory dispersion detected is, therefore, of the type already observed for tartaric acid by Biot, and it is not caused by a Cotton effect in this spectral region. The red dithiourethanes, however, do show an absorption band at ∼520 nm, and the rotatory dispersion features indeed result from a Cotton effect associated with this absorption.

17

18

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

It must be admitted that Tschugaev was only able to detect the ﬁrst maximum of the ORD curve, because of the onset of high absorption near 450 nm precluding measurements of the short-wavelength part of the dispersion curve. Therefore, it might be considered stretching a point rather far to speak of his truly having detected Cotton effects, since this depends on quite some interpolation and interpretation. But his contention was much strengthened by G. Bruhat, to whom he had sent samples of both the d - and the l -bornyl dithiourethane mentioned above. Bruhat was able to measure the circular dichroism, both in toluene solution and in the melt [35, 64]. The CD maximum at 520–530 nm coincided with the absorption maximum at 520 nm. Bruhat could conﬁrm, but not extend, Tschugaev’s ORD data, by the way. This then not only proved that Tschugaev’s interpretation of his rotatory dispersion curves had been correct, but also provided the ﬁrst example of circular dichroism observed in a well-deﬁned (organic) compound. With Tschugaev’s papers at hand, it is interesting to follow the progress of his search for a ﬁtting technical term to describe what he initially calls an anomalous rotatory dispersion “in the sense of Cotton,” until he ﬁnally in 1912 arrived at the “Cotton phenomenon.” This then became the internationally accepted term to be used for two decades, until it lost ground against Lifschitz’s “Cotton effect.” T. M. Lowry revisited Tschugaev’s compounds in 1932, conﬁrming the earlier data and extending the wavelength range of the observations, thanks to improvements in the optical instrumentation and the introduction of photography. He also performed calculations on a more advanced basis in order to analyze and simulate the data [65]. Now it had become possible to take photographic readings at many points of the wavelength scale down to 325 nm. Lowry not only supplemented and slightly extended Tschugaev’s earlier ORD results, but now he could additionally make available circular dichroism data for the xanthates. These compounds, with a weak absorption at 360 nm (which Tschugaev had missed) and a strong one at 280 nm, exhibit a CD maximum at ∼355 nm. The steep rise of the absorption toward shorter wavelengths still precluded the precise observation of the second ORD maximum at about 330 nm. It is notable that Lowry found his photographic CD measurements of the dithiourethanes in the year 1932 less exact than Bruhat’s visual measurements of 1911. Colored organic compounds were not unknown apart from the sulfur-containing derivatives discussed above, but were not easily available in optically active form. For chiroptical studies they should advantageously stem from the pool of optically active natural compounds or their derivatives, because the organic chemists of those days seem to have tended to avoid resolutions, contrary to their colleagues in the ﬁeld of Werner complexes. Therefore, it is not surprising that the yellow camphor quinone attracted the attention of physical chemists and physicists alike. The Russian physicist Nina Wedeneewa detected in this compound a Cotton effect near 490 nm by ellipticity and rotatory dispersion measurements; although the work had been done in 1919, its publication was delayed until 1923 because of the political turmoil in Russia [20]. Her main interest was the analysis of the data in terms of the Drude theory of optical dispersion. Slightly later, in 1925, Israel Lifschitz also had tried to measure the optical rotation of camphor quinone near the absorption region, but could reach the ﬁrst maximum only [53]. Lowry later repeated, conﬁrmed, and extended Nina Wedeneewa’s ﬁndings, thereby tacitly correcting the sign of her circular dichroism data, as he studied camphor quinone both in solution and in the vapor phase [66]. Also this second case, in which the circular dichroism of an organic molecule has been measured successfully, in addition to Tschugaev’s/Bruhat’s dithiourethane,

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

stems from Russia. Unfortunately, the First World War, the Russian Revolution, and Tschugaev’s premature death interrupted and eventually ended a period of intensive research on chiroptics in that country. Intriguing colored organic compounds are the “nitrosites” (1,2-nitroso-nitrites) and “pseudo-nitrosites” (1,2-nitroso-nitro compounds), which can be made by the addition of (formally) N2 O3 to oleﬁns. These compounds exhibit green or blue colors in solution, provided that the nitroso group is monomeric. E. Deussen, working with the blue optically active caryophyllene nitrosite, had noticed an anomalous rotatory dispersion, minimalistically based on readings at just two wavelengths [67]. The suggestion by Tschugaev, that this might be caused by a Cotton effect in the visible [68], led Stotherd Mitchell in Glasgow in 1928 to check the chiroptical properties of this compound [69], but with readings taken at eight wavelengths between 691 and 436 nm. Indeed, a Cotton effect was found in the CD at ∼680 nm. It might be remarked that the absorption curves were measured separately with right- and left-circularly polarized light, whereas in practically all earlier cases data on ellipticities had been collected. In summary, the circular dichroism of only three (colored) organic compounds had become known by 1928. Although in some cases the claim has been put forward to have seen in such compounds a Cotton effect by rotatory dispersion, this should be taken with a grain of salt, because this cannot normally be substantiated by presenting the whole sigmoidal dispersion curve. Usually one wing of the curve is missing (commonly on the high-energy side of the absorption band), for reasons discussed above. It may come as a surprise, therefore, that as early as 1910, Eug`ene Darmois, one of Cotton’s students, published the ORD Cotton effect of even a colorless organic compound, namely, camphor, at ∼300 nm [70]. It is true that he had not been able to obtain rotation data at the absorption maximum itself, but only on either side of it (with a gap between 313 and 265 nm), but the dispersion curve can be easily completed by interpolation. This ﬁnding is all the more remarkable, because it demonstrated for the ﬁrst time the possibility of taking rotation values down to ∼250 nm in the ultraviolet. Regrettably, the response of the chemical community was slow, probably because of the technical difﬁculties involved in the construction of a suitable spectropolarimeter. It took some 20 years before Darmois’s work could be taken up again; then, W. Kuhn published ORD, CD, and UV data of camphor, taken right through the absorption band [71], and T. M. Lowry likewise reported data of the related camphor-β-sulfonic acid [72]. It seems rather daring that Darmois [70] also tried to measure the chiroptical properties of oleﬁns like α- and β-pinene or limonene, but—not unexpectedly—without much success. This chromophore still resisted the efforts of R. Servant in 1932, but at least this time some indication of a ﬁrst ORD maximum seemed to be suggested at around 280 nm in the case of the pinenes [73]. From the foregoing it appears that rather suddenly, by around 1930, many more types of compounds were studied by chiroptical techniques. This resulted from the progress in instrumentation, to be related brieﬂy in the next section. Now, a great many additional chromophores in colorless organic compounds, like nitro, azido, nitrito, and particularly carbonyl groups, opened the way for studies of the circular dichroism and the rotatory dispersion. Nevertheless, just a few typical samples were used to be investigated for physical–chemical purposes. It would take many more years before organic chemists made use of the now accessible optical techniques for stereochemical correlations and the determination of the absolute conﬁguration. But this then is far outside the scope of this overview. These newer developments since the early 1930s will not be treated here, because many summaries are already available. The major source of information

19

20

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

for the organic chemist on these later chiroptical investigations is Carl Djerassi’s book [74], which has become a “classic” in the ﬁeld to which the reader should turn.

1.9. ADVANCES IN INSTRUMENTATION AND THEORY; THE WAY INTO THE ULTRAVIOLET Very brieﬂy, the technical advances shall be sketched here that, around 1930, led to the (short-lived) upsurge in chiroptical activity mentioned in the foregoing section. Details shall be omitted, because comprehensive reviews are readily available. In addition to Lowry’s encyclopedic monograph of 1935 [21] and Mitchell’s book of 1933 [22], one should pay particular attention to Bruhat’s 1930 treatise on polarimetry, since most of the “newer” development of spectropolarimeters and ellipsometers resulted from the efforts at his laboratory in Paris [75]. Another active center was at the Technische Hochschule (Institute of Technology) Karlsruhe, Germany, where the Swiss physical chemist Werner Kuhn not only worked on theoretical and experimental chiroptics, but also developed his own apparatus [76]. Relevant optical instruments have also been reviewed by R. Descamps in Brussels, Belgium, who had himself constructed and perfected a spectropolarimeter for the UV region [77]. Of course, Lowry’s important contributions from his Laboratory of Physical Chemistry in the University of Cambridge, UK, are by no means to be forgotten. Progress in instrumentation for chiroptical studies meant, besides the obvious improvement of sensitivity and reliability, by and large the extension of the wavelength range into the ultraviolet. It was evident that the absorption bands of the vast majority of chemical compounds are located in the UV. It may come as a surprise that polarimetric measurements in the ultraviolet have been known for a considerable time and can be traced back to the nineteenth century. Various instruments for this purpose are described by Lowry [21], but the operation of the apparatus was laborious, the accuracy of the results questionable, and the accessible wavelength range rather limited. It does not seem, moreover, that these techniques have been applied to the detection of Cotton effects before the exceptional pioneering work of E. Darmois in 1911 [70]. One of the earlier attempts at measuring optical rotations in the UV was by Lowry himself, who in 1908 combined a half-shadow polarimeter with a UV spectrograph [78]. Darmois could collect data to the wavelength limit of 250 nm. But in order to make this wavelength range more generally accessible, and even extend it toward higher energy, three different problems had to be solved: ﬁrst, with regard to the light sources; second, with the transparency of optical components to the UV light; and third, in connection with the efﬁciency of detectors. 1. The solar spectrum provided light in the laboratory only down to the limit of 300 nm, because of atmospheric absorption. Mercury vapor lamps allowed readings to be taken for another 50 nm, down to 250 nm. Beyond that wavelength, various other light sources have been used, in all cases providing an array of separate spectral lines—for example, the iron arc to 233 nm, or the cadmium spark to 210 nm. 2. The transparency limits of glass preclude its use in UV instruments for lenses, prisms, or similar optical devices, except for the near-UV range. Materials like Iceland spar (cutoff at 250 nm), ﬂuorspar, or quartz had to be introduced to improve the transmittance of UV radiation. Even the Canada balsam used in

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

conventional Nicol prisms had to be replaced by, for example, glycerol, in order to allow their utilization below 340 nm. 3. As to the detectors, the ﬂuorescent screens of the early days were soon replaced by photographic implements, but the evaluation of the developed plates posed new problems in its own right. The improvements in all these areas led to the independent construction of two similar recording spectropolarimeters in the year 1926, giving access to UV measurements up to the high energy limit of 250 nm [79]. An important objective in the construction of these instruments was the desire to hold to a minimum the number of optical parts that had to be transparent to UV light. In Cotton’s photographic spectropolarimeter, the optical density of the photographs had to be evaluated by means of a photometer or, better, by a microphotometer. Although the recording was automatic, it was not continuous, but consisted of discrete exposures. Bruhat, on his part, used a photoelectric device. The data analysis could still be cumbersome, but the measurement of rotations in the UV had nevertheless become relatively easy. Bruhat’s polarimeter was perfected still further [80], to the extent that the photoelectric measurements became superior in precision to those obtained by the photographic method applied by Servant [73]. All the instruments mentioned so far are polarimeters. Similar advances had taken place in the construction of ellipsometers. Again, Bruhat was in the forefront with the development of a polarimeter–ellipsometer [81]. This widely used visual instrument consisted of an ordinary polarimeter, ﬁtted with a mica λ/4 plate. Werner Kuhn developed an even more advanced photographic device for use in the ultraviolet [82]. It contained optical parts of quartz and ﬂuorite only, and it allowed measurements to be taken all the way to 190 nm. This apparatus was the preferred instrument for many years to come and was marketed by the well-known makers of optical instruments in Berlin, Schmidt & Haensch. A similar ellipsometer was described by Mathieu in Paris [83]; it was designed particularly for the wavelength range of 450–280 nm. In conclusion, in the beginning of the 1930s, the development of instrumentation to measure the optical rotatory dispersion and the circular dichroism (in terms of ellipticities) had progressed to a state of perfection that was hard to improve upon in the decades to come. But, although it had been convincingly shown now that the Cotton effects of a wealth of optically active (organic) compounds were accessible, the chemical community at large was slow to make use of the chiroptical techniques. This resulted probably in part from the fact that the importance of stereochemistry had not yet been realized widely, especially among organic chemists. Also, the necessary apparatus still had to be built individually, and the measurement of ORD and CD was still far from being routine. Only with the advent of commercial recording instruments for ORD [84] and CD [85] many years later was the ﬁeld of chiroptical investigations opened to the general chemist. Very few sentences must sufﬁce on the contemporary development of the theory of optical activity, because this topic lies far away from the present overview. Paul Drude’s theory of optical activity in isotropic media, as expanded in his famous book of 1900 [26], has been the standard with which most physical(–chemical) research had to contend for the ﬁrst three decades of the twentieth century. Many experimental investigations, including Lowry’s, have been motivated by the search for an improved version of the “Drude equation” of optical activity. An important step forward in this line was taken by L. Natanson, who in the year 1909 succeeded in deriving an equation that could approximate the optical dispersion within the absorption band, a problem that Drude’s original equation could not handle [17, 27, 28]. An advanced theory of optical activity

21

22

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

was published by Max Born (1882–1970, Nobel Prize 1954) in 1915 [86] and at about the same time by C. W. Oseen [87] and F. Gray [88]. Although the importance of these papers, especially Born’s, was immediately recognized, practically no direct inﬂuence on the study of chiroptical properties was engendered. Likewise, the important paper by the Belgian physicist L´eon Rosenfeld (1904–1974) on the quantum-mechanical theory of optical activity did not affect the chemical community at the time [89]. It took Werner Kuhn’s (1899–1963) simpliﬁcation of Born’s theory into his coupled-oscillator model to attract the chemists’ attention [76, 90]. Now the application of theoretical methods to actual problems of stereochemistry seemed to become realistic. Even the determination of the absolute conﬁguration of molecules might come into reach. Indeed, Kuhn’s conclusion that (−)-butan-2-ol had the (R) conﬁguration was the culminating point of his theoretical work [91]. A comprehensive overview of these developments in theory is found in Mathieu’s monograph on the molecular theories of natural optical activity [92].

1.10. SOME WORDS ON NOMENCLATURE: COTTON EFFECT, OPTICAL ROTATORY DISPERSION, CD, ORD The expression “Cotton effect” (originally in German) was introduced by Israel Lifschitz in 1922. The earlier technical term, ﬁrst used by Tschugaev in 1912 (also originally in German), is “Cotton’s phenomenon.” This latter expression developed gradually by the contraction of phrases like “anomale Rotationsdispersion im Sinne A. Cottons” (anomalous rotation dispersion in the sense of A. Cotton) [93] or “la mani`ere de voir de M. Cotton” (literally: Mr. Cotton’s way of viewing) and “ph´enom`ene de la dispersion anormale” (phenomenon of anomalous dispersion) [94]. Tschugaev never used the word “Cotton effect” in whatever language. It is slightly confusing, therefore, to ﬁnd this term in his Russian collected works [95], but on closer inspection this turns out to appear in a posthumous translation of his German papers into Russian. He himself always wrote ” (i.e., Cotton’s phenomenon). in Russian “ As can be seen from Table 1.1, both expressions were used side by side for a period of 10 years, but eventually “Cotton effect” was victorious. The situation is slightly simpler with “rotatory dispersion.” From the beginning in 1877, this term had become established, with minor variations as shown in Table 1.2. The only point meriting some attention is the question, Why has rotatory dispersion nowadays become “optical rotatory dispersion”? The “optical” was originally added in order to differentiate between the two effects of magnetic and optical rotatory dispersion, both of which were the subject of a series of papers by Lowry, with the ﬁrst one appearing in 1913. Usually, however, it was considered unnecessary to point to this difference, because magnetic rotatory dispersion rarely plays a role in chemistry. The modern habit of always referring to “optical rotatory dispersion”, which also led to the common abbreviation “ORD,” seems to originate with Carl Djerassi, who used it since 1955 in a great many papers and who publicized it further by his textbook on ORD. The counterpart expression with respect to circular dichroism should be “optical circular dichroism,” in order to likewise differentiate between optical and magnetic circular dichroism. Indeed, the French “fathers” of modern CD instruments and their application in chemistry have used the term “dichro¨ısme circulaire optique” [96], but for reasons unknown, in this case the chemical community has continued to ignore the “optical.” For the ﬁrst half of the twentieth century, abbreviations were not much used in (physical) chemistry. With fashions changing, this became a craze, however, in the second half, especially in the United States and the Soviet Union. Typical is the general use

23

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

TAB L E 1.1. The Change from Cotton’s Phenomenon to the Cotton Effect The Cotton Phenomenon 1912 1913 1914

1915 1922 1922 [1925]

1926

1930, 1933

L. Tschugaeff: Cottonsches Ph¨anomena L. Tchougaeff: ph´enom`ene Cottonb L. Tschugaeff: Cotton’s phenomenonc M. Del´epine: les ph´enom`enes d´ecouverts par M. Cottond H. Rupe: Cottonsches Ph¨anomene H. Grossmann: Cotton-Ph¨anomenf F. M. Jaeger: ph´enom`ene de Cottonh

The Cotton Effect

1922

J. Lifschitz: Cottoneffektg

1922 [1925] 1923

F. M. Jaeger: l’effet de M. Cottonh F. M. Jaeger: l’effet Cottoni

1928 1929 1930

S. Mitchell: Cotton Effectk W. Kuhn: COTTON-Effektl G. Bruhat: l’effet Cottonn

1935

T. M. Lowry: Cotton Effecto

W. Pﬂeiderer: Cottonsches Ph¨anomenj

T. M. Lowry: Cotton phenomenonm

Cotton effect (Engl.); Cottoneffekt (German, Danish); effet Cotton (French); Effeto Cotton (Italian); Efekt Cottona (Polish); (Russian) a L. Tschugaeff, J. Prakt. Chem. 1912, [2] 86 , 545–550; L. Tschugaeff, G. Glinin, Ber. Dtsch. Chem. Ges. 1912, 45 , 2759–2764. b L. Tchougaeff, A. Kirpitcheff, Bull. Soc. Chim. Fr. 1913, [4] 13 , 796–803. c L. Tschugaeff, Trans. Faraday Soc. 1914, 10 , 70–79. d M. Del´ epine, C. R. H. Acad. Sci . 1914, 159 , 239–241. e H. Rupe, Liebigs Ann. Chem. 1915, 409 , 327–357. f H. Grossmann, M. Wreschner, Die anomale Rotationsdispersion, Sammlung chem. u. chem.-techn. Vortr¨ age, W. Herz, ed., Enke, Stuttgart, 1922, 26 , 259–314. g J. Lifschitz, Rec. Trav. Chim. Pays-Bas 1922, 41 , 627–636. h F. M. Jaeger, Rapp. Disc. Inst. Int. Chimie Solvay (Conseil Chim. 1922, Bruxelles), Gauthier-Villars, Paris, 1925, 199–202. i F. M. Jaeger, Bull. Soc. Chim. Fr. 1923, [4] 33 , 853–889. j W. Pﬂeiderer, Z. Phys. 1926, 39 , 663–685. k S. Mitchell, J. Chem. Soc. London 1928, 3258–3260. l W. Kuhn, Z. Phys. Chem. 1929, B4 , 14–36. m T. M. Lowry, Trans. Faraday Soc. 1930, 26 , 266–271; T. M. Lowry, H. Hudson, Philos. Trans. 1933, A232, 117–154. n G. Bruhat, Trait´e de Polarim´etrie, Editions de la Revue d’Optique, Paris, 1930. o T. M. Lowry, Optical Rotatory Power, Longmans, Green and Co., London, 1935; reprint: Dover Publications, New York, 1964.

of “RD” in C. Djerassi’s book published in 1960, incidentally one of the ﬁrst major applications of this abbreviation. It is amusing to note that here “RD” still refers to “rotatory dispersion,” of course, while the same author had already adopted “optical rotatory dispersion” for some years. It was consistent with adding “optical” to “rotatory dispersion,” to also add the letter “O” to the abbreviation “RD”: The birth of “ORD” took place in Djerassi’s environment in the early 1960s.

24

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

TAB L E 1.2. From Rotatory Dispersion to Optical Rotatory Dispersion Rotatory Dispersion, Dispersion Rotatoire, Rotationsdispersion

1913 1914 1915 1917 1935

H. Landolta: (anomale) Rotationsdispersion [German] A. Cottonb: dispersion rotatoire (anomale) [French] H. Grossmannc: (anormale) Rotationsdispersion [German] L. Tschugaeffd: anomale Rotationsdispersion im Sinne A. Cottons [German] L. Tschugaeffe: dispersion rotatoire ano(r)male [French] T. M. Lowryf : rotatory dispersion [English] T. S. Pattersong: (normal/abnormal) rotation-dispersion [English] T. M. Lowryh: an exact deﬁnition of normal and anomalous rotatory dispersion F. M. Jaegeri : rotatie-dispersie [Dutch] T. M. Lowryj : (normal and anomalous) rotatory dispersion [English]

1913 1926 1955 1959 1960

T. M. Lowryk: magnetic and optical rotatory dispersion [English] W. Pﬂeidererl : optische/magnetische Rotationsdispersion [German] C. Djerassim: Optical Rotatory Dispersion Studies [1st paper] [English] F. Woldbyen: optical rotatory dispersion [English] C. Djerassi: Optical Rotatory Dispersion, McGraw-Hill, New York [English]

1877 1895 1908 1911

Optical Rotatory Dispersion

a

H. Landolt, Liebigs Ann. Chem. 1877, 189 , 241–337 (p. 274). Cotton, C. R. H. Acad. Sci . 1895, 120 , 989–991. c H. Grossmann, H. Loeb, Z. Ver. Deutsch. Zuckerind., Allg. Teil 1908, 58 , 994–1009. d L. Tschugaeff, Z. Phys. Chem. 1911, 76 , 469–483. e L. Tschugaeff, A. Ogorodnikoff, Ann. Chim. Phys. 1911, [8] 22 , 137–144. f T. M. Lowry, J. Chem. Soc. London 1913, 103 , 1062–1067. g T. S. Patterson, Trans. Faraday Soc. 1914, 10 , 111–117. h T. M. Lowry, J. Chem. Soc. London 1915, 107 , 1195–1202. i F. M. Jaeger, Chem. Weekbl . 1917, 14 , 706–732. j T. M. Lowry, Optical Rotatory Power, Longmans, Green and Co., London, 1935. k T. M. Lowry, T. W. Dickson, J. Chem. Soc. London 1913, 103 , 1067–1075. l W. Pﬂeiderer, Z. Phys. 1926, 39 , 663–685. m C. Djerassi, E. W. Foltz, A. E. Lippman, J. Am. Chem. Soc. 1955, 77 , 4354–4359. n F. Woldbye, Acta Chem. Scand . 1959, 13 , 2137–2139. b A.

The abbreviation “CD” for circular dichroism was introduced at about the same time. Whereas “OCD” has never become popular, “MCD” for magnetic circular dichroism is a logical and accepted abridgment. In the same vein, a reasonable abbreviation for magnetic rotatory dispersion should be “MRD”. The misnomer “MORD” (which in German means “murder,” by the way) should fall into disuse. Details concerning these various abbreviations are collected in Table 1.3. It remains to remind the reader that the terms “rotatory dispersion” and “circular dichroism” as well as their abbreviations might change, if languages other than English are used. Some examples are collated in Table 1.4.

1.11. BIOGRAPHICAL NOTICES: G. BRUHAT, A. COTTON, W. KUHN, I. LIFSCHITZ, T. M. LOWRY, L. NATANSON, AND L. TSCHUGAEV It is a moot question how far a treatise on the history of natural sciences ought to be supplemented by a personalized account. It is the present author’s contention that such an approach is helpful to provide a balanced background for the scientiﬁc results

25

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

TAB L E 1.3. From RD to ORD (but not from CD to OCD) 1960

C. Djerassi,a Optical Rotatory Dispersion: “abbreviated RD (curves)”

1960

W. Klyne, Adv. Org. Chem. 1 , 239–348: “abbreviated R.D. curves”

1965

P. Crabb´e,b Optical Rotatory Dispersion and Circular Dichroism in Organic Chemistry: “circular dichroism, abbreviated CD; rotatory dispersion curves, abbreviated RD”

1965

L. Velluz, M. Legrand, M. Grosjean,c Optical Circular Dichroism: “CD-curves”

1963

D. Lightner,d PhD Thesis: ORD

1966

K. Mislow,e Introduction to Stereochemistry: ORD, CD [preface September 1964, © 1965, published 1966]

1967

G. Snatzke,f ed., Optical Rotatory Dispersion and Circular Dichroism in Organic Chemistry [Summer School, Bonn, 24 September–1 Oct., 1965]: ORD, CD; “MORD”, MCD

1972

P. Crabb´eg, ORD and CD in Chemistry and Biochemistry

a

McGraw-Hill, New York. Holden-Day, San Francisco. c Verlag Chemie, Weinheim/Academic Press, New York. d Stanford, CA. e W. A. Benjamin, New York. f Heyden & Son, London. g Academic Press, New York. b

TAB L E 1.4. ‘CD’’ and ‘‘ORD’’ [and ‘‘CE’’] in English, French, German, and Russian CD RD ORD

DC DR DRO

КД

circular dichroism, Circulardichroismus rotatory dispersion, Rotationsdispersion optical rotatory dispersion, optische Rotationsdispersion Круговой дихроизм

dichro¨ısme circulaire dispersion rotatoire dispersion rotatoire optique

ДОВ

Дисперсия оптического вращения

[CE

Cotton effect, Cotton-Effekt]

[EC

effet Cotton]

discussed in the foregoing sections. After all, the common expression “it was found that” is deceptive insofar as it tends to obscure the fact that these results were not “found” by anonymous agencies, but earned by individual scientists, working at a speciﬁc time and under speciﬁc circumstances that pertain to both their professional and their private lives. Among the many actors in the ﬁrst decades after the discovery of the Cotton effect, who could be considered for some biographical notices, here seven scientists are selected, as listed in the section title in alphabetical order. These names may reﬂect some personal predilection, but it is not too difﬁcult to defend these choices impartially: Cotton is, of course, of outstanding importance. Natanson opened the route to an understanding of the interconnection of CD, ORD, and absorption generally. Tschugaev pioneered the investigation of chiroptical properties of organic molecules, while Lifschitz did the same with coordination compounds and has, moreover, coined the technical term “Cotton effect.” Bruhat was of particular importance for the development of instruments for chiroptical studies, together with Lowry and Kuhn, while the two latter scientists (both of them chemists, by the way) were instrumental also for the progress in the application

26

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

of the theory of optical activity to chemical problems. All of these persons earned universal recognition, and even fame, also in other ﬁelds of chemistry or physics. All of them would merit attention, even if their part in the progress of chiroptics were to be neglected. Actually, for some among them, their activity in this ﬁeld was rather incidental, if seen against their lifetime achievements. In the following, summary biographies of these scholars will be related in sequence according to their date of birth, with Cotton’s biography at the end, however. The theoretical physicist Ladisla(u)s [Władysław] Natanson (1864–1937) was born in Warszawa (Warsaw), Poland, the son of a medical doctor. Warsaw was under Russian rule at the time, and Natanson was a Russian subject. After graduation from a classical school in 1882, he enrolled as a student at the Faculty of Sciences of the University of St. Petersburg, where he became a “Candidate” (Licentiate) in 1886. After a few months at the Cavendish Laboratory at Cambridge, U.K., he returned to Imperial Russia in order to fulﬁl the requirements necessary for Russian subjects who wanted to embark on an academic career. Therefore, in 1887 he moved to Dorpat, Livonia (now Tartu, Estonia), to work for his doctorate with the physicist Professor A. von Oettingen, which he obtained in 1888. Incidentally, the ofﬁcial language at the University of Dorpat was German. After some postdoctoral studies with L. Boltzmann at the University of Graz, Austria, he returned to Warsaw to write an “Introduction to Theoretical Physics” [97]. This book was met with much acclaim, which helped him to be granted a position at the Jagiellonian University Krak´ow (Cracow), Poland (then under Austrian rule). There he moved up through the academic ranks: 1894 Titular Professor, 1899 Extraordinary Professor, and ﬁnally 1902 Professor of Theoretical Physics. Later he was also appointed Dean of the Faculty, and in 1922/23 he became Rector of the University. Elected to the Academy of Sciences in 1893, he became President of the Section of Mathematical and Natural Sciences in 1926, until he resigned from both the Academy and the University in 1935 for health reasons. His professional achievements are expanded upon in the obituary notice by L´eon Klecki, which includes a bibliography [98]. The chemist Lev Aleksandrovi(ts)ch Tschugaev (Chugaev) (1873–1922) was born in Moscow, Russia. After completing his studies at the University of Moscow in 1894, he became Assistant at the Bacteriological Institute of the University, where he started his research on the optical activity of organic compounds. From the beginning, Tschugaev strove to be competent in organic as well as inorganic chemistry. His master’s thesis in 1903 dealt with studies in the terpene and camphor series, while in his doctoral thesis (habilitation) of 1906 he presented results from coordination chemistry. Throughout his life, he successfully followed this dichotomy in his research, with some excursions into physical chemistry. In 1906 he was appointed Professor at the Technical University in Moscow, and in 1908 he was called to the Chair of Inorganic Chemistry at the Imperial University of St. Petersburg (later: University of Petrograd). He held this position at the time of his death. During World War I and the Russian Revolution he was mainly active in the ﬁeld of technical and applied chemistry. He was one of the founders of the Institute of Applied Chemistry in Petrograd and became its director. In the aftermath of the revolution, he died of typhoid fever at Wologda, Russia, at the age of not yet 50. Obituary notices have appeared in England [62] and in Germany [99], acknowledging his abundant contributions to chemistry. Attention should also be paid to his “Selected Works” [95]. The physical chemist Thomas Martin Lowry (1874–1936) was born at Low Moor, Bradford, Yorks., U.K., the son of a Wesleyan Chaplain. He studied at the Central Technical College, South Kensington, London. From 1896 to 1913 he was assistant to Professor

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

H. E. Armstrong and from 1904–1913 he was Lecturer in Chemistry at the Westminster Training College. In 1913 he became Head of the Chemical Department in Guy’s Hospital Medical School and Professor of the University of London. In 1920, ﬁnally, he was appointed to the newly created Chair of Physical Chemistry at Cambridge University. He continued at Cambridge for the rest of his life. In his extensive obituary notice, W. J. Pope wrote [100]: “[The book] on “Optical Rotatory Power” was issued in 1935 and will long remain a standard work on the subject. The immense amount of accurate experimental work which Lowry has left on record secures him a permanent place in the history of the science to which he was devoted.” But, although chiroptical methods are central to this overview, Lowry’s important contributions to other areas should not be forgotten. Mention may be made of his studies of the polymorphism of inorganic salts, his “Studies of Valency” and of the nitrogen oxide/water system, the nature of the sulfur halides, and the stereochemistry of tellurium compounds. In all cases he tried to apply modern physical concepts to chemical problems. Lowry was a member of the Faraday Society from its beginnings in 1903 and acted as its president in 1928–1930. He became a Fellow of the Royal Society in 1914. The physicist Georges Bruhat (1887–1945) was born in Besanc¸on, France, the ´ son of a civil servant. He was admitted with honors into the renowned Ecole Normale Sup´erieure (ENS) in Paris in 1906. After having obtained his B.Sc. (“licence e` s sciences physiques”) at the University of Paris, he acquired the qualiﬁcation to teach at secondary schools (as “Professeur agr´eg´e ”) and taught at a high school in Paris for one year. Perhaps this interval in his scientiﬁc career awakened his interest in teaching and in writing textbooks for this purpose. He then got a position as laboratory assistant (“pr´eparateur”) at the ENS, which enabled him to work under the guidance of A. Cotton on his doctoral dissertation: “La dispersion anormale du pouvoir rotatoire mol´eculaire” (the anomalous dispersion of the molecular rotatory power). After an interruption by his military service during the First World War, he could continue his academic career in 1919 in Lille, France, where he was Professor of General Physics from 1921 to 1927. His successor was Marcel Pauthenier, by the way, who had been his partner in the construction of UV spectropolarimeters [79]. Bruhat returned to the University of Paris in 1929 as Lecturer and Professor Extraordinary. In 1938 he was promoted to the Chair of Theoretical and Celestial Physics. During this time he published four compendious textbooks on general physics: Electricity (1924), Thermodynamics (1926), Optics (1930), and Mechanics (1934). These books have become standard texts in French universities, with many editions; Optics, for example, has been reedited as recently as 2004. Bruhat also continued his association with the ENS, serving as “Sub-Director” from 1935, and as acting director during World War II. In the beginning of August 1944, he was arrested by the political police (“Gestapo”) of the German occupation powers and held prisoner in lieu of a student accused of activities in the French Resistance Movement. He was taken to Germany into a concentration camp and died there on January 1, 1945, of pneumonia and exhaustion. The chemist Israel Lifschitz (1888–1951) was born in Shklov, Russia (now Belarus) (see Figure 1.4). His family was German and lived in Leipzig, Germany. His mother had moved to Shklov just for her conﬁnement, motivated by family regards. He studied chemistry at the University of Leipzig and worked there for his doctoral degree under the guidance of A. Hantzsch. His dissertation of 1911 dealt with the spectroscopic properties of various organic nitrogen compounds. Faithful to the ideas already developed in this dissertation, his main interest became the correlation of chemical constitution and bonding with electronic absorption. Although by training he was an organic chemist, his research

27

28

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 1.4. Israel Lifschitz. (Photograph provided by Ms. E. H. Lifschitz, Haifa, and reproduced with her permission.)

led him far into physical chemistry and, later, also inorganic chemistry. Leaving Leipzig in 1911, he moved to the University of Zurich, Switzerland, for his habilitation. To this end, he submitted a second dissertation on the changes of light absorption by organic acids on salt formation. As a result, the title of “Private Docent of Chemistry” was conferred on him in 1914, allowing him to do independent research, but not connected with any paid position. The life of a Private Docent was difﬁcult, unless one was privately afﬂuent. The economic troubles after World War I also forced Lifschitz to interrupt his work at the university in 1920, in order to look for a source of income outside of Academia, to support his growing family. Lifschitz extended his investigations to include the optical rotatory dispersion of transition metal coordination compounds (see, e.g., paper V of a series on the function of chromophores [101]); after 1920 this became his central area of research. From these papers it can be deduced that he had established a cooperative arrangement with the Dutch inorganic stereochemist F. M. Jaeger at the Rijksuniversiteit Groningen (RUG) in the Netherlands, by 1919 at the latest. Jaeger invited Lifschitz to join his laboratory and offered him a tenured staff position as “Conservator”. This induced Lifschitz to move to Groningen, although he had become a Swiss citizen and probably had intended to stay in that country. As a result, in the summer of 1921, Lifschitz became Private Docent

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

of Electrochemistry and Photochemistry, as well as Conservator at the Laboratory of Inorganic and Physical Chemistry at RUG. He helped to develop the department into a center of chiroptical spectroscopy. RUG became the birthplace of the “Cotton effect,” since Lifschitz coined this technical term there. The relevance of his research was readily acknowledged by stereochemists and physical chemists at other institutions. Already in 1935, J.-P. Mathieu in Paris had pointed out that Lifschitz was the ﬁrst to systematically investigate the relations between the experimental data on the Cotton effect and the chemical bonds in optically active compounds [83]. Lifschitz held his position at RUG until he was dismissed in November 1940 for political reasons. Under the German occupation of the Netherlands in World War II, the situation of Jewish people deteriorated continually. Since Lifschitz was unable to continue his research at RUG, he turned again to his private studies of the mystical movements in Judaism, of which he was a dedicated and competent scholar. He was a deeply religious person, who had also published on related subjects. Now he became absorbed again in his studies of Chassidism, and he gave even private lessons on the Zohar. All this ended, when he, his wife, and their ﬁve children were detained in February 1943. He was eventually deported to the Theresienstadt Concentration Camp in Bohemia in September 1944, separated from his three sons, two of whom did not survive the ordeal. Liberated at the end of the war, he, his wife, and the three children left returned to Groningen, where he was reinstalled at RUG. He died in the Netherlands in 1953 and was buried in Haifa, Israel. The dire story of the family’s fate has been reported by his elder daughter Esther Hadassa Lifschitz [102]. It had proved to be fatal that the Dutch authorities had insisted, shortly before the war, that the family accept Dutch citizenship while relinquishing their Swiss citizenship. An attempt to regain Swiss nationality during World War II failed. It is likely that they could have otherwise left for Switzerland. The present paper may perhaps draw some attention to Lifschitz, who has been undeservedly forgotten. The physical chemist Werner Kuhn (1899–1963) was born at Maur am Greifensee, Switzerland, the son of a pastor. He studied chemistry at the ETH Z¨urich and obtained his doctorate at the University of Zurich in 1923, with a dissertation on a photochemical topic. After working with Niels Bohr in Copenhagen, Denmark, for two years, he returned to Zurich for his habilitation in physical chemistry in 1927. He moved to Germany thereafter. For three years he worked at the University of Heidelberg, where he started his research on optical activity; in 1930–1936 he worked at the Technische Hochschule Karlsruhe, and in 1936 he occupied the Chair of Physical Chemistry at the University of Kiel. In the year 1939 he returned to his home country, when he was called to the Chair of Physical Chemistry at the University of Basel. Later he was also elected rector of the University. He remained there until the end of his life. A report on “Leben und Werk von Werner Kuhn” (life and work of W. K.), including a bibliography, appeared in 1984 in connection with a “Werner Kuhn Symposium” of the Swiss Chemical Society [103]. The physicist Aim´e Auguste Cotton (1869–1951) was born in the French provincial town of Bourg-en-Bresse, where his father taught mathematics (see Figure 1.5). He was ´ a student of physics at the Ecole Normale Sup´erieure (ENS) in Paris from 1890 to 1893. He completed his doctoral studies, in the course of which he discovered the “Cotton effect” at the Laboratory of Physics with M. Brillouin and J. Violle and earned the title “Docteur e` s-sciences physiques” in 1896. It is noteworthy that he included in his thesis his ﬁrst attempts at measuring magnetic optical activity, since his future scientiﬁc career was to be centered on the physics of magnetism and magneto-optics. It would be in these ﬁelds, rather than in research on natural optical activity, that he would rise to eminence.

29

30

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 1.5. Aime´ Auguste Cotton. (From L. de Broglie, Notice sur la vie et l’œvre de Aime´ ´ Cotton, Academie des sciences—Institut de France, Paris, 1953; reproduced with permission.)

After a few years at the Faculty of Sciences of the University of Toulouse as Assistant/Associate Professor (Maˆıtre de Conf´erences, Professeur adjoint), he returned to the ENS in Paris, entrusted with the substitution of the Academician J. Violle, his teacher, for the period of 1900–1904. From 1904 to 1922 he served as Lecturer (Charg´e de cours) at the Faculty of Sciences of the University of Paris, delegated to the ENS. He was promoted to Professeur-adjoint in 1910 and became Professor of Theoretical and Celestial Physics at the Sorbonne in 1920. Finally, he was called to the Chair of General Physics there in 1922, succeeding G. Lippmann (Nobel Prize 1908). In 1923 he was elected a member of the illustrious Academy of Sciences as successor to J. Violle. He even became President of the Academy in 1938. Cotton retired from the Sorbonne in 1941, but continued until his death as Director of the Laboratory for Magneto-Optical Studies that he had founded in 1927. In France, A. Cotton is considered to be one of the eminent physicists of the twentieth century. He himself described his scientiﬁc aims and accomplishments in a 1923 pamphlet on the occasion of his election to the Academy [104]. As he pointed out, he had worked extensively on the Zeeman effect, but even more on the magneto-optical properties of colloids and molecular solutions. Many of the latter investigations were performed in

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

cooperation with the biologist Henri Mouton (1869–1935) of the Institute Pasteur (later Professor of Physical Chemistry at the University of Paris). Together they discovered in 1907 the important “Cotton-Mouton-Effect” (magnetic ﬁeld induced linear birefringence) [105]. Cotton’s fame in physics rests, arguably, more on this discovery than on the “Cotton effect” in natural optical activity. Detailed information on his life and work is easily available thanks to various obituaries [106] and to a notice published on the occasion of the centenary of his birth [107]. Cotton received many honors and was awarded many prizes. He was nominated for the Physics Nobel Prize in the years 1915, 1916, 1920, 1922, 1925, 1927, 1928, 1929, 1930, 1931, 1932, 1933, 1934, 1944, and 1949, but without success [108]. His memory is kept alive in France with the “Laboratoire Aim´e Cotton,” as his former Laboratory for Magneto-Optical Studies has been renamed. It is now the Atomic and Molecular Physics Laboratory of CNRS, associated with the University Paris XI and situated on its campus in Orsay. Also the “Prix Aim´e Cotton” should be mentioned, which was established by the French Physical Society in 1953 in memory of A. Cotton and is awarded annually. In what esteem he is held by the French physicists is also apparent by the fact that he is one of the twelve “most eminent” physicists chosen by the Academy on the occasion of the World Year of Physics 2005 (WYP2005/UNESCO). Among chemists of countries other than France, and especially among the younger fold, the name “Cotton” rarely brings to their minds memories of a speciﬁc person, unfortunately, unless they mistake it for the name of the distinguished American inorganic chemist A. Albert Cotton (1930–2007) of textbook fame. However, whether they know anything about the person or not, for the chemists working in chiroptics or stereochemistry, T. M. Lowry’s words of 1935 still hold true [21]: Cotton’s discovery in absorbing optically-active media of the twin phenomena of circular dichroism and of anomalous rotatory dispersion, which are indissolubly associated with his name, is [. . .] amongst the “classics.”

ACKNOWLEDGMENTS It is a pleasure to acknowledge the kind cooperation of Ms. E. H. Lifschitz in Haifa, Israel, who has provided important information on her father, Israel Lifschitz, and has permitted the publication of his photograph. It is a privilege to acknowledge also the untiring help of Dr. Henry Joshua of New York City, who has overcome many difﬁculties in his endeavors to establish contacts with Ms. Lifschitz, whose whereabouts had been unknown. The author is also grateful to Professor Jerome Gurst of Pensacola, Florida, for language counseling and for his helpful critique of the manuscript.

REFERENCES 1. D. F. Arago, M´em. Cl. Sci. Math´em. Phys. Inst. France 1811, 12 , I, 1–16, 113–134 (published 1812); [J. B.] Biot, Phys. Inst. France 1812, 13 , I, 1–372 (May 1813). 2. J. B. Biot, Bull. Sci. Soc. Philomath. 1815, [3] 2 , 190–192. 3. J. B. Biot, M´em. Acad. Roy. Sci. Inst. France 1817, [2] 2 , 41–136 (September 1818). 4. A. Arndtsen, Ann. Chim. Phys. 1858, [3] 54 , 403–421; Pogg. Ann. 1858, [2] 105 , 312–317.

31

32

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

5. H. Landolt, Liebigs Ann. Chem. 1877, 189 , 241–337 (p. 274). 6. J. B. Biot, Ann. Chim. Phys. 1860, [3] 59 , 206–326; appendix 326–345 (page 266/7, Section III, 8). In this appendix, by the way, Biot refutes claims that the Baltic-German physicist Thomas Johann Seebeck (1710–1831) might have been the ﬁrst to observe optical activity in solutions. 7. H. Landolt, Das optische Drehungsverm¨ogen organischer Substanzen und die praktischen Anwendungen desselben, Vieweg und Sohn, Braunschweig, 1879; English translation: Handbook of the Polariscope, D. C. Robb, V. H. Veley, translators, Cambridge, 1882 (reprint: BiblioLife, Charleston, SC, 2009). 2nd ed. Vieweg und Sohn, Braunschweig, 1898; English translation: The Optical Rotating Power of Organic Substances, J. H. Long, translator, Chemical Publishing Co., Easton, PA, 1902. 8. H. Fehling, Arch. Physiol. Heilkunde 1848, 7 , 64–73; Ann. Chem. Pharm. 1849, 72 , 106–113. 9. F. (-)P. Leroux, C. R. H. Acad. Sci . 1862, 55 , 126–128; Pogg. Ann. 1862, [2] 117 , 659–660. 10. C. Christiansen, Pogg. Ann. 1870, [2] 141 , 479–480; Pogg. Ann. 1871, [2] 143 , 250–259. A. Kundt, Pogg. Ann. 1870, [2] 142 , 163–171; Pogg. Ann. 1871, [2] 143 , 149–152; Pogg. Ann. 1871, [2] 143 , 259–269. J.-L. Soret, Arch. Sci. Phys. Natur. 1871, [2] 40 , 280–283. 11. A. Cotton, Absorption in´egale des rayons circulaires droit et gauche dans certain corps actifs, C. R. H. Acad. Sci . 1895, 120 , 989–991. 12. A. Cotton, Dispersion rotatoire anomale des corps absorbants, C. R. H. Acad. Sci . 1895, 120 , 1044–1046. 13. A. Cotton, Recherches sur l’absorption et la dispersion de la lumi`ere par les milieux dou´es du pouvoir rotatoire, Th`ese de Doctorat, Paris, 1896; Ann. Chim. Phys. 1896, [7] 8 , 347–432; summary: J. Phys. Th´eor. Appl . 1896, [3] 5 , 237–244. 14. [J. B.] Biot, Bull. Sci. Soc. Philomath. 1815, [3] 2 , 26–27; W. Haidinger, Pogg. Ann. 1845, [2] 65 , 1–30. 15. W. Haidinger, Pogg. Ann. 1847, [2] 70 , 531–544. 16. It would be interesting to know if Fehling’s solution was indeed Cotton’s very ﬁrst sample. Apparently his papers, including his laboratory notebooks 1895–1920, are at the Niels Bohr Library, American Center for Physics, College Park, MD. 17. L. Natanson, Bull. Int. Acad. Sci. Cracovie, Cl. Sci. Math. Nat., Jan. 1909, 25–37. 18. A. Cotton, J. Phys. Th´eor. Appl . 1898, [3] 7 , 81–85. 19. F. Billet, Trait´e d’optique physique, Mallet-Bachelier, Paris, 1858/59 (2 vols.). 20. N. Wedeneewa, Ann. Phys. 1923, [4] 72 , 122–140. 21. T. M. Lowry, Optical Rotatory Power, Longmans, Green and Co., London, 1935; reprint: Dover Publications, New York, 1964. 22. S. Mitchell, The Cotton Effect and Related Phenomena, G. Bell & Sons, London, 1933. 23. W. Kuhn, A. Szabo, Z. Phys. Chem. 1931, B15 , 59–73. 24. W. O. [i.e., W. Ostwald], Z. Phys. Chem. 1896, 19 , 383, no. 87 and 88. 25. W. O. [i.e., W. Ostwald], Z. Phys. Chem. 1896, 21 , 158–163. 26. P. Drude, Lehrbuch der Optik , Hirzel, Leipzig, 1900; English translation: The Theory of Optics, C. R. Mann, R. A. Millikan [Nobel Prize 1923], translators, Longmans, Green & Co., London, 1902; reprint: Dover Publications, New York, 1959. 27. L. Natanson, Bull. Int. Acad. Sci. Cracovie, Cl. Sci. Math. Nat., O eliptycznej polaryzacyi s´wiatła, przepuszczonego przez ciało naturalnie , Oct. 1908, 764–783. 28. L. Natanson, J. Phys. Th´eor. Appl . 1909, 8 , 321–347. 29. A. Cotton, C. R. H. Acad. Sci . 1911, 153 , 245–247.

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

30. W. Kuhn, Proc. [2 nd] Conf. Rotatory Dispersion, Santa Fe, CA, 20–22 Jan., 1960; Tetrah. 1961, 13 , 1–12. 31. M. F. McDowell, Phys. Rev . 1905, 20 , 163–171. 32. L. B. Olmstead, Phys. Rev . 1912, 35 , 31–46. 33. H. Grossmann, A. Loeb, Z. Ver. Deutsch. Zuckerind., Allg. Teil 1908, 58 , 994–1009. 34. H. Volk, Ber. Dtsch. Chem. Ges. 1912, 45 , 3744–3748. 35. G. Bruhat, Th`ese de Doctorat, Paris, June 1914; Ann. Physique 1915, [9] 3 , 232–282, 417–489. 36. G. Bruhat, Ann. Physique 1920, [9] 13 , 25–48. 37. R. de Mallemann, P. Gabiano, C. R. H. Acad. Sci . 1927, 185 , 350–352. 38. J.-P. Mathieu, C. R. H. Acad. Sci . 1931, 193 , 1079–1081; Ann. Physique 1935, [11] 3 , 371–460. 39. J.-P. Mathieu, C. R. H. Acad. Sci . 1932, 194 , 1727–1729. 40. W. Pﬂeiderer, Z. Phys. 1926, 39 , 663–685. 41. R. D. Gillard, Progr. Inorg. Chem. 1966, 7 , 215–276. Oddly, this author maintains that “Aim´e Cotton (1872–1944) discovered the [Cotton] effect while Professor of Physics at the Sorbonne in Paris.” Both the dates (recte: 1869–1951) and the professorship are erroneous (Cotton became a Professor at the Sorbonne in 1910, but had discovered the effect in 1895). 42. A. Werner, Z. Anorg. Allg. Chem. 1893, 3 , 267–330; see also: Vierteljahresschrift der Z¨urch. Naturf. Ges. 1891, 36 , 129–169. 43. A. Werner, Ber. Dtsch. Chem. Ges. 1911, 44 , 1887–1898. 44. M. Del´epine, C. R. H. Acad. Sci . 1914, 159 , 239–241; Bull. Soc. Chim. Fr. 1917, [4] 21 , 157–172. 45. G. Bruhat, Bull. Soc. Chim. Fr. 1915, [4] 17 , 223–227. 46. A. Werner, Ber. Dtsch. Chem. Ges. 1912, 45 , 121–130; 3061–3070. 47. A. Werner, J. Poupardin, Ber. Dtsch. Chem. Ges. 1914, 47 , 1954–1960. 48. A. Werner, Ber. Dtsch. Chem. Ges. 1914, 47 , 3087–3094. 49. A. Werner, Helv. Chim. Acta 1918, 1 , 5–32. 50. Important reviews of stereochemistry and optical activity at Groningen: F. M. Jaeger, Proc. Akad. Wet. Amsterdam 1915, 17 , 1217–1236; Lectures on the Principle of Symmetry, Elsevier, Amsterdam, 1917 (2nd enlarged edition 1920); Bull. Soc. Chim. Fr. 1923, [4] 33 , 853–889; Spatial Arrangements of Atomic Systems and Optical Activity, McGraw-Hill, New York, 1930. 51. F. M. Jaeger, Rec. Trav. Chim. Pays-Bas 1919, 38 , 171–314. 52. J. Lifschitz, Z. Phys. Chem. 1923, 105 , 27–54. It may be pointed out that the author abbreviated his ﬁrst name, Israel, in print usually by the letter “J”; this was not uncommon in those days. 53. J. Lifschitz, Z. Phys. Chem. 1925, 114 , 485–499. 54. F. M. Jaeger, H. B. Blumendal, Z. Anorg. Allg. Chem. 1928, 175 , 161–230. 55. W. Kuhn, K. Bein, Z. Anorg. Allg. Chem. 1933/34, 216 , 321–348 (ORD); Z. Phys. Chem. 1934, B24 , 335–369 (CD). 56. R. E. Ballard, A. J. McCaffery, S. F. Mason, Proc. Chem. Soc. London 1962, 331–332. 57. A. J. McCaffery, S. F. Mason, Proc. Chem. Soc. London 1962, 388–389. 58. H. Rupe, Zeltner, W. Lotz, M. Silberberg, Liebigs Ann. Chem. 1903, 327 , 157–200 (ﬁrst paper of the series); H. Rupe, L. Silberstrom, Liebigs Ann. Chem. 1918, 414 , 99–111 (ninth paper). 59. T. M. Lowry, J. Chem. Soc. London 1899, 75 , 211–244.

33

34

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

60. H. Rupe, A. Krethlow, K. Langbein, Liebigs Ann. Chem. 1921, 423 , 324–342 (thirteenth paper of the series [58]). 61. T. M. Lowry, ed., Trans. Faraday Soc. 1914, 10 , 44–138. 62. So transliterated by T. M. Lowry in the obituary note (J. Chem. Soc. London 1923, 123 , himself, who often published in Germany, usually spells his name 956–958). . A. there Tschugaeff, and sometimes Tch´ugaeff or Tschugajew, in France also Tchougaeff. Chemical Abstracts always list him as L[ev] A[leksandrovich] Chugaev. In the references to this overview, the original spelling as given in the papers has been chosen. 63. L. Tschugaeff, Ber. Dtsch. Chem. Ges. 1909, 42 , 2244–2247 (paper I of the series); L. Tschugaeff, A. Ogorodnikoff, Z. Phys. Chem. 1910, 74 , 503–512 (II); L. Tschugaeff, Z. Phys. Chem. 1911, 76 , 469–483 (III); L. Tschugaeff, A. Ogorodnikoff, Z. Phys. Chem. 1912, 79 , 471–480 (IV); L. Tschugaeff, A. Ogorodnikoff, Z. Phys. Chem. 1913, 85 , 481–510 (V); L. Tschugaeff, W. Pastanogoff, Z. Phys. Chem. 1913, 85 , 553–572 (VI). See also: L. Tschugaeff, A. Ogorodnikoff, Ann. Chim. Phys. 1911, [8] 22 , 137–144; L. Tschugaeff, J. Prakt. Chem. 1912, [2] 86 , 545–550; L. Tchougaeff, Bull. Soc. Chim. Fr. 1912, [4] 11 , 718–722; L. Tschugaeff, G. Glinin, Ber. Dtsch. Chem. Ges. 1912, 45 , 2759–2764; L. Tchugaeff, A. Kirpitcheff, Bull. Soc. Chim. Fr. 1913, [4] 13 , 796–803; L. Tschugaeff, , A. A. , , B. [L. Trans. Faraday Soc. 1914, 10 , 70–79; . A. 1915, A. Tschugaev, A. A. Glebko, G. V. Pigulevskii], J. Russ. Phys. Chem. Soc. 47 , 774–775. 64. G. Bruhat, C. R. H. Acad. Sci . 1911, 153 , 248–250. 65. T. M. Lowry, H. Hudson, Philos. Trans. 1933, A232 , 117–154. 66. T. M. Lowry, H. K. Gore, Proc. Roy. Soc. 1932, A135 , 13–22. 67. E. Deussen, J. Prakt. Chem. 1912, [2] 85 , 484–488. 68. L. Tschugaeff, J. Prakt. Chem. 1912, [2] 86 , 545–550. 69. S. Mitchell, J. Chem. Soc. London 1928, 3258–3260. 70. E. Darmois, Ann. Chim. Phys. 1911, [8] 22 , 247–281, 485–590; Th`ese de Doctorat, Paris, 1910. 71. W. Kuhn, H. K. Gore, Z. Phys. Chem. 1931, B12 , 389–397. 72. T. M. Lowry, H. S. French, J. Chem. Soc. London 1932, 2654–2658. 73. R. Servant, C. R. H. Acad. Sci . 1932, 194 , 368–369. 74. C. Djerassi, Optical Rotatory Dispersion. Applications to Organic Chemistry, McGraw-Hill, New York, 1960. ´ 75. G. Bruhat, Trait´e de Polarim´etrie [with a preface by A. Cotton], Editions de la Revue d’Optique th´eorique et instrumentale, Paris, 1930. 76. Review: W. Kuhn, Theorie und Grundgesetze der optischen Aktivit¨at [theory and fundamental principles of optical activity], in Stereochemie, K. Freudenberg, ed., Deuticke, Leipzig, 1933, pp. 317–434. 77. R. Descamps, Trans. Faraday Soc. 1930, 26 , 357–371. 78. T. M. Lowry, Proc. Roy. Soc. 1908, A81 , 472–474; Philos. Trans. 1912, A212 , 261–297; T. M. Lowry, W. H. C. Coode-Adams, Philos. Trans. 1927, A226 , 391–466. 79. A. Cotton, R. Descamps, C. R. H. Acad. Sci . 1926, 182 , 22–26; R. Descamps, Rev. Opt. Th´eor. Instr. 1926, 5 , 481–501; G. Bruhat, M. Pauthenier, C. R. H. Acad. Sci . 1926, 182 , 888–890 (with comments by A. Cotton, C. R. H. Acad. Sci . 1926, 182 , 890–891); G. Bruhat, M. Pauthenier, Rev. Opt. Th´eor. Instr. 1927, 6 , 163–184. 80. G. Bruhat, P. Chatelain, C. R. H. Acad. Sci . 1932, 195 , 462–465; J. Physique 1932, [7] 3 , 501–511; G. Bruhat, A. Guinier, C. R. H. Acad. Sci . 1933, 196 , 762–764; Rev. Opt. Th´eor. Instr. 1933, 12 , 396–416. 81. G. Bruhat, Bull. Soc. Chim. Fr. 1930, [4] 47 , 251–261.

T H E F I R S T D E C A D E S A F T E R T H E D I S C O V E RY

82. W. Kuhn, Ber. Dtsch. Chem. Ges. 1929, 62 , 1727–1731; W. Kuhn, E. Braun, Z. Phys. Chem. 1930, B8 , 445–454. 83. J.-P. Mathieu, Ann. Physique 1935, [11] 3 , 371–460. 84. H. Rudolph, J. Opt. Soc. Am. 1955, 45 , 50–59; Rudolph Photoelectric Spectropolarimeter®, O. C. Rudolph & Sons, Caldwell, NJ. 85. M. Grosjean, M. Legrand, C. R. H. Acad. Sci . 1960, 251 , 2150–2152 (Centre de Recherche Roussel-Uclaf); Dichrograph®, Societ´e Jouan, Paris. 86. M. Born, Phys. Z . 1915, 16 , 251–258; Ann. Phys. 1918, [4] 55 , 177–240. 87. C. W. Oseen, Ann. Phys. 1915, [4] 48 , 1–56. 88. F. Gray, Phys. Rev . 1916, [2] 7 , 472–488. 89. L. Rosenfeld, Z. Phys. 1928/29, 52 , 161–174; Rosenfeld was at that time a postdoctoral associate with M. Born in G¨ottingen, Germany. 90. W. Kuhn, Z. Phys. Chem. 1929, B4 , 14–36. 91. W. Kuhn, Z. Phys. Chem. 1935/36, B31 , 23–57. 92. J.-P. Mathieu, Les Th´eories Mol´eculaires du Pouvoir Rotatoire Naturel , Gauthier-Villars, Paris, 1946. 93. L. Tschugaeff, Z. Phys. Chem. 1911, 76 , 469–483. 94. L. Tschugaeff, A. Ogorodnikoff, Ann. Chim. Phys. 1911, [8] 22 , 137–144. , , Publishing House of the Academy of Sciences of the SSSR, 95. . A. Moscow, 1955. 96. L. Velluz, M. Legrand, M. Grosjean, C. R. H. Acad. Sci . 1963, 256 , 1878–1881. do ﬁzyki teoretycznej , Redakcye “Prac Matematyczno-Fizycznych”, War97. L. Natanson, saw, 1890. 98. L. Klecki, Prace Matematyczno-Fizyczne 1939, 46 , 1–18 (in French). 99. J. Salkind, Ber. Dtsch. Chem. Ges. 1922, 55 , 141A–142A. 100. W. J. Pope: Thomas Martin Lowry, Obituary Notices of Fellows of the Royal Society 1938, 2 , 287–293. 101. J. Lifschitz, E. Rosenbohm, Z. Wiss. Photogr., Photophys. Photochem. 1920, 19 , 198–214. 102. E. H. Lifschitz, Er was weinig begrip voor de joden [there was not much sympathy with the Jews], in Terug van weggeweest, J. van Gelder, ed., Stichting Geldersboek, Groningen, 1993, pp. 141–148 (Chapter 15); also: private communication to the author. 103. H. Kuhn, Chimia 1984, 38 , 191–211. 104. A. Cotton, Notice sur les Travaux Scientiﬁques, Presses Universitaires de France, Paris, 1923. 105. A. Cotton, H. Mouton, Ann. Chim. Phys. 1907, [8] 11 , 145–203, 289–339. 106. M. Javillier, C. R. H. Acad. Sci . 1951, 232 , 1521–1527; J. Cabannes [Cotton’s successor at the Sorbonne], Ann. Physique 1951, [12] 6 , 895–898; Louis [Prince] de Broglie [Nobel Prize 1929], Notice sur la vie et l’œvre de Aim´e Cotton, Institut de France, Acad´emie des Sciences, Paris, 1953 [30 pages]. 107. A. Kastler, C. R. H. Acad. Sci . 1969, 269 , 70–74. 108. E. Crawford, The Nobel Population 1901–1950: A Census of the Nominators and Nominees for the Prizes in Physics and Chemistry, Universal Academy Press, Tokyo, 2002. (Nomination 1915, page 62 (C. Fabry, nominator). Similarly: 1916, p.64 (M. Brillouin); 1920, p.80 (J. Bordet); 1922, p.84 (J. Bordet); 1925, p.94 (J. Bordet); 1927, p.102 (J. Bordet, C.E. Guillaume, C.E. Guye, A.Schidlof); 1928, p.106 (C.E. Guillaume); 1929 , p.110 (C.E. Guillaume); 1930, p.116 (C.E. Guillaume, H. Villat); 1931, p.120 (C.E. Guillaume, R. de Mallemann); 1932, p.124 (C.E. Guillaume, G. Reboul); 1933, p.128 (C.E. Guillaume); 1934, p.132 (H. Buisson, C.E. Guillaume, C.E. Guye, J. Perrin, V. Posejpal, P. S`eve); 1944, p.170 (H. Beghin); 1949, p.186 (M. Pauthenier)).

35

PART II ORGANIC STEREOCHEMISTRY

2 SOME INHERENTLY CHIRAL CHROMOPHORES—EMPIRICAL RULES AND QUANTUM CHEMICAL CALCULATIONS Marcin Kwit, Pawel Skowronek, Jacek Gawronski, Jadwiga Frelek, Magdalena Woznica, and Aleksandra Butkiewicz

Molecules that do not contain center of chirality can still be chiral if their structure does not contain a plane of symmetry. Typical examples are hexahelicene and biphenyl (Figure 2.1) in which the chromophoric system (in these cases π electron system) is helical due to steric reasons preventing enantiomerization of the molecule at ambient temperature. Examples discussed in this chapter include (a) conjugated chromophoric systems of dienes, (b) enones, (c) helical chalcogenides, and (d) nonplanar amide chromophores. Many molecules of natural origin contain chromophores of either of this type (e.g. steroids, terpenes, amino acids, antibiotics, and metabolites); hence interpretation of their ECD may be useful for their absolute structure elucidation.

2.1. DIENES AND TRANS-ENONES Dienes and enones form a class of chromophores that can be helical due to nonplanarity of the π -bond system. Their chiroptical properties have been a subject of numerous studies over the past 50 years [1], aimed at clarifying the role of twist of the π -bond system and chirality of the molecule on the ECD spectra and optical rotation. There are other types of helical chromophoric systems; among these are dichalogenides (disulﬁdes, diselenides, ditellurides), which are distinctly different due to their non-π -electron chromophore, and these will be discussed in detail in the following section (Figure 2.2). The conformations of these molecules may be conveniently described by considering their helicities. 1,3-Dienes can exist in two planar conformations deﬁned as s-trans and s-cis and an inﬁnite number of nonplanar, skewed forms, traditionally called cisoid or Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

39

40

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 2.1. Hexahelicene and biphenyl—examples (M)-Hexahelicene

of inherently chiral (helical) chromophores.

(M)-Biphenyl

(a)

(b) C C

C

C

C

C

O

w=C=C−C=C

C

Ch Ch

C

C

C

C

w=O=C−C=C

R

C

w = C − Ch − Ch − C (Ch = S, Se, Te)

t=C−C=C−R

Figure 2.2. Simple helical chromophores and deﬁnitions of torsion angles ω (a) and deﬁnition of angle τ that describes distortion of the C=C bond from planarity (b).

transoid , depending on the conformation of the nearer planar form. The conformation can otherwise be described by torsion angle ω, which can be either positive (P ) or negative (M ) (see Scheme 2.1). The same is true for enones and dichalcogenides; however, in the last case both planar s-trans and s-cis conformations represent the transition states, not any stable structures. Additional classiﬁcation of chromophore structures may be accomplished on the basis of molecular symmetry [1b]. Both butadiene and symmetrically disubstituted dichalcogenides in planar forms belong to either C2h or C2v symmetry point group, respectively, for s-trans and s-cis conformers. Rotation around the C–C or Ch–Ch bonds reduces the

(a) X

X S-cis (sp)

X Cisoid

Planar

P-helical

Cisoid

X Transoid

X S-trans (ap)

X Transoid

(b)

M-helical

P-helical

Planar

M-helical

X = CH2 (1,3-dienes) X = O (enones)

Scheme 2.1. (a) Possible planar and nonplanar conformations of 1,3-dienes and enones and (b) deﬁnition of M- and P-helicities in the cases of cisoid and transoid 1,3-dienes.

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

symmetry to C2 in skewed, chiral conformations. Unlike butadiene and dichalcogenides, the simplest enone (acrolein) exhibits only trivial C1 symmetry when skewed and Cs symmetry for both planar forms. The electronic transitions of 1,3-dienes and enones are reﬂections of their structures. The degree of electron delocalization determines intensity and energy of the low-energy π –π * electronic transition in the diene chromophore. For the most populated s-trans conformation of butadiene, the most intense band is found at 210 nm with εmax > 20,000 [2], whereas for s-cis planar conformation, both the intensity and energy of the π –π * electronic transition is signiﬁcantly lower. In the case of cyclopentadiene the long-wavelength absorption band appears at 240 nm (εmax = 3500) [2]. The extinction coefﬁcients and energies of the low-energy π –π * electronic transitions for skewed dienes are higher for transoid compared to cisoid structures and reach a minimum (λmax = 190 nm) for perpendicular diene confomation (ω = ±90◦ ) [2, 3]. In the case of 1,3-cyclohexadiene the observed UV maximum appears at 256 nm (εmax = 9000) [4]. The lowest-energy π –π* electronic transition has been used as a diagnostic band for solving stereochemical problems with the use electronic circular dichroism (ECD) spectroscopy. Contrary to dienes, α, β-unsaturated ketones typically exhibit two absorption bands: one very weak near 330 nm (R-band) and the second at around 230 nm (K-band). The lowest-energy band originates from a forbidden transition from nonbonding 2py orbital into antybonding π * orbital; thus this electronic transition is deﬁned as n –π * [5, 6]. In the case of 2-cyclohexenone the lowest-energy n –π * transition is found at 320 nm (ε = 36) [7]. For chiral enones the dissymmetry factor (g) for the lowest-energy n –π * electronic transition is in the range 10−1 to 10−2 . The second UV absorption band in enone chromophore has the character of an intramolecular charge transfer transition from the vinyl to the carbonyl group and exhibits much higher intensity compared to the lowest-energy n –π* electronic transition. Oscillator strengths for transoid enones are usually higher than for cisoid . The dissymmetry factor for π –π * electronic transition is in the range 10−3 to 10−4 . In the case of 2-cyclohexenone, this band appears at 225 nm (εmax = 13800) [7]. Although only two bands are observed in the UV spectra of enones, ECD spectra of α, β-unsaturated ketones exhibit three (sometimes four) Cotton effects between 360 and 185 nm. The weak UV absorption at about 300 nm is responsible for the long-wavelength Cotton effect of moderate intensity. This band is followed by a π –π * transition Cotton effect corresponding to the UV band placed between 220 and 250 nm. Another Cotton effect that does not have the corresponding UV maximum appears at around 200 nm and usually exhibits a high rotatory strength, sometimes even higher than that for the π –π * electronic transition. The origin of this Cotton effect is not clear. Earlier calculations by Liljefors and Allinger suggested that this is another π –π * transition of low intensity in nearly planar enones [8]. Snatzke suggested that this is the second forbidden n –π * electronic transition [1i, 9]. The direction of polarization of this transition has been determined by linear dichroism measurements [10]. The fourth electronic transition is observed between 195 and 185 nm in ECD spectra of some enones having an axial substituent in α or β positions. This transition was considered as n –σ * excitation [1d–1f]. During the last few decades, various empirical rules have been proposed to correlate the signs of the Cotton effects of 1,3-dienes and α, β-unsaturated ketones with their stereochemistry (Table 2.1). Whereas some of the correlation rules are of historical value, two basic stereochemical concepts underlying the development of such correlations are brieﬂy considered below.

41

42

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

TAB L E 2.1. Empirical Rules Correlating Cotton Effect for a Given Electronic Transition with the Chirality of Compounds Containing 1,3-Diene and Enone Chromophores Electronic Transition

Rule

π –π * (dienes) π –π * (dienes) π –π * (dienes) π –π * (dienes) π –π * (planar s-trans dienes) π –π * (enones) n –π * (enones) n –π * and π –π * (enones) n –π * (planar s-trans enones) n –π * and π –π * (enones) 210-nm transition (cyclic enones)

References

Diene Helicity Rule (Moscowitz et al., 1961) Allylic Axial Chirality Rule (Burgstahler and Barkhurst, 1970; Burgstahler et al., 1976) Quadrant Rule (Moriarty et al., 1979) Sector Rule (Weigang, 1979) Planar Diene Rule (Duraisamy and Walborsky, 1983)

11 12, 13

Enone Helicity Rule (Djerasi et al., 1962, Whalley, 1962) Enone Helicity Rule (Snatzke, 1965) Orbital Enone Helicity Rule (Kirk, 1986)

17, 18 19 1e

Sector Rule (Snatzke, 1979; Snatzke, 1965)

1h, 19a

Allylic Axial Chirality Rule (enones) (Burgstahler and Barkhurst, 1970) Carbon–Carbon Bond Chain in Cyclic Enones Rule (Gawronski, 1982)

14 15 16

12 1d

For correlating the chiroptical phenomena with the diene or enone structure, two fundamental effects had to be taken into account: (a) the contribution from the helicity of the chromophore and (b) the effect of extrachromophoric perturbation. Whereas the ﬁrst of these effects correlates conformation of the chromophore with its spectroscopic properties, the second is related to the conﬁguration at the stereogenic center(s). Both effects may act in the same or in opposite direction, so determining the dominant contribution is an important yet not always easy task. In the case of rigid 1,7,7-trimethyl2,3-dimethylene-bicyclo[2.2.1]heptane (1), chirality originates from the presence of the methyl group connected to C1 carbon atom [20]. As long as the chromophore is planar, the effect of extrachromophoric perturbation dominates. On the other hand, in the case of highly ﬂexible (3S,8S,E,E )-dimethyl-deca-4,6-diene (2), both effects contribute to the π –π * transition Cotton effect [21], whereas in the case of the rigid structure of ergosterol (3) a negative twist of diene moiety is considered as the origin of the long-wavelength Cotton effect [22]. C9H17

H HO 1

2

De = −0.6 (250 nm)

De = +3.0 (230 nm)

3 w = −11° De = −18.0 (269 nm)

According to the Diene Helicity Rule (DHR) [11a], the sign of the long-wavelength Cotton effect reﬂects directly the sense of helicity of the chromophore; that is, for positive twist of the 1,3-diene system a positive long-wavelength π –π * transition Cotton effect is expected (Figure 2.3a).

43

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

(a)

(b)

Figure 2.3. Empirical correlating P +C.e. w > 0°

M −C.e. w < 0°

+C.e. for right-handed angle Caxial −Callyl −C = C

−C.e. for left-handed angle Caxial − Callyl −C = C

rules for 1,3-dienes: (a) Diene Helicity Rule and (b) Allylic Axial Chirality Rule.

The validity of DHR has been challenged by calculations at different levels of sophistication [3, 23–31]. Most of them were performed for simple models like 1,3butadiene and generally conﬁrmed the validity of DHR, but it should be noted that some of the calculations gave opposite results [28, 31]. To overcome the inadequacy of the Diene Helicity Rule for certain types of diene structures, the concept of allylic axial substituent contributions to optical activity of conjugated dienes was proposed by Burgstahler and co-workers (Allylic Axial Chirality Rule, hereafter referred to as AACR) [12, 13]. According to this concept, the sign of the long-wavelength Cotton effect of conjugated dienes is primarily due to the contribution of allylic axial substituents, such as alkyl groups. The sign of the contribution is determined by the helicity (+ or −) of the Caxial –Callyl –C=C bond system (Figure 2.3b). Thus, antipodal CD curves of 15-methylene-5α-cholest-8(14)-en-3β-ol acetate (4) and 3-isopropylidene-A-norcholest-5-ene (5) were explained as due to contribution from the allylic axial substituents; these dienes exhibit Cotton effects of opposite signs to those required by the DHR. + AcO

H + + – H H H w = +20° 4 De = +4 (250 nm)

H – H

–

w = +20°

– H

5 De = −5 (250 nm)

A corollary to this rule is a low contribution of the diene chromophore to the rotational strength of the π –π * transition. Certain substituents attached to one of the diene carbon atoms (e.g., the CN group) can cause sign reversal of the long-wavelength Cotton effect, in the absence of any obvious structural change [32], and this is another example of inconsistency with the Diene Helicity Rule. An experimental and theoretical study of α-phellandrene (6) and other 5-alkyl-1,3cyclohexadienes (7, 8) by Lightner et al. [30] for the ﬁrst time provided a dissection of the contributions of various structural elements to the cyclohexadiene 260 nm Cotton effect. In the case of (5R)-5-methyl-1,3-diene (7), variable-temperature ECD spectra were invariant, due to a low energy difference between equatorial and axial conformers, estimated as 0.05 kcal mol−1 (Figure 2.4a,b), and exhibited a positive long-wavelength Cotton effect ( ε = +5 at 260 nm). A dramatic change of ECD spectra with temperature was observed when a tert-butyl group was attached to the 1,3-cyclohexadiene skeleton. At slightly elevated temperature, the sign of the long-wavelength Cotton effect was positive ( ε = +2.5 at +31◦ C) and stepwise decreased upon lowering the temperature. At −180◦ C the value of the long-wavelength Cotton effect was ε = −3.0, which corresponds to energy difference Gax-eq = 0.4 kcal mol−1 .

44

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The rotatory strengths estimated from ECD spectra for 8 are positive for axial P helical conformer (R = +55.6 × 10−40 cgs units) and negative for equatorial M -helical conformer (R = −15.0 × 10−40 cgs units). The sign of the contribution due to the s-cisdiene moiety alone is calculated opposite to that predicted by the Diene Helicity Rule for 5-alkyl-1,3-cyclohexadienes, and the sign and magnitude of the Cotton effect are apparently dominated by the contributions of the axial allylic bonds (groups) (Figure 2.4c). Recently, a more advanced ab initio calculation by Hansen and Bak [33] in the random phase approximation (RPA) using Aug-cc-pVTZ atomic basis set provided important conﬁrmation of the earlier ﬁndings on the role of allylic substituents. The effects of the allylic methyl groups were found to follow a quadrant rule being almost additive, and the contributions from axial substituents were calculated signiﬁcantly larger than those from the equatorial groups. Analysis of chiroptical properties of compounds having enone chromophore is usually much more laborious. The ﬁrst difﬁculty is due to small energy differences between conformers, and the second is the origin of the electronic transitions. While the n –π * electronic transition has origin similar to that of saturated ketones, higher energy transitions in enones are not pure and involve various types of orbitals. Estimated inversion barrier for 2-cyclohexenone is lower than the inversion barrier experimentally determined for cyclohexene [34]; and full inversion cycle involves a number of structures, each characterized by out-of-plane position of at least one saturated carbon atom. The diversity of possible conformation does not vanish even if a 2-cyclohexenone ring is a part of polycyclic systems, as in steroids and terpenoids. For example, for testosterone (9) molecule in the crystal lattice the dihedral angle ω that characterizes chromophore helicity takes different values, both positive and negative,

(a)

(c)

(b)

Figure 2.4.

Conformational drawings of (R)-α-phellandrene (6) (a) and (5R)-5-alkyl-1,3-

cyclohexadienes 7 and 8 (b) showing the axial or equatorial position of the alkyl group in relation to diene helicity and estimated group contributions to the rotatory strength (R (×10−40 cgs units) for the lowest-energy π −π ∗ electronic transition of (P)-1,3-cyclohexadiene (c).

45

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

depending on the space group and on the presence or not of other molecules forming hydrogen bonds or complexes [35–38].

Almost 50 years ago, Djerassi et al. [17], Whalley [18], and Snatzke [19] proposed rules for the ﬁrst two Cotton effects: positive enone helicity was correlated with a negative n –π * and a positive π –π * Cotton effect. This rule was thereafter applied by Kirk to s-cis and s-trans enones unperturbed by polar substituents, and (P ) π orbital helicity at C1 and C2 carbon atoms was correlated with a positive n –π * and a negative π –π * Cotton effect (Figure 2.5a,b) [1e]. These rules worked well for cyclohexenones, but for cyclopentenones inverse rules were proposed [1d, 1f, 19b]. Snatzke proposed a modiﬁcation of the octant rule that correlates the sign of the n –π * Cotton effects with the stereostructure of planar enones [1h,1i,19c]. The n –π * Cotton effects of steroidal enones in oriented (anisotropic) systems were later studied by Kuball et al. [39]. Their studies demonstrated that sector or helicity rules can be applied, provided that vibronic progressions originating from various conformers are taken into account. On the other hand, Burgstahler and Barkhurst [12] has shown for the ﬁrst time the importance of allylic axial substituents that led to a breakdown of the enone helicity rule for the π –π * transition of some s-cis steroid enones. The electronic transition observed between 220 and 200 nm in ECD spectra of some enones [9, 40] is characterized by a low oscillator strength, making it difﬁcult to detect in (a)

(b)

O

O

s-cis (P)

O

s-trans (P)

s-cis (P)

−C.e. (n−p*), +C.e. (p−p*) (c)

O

O

s-trans (M )

+C.e. (n−p*), −C.e. (p−p*) (d)

O

R b′ O

O

+C.e. (~220 nm)

−C.e. (~220 nm)

a′ R

b′

b

a′

a O

Figure 2.5. (a) The ﬁrst Enone Helicity Rule. (b) Kirk’s Enone Orbital Helicity Rule. (c) correlation rule for the third Cotton effect. (d) Correlation rule for the sign of the fourth Cotton effect in 2-cyclohexenones.

46

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

the UV spectra, and a high rotatory strength, resulting in some cases in overlap with the Cotton effect originating from the π –π * electronic transition. Attempts to correlate the sign of the third Cotton effect with structural parameters were met with many difﬁculties. Burgstahler correlated the sign of the third Cotton effect with the chirality of α carbon atom [12]. In the case of 4 -3-ketosteroids, a positive value of the third Cotton effect is in agreement with the chirality due to an axial 2β-substituent. This relationship can only be treated as tentative due to numerous exceptions. Gawronski suggested that the sign of the third Cotton effect is connected with the presence of saturated carbon–carbon bond chain, especially in polycyclic enones. It shows a clear relation to the absolute conﬁguration at the allylic position (Figure 2.5c) [1d, 1f]. The fourth electronic transition, thought to involve n and σ ∗ orbitals, has been found in enones having α or β substituents in an axial position. The sign of this Cotton effect is dominated by the orientation of substituent, with assumption of a sofa(5) conformation of 2-cyclohexenone skeleton (Figure 2.5d).[1d–1f]. In the light of the above-mentioned facts, proper determination of the nature of electronic transition(s) involved is mandatory for rational stereochemical analysis of compounds containing the enone chromophore. This is of special interest in the case of molecules containing polar groups, which may inﬂuence strongly both the sign and the magnitude of the Cotton effects [41]. On the basis of DFT calculation at the B2LYP/6-311 + +(2d,2p) level and with the use of NBO method [42], it was possible to determine the origin of the ﬁrst three electronic transition for a group of model compounds. For s-trans-acrolein, 2-cyclohexenone, (4S )-4-hydroxy-2-cyclohexenone (10), and (5R)-5-hydroxy-2-cyclohexenone (11), characterized by planar conformation of the chromophore (ω = 180◦ , τ = 0◦ ) and sofa(5) conformation of the cyclohexenone skeleton, the origin of the ﬁrst two electronic transitions did not raise doubts (Figure 2.6). The long-wavelength electronic transition with a small oscillator strength involves HOMO(−1)–LUMO orbitals (nC=O –πC=C * type), whereas the second electronic transition with the highest oscillator strength is of CT character, involving HOMO and LUMO orbitals (πC=C –πC=O * type). The third electronic transitions, responsible for the third strong Cotton effect observed in the ECD spectra of enones, is characterized by a small oscillator strength. For acrolein the third electronic transition involves both πC=O –πC=O * and nC=O –σ * transitions, while in the case of 2-cyclohexenone it is a mixture of electronic transitions of the πC=C –πC=C * type (HOMO–LUMO(+1), the main contribution) and πC=C –σ * [HOMO–LUMO(+2)]. This supports the earlier suggestion [1d] that the conﬁguration of saturated C–C bond chain in cyclic enones may be, in the absence of polar substituents, the controlling factor of the sign of the short-wavelength Cotton effect. The presence of a hydroxy group in enones 10 and 11 causes a change of the character of the third electronic transition, compared to 2-cyclohexenone. This electronic transition appearing at ∼190 nm involves the lone pairs of the hydroxy group [HOMO(−2)] and LUMO orbitals and thus may be referred to as nOH –πC=O * type. This may suggest that the third, short-wavelength Cotton effect in 2-cyclohexenones having polar substituents at C4 and/or C5 position depends on the helicity of the (H)O–C · · · C=O bond system [41]. In the analysis of the origin of optical activity of cyclic 1,3-dienes and enones the effect of nonplanarity of the C=C bond(s) is usually neglected. Recently, Diedrich and Grimme [43] performed an advanced quantum chemical calculation of the rotatory strength for the electronic transitions of twisted (−10 deg) ethylene. The calculated rotatory strength for the π –π * transition was of the order 75–198 × 10−40 cgs units, depending on the method used. Since nonplanar ethylene generated a high rotatory

47

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

(a)

(b)

O 10

IIIb

OH 11

IIId

IIIc

LUMO(+1)

LUMO

(d)

O

O

O

LUMO(+2)

OH

(c)

I II IIIa

I II

I II IIIe

I II IIIe

HOMO

HOMO(−1)

HOMO(−2)

Figure 2.6. Molecular orbitals obtained with the use of the NBO method for representative planar enones: (a) Acrolein and (b) I nC=O –πC=O ∗, II πC=C –πC=O *, IIIa πC=O –πC=O *, IIIb nC=O –σC(O) – H *, IIIc πC=C –πC=C ∗, IIId πC=C –σC – C ∗ /σC – H *, IIIe nOH –πC=O *.

strength, neglecting such an effect in the case of π –π * electronic transitions of nonplanar enone and diene seems unjustiﬁed. Nonplanar conformations of simple 1,3-butadiene and acrolein may be chiral due to nonplanarity of the C=C bonds, deﬁned as the sign and value of torsion angle τ (Figure 2.2). Thus, nonplanar molecular conformations may be due to nonzero values of either angle ω or angle(s) τ , or both. If both torsion angles ω and τ are considered, chromophore conformation may be deﬁned as homohelical (if the signs of angles ω and τ are the same) or as heterohelical , if the signs of angles ω and τ are opposite [44]. In the case of s-trans acrolein and 2-cyclohexenone, deformation of the C=C bond from planarity results in the appearance of nonzero rotatory strength. Enone Helicity Rule is obeyed as long as the twist of the C=C bond and enone helicity are of opposite sense (heterohelical [41]). We will now discuss the effect of substitution of 1,3-cyclohexadiene and 2-cyclohexenone, which sometimes leads to contrasting effects. (S , S )-trans-1,2-Dihydroxy-3,5-cyclohexadiene (12) in the crystal phase forms an M -helical conformer (ω = −11.5◦ ), with the hydroxy groups occupying equatorial positions. Variable-temperature ECD spectra measured in methanol reveal that the rotatory strengths for both P- and M-helical conformers are positive, in accordance with the helicity of O–Callyl –C=C system, and the rotatory strength for diaxial P -conformer is one order of magnitude higher than that for the diequatorial one (Scheme 2.2) [45]. Signiﬁcant solvent effect can be observed in the ECD spectra of 12. ECD spectrum in methanol solution at room temperature shows a Cotton effect ( ε = +11.2 at 258 nm),

48

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

w = −11.5° (X-ray)

OH

H

OH O

OH

H

12

O

methanol

H

cyclohexane

H

H

Δep−p* +6 (cyclohexane) +11 (methanol)

H M (more abundant) R p−p* > O

OH P R p−p* >> O

Scheme 2.2. Conformational equilibrium of dihydroxydiene 12.

due to increased population of P -helical conformer, whereas in nonpolar solvent the presence of intramolecular OH· · ·O bond favors the diequatorial M-conformer, with concomitant decrease of the long-wavelength Cotton effect ( ε = +6 at 259 nm). According to recent calculation performed for 12 at the DFT/B3LYP/6-311++G(d,p) level, diaxial P conformer is of 0.7 kcal mol−1 higher energy relative to the diequatorial one. Rotatory strength of the π –π * electronic transition, calculated at the mPW1PW91/6311++G(2d,p) level for P -helical 12 is almost four times higher than that for the lowest-energy diequatorial M conformer [46], consistent with the experimental ECD data. More complex systems are represented by arene metabolites 13–17 [41, 46–48]. These compounds, being valuable chiral building blocks and ligands in organic synthesis [49], are characterized by the presence of a vicinal cis-diol system and a 1,3-diene or enone chromophore. All of them can exist in solution in an equilibrium of diastereoisomeric diene (enone) conformers of P or M helicity with one of the OH group in an axial and the other in an equatorial position.

The main problem with stereochemical analyses of compounds of this type is reliable determination of conformer population. In general, the number of available conformers is not limited to M and P diastereoisomeric structures. Intramolecular hydrogen bond patterns of the hydroxy groups further increase the number of distinct stereoisomeric structures (Scheme 2.3). Thus, either C1(or C4)–OH or C2(or C5)–OH can be a hydrogen bond donor, and the orientation of the O–H bond against the vicinal C–H bond

49

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

4

R

3

H

R

H O R2

O

3

R4

1

R

R1 2

H

R

4

R

2 1

R

R

4

H

3 R O

H

2

R H

H

R

1

R

4

3 R O

R2

R1

O

H

O

H

H P3

H P1

H M3

H

H a OH

3

R

O H

O

H H M1

R w

R

H

enones, R3

b OH

==O

4

R

R

3

H

O H R2

O

1

R H

H M2

4

R

3

H

R

O

2

R

R H

O H H M4

1

4

R

H 3 O R

R2 H

H H P2

O

1

R

R

4

3 R O

H

H

R2 O

1

R

HH P4

Scheme 2.3. Diastereomeric Pand M conformers of arene metabolites 13–17 and the deﬁnition of torsion angles α, β, and ω.

can be either syn or anti . This makes the number of available conformers up to eight (M 1–M 4, P 1–P 4), and the number can still be higher if one includes the rotamers due to the presence of nonspherical substituents in the ring. Contrary to 1,3-cyclohexadiene derivatives, in the case of cis-ketodiols 15–17 the conformational equilibrium is strongly affected by solvent polarity. Calculations performed at MP2/Aug-cc-pVTZ//B3LYP/6-311++G(2d,2p) level and with the use of a PCM model led to a conclusion that in polar surroundings a conformer of type P 4 dominates, in contrast to 1,3-cyclohexadiene derivatives, where conformer P 4 does not participate in conformational equilibrium even to a small extent [46–48]. It should be noted that X-ray diffraction analysis may provide quite different results, since in the crystal lattice the most important structure-determining factor is the possibility of formation of intermolecular hydrogen bonds. Thus, in the case of ﬂuorinated ent-14e the molecular structure found in the crystal corresponds to conformer M 4. In contrast to M 1, M 2, and P 1 are calculated at the B3LYP/6-311++G(d,p) level as the lowest-energy

Figure 2.7. Potential energy surface calculated at the B3LYP/6-311++G(d,p) level for P- and M-helical conformers of ﬂuorinated derivative 14e (a) and X-ray diffraction determined structure of ent-14e (b).

50

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a)

(b)

13b 13d

2 OH OH

Dioxygenase TDO-O2

OH

Dioxygenase TDO-O2

OH

0 Δe −2 13c

−4 minor

major

−6

13e 220

13a 13b 13c 13d 13e

13a 240

260

280

300

320 nm

Figure 2.8. (a) Empirical rule correlating the stereochemical course of arene cis-dihydroxylation with the relative size of substituents. (b) ECD spectra of cis-diols 13a-13e measured in cyclohexane solution.

conformers (Figure 2.7) for isolated molecule 14e [47]. The calculated preferred conformations of dihydrodiols substituted at the positions C3 or C6 by triﬂuoromethyl group are strongly affected by the presence of sequence of hydrogen bonds Oeq H · · · Oax H · · · FCF2 , which may shift the conformer equilibrium into such a conformer. The CD spectra measured in nonpolar solvent for a number of 3-substituted cisdihydrodiols 13a–13e differ markedly in the sign of the diagnostic long-wavelength spectral region (Figure 2.8b). This suggests that either the absolute conﬁguration assigned according to Figure 2.8a is wrong, or the absolute conﬁguration at C1 carbon atom is the same for the whole series and the cis-dihydrodiols 13a–13e differ in chiroptical properties [46]. Extensive computational study on the structure and chiroptical properties of arene metabolites led to a conclusion that the presence of substituents X and/or Y, as well as the hydroxy groups, is the decisive factor in shifting the P ↔ M conformer equilibrium in one or another direction, and this determines the sign and magnitude of the longwavelength diene Cotton effect. The substituent effect on the ECD spectra of cis-diols 13–14 may be ordered as follows: CN > Br > CH3 > CF3 > F = H. This study indicates that no simple empirical model, including the DHR and the AACR, can fully account for the experimental CD data of all cis-dihydrodiols. In relation to the Diene Helicity Rule, rotatory strength contribution from helical cyclohexadiene chromophore is in any case weak and, in certain cases (Br, CN substituents at C3), does not correlate with the sign of diene torsion angle. The failure of the AACR results from mutually canceling contributions due to allylic hydroxy groups, both axial and equatorial, in conformers of P and M helicity. Deceptively simple cis-ketodiols 15–17 are in fact complex systems, due to their conformational equilibria. Circular dichroism spectra measured in acetonitrile solutions and calculated at the PCM/B2LYP/Aug-cc-pVTZ level are quite similar, regardless the enone structure and in the case of cis-ketodiols 15 exhibit a positive/negative/positive sequence of Cotton effects. However, in the case of 16 and 17 the sign of the longwavelength nC=O –πC=O * Cotton effect appears affected by the substituent in C6 position, but the patterns of the second πC=C –πC=O * and the third nOH –πC=O * Cotton effects are similar to that measured and calculated for 15. Surprisingly, detailed inspection of the ECD spectra calculated for individual conformers of 15a shows the same pattern, but not the magnitudes, of Cotton effects [41]. Rotatory strengths calculated for M -helical conformers remain in agreement with DHR and also with the alternative AACR, which

51

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

assigns a dominant contribution to the axial hydroxy group at the allylic position. On the other hand, ECD spectra calculated for the P conformers, which dominate in the conformational equilibrium, are not in agreement with the empirical rules and show much higher magnitudes than those calculated for their M -helical counterparts. In the case of P -helical conformers a dominant contribution is assigned, rather unexpectedly, to the equatorial hydroxy group [41]. With the use of modern computational approach, Kwit et al. estimated the effect of structural factors on the rotatory strengths of electronic transitions by replacing systematically the hydrogen atoms at C3–C6 in (P )-2-cyclohexenone with either a polar hydroxy group or a nonpolar methyl group. As expected, the dominant role of substituent at C4 over substituents in either C5 or C6 positions is clearly visible, regardless their electronic nature and orientation. For the πC=C –πC=O * transition of monosubstituted (P )2-cyclohexenones the substituent contributions, including signs, are as follows: 4eq-OH (−), 4ax-OH (+), 4ax-Me (+) > 4eq-Me (−), 5ax-OH (−), 6eq-OH (+) > 5ax-Me (+), 6ax-Me (−) > 5eq-OH (+), 5eq-Me (−), 6eq-Me (+), 6ax-OH (+). Another striking result revealed for the ﬁrst time is that the sign of the principal πC=C –πC=O * transition Cotton effect is less dependent on the enone nonplanarity (angle ω) and more dependent on nonplanarity of the C=C bond [41, 46b]. The models of ECD contributions for arene metabolites of mono-substituted 1,3cyclohexadiene and cis-ketodiol type are proposed shown in Figure 2.9. These models redeﬁne the importance of structural factors previously considered as responsible for chiroptical properties of dienes and enones. Nonplanarity of the C=C bonds, usually neglected, is of equal or even higher importance compared to the distortion of the whole conjugated chromophore from planarity.

2.2. DICHALCOGENIDES: MOLECULES WITH INHERENTLY CHIRAL CHROMOPHORES Dichalcogenides are unique examples of molecules with an inherently chiral chromophore. The dichalcogenide Ch–Ch moiety (Ch = S, Se, or Te) exists in the form of two helical enantiomeric P /M conformations. Both enantiomeric conformers give opposite Cotton effects that cancel in the experiments. When the dichalcogen moiety is placed in a chiral surrounding, the conformers are no longer of equal energy and this (a)

(b)

Figure 2.9. Estimated bond contributions to the rotatory strength R (×10−40 cgs units) of the lowest-energy π − π * electronic transition of arene metabolites of (a) the type (P)-1,3cyclohexadiene derivative (a) and (b) the type (P)-2-cyclohexenone derivative.

52

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Scheme 2.4. Full interconversion scheme for dichalcogenide R–Ch–Ch–R conformers.

gives rise to the Cotton effects in the CD spectra. The conformer equilibrium involves transition states (TS) of either s-cis or s-trans structure in which the Ch–Ch moiety is planar (Scheme 2.4). C–S–S–C torsion angle (ω) for dimethyl disulﬁde in the gas phase was reported in the review as 90.44◦ (±0.45◦ ) [50]. Values 85◦ , 82◦ , and 87◦ were found for dimethyl diselenide based on the microwave [51] and vibrational spectra [52] as well as on the electron diffraction data [53]. The data from low temperature X-ray diffraction of dimethyl dichalcogenides are shown in Table 2.2. For benzene solution the values of torsion angle ω calculated from the dipole moment measurements (68.5 ± 2.0◦ for Me2 S2 , 65.4 ± 4.0◦ for Me2 Se2 , 35.9 ± 10.0◦ for Me2 Te2 ) [55] are much smaller than those found in the gas phase. This difference was explained by interactions between the solvent and the solute molecules. Data found for the diphenyl dichalcogenides in the crystal state and for benzene solution are more coherent— that is, 96.2◦ vs. 81.3◦ for Ph2 S2 , 97.1◦ vs. 62.6◦ for Ph2 Se2 , 88.5◦ vs. 84.0◦ for Ph2 Te2 [55]. Skewed structure of dichalcogenides is explained by the gauche effect, which was introduced for disulﬁdes [50] but which can be extended to diselenides and ditellurides. In the dichalcogenide molecule the nonbonding electron pairs that reside on the p orbitals perpendicular to the Ch–Ch bond are partially overlapping. The destabilization due to the lone-pair–lone-pair repulsion in orthogonal position is reduced in nonplanar cisoid and transoid conformations of the dichalcogenide molecule. Hyperconjugation by which the p lone pairs of the chalcogen atoms are overlapping with the σ * molecular orbitals of the R–Ch bonds located in the same plane provides additional stabilizing effect [56]. Rotation barrier for the interconversion between enantiomeric conformers of dichalcogenides is low enough to allow free rotation of the Ch–Ch bond in acyclic molecules at room temperature (Scheme 2.4). Due to the repulsion between substituents on the chalcogen atoms, s-trans transition state is of lower energy. For S–S bond the rotation barrier is estimated in the range 6.8–13.2 kcal mol−1 [57–61]. Calculations (including MP2 method) give the value ca. 5 kcal mol−1 through the s-trans transition TAB L E 2.2. X-ray Data for Dimethyl Dichalcogenides Me2 Ch2

˚ Ch–Ch [A]

˚ C–Ch (A)

ω (degrees)

Me2 S2 Me2 Se2 Me2 Te2

2.03 2.31 2.71

1.80 1.94 2.15

86 85 90

Source: Data taken from reference 54.

53

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

state and 9 kcal mol−1 through the s-cis transition state. [50, 62, 63]. Calculated rotation barrier for H2 Se2 is lower (4.47 and 6.21 kcal mol−1 , respectively). Substitution with the phenyl group does not change signiﬁcantly the magnitude of the rotation barrier which was calculated for diphenyl disulﬁde, diselenide, and ditelluride as 8.6, 8.2, and 5.3 kcal mol−1 through a s-cis transition state and 5.4, 5.2, and 3.7 kcal mol−1 through a s-trans transition state [64]. It is seen that the rotation barrier is lowered with the increase of the length of the Ch–Ch bond (see Table 2.2). A model for interpretation of the CD spectra of disulﬁde chromophore was proposed by Bergson [65, 66] and later veriﬁed by Linderberg and Michl [67] and by Woody [68]. Woody has shown the dependence of the rotation strength on the torsion angle of disulﬁde chromophore. In this model the sign of the long-wavelength Cotton effect obeys a quadrant rule (Figure 2.10). Bergson’s model due to its simplicity has been very successful. Recent study of the disulﬁde transitions using advanced computational methods led to similar conclusions [69–71]. As molecular structures of dichalcogenides are similar, this rule can be easily extended to diselenides and ditellurides [70–73]. It has to be noted that the longwavelength Cotton effect is of differing nature at different values of the torsion angle. The two lowest-energy electronic transitions are described as nA –σ * and nB –σ *. In the range 0◦ –90◦ , nA is a HOMO whereas nB is HOMO(−1). The highest energy difference between nA and nB orbitals is calculated for values of the dichalcogenide torsion angle. This difference diminishes at around 90◦ where nA and nB orbitals are almost degenerate. For the torsion angle greater than 90◦ , nB becomes HOMO and this leads to a change of the rotational strength. Similar switch of orbital positions takes place at −90◦ (Figure 2.11) [71, 72]. The ﬁrst attempt to experimental study of chiroptical properties of dichalcogenide chromophore was made by Djerassi et al. [74], who recorded the ORD spectra for cystine and selenocystine. They observed negative Cotton effects at 250 and 290 nm, respectively. Ringdahl et al. [75] have found a long-wavelength Cotton effect at around 320 nm in the ECD spectra of selenocystine Other examples of CD spectra are provided by 2,2-dithioand 2,2-diseleno derivatives of propionic acid [76, 77]. It was expected that the CD spectroscopy based on the disulﬁde chromophore will ﬁnd application in the analysis of protein structure. However disulﬁde CD bands in proteins are difﬁcult to identify since the average number of disulﬁde bonds in proteins is low and the CD band originating from

0° −C.e.

+C.e.

180°

90°

+C.e.

Δe 0

−C.e. −90° −180

−90

0

90

180

Torsion angle R-S-S-R

Figure 2.10. Relation between the torsion angle in R–S–S–R disulﬁdes and observed sign of the long-wavelength Cotton effect.

54

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

l [nm] 480

Rotational strength (R)

20

460 10

440 420

0

400

−10

360 340

−20

Figure 2.11. Rotatory strength (R ×10−40 cgs units,

320

solid line) and wavelength (dashed line) of the long-wavelength transition for dimethyldiselenide. (Data from reference 72.)

380

0

30

60

90

120

150

Dihedral angle C-Se-Se-C

180

the disulﬁde chromophore is often overlapped by the CD bands originating from aromatic residues, as well as from the amide chromophore. Laur [78] presented the CD spectra of rigid cyclic (9S , 10S )-trans-2,3-dithiadecalin (18a) and (9S ,10S )-trans-2,3-diselenadecalin (18b) and open-chain disulﬁde (19a), diselenide (19b), and ditelluride (19c) substituted by (S )-2-methylbutyl groups (Figure 2.12). For rigid trans-decaline derivatives (18a, 18b) in a chair–chair conformation, C–Ch–Ch–C bond forms a left-handed (M ) helix. The long-wavelength Cotton effects for 18a and 18b are negative, in direct correlation with the twist direction of the helix. Due to distortion of the dichalcogenide chromophore from the optimum 90◦ value to ∼60◦ , the position of the UV absorption bands and consequently of the Cotton effects is red-shifted by around 40 nm. The torsion angle of the open-chain derivatives 19 is not conﬁned to just one value, so it is expected that these compounds exist as a mixture of freely interconverting diastereoisomers with the torsion angle C–Ch–Ch–C around 90◦ or −90◦ . A small energy difference between distereoismers due to steric repulsion between the alkyl chains leads to a small difference in their population and to the rise of small, but nonzero, Cotton effects. Indeed, the observed Cotton effects are about 30–50 times lower than those observed for trans-decalin 18.

0.4

4

0.2

2

H Ch Ch H 18 a, Ch = S b, Ch = Se

Ch Ch

19 a, Ch = S b, Ch = Se c, Ch = Te

0 Δe

Δe 0.0 −0.2

18b 19b

−0.4 250

300

350

400

450

−2 −4

500 l[nm]

Figure 2.12. ECD spectra of the diselenadecalin (18b) (right-hand scale) and diselenide (19b) (left-hand scale) in acetonitrile. (Redrawn using the data from reference 78.)

55

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

Further examples are the CD spectra of sugar-substituted diselenides and ditellurides [79]. The authors characterize the observed bands but do not consider the effect of chiral substituents on the dichalcogenide conformation. Recently, the CD spectra of diglycosyldisulﬁdes and diselenides were reported, and a good agreement between the experimental and the calculated CD spectra of di(teraacetylated glycosyl)diselenide using the TDDFT method was obtained [71]. The CD spectra of several symmetrical diselenides with chiral alkyl substituents allowed us to determine the effect of chirality of the alkyl groups on the chirality of the diselenide moiety with the aid of TDDFT calculation. Thus, the S conﬁguration of the carbon atom adjacent to the Se atom correlates with positive long-wavelength Cotton effect. This effect was explained by a model based on steric repulsion between bulky substituents [72]. The CD spectra measured for the dichalcogenides substituted with the same chiral alkyl groups show a shift of the position of the long-wavelength n –σ * Cotton effects toward a longer wavelength and a signiﬁcant reduction of the Cotton effect intensity on going from S to Te. This trend is observed in the CD spectra of di[(S )2-methylbutyl] derivatives of dichalcogenides [78] and also for dineomenthyl derivatives 20 (Figure 2.13). The decrease of the Cotton effect intensity in this series is a result of elongation of the chalcogen–chalcogen and chalcogen–carbon bonds, as shown in Table 2.2, which causes signiﬁcant reduction of steric repulsion between the substituents and hence lowering of free energy difference between the diastereomeric conformers of dichalcogenides of M and P helicity. Contemporary procedure for the analysis of ECD spectra of chiral dichalcogenides requires computation of conformer distribution and then the average CD spectra [71, 72, 80]. However, from a practical point of view the information about helicity of the dichalcogenide moiety can be obtained directly from the sign of the Cotton effect due to the long-wavelength nA –σ * transition. Higher-energy nB –σ * transition usually overlaps with other transitions and may be not distinguishable in the observed CD spectra. Orbitals nA and σ * involved in the long-wavelength transition are localized on the dichalcogenide moiety and are not affected by alkyl substituents. A further simpliﬁcation of the analysis of rotational strength of complex chiral dichalcogenides can be based on the data obtained for dimethyl dichalcogenide. Calculation of the whole system seems justiﬁed only for structures with the C–Ch–Ch–C torsion angle very close to 90◦ (or −90◦ ) where nA and nB are almost degenerate and their exact energy may depend on the substituents in the molecule.

1.6 Ch Ch

1.2 Δe

20 a, Ch = S b, Ch = Se c, Ch = Te

20a 20b 20c

0.8 0.4 0.0 250

300

350

400

450

500

550 l[nm]

Figure 2.13. ECD spectra of the dineomenthyl disulﬁde 20a, diselenide 20b, and ditelluride 20c in cyclohexane solution.

56

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

2.3. THE AMIDE CHROMOPHORE IN BI- AND POLICYCLIC β-LACTAM RING SYSTEMS The amide chromophore has been recognized as a basic unit in a variety of bioactive compounds. Among them the β-lactam family of antibiotics represents some of the most clinically relevant antibiotics known [81–84]. This is attributable to their broad spectrum of antibacterial activity and a relatively low level of toxicity [83, 85]. To connect the biological activity with a deﬁned stereochemistry of β-lactams, many research groups were engaged in the analysis of their chiroptical properties [86–99]. In the case of monocyclic β-lactams the planarity of the amide chromophore was demonstrated [90, 96, 99, 100]. The assumption of planarity of this chromophore has allowed to explain many features of protein structure. The (3R, 4S )-3-phenoxy-4-vinylazetidin-2-one presented below is an example of monocyclic β-lactam with a planar β-lactam moiety (Figure 2.14). The O=C2–N1–C4 torsion angle equal to 177.28◦ , derived from its X-ray structure, clearly demonstrates the planarity of amide chromophoric system. According to Moscowitz [101–103], such a planar amide chromophore has to be categorized as inherently achiral but chirally perturbed by its neighborhood. Therefore, one has to take into consideration the chiral perturbation (mainly through space), and this is usually achieved by the help of different sector rules. However, based on both the X-ray and computational studies, it has been demonstrated that in small cyclic peptides and medium-sized lactams the amide chromophore can be slightly nonplanar [94, 97, 104–108]. The skewness of the amide unit causes its inherent chirality, and thus such a nonplanar chromophore now belongs to the Moscowitz inherently chiral class of chromophores [101–103]. These chromophores are characterized by very strong CD effects mainly governed by their chirality. It means that the contributions from all the other atoms and groups may be neglected, so that the rules correlating the stereostructure with CD data can be classiﬁed as “chirality rules” or “helicity rules.” As a result of the extensive studies of lactam chromophores, several sector and helicity rules for the correlation between the structure and Cotton effect (CE) signs of n –π* transition have been established. Among them the β-lactam octant rule [90, 97, 109], Weigang’s sector rule [100], a modiﬁcation of Weigang’s lactam rule [99], and Ogura’s [91] and Wolf’s [96] helicity rules can be mentioned. These empirical rules correlate the sign of the CE designated as the n –π * transition of the β-lactam chromophore with the absolute conﬁguration of monocyclic azetidinones. The 4π electrons and two free electron pairs of the amide chromophore are located on the carbonyl oxygen atom. The molecular orbital (MO) occupied by the highest energy electron pair is largely (80–90%) located on the carbonyl oxygen and is a 2p orbital. Its axis is located in the plane of the amide group and is perpendicular to the direction of the

Figure 2.14. Crystal structure of (3R, 4S)-3-phenoxy-4-vinylazetidin-2-one. Thermal ellipsoids are shown at 50% probability level.

57

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

C=O bond. The second pair of free electrons, with a much lower energy, occupies orbitals extensively overlapping with d orbitals. This orbital is of both 2s and 2p character, and its axis is directed along the C=O bond (Figure 2.15). Four electronic transitions can theoretically occur within the amide chromophore, namely π0 –π * (around 190 nm), π+ –π * (around 140 nm), n –π * (around 220 nm), and n –π *. The last transition is predicted theoretically but has not been observed experimentally. Amide transition π0 –π * is electrically allowed (εmax ∼ 104 M−1 cm−1 ), and the direction of the electric transition moment μ approximately deﬁnes a line connecting the N and O atoms (Figure 2.15). The n –π* transition is electrically forbidden (εmax ∼ 102 M−1 cm−1 ); however, it has a large magnetic transition moment m directed along the line passing through the C=O bond (Figure 2.15). In penicillins and cephalosporins the β-lactam unit is nonplanar and its nitrogen atom is pyramidal. The relationship between an (R) AC at the ring junction carbon atom and the positive sign of the lowest energy CE attributed to the amide n –π * transition which occurs in penicillins and cephalosporins at around 230 nm and 260 nm, respectively, was well-documented [86, 89, 93, 104, 110–112]. There is a nontrivial question whether the same regularity is valid for the oxa- and carbaanalogues of penicillins and cephalosporins. To clarify this, a study was undertaken to establish a correlation of the absolute stereostructure of a variety of β-lactam derivatives with the sign of the amide n –π * transition in their CD spectra. In addition, to rationalize the experimental results and to ﬁnd out the scope and limitations of observed regularities, the calculation of chiroptical properties of β-lactam antibiotic analogues at B3LYP/TZVP level of theory were performed.

2.3.1. General Structures of Penicillins and Cephalosporins The chiroptical properties of oxacephams have been studied on variety of bi-, tri-, and tetracyclic oxacephams [113, 114]. Some of them, namely oxacephams 21–29, are presented in Chart 2.1 while the CD spectra of representative members of this group are shown in Figure 2.16. As can be seen in Figure 2.16, the investigated compounds exhibit, generally, two CD bands at around 220 and 190 nm. The 220-nm CD band can be assigned to the n –π * electronic transition of β-lactam unit, whereas the band at around 190 nm corresponds (a)

(b) π* O C μπ→π*

mn→π*

π*

π0

n→π* π0→π* n π0 n′

N

C N

n

C N

n′

π+ π+

Figure 2.15. (a) Molecular orbitals and electronic transitions within the amide chromophore. (b) Directions of the magnetic m and electric μ transition moments deﬁning the electronic transitions within the amide system.

58

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Chart 2.1.

(a)

(b)

Figure 2.16. (a) CD spectra of selected oxacephams 21 (——), 26 (· · ·), and 27 (– – –) recorded in acetonitrile. (b) Crystal structures of compounds 26 and 28 with the crystallographic numbering scheme. Thermal ellipsoids are shown at 50% probability level.

to the π –π * excitation of the same unit. With respect to the sign of band at 220 nm, the oxacephams fall into two different groups. In the ﬁrst group, consisting of compounds with (6R) absolute conﬁguration, the sign of this CD band is positive, whereas in the second group, represented by compounds with (6S ) absolute conﬁguration, this band is negative [113]. Based on that data, additionally corroborated by the speciﬁcs of their synthetic pathway and the X-ray analysis obtained for compounds 26 and 28 [113], it can be unambiguously established that the (6R) AC corresponds to a positive CE at around 220 nm whereas (6S ) AC corresponds to a negative sign of the same CE.[113, 114] The data indicate the nonplanarity of the amide chromophore and the pyramidal conﬁguration of the amide nitrogen in the studied oxacephams. Thus, the β-lactam chromophore is inherently chiral, and the sense of its chirality is expressed as a right or left helicity that correlates well with the AC at C6 carbon atom. The helicity appears to be independent of the kind and the position of other substituents present in the oxacepham moiety [113, 115, 116].

59

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

The nonplanarity of the chromophore excluded application of the known sector rules for the prediction of the n –π * CE sign, since these rules were developed for the planar amide chromophore only [90, 97, 100]. In such case, the helicity rule should, in principle, be able to correlate the chiroptical properties and structure. However, for the studied oxacephams a breakdown of Ogura’s [91] and Wolf’s [96] helicity rules was found [113]. The spiral rule [107, 108] based mostly on the CD results obtained for the nonplanar α-lactams and monocyclic β-lactams that correlates the positive/negative torsional angle O=C–N–C with a negative/positive n –π * CE, respectively, in general, was valid for oxacephams studied. However, it was found that the absolute conﬁguration of the bridgehead carbon atom determines the sign of the O=C–N–C torsional angle. Therefore, to connect directly the AC of this carbon atom with the sign of the CE due to n –π * transition observed around 220 nm in oxacephams, a simple helicity rule has been proposed [113]. According to this rule, a positive sign of the 220 nm CE corresponds to an (R) AC at the bridgehead carbon atom, whereas a negative sign of the same CE is related to a (6S ) AC. The rule was experimentally demonstrated to be correct for a variety of oxacephams [114–119]. The DFT conformational analysis of oxacepham 21 at B3LYP/TZVP level of theory indicates the skewness of the β-lactam unit by negative O9–C8–N1–C2 and O9–C8–N1–C6 torsion angles of −24.1◦ and −175.7◦ , respectively. The six-membered ring in the computed lowest energy conformer of oxacepham 21 is in a chair conformation with β-lactam ring in energetically favorable equatorial position. Furthermore, it has been found that the positive long-wavelength CE around 220 nm (Figure 2.16) is in an excellent agreement with the TD-DFT simulated positive CD band of 21 (Figure 2.17a) and also follows the helicity rule. The band at 220 nm has mainly the character of an amide n(O)–π * transition. In summary, the agreement between simulated and experimental CD spectra conﬁrms not only the absolute conﬁguration and conformation of 21 but also the validity of helicity rule for this oxacepham. The same helicity rule works very well also for cephams [118], as demonstrated in Figure 2.17b with cepham 30 as representative example of this group of compounds. As expected for the (R) AC at the ring junction, cepham 30 exhibits, similarly to oxacephams, a positive CE at around 220 nm. The six-membered ring of 30 is in a chair conformation, equally as in 21. Similarly to previously described cephams [118], the conformational ﬂexibility in 30 is limited to the side-chain substituent at C7. The simulated CD spectrum (a)

(b)

Figure 2.17. (a) Simulated at B3LYP/TZVP theory level CD spectrum of 21 (– – –) compared with its experimental CD spectrum (——) and calculated structure of the lowest-energy conformer of 21. (b) Simulated at B3LYP/TZVP theory level CD spectrum of 30 (– – –) compared with its experimental CD spectrum (——) and calculated structure of the lowest energy conformer of 30. The vertical bars represent calculated at B3LYP/TZVP level rotatory strengths. Protons are omitted for clarity.

60

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

of lowest energy conformer is in a good agreement with experiment (Figure 2.17b). The positive band at 212 nm is caused by transitions out of the amide n(N) into the amide n(O)π* orbital, as well as by transitions out of the sulfur lone pair into the same acceptor orbitals. The negative excitations at around 235 and 207 nm result mainly from the sulfur lone pair transitions, with some admixture of the amide n(O) as donor orbital. With a torsion angle O=C–N–C6 equal to −175◦ for 30, the helicity rule is satisﬁed. The applicability of the helicity rule was also tested for carbacephams [118]. Similarly to the earlier discussed cephams and oxacephams, two bands are present in their CD spectra in the 280- to 190-nm spectral range. The long-wavelength band assigned to the n –π * excitation occurs at around 220 nm while the second band, of the π –π * origin, appears at around 200 nm. As shown in Figure 2.18, carbacephams with (6R) conﬁguration possess a positive CE at around 220 nm while carbacephams with an (S ) AC at C6 have a negative CE in the same spectral range. The signs are exactly what one would expect to see applying the helicity rule. Thus, the CD spectra of discussed compounds conform to the rule developed for oxacephams. The presence of isolated double bond in compounds 32 and 33 does not inﬂuence their CD spectra, which appear similar to their saturated counterpart 31 (Figure 2.18a). Completely different situation occurs in the case of compounds 34 and 35 where a conjugated double bond is present. The presence of such an α,β-unsaturated amid chromophore, deﬁned also as cephem, causes a red shift (by about 30–40 nm) and the crucial CD band arises, depending on actual substitution, at 250 nm or 260 nm (Figure 2.18b). Except for other differences, compounds with substituents at C2 absorb at longer wavelength when compared to those unsubstituted at this carbon atom. This regularity is independent of the type of heteroatom present at position 5 in a six-membered ring fused with an azetidinone ring. Additional examples, including the 7-aminocephalosporanic acid, the active nucleus for the synthesis of cephalosporins and intermediates, and its derivatives, can be found in the recently published review [119]. The model carbacepham 31 demonstrates the impact of conformational factors on the CD spectra. As predicted by DFT calculations, the six-membered ring of the lowestenergy conformer of this carbacepham exists in a chair conformation [118, 119]. Note also that the only chromophore of 31 that absorbs above 180 nm is the amide group of the azetidin-2-one ring. It is thus very unexpected that despite this apparent structural simplicity, the computed CD spectrum of the lowest-energy conformer of 31 displays little, if any, resemblance with the experimental CD spectrum, shown in Figure 2.19. The computed amide n(O)–π * CD band is considerably blue-shifted, while the amide n(N)–π ∗ band is not only notably shifted to a higher-energy region but even has the wrong sign.

Figure 2.18. (a) CD spectra of carbacephams 31 (——), 32 (– – –), and 33 (· · ·) recorded in acetonitrile. (b) CD spectra of carbacephems 34 (——), 35 (·– · – · –·) and oxacephem 36 (– – –) recorded in acetonitrile.

61

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

Figure 2.19. Simulated CD spectrum of carbacepham 31 at 0 K (– – –) and

5

OH H H

20

N

10

O

31

0

0 −5

−10

−10

−20 190

Rvel⋅10−40 cgs

Δε (M−1cm−1)

10

350 K (· · ·) compared to experiment (——). The 0 K curve corresponds to the optimized PBE0/SV(P) conformer of 31 shown on the right. (From J. Frelek, P. Kowalska, M. Masnyk, A. ´ Kazimierski, A. Korda, M. Woznica, M. Chmielewski, F. Furche, Circular dichroism and conformational dynamics of cephams and their carba and oxa analogues, Chem. Eur. J. 2007, 13, 6732–6744. Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission.)

230 210 250 Wavelength (nm)

In order to investigate the impact of thermal effects on the CD spectrum, molecular dynamics (MD) simulations of 31 at 350 K were carried out and demonstrated that the saturated six-membered ring system shows a considerable ﬂexibility. The results of the MD simulations revealed that the CD spectrum is highly dependent on the conformation of both the four- and the six-membered rings, with sign inversions occurring for both bands at various points of the MD simulation [118]. The simulated CD spectrum at 350 K compares much better with the experiment than the 0 K spectrum (Figure 2.19). Both bands are broadened and red-shifted compared to the 0 K spectrum and have the correct sign. This result is not surprising because the experimental CD spectra were measured at room temperature and thus represent a thermal average over the CD spectra of many different conformations. This effect more clearly reﬂects the spectrum at 350 K, taking into consideration all possible conformations. On the basis of the aforementioned discussion, it can be stated that regardless of the presence of carbon, oxygen, or sulfur atom at 5 position of the six-membered ring, all cepham analogues absorb in the same absorption range. The sign of the decisive CE at around 220 nm depends on the AC at C6 only. Thus, oxacephams, cephams, and carbacephams with (6R) AC display a positive CD band in this spectral region, whereas

(a)

(b)

Figure 2.20. (a) CD spectra of penicillin V (38) (– – –) and penam 39 (——) recorded in water and acetonitrile, respectively. (b) Simulated at B3LYP/TZVP theory level CD spectrum of lowestenergy conformer of 39 (– – –) compared with its experimental CD spectrum (——) and calculated structure of lowest energy conformer. The vertical bars represent calculated at B3LYP/TZVP level rotatory strengths. Protons are omitted for clarity.

62

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

their counterparts with (6S ) AC exhibit a negative sign of the n –π * CE in the same spectral range. Therefore, it can be concluded that all those groups of compounds are subject to the helicity rule. In the case of conformationally ﬂexible compounds, however, the additional theoretical support for the experimental data is recommended. The penams also conform perfectly to the helicity rule. As can be seen in Figure 2.20, the penam representatives 38 (penicillin V) and 39 possess a positive CE in the longwavelength part of the spectrum in accordance with the (R) stereochemistry at the ring junction. The shift of the CD maxima into the lower-energy spectral region by ≈10 nm is related to the penam-constrained backbone. The bicyclic system of penams is relatively rigid, and its conformational lability is largely restricted to the side-chain substituent at C6 carbon atom. Its relatively unrestricted mobility is evident, considering the results of conformational search for penam 39 for which seven conformers in the energy range of 2.8 kcal mol−1 are found. In respect to the ﬁve-membered ring conformation space, however, only two conformers exist. In the ﬁrst one, populated at the conformational equilibrium over 88%, the ﬁve-membered ring is in an envelope conformation with C2 carbon atom below the average plane passing through the ring (Figure 2.20). In the second conformer, the ﬁve-membered ring adopts again an envelope conformation but with the sulfur atom above the ring plane. Beyond these two, the other individual conformers show differences only in the conformation of substituent at C6 carbon atom which demonstrates its relatively substantial ﬂexibility. The computed O=C–C7–N1–C2 and O=C–C7–N1–C5 torsion angles for all conformers of penam 39 are calculated to be negative, thus providing corroborating evidence for the nonplanarity of amide chromophore. In simulated at B3LYP/TZVP theory level ECD spectra the sign of the lowest-energy excitations is positive for all conformers, as predicted by the helicity rule for (5R) AC. The positive CE at around 240 nm is in accordance with both the helicity rule and the calculations. Based on this we can conclude that for penams the requirements of the helicity rule are met, and therefore the rule can be successfully applied to this group of compounds. Oxaanalogues of penicillins, commonly referred to as clavams, exhibit the same shape of the CD spectra as oxacephams and penams (Figure 2.21). The positive sign of the decisive CD band arising at around 240 nm in clavam 40 corresponds to its (R)conﬁguration at C5, and the negative one corresponds to the (S )-conﬁguration of the ring junction in its local enantiomer 41. This ﬁnding validates the proposed rule for clavams too. As can be seen in Figure 2.21a, the shape of CD spectra of clavams corresponds very well with the shape of spectra of natural penicillins. (a)

(b)

Figure 2.21. (a) CD spectra of clavams 40 (——) and 41 (· · ·) recorded in acetonitrile compared with the CD spectrum of penicillin V (38) (– – –). (b) Simulated at B3LYP/TZVP theory level CD spectrum of lowest energy conformer of 42 (– – –) compared with experimental CD spectrum (——) and calculated structures of its two lowest-energy conformers. Protons are omitted for clarity.

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

A more complex situation is encountered when considering the spectra of clavams with an additional, interfering chromophore in the molecule such as phenyl, which substantially perturbs the electronic structure of the system and therefore signiﬁcantly inﬂuences the CD spectra. However, the recent study on this subject demonstrated that even in the case like that, the requirements of helicity rule have been met and the rule can be successfully applied to correlate the structure and the respective chiroptical properties for unsubstituted and substituted clavams as well [120]. In the ECD spectra of the representative members of carbapenams 43–45, the 240nm ECD band is negative for compounds 43 and 45 with (5S ) AC and positive for 44 with (5R) AC, as predicted by helicity rule (Figure 2.22). Even the presence of interfering chromophores in compound 45 (i.e., phenoxy group and double bond) does not alter the relationship between the sign of 240-nm CD band and absolute conﬁguration of the C5 carbon atom [121]. Similarly to clavams and penams, the bicyclic system of carbapenams is relatively rigid and its conformational lability is largely restricted to substituents at the C6 carbon atom and in the ﬁve-membered ring. Although for carbapenam 43 nine conformers were found in the energy range of 2.4 kcal mol−1 , in respect to the ﬁve-membered ring conformation space, however, there were only two distinct conformers present. In the ﬁrst one, populated in the conformational equilibrium over 75%, the ﬁve-membered ring is in an envelope conformation with C4 carbon atom below the average plane passing through the ring. In the second conformer the ﬁve-membered ring adopts a halfchair conformation with C2 and C3 carbon atoms located above and below the ring plane, respectively (Figure 2.22). Beyond these two, the other individual conformers show differences in the substituent at C6 carbon atom which demonstrates its relatively substantial ﬂexibility. The average ECD spectrum shows very close agreement between experiment and theory, thus providing evidence that these conformers are present in solution under given conditions (Figure 2.22). The negative CE at around 240 nm is in accordance with both the helicity rule and the calculations [121]. In order to show that the nonempirical correlation between chiroptical properties and stereochemistry is not restricted speciﬁcally only to very simple cases, more complex

Figure 2.22. CD spectra of carbapenams 43 (——), 44 (·–·–·–), and 45 (– – –) recorded in acetonitrile and simulated at B3LYP/TZVP theory level CD averaged spectrum of 43 (· · ·) as well as lowest-energy envelope (E) and half-chair (H-C) conformers of 43. Protons are omitted for clarity.

63

64

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

compounds such as carbapenam 46 which incorporates a conjugated diene chromophore at C4 in addition to the amide chromophore have been studied. Although the sign of observed at 240-nm CD band in 46 apparently is consistent with the helicity rule, CD contributions to this band of other electronic transitions can play a decisive role (Figure 2.23). In fact, the TDDFT calculations brought some more insight about the complex UV and CD relationship in carbapenam 46. The ﬁve-membered ring in this compound is nearly planar with a nitrogen atom slightly deviated from the plane formed by remaining four carbon atoms. The conformational differences in 46 are limited mostly to the conformation of the side chain. In all four conformers calculated for 46, the O8–C7–N1–C2 and O8–C7–N1–C5 torsion angles are negative, and thus helicity rule requirements are met. Consequently, the simulated CD spectrum for these geometries showed an expected positive band at ∼240 nm (Figure 2.23) [121]. The analysis show that the electronic transitions from the amide and diene chromophores are mixed and appear in approximately the same energy range. Thus, the longwavelength CD band is an admixture of the amide n –π * and diene π –π * excitations occurring at 240 nm and 252 nm, respectively (MO49 → MO∗ 52 and MO49 → MO∗ 51, respectively). Regardless of this complexity within the band at 240 nm, the positive sign of its component related to the amide n –π* excitation and its decisiveness in terms of the helicity rule are in accord with the rule. In addition, the agreement between the experimental and Boltzmann-averaged ECD spectra is very accurate and conﬁrms both absolute conﬁguration and conformation of carbapenam 46 [121]. It should be added that very recently the helicity rule was reformulated. The reason for this was the fact that presence of substituents in the vicinity of the ring junction may cause the change of AC descriptors of the bridgehead carbon atom from R to S , and vice versa. According to the CIP priority rules, such a change causes an allyl group at C4 in 46 whereas propionic acid methyl ester substituent attached to the same carbon atom in 45 does not. Therefore, to avoid misunderstandings, after reformulation the rule connects the sign of the n –π * amide transition with the d or l conﬁguration, as deﬁned by Wo´znica et al. [121].

(a)

(b)

Figure 2.23. (a) Simulated ECD and UV spectra of lowest energy conformer of carbapenam 46 (——) and experimental ECD spectrum of 46 (– – –). (b) Dominant contributions of MOs of particular excitations of carbapenam 46 lowest-energy conformer.

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

As was mentioned before, the molecular dynamics (MD) simulations carried out for some carba- and oxacephams revealed a considerable ﬂexibility of the saturated sixmembered ring and high dependency of the ECD spectra on the conformation of both the four- and the six-membered rings [113, 118]. In some cases this dependency resulted in a breakdown of the helicity rule caused by a change of the conformation of the pyranose ring from a chair to a boat. Therefore, the question arose about scope and limitations of the helicity rule. To answer this question, chiroptical properties of several model compounds with β-lactam ring fused to a seven-member ring were studied. These ringexpanded cephalosporin analogues 47–50 (Figures 2.24 and 2.25) were chosen due to their increased conformational ﬂexibility in comparison with investigated earlier bicyclic β-lactams with six- and four-membered rings condensed together. In addition, a decrease of the skeletal strain energy expected for compounds in question may result in overall ﬂattening of the system. Therefore, a subsequent breakdown of the helicity rule cannot be excluded. Up to four absorption bands are present in the ECD spectra of investigated compounds 47–49 in the spectral range of 185–300 nm (Figure 2.24). The band occurring around 220 nm attributed to the amide n(O)–π ∗ transition is of a particular interest because this band is the subject of the helicity rule. According to the helicity rule for compounds 47–49 belonging to the (7R)conﬁgurational series, a positive CD is expected, whereas for β-lactam ent 48 with (7S )-conﬁguration a negative one is expected. In fact, in both cases opposite bands at around 215 nm were observed in disagreement with the helicity rule (Figure 2.24). The question arises whether compounds 47–49 constitute an exception or the rule itself is imperfect. A ﬁnding the reasons for this inconsistency appears undoubtedly very important considering the future applicability of the helicity rule. Among the factors that may play a role are the higher conformational ﬂexibility of compounds 47–49 and/or a signiﬁcant change in the geometry of β-lactam chromophore by adopting a planar conformation that does not obey the rule [113]. In order to throw light into the origin for observed deviations from the rule, TDDFT calculations were carried out for β-lactams 47 and 48 (Figure 2.24). In these cases, depending on the compound, the main ECD band is a composite of excitations out of the amide n –π * transition and transitions out of the sulfur or oxygen lone pairs into the same acceptor orbitals. In addition, the transitions out of the double-bond orbitals mix strongly with the amide n –π * transition. Some of the structures of conformers of (a)

(b)

Figure 2.24. (a) ECD spectra of β-lactams: 47 (· · ·), ent 48 (– – –); and 49 (——) recorded in acetonitrile. (b) Simulated at B3LYP/TZVP theory level average CD spectrum of 47 and 48 compared with experimental spectra (ε is the molar decadic absorption coefﬁcient. Protons are omitted for clarity.

65

66

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

Figure 2.25. (a–g) Computed structures of conformers of β-lactam 50 and their simulated ECD spectra. (h) The Boltzmann-averaged spectrum compared to experiment (ε is the molar decadic absorption coefﬁcient). Protons are omitted for clarity.

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

β-lactams 47 and 48, calculated within the range of 3 kcal mol−1 , demonstrate presence of a small deviation from planarity of the amide chromophore manifested by a slight pyramidality of the amide nitrogen. Nevertheless, the negative sign of the decisive ECD band in both experimental and Boltzmann-averaged ECD spectra clearly demonstrates the breakdown of the helicity rule for these compounds. For saturated carbaanalogue 50, seven conformers within the energy range of 2.4 kcal mol−1 were found. To simplify and speed up the calculations, the large substituent at C8 carbon atom (OSit BuMe2 ) was substituted by a smaller OSiMe3 group. Both groups have very similar electronic properties, and the change should not signiﬁcantly inﬂuence the electronic spectra. As can be seen in Figure 2.25, two of seven conformers obtained for 50, namely 3 and 4, have slightly skewed amide chromophore, whereas in the remaining ﬁve conformers the chromophoric system is planar. The population ratio of conformers with a planar and with a nonplanar chromophore in β-lactam 50 is approximately 4:1. Therefore, because the average ECD spectrum is, by deﬁnition, the sum of weighted contributions of all conformers, the negative sign of the decisive band should predominate, which is indeed the case (Figure 2.25h). However, a small positive ECD band at around 240 nm, originating from the twisted conformers 3 and 4, is present in both experimental and simulated spectra. The band at around 220 nm has primarily the character of an amide n –π * transition. The average ECD spectrum shows a very close agreement between experiment and theory, thus providing strong evidence that these conformers are present in solution under given conditions. The seven-membered ring in conformers of 50 is in a chair or a twist–chair conformation, and the azetin-2-one ring is in equatorial position at C7 (Figure 2.25). Beyond that, the individual conformers show conformational differences mostly within the substituent at C8 carbon atom. Independent evidence comes from the X-ray diffraction data for a thioanalogue of β-lactam 50, namely thiolactam 50a, the only one forming crystals suitable for such an analysis. The solid-state structural data clearly demonstrates the planarity of the amide chromophore and the chair form of seven-membered ring. Additionally, the dihedral angle C9–C7–C2–N1 equal to −2.0◦ points to the sp 2 hybridization of the amide nitrogen atom.

50

50a

In conclusion, the combined experimental and theoretical studies have shown that the ring-expanded β-lactam analogues do not follow the helicity rule since they do not belong to β-lactam type with nonplanar amide chromophore, for which the rule is valid. Therefore, these β-lactam analogues belong to the second class of Moskowitz’s chromophores [102, 103]—that is, to locally achiral but chirally perturbed. Thus, depending on the type of chiral perturbation originating from either chiral conformation of the ring incorporating achiral chromophore or from bonds closely connected to the chromophore, the chirality of the second or third sphere, respectively, should govern the CD of these compounds, as proposed by Snatzke [122]. According to this view, chirality or sector rules can correlate the stereochemistry around the chromophore with the sign of respective CD bands.

67

68

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

ACKNOWLEDGMENTS The preparation of this chapter and the work of our research group (JF, AB, MW) described herein has been supported by the Ministry of Science and Higher Education, grants N N204 123507 and N N204 092935.

REFERENCES 1. Reviews and book chapters: (a) J. Gawronski, Houben-Weyl Methods of Organic Chemistry, Vol. E21a, 4th ed., G. Thieme, Stuttgart, 1995, pp. 499–533. (b) P. Salvadori, C. Rosini, L. Di Bari, in The Chemistry of Dienes and Polyenes, Vol. 1, Z. Rapoport, ed., John Wiley & Sons, New York, 1997, pp. 111–147. (c) J. Gawronski, in Circular Dichroism: Principles and Applications, N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, pp. 305–335. (d) J. K. Gawronski, Tetrahedron 1982, 38 , 3–26. (e) N. Kirk, Tetrahedron 1986, 42 , 777–818. (f) J. Gawronski, Conformations, Chiroptical and Related Spectral Properties of Enones, in The Chemistry of Enones, S. Patai and Z. Rappoport, eds., John Wiley & Sons, New York, 1989, Chapter 3. (g) D. A. Lightner, J. E. Gurst, Organic Conformational Analysis and Stereochemistry from Circular Dichroism Spectroscopy, WileyVCH, New York, 2000, pp. 337–409. (h) G. Snatzke, Angew. Chem. Int. Ed . 1979, 18 , 363–377. (i) G. Snatzke, in Fundamental Aspects and Recent Developments in Optical Rotatory Dispersion and Circular Dichroism, F. Ciardelli and P. Salvadori, eds., Heyden, London, 1973, pp. 109–124. 2. H. Suzuki, Electronic Absorption Spectra and Geometry of Organic Molecules, Methuen, London, 1963. 3. A. Rauk, H. A. Peoples, J. Comput. Chem. 1980, 1 , 240–256. 4. N. J. Bunce, J. P. Bergsma, M. D. Bergsma, W. De Graaf, Y. Kumar, L. Ravanal, J. Org. Chem. 1980, 45 , 3708–3713. 5. L. H. Hall, Group Theory and Symmetry in Chemistry McGraw-Hill, New York, 1969. 6. G. Herzberg, Molecular Spectra and Molecular Structure Van Nostrand, Princeton, NJ, 1966. 7. E. S. Stern, C. J. Timmons, Electronic Absorption Spectra in Organic Chemistry, St. Martin’s Press, New York, 1971. 8. T. Liljefors, N. L. Allinger, J. Am. Chem. Soc. 1976, 98 , 2745–2749. 9. H. Ziffer, C. H. Robinson, Tetrahedron 1968, 24 , 5803–5816. 10. J. K. Gawronski, T. Liljefors, B. Nord´en, J. Am. Chem. Soc. 1979, 101 , 5515–5522. 11. (a) A. Moscowitz, E. Charney, U. Weiss, H. Ziffer, J. Am. Chem. Soc. 1961, 83 , 4661–4663. (b) G. Wagniere, W. Hug, Tetrahedron Lett. 1970, 4765–4768 12. A. W. Burgstahler, R. C. Barkhurst, J. Am. Chem. Soc. 1970, 92 , 7601–7603. 13. A. W. Burgstahler, L. O. Weigel, J. K. Gawronski, J. Am. Chem. Soc. 1976, 98 , 3015–3016. 14. R. M. Moriarty, H. E. Paaren, U. Weiss, W. B. Whalley, J. Am. Chem. Soc. 1979, 101 , 6804–6810. 15. O. E. Weigang, Jr., J. Am. Chem. Soc. 1979, 101 , 1965–1975. 16. M. Duraisamy, H. M. Walborsky, J. Am. Chem. Soc. 1983, 105 , 3264–3269. 17. C. Djerassi, R. Records, E. Bunnenberg, K. Mislow, A. Moscowitz, J. Am. Chem. Soc. 1962, 84 , 870–872. 18. W. B. Whalley, Chem. Ind . 1962, 1024–1027. 19. (a) G. Snatzke, Tetrahedron 1965, 21 , 413–419. (b) G. Snatzke, Tetrahedron 1965, 21 , 421–438. (c) G. Snatzke, Tetrahedron 1965, 21, 439–448. (d) G. Snatzke, in Optical Rotatory Dispersion and Circular Dichroism in Organic Chemistry, G. Snatzke, ed. Sadtler Research Labs, Philadelphia, 1967, pp. 208–223.

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

20. A. W. Burgstahler, D. L. Boger, N. C. Naik, Tetrahedron 1976, 32 , 309–315. 21. G. Giacomelli, L. Lardicci, C. Bertucci, A. M. Caporusso, Tetrahedron, 1978, 34 , 2015–2020. 22. H. J. C. Jacobs, E. Havinga, Rec. Trav. Chim. 1965, 84 , 932–941. 23. K. K. Cheong, A. Oshita, D. J. Caldwell, H. Eyring, Proc. Natl. Acad. Sci. USA 1970, 67 , 1727–1733 24. M. C. A. Donkersloot, H. M. Buck, J. Mol. Struct. 1988, 180 , 389–390. 25. H. Weiss, H. Ziffer, E. Charney, Tetrahedron 1965, 21 , 3105–3120. 26. E. Charney, H. Ziffer, U. Weiss, Tetrahedron 1965, 21 , 3121–3126. 27. W. Hug, G. Wagniere, Helv. Chim. Acta 1971, 54 , 633–649. 28. J. S. Rosenﬁeld, E. Charney, J. Am. Chem. Soc. 1977, 99 , 3209–3214. 29. E. Charney, C.-H. Lee, J. S. Rosenﬁeld, J. Am. Chem. Soc. 1979, 101 , 6802–6804. 30. D. A. Lightner, T. D. Bouman, J. K. Gawronski, K. Gawronska, J. L. Chappuis, B. V. Crist, A. E. Hansen, J. Am. Chem. Soc. 1981, 103 , 5314–5327. 31. A. Rauk, J. O. Jarvie, H. Ichimura, J. M. Barriel, J. Am. Chem. Soc. 1975, 97 , 5656–5664. 32. J. Lessard, L. Ruest, Ch. R. Engel, Can. J. Chem. 1972, 50 , 1433–1437. 33. A. E. Hansen, K. L. Bak, J. Phys. Chem. A 2000, 104 , 11362–11370. 34. R. Bucourt, Topics in Stereochemistry N. L. Allinger and E. L. Eliel, eds., 1974, 8 , 159. 35. P. J. Roberts, R. C. Patterson, G. M. Sheldrick, N. W. Isaacs, O. Kennard, J. Chem. Soc. Perkin II , 1973, 1978–1984. 36. B. Busetta, C. Courseille, F. Leroy, M. Hospital, Acta Crystallogr. Sec. B 1972, 28 , 3293–3299. 37. A. Cooper, G. Kartha, E. M. Gopalakrishna, D. A. Norton, Acta. Crystallogr. Sec. B 1969, 25 , 2409–2411. 38. A. Cooper, E. M. Gopalakrishna, D. A. Norton, Acta. Crystallogr. Sec. B 1968, 24 , 935–941. 39. H.-G. Kuball, S. Neubrech, A. Sch¨onhofer, Chem. Phys. 1992, 163 , 115–132. (b) H.-G. Kuball, B. Schultheis, M. Klasen, J. Frelek, A. Sch¨onhofer, Tetrahedron: Asymmetry, 1993, 4 , 517–528. (c) J. Frelek, W. J. Szczepek, H. P. Weiss, G. J. Reiss, W. Frank, J. Brechtel, B. Schultheis, H.-G. Kuball, J. Am. Chem. Soc. 1998, 120 , 7010–7019. (d) J. Frelek, W. J. Szczepek, S. Neubrech, B. Schultheis, J. Brechtel, H.-G. Kuball, Chem. Eur. J . 2002, 8 , 1899–1907. 40. L. Velluz, M. Legrand, R. Viennet, C. R. Acad. Sci. Paris. 1965, 261 , 1687–1691. 41. M. Kwit, J. Gawronski, D. R. Boyd, N. D. Sharma, M. Kaik, Org. Biomol. Chem. 2010, 8 , 5635–5645. 42. (a) F. Weinhold, in, Encyclopedia of Computational Chemistry, Vol. 3, P. v. R. Schleyer, N. L. Allinger, T. Clark, J. Gasteiger, P. A. Kollman, H. F. Schaefer III, and P. R. Schreiner, eds., John Wiley & Sons, Chichester, UK, 1998, pp. 1792–1811. (b) A. E. Reed, L. A. Curtiss, F. Weinhold, Chem. Rev . 1988, 88 , 899–926. (c) F. Weinhold, J. E. Carpenter, in, The Structure of Small Molecules and Ions, R. Naaman and Z. Vager, eds., Plenum, New York, 1988, pp. 227–236. (d) F. Weinhold, Computational Methods in Organic Photochemistry: Molecular and Supramolecular Photochemistry, A. G. Kutateladze, ed., Taylor & Francis/CRC Press, Boca Raton FL, 2005, pp. 393–476. 43. C. Diedrich, S. Grimme, J. Phys. Chem. A 2003, 107 , 2524–2539. 44. For the other deﬁnition of homo- and heterohelicity see, for example: M. Kwit, U. Rychlewska, J. Gawronski, New J. Chem. 2002, 26 , 1714–1717. 45. J. Gawronski, K. Gawronska, G. Buczak, A. Katrusiak, P. Skowronek, H. Suemune, Tetrahedron: Asymmetry 1996, 7 , 301–306. 46. (a) J. Gawronski, M. Kwit, D. R. Boyd, N. D. Sharma, J. F. Malone, A. F. Drake, J. Am. Chem. Soc. 2005, 127 , 4308–4319. (b) M. Kwit, J. Gawronski, Chirality 2011, 23 , 744–751.

69

70

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

47. M. Kwit, N. D. Sharma, D. R. Boyd, J. Gawronski, Chem. Eur. J . 2007, 13 , 5812–5821. 48. M. Kwit, N. D. Sharma, D. R. Boyd, J. Gawronski, Chirality 2008, 20 , 609–620. 49. (a) D. R. Boyd, G. N. Sheldrake, Nat. Prod. Rep. 1998, 15 , 309–324. (b) T. Hudlicky, D. Gonzalez, D. T. Gibson, Aldrichimica Acta 1999, 32 , 35–62. (c) R. A. Johnson, Org. React. 2004, 63 , 117–264. (d) D. R. Boyd, T. D. H. Bugg, Org. Biomol. Chem. 2006, 4 , 181–192. 50. E. Zysman-Colman, D. N. Harpp, J. Sulfur. Chem. 2004, 25 , 291–316. 51. P. Groner, C. W. Gillies, J. Z. Gillies, Y. Zhang, E. Block, J. Mol. Spectrosc. 2004, 226 , 169–181. 52. W. H. Green, A. B. Harvey, J. Chem. Phys. 1968, 49 , 3586–3595. 53. P. D’Antonio, C. George, A. H. Lowrey, J. Karle, J. Chem. Phys. 1971, 55 , 1071–1075. 54. O. Mundt, G. Becker, J. Baumgarten, H. Riffel, A. Simon, Z. Anorg. Allg. Chem. 2006, 632 , 1687–1709. 55. H. Lumbroso, J. M. Catel, G. Le Coustumer, C. G. Andrieu, J. Mol. Struct. 1999, 513 , 201–218. 56. R. Steudel, Y. Drozdova, K. Miaskiewicz, R. H. Hertwig, W. Koch, J. Am. Chem. Soc. 1997, 119 , 1990–1996. 57. W. N. Hubbard, D. R. Douslin, J. P. McCullough, D. W. Scott, S. S. Todd, J. F. Messerly, I. A. Hossenlopp, A. George, G. Waddington, J. Am. Chem. Soc. 1958, 80 , 3547–3554. 58. D. W. Scott, H. L. Finke, M. E. Gross, G. B. Guthrie, H. M. Huffman, J. Am. Chem. Soc. 1950, 72 , 2424–2430. 59. H. A. Khwaja, S. Walker, Can. J. Chem. 1983, 61 , 2107–2111. 60. R. L. Redington, J. Mol. Spectrosc. 1962, 9 , 469–476. 61. J. Hahn, P. Schmidt, K. Reinartz, J. Behrend, G. Winnewisser, K. M. T. Yamada, Z. Naturforsch. 1991, 46 , 1338–1342. 62. V. Aquilanti, M. Ragni, A. C. P. Bitencourt, G. S. Maciel, F. V. Prudente, J. Phys. Chem. A, 2009, 113 , 3804–3813. 63. D. Nori-Shargh, H. Yahyaei, J. E. Boggs, J. Mol. Graphics Modell . 2010, 28 , 807–813. 64. T. Shimizu, H. Isono, M. Yasui, F. Iwasaki, N. Kamigata, Org. Lett. 2001, 3 , 3639–3641. 65. G. Bergson, Ark. Kemi 1958, 12 , 233–237. 66. G. Bergson, Ark. Kemi 1962, 18 , 409–434. 67. J. Linderberg, J. Michl, J. Am. Chem. Soc. 1970, 92 , 2619–2625. 68. R. W. Woody, Tetrahedron 1973, 29 , 1273–1283. 69. M. Pericou-Cayere, M. Rerat, A. Dargelos, Chem. Phys. 1998, 226 , 297–306. 70. J. Seino, Y. Honda, M. Hada, H. Nakatsuji, J. Phys. Chem. A 2006, 110 , 10053–10062. ´ Ken´ez, S. Antus, L. Szil´agyi, T. Z. Illy´es, I. Szab´o, 71. T. Kurt´an, G. Pescitelli, P. Salvadori, A. Chirality 2008, 20 , 379–385. ´ 72. P. Skowronek, J. Scianowski, J. Gawronski, Tetrahedron: Asymmetry, 2006, 17 , 2408–2412. ´ 73. P. Skowronek, J. Gawronski, J. Scianowski, 12th International Conference on Circular Dichroism, Brescia 2009. 74. C. Djerassi, A. Fredga, B. Sj¨oberg, Acta Chem. Scand . 1961, 15 , 417–426. 75. B. Ringdahl, J. C. Craig, G. Zdansky, A. Fredga, Acta Chem. Scand . 1980, 34 , 735–738. 76. P. M. Scopes, R. N. Thomas, M. B. Rahman, J. Chem. Soc. C 1971, 1671–1676. 77. B. Ringdahl, J. C. Craig, A. Fredga, W. A. Bonner, Acta Chem. Scand . 1981, 35 , 467–471. 78. P. H. A. Laur, Proceedings of Third International Symposium on Organic Selenium and Tellurium Compounds, 1981, 217–229. 79. M. Michalska, G. Snatzke, Liebigs Ann. Chem. 1987, 179–181. 80. Y. Honda, A. Kurihara, Y. Kenmochi, M. Hada, Molecules 2010, 15 , 2357–2373.

S O M E I N H E R E N T LY C H I R A L C H R O M O P H O R E S

81. J. Mann, M. J. C. Crabbe, Bacteria and Antibacterial Agents, Oxford University Press, New York, 1996. 82. R. D. G. Cooper, Topics in Antibiotics Chemistry, Ellis Horwood, Chichester, 1980. 83. G. I. Georg, Bioorg. Med. Chem. Lett. (Symposia-in-print No 8) 1993, 3 , 2159. 84. A. G. Brown, D. Butterworth, M. Cole, G. Hanscomb, J. D. Hood, C. Reading, G. N. Rolinson, J. Antibiot. 1976, 29 , 668–669. 85. R. M. Sweet, L. I. Dahl, J. Am. Chem. Soc. 1970, 92 , 5489–5507. 86. D. B. Boyd, J. Am. Chem. Soc. 1972, 94 , 6513–6519. 87. D. B. Boyd, J. Med. Chem. 1973, 16 , 1195–1199. 88. R. Busson, H. Vanderhaeghe, J. Org. Chem. 1978, 43 , 4438–4441. 89. R. Busson, E. Roets, H. Vanderhaeghe, J. Org. Chem. 1978, 43 , 4434–4437. 90. H. Rehling, H. Jensen, Tetrahedron Lett. 1972, 27 , 2793–2796. 91. H. Ogura, H. Takayanagi, K. Kubo, K. Furuhata, J. Am. Chem. Soc. 1973, 95 , 8056–8059. 92. H. Ogura, K. Takeda, H. Takahashi, Chem. Pharm. Bull . 1975, 23 , 2469–2473. 93. V. Toome, B. Wegrzynski, G. Reymond, Biochem. Biophys. Res. Commun. 1976, 69 , 206–211. 94. J. McCann, A. Rauk, G. V. Shustov, H. Wieser, D. Yang, Appl. Spectros. 1996, 50 , 630–641. 95. D. Galle, M. Tolksdorf, M. Braun, Tetrahedron Lett. 1995, 36 , 4217–4720. 96. H. Wolf, Tetrahedron Lett. 1966, 5151–5156. 97. G. Barbaro, A. Battaglia, A. Guerrini, C. Bertucci, S. Geremia, Tetrahedron: Asymmetry 1998, 9 , 3401–3409. 98. Y. Chen, A. G. Petrovic, M. Roje, G. Pescitelli, M. M. Kayser, Y. Yang, N. Berova, G. Proni, Chirality 2010, 22 , 140–152. 99. T. Polonski, M. J. Milewska, Croat. Chim. Acta 1989, 62 , 129–134. 100. E. Ong, L. Cusachs, O. Weigang, J. Chem. Phys. 1977, 67 , 3289–3297. 101. A. Moscowitz, Advan. Chem. Phys. 1962, 4 , 67–112. 102. A. Moscowitz, Dansk Kemi 1962, 43 , 113–117. 103. A. Moscowitz, K. Mislow, M. A. W. Glass, C. Djerassi, J.Am.Chem.Soc. 1962, 84 , 1945–1955. 104. D. B. Boyd, J. P. Rhiel, F. S. Richardson, Tetrahedron 1979, 35 , 1499–1508. 105. I. Fric, P. Malon, M. Tichy, K. Blaha, Collect. Czechoslov. Chem. Comm. 1977, 42 , 678–686. 106. K. Blaha, M. Budesinsky, I. Fric, Z. Koblicova, P. Malon, M. Tichy, Tatrahedron Lett. 1978, 41 , 3449–3452. 107. G. Shustov, A. Kachanov, I. Chervin, R. Kostyanovsky, A. Rauk, Can. J. Chem. 1994, 72 , 279–286. 108. G. V. Shustov, G. K. Kadorkina, S. V. Varlamov, A. V. Kachanov, R. G. Kostyanovsky, A. Rauk, J. Am. Chem. Soc. 1992, 114 , 1616–1623. 109. M. Braun, D. Galle, Synthesis 1996, 819–820. 110. D. B. Boyd, C.-Y. Yeh, F. S. Richardson, J. Am. Chem. Soc. 1976, 98 , 6100–6106. 111. D. B. Boyd, J. P. Riehl, F. S. Richardson, Tetrahedron 1979, 35 , 1499–1508. 112. J.-C. Mueller, V. Toome, D. L. Pruess, J. Antibiotics 1983, XXXVI , 217–225. 113. R. Łysek, K. Borsuk, M. Chmielewski, Z. Kału˙za, Z. Urba´nczyk-Lipkowska, A. Klimek, J. Frelek, J. Org. Chem. 2002, 67 , 1472–1479. 114. J. Frelek, R. Łysek, K. Borsuk, J. Jagodzi´nski, B. Furman, A. Klimek, M. Chmielewski, Enantiomer 2002, 7 , 107–114. 115. A. Kozioł, J. Frelek, M. Wo´znica, B. Furman, M. Chmielewski, Eur. J. Org. Chem. 2009, 338–341.

71

72

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

116. T. T. Danh, W. Bocian, L. Kozerski, P. Szczukiewicz, J. Frelek, M. Chmielewski, Eur. J. Org. Chem. 2005, 429–440. 117. A. Kozioł, B. Furman, J. Frelek, M. Wo´znica, E. Altieri, M. Chmielewski, J. Org. Chem. 2009, 74 , 5687–5690. 118. J. Frelek, P. Kowalska, M. Masnyk, A. Kazimierski, A. Korda, M. Wo´znica, M. Chmielewski, F. Furche, Chemistry—Eur. J . 2007, 13 , 6732–6744. 119. M. Wo´znica, P. Kowalska, R. Rysek, M. Masnyk, M. G´orecki, M. Kwit, F. Furche, J. Frelek, Curr. Org. Chem. 2010, 14 , 1022–1036. 120. M. Chmielewski, M. Cierpucha, P. Kowalska, M. Kwit, J. Frelek, Chirality 2008, 20 , 621–627. 121. M. Wo´znica, M. Masnyk, S. Stecko, A. Mames, B. Furman, M. Chmielewski, J. Frelek, J. Org. Chem. 2010, 75 , 7219–7226. 122. G. Snatzke, Angew. Chem., Int. Ed. Engl . 1979, 18 , 363–377.

3 ELECTRONIC CD OF BENZENE AND OTHER AROMATIC CHROMOPHORES FOR DETERMINATION OF ABSOLUTE CONFIGURATION ´ Sandor ´ Tibor Kurtan, Antus, and Gennaro Pescitelli

3.1. BENZENE DERIVATIVES WITH CONTIGUOUS CHIRALITY CENTER; SECTOR AND CHIRALITY RULES The benzene chromophore is a common structural feature in numerous optically active synthetic and natural products, and hence Cotton effects (CEs) of the characteristic π –π ∗ electric transitions are regularly utilized for the determination of their absolute conﬁgurations. Above 175 nm, benzene shows three π –π ∗ electronic absorption bands, centered at 184, 204, and 254 nm and designated as 1 Ba,b (E1u ), 1 La (B1u ), and 1 Lb (B2u ), respectively [1–3]. Depending on the substitution pattern of the aromatic ring, the position and intensities of these absorption bands can be somewhat altered, but the spectrum is essentially unchanged. It is the longest wavelength CE belonging to the 1 Lb (B2u ) band, which is most often used to determine the absolute conﬁguration of benzene derivatives. The 1 Lb transition is both electronically and magnetically forbidden in benzene, and its electronic absorption intensity derives from vibronic borrowing from the allowed 1 Ba,b transition. The 1 Lb band shows well-deﬁned vibrational ﬁne structure, and its CEs are associated with allowed transitions from the lowest-energy vibrational mode in the ground state to totally symmetric vibrational modes in the lowest-energy electronically excited state, and the lowest-energy CE corresponds to the 0–0 vibrational transition 1 Lb CE. Sometimes two distinct vibrational progressions appear with similar spacing and a small separation. The two progressions may have CEs with opposite signs, and in this situation the CD in the 1 Lb region appears as a succession of a sequence of minima and maxima or of maxima of alternating sign [4–6]. Upon substitution of the benzene ring, an additional intensity is attributed to the induced electric transition moment due to the substituent that destroys the symmetry of benzene [7]. In optically active benzene Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

73

74

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

TAB L E 3.1. 1 Lb Band CEs of (R)-Phenylmethylcarbinols in Methanola H

R C

OH CH3

Compound (R)-1 (R)-2 (R)-3 (R)-4 (R)-5 (R)-6 (R)-7 a Spectroscopic

CD [λ, nm (ε × 102 )

R H p-Cl p-CF3 m-Cl m-CF3 o-Cl o-CF3

268 276 268 274 271 273 270

Reference

(−17) (+2.5) (−12) (−28) (−15) (+6.7) (−13)

9 9 9 10 10 10 11

moments [8]: qCl = +6, q(CF3 ) = −9 [(cm mol)/L]−1/2 .

derivatives, the rotational strength of the 1 Lb CEs is inﬂuenced by the 1 Ba,b transitions through vibronic coupling as well as by the induced transition moments of the aromatic substituents. In previous papers [10, 12–16] and a review [17] on the ECD study of phenyl- and benzylcarbinols, phenyl- and benzylcarbinamines, and 1-substituted indans and tetralins, Smith applied an empirical sector rule to describe the vibronic contribution to the 1 Lb CE in monosubstituted benzene derivatives with a contiguous chirality center. This sector rule divides the space around the benzene chromophore into 12 sectors, but it is simpliﬁed to a quadrant rule for monosubstituted benzenes, since only sectors surrounding the chirality centers have to be considered (Figure 3.1). In the quadrant rule, the plane of the benzene ring deﬁnes a nodal plane, while the other one, perpendicular to the former, is allocated by the attachment bond of the benzylic carbon as shown in Figure 3.1. In substituted benzene derivatives with hydrogen atom at the contiguous chirality center, the benzylic hydrogen eclipses or nearly eclipses the plane of the benzene ring as supported by X-ray [18, 19], 1 H NMR [20, 21], and molecular modeling [22]. Since the observed chiroptical properties are dependent on both the conformation and absolute conﬁguration, the knowledge of the proper conformation regarding the rotation about the benzylic

(a)

(b)

(c)

H C* R1

H

H C*

R2

R1

* R2

R1

R2

Figure 3.1. (a) Sector rule for third sphere contribution to the 1 Lb CE in monosubstituted benzene derivatives with contiguous chirality center. The plane of the benzene ring is also a nodal plane; signs are for the upper sectors. (b) Quadrant sector rule of monosubstituted benzene derivatives with benzylic chirality center. The plane of the benzene ring is also a nodal plane; signs are for the upper sectors. (c) Quadrant sector rule with signs of all the four sectors viewed from the direction of the benzylic carbon. Thick line represents the benzene ring, deﬁning a nodal plane. Benzylic hydrogen eclipces or nearly eclipses the plane of the benzene ring.

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

attachment bond is crucial for the correct application of the sector rule. If the benzylic hydrogen is not totally eclipsed with the benzene ring as observed by Pescitelli et al. [23] (vide infra) in the ECD analysis of PhCH(Me)t-Bu supported by TDDFT-ECD calculation, and by Butz and co-workers for α-phenylethyl alcohol (1) from gas-phase data [24], the application of the benzene sector rule becomes ambiguous (Table 3.1). If the benzylic hydrogen lies in the plane of the benzene ring (nodal plane), it does not have signiﬁcant contribution, while the R1 and R2 groups are in front and rear sectors with negative and positive contributions, respectively. Based on experimental ECD data, a sequence of magnitudes of the contribution to the 1 Lb CE of various groups (related to group polarizabilities) has been determined [13]. For example, in (R)-α-phenylethyl alcohol [(R)-1, R1 = Me, R2 = OH, Figure 3.1], the methyl group, located in the negative lower front sector, has larger rotatory contribution than the hydroxyl group, which implies a negative 1 Lb band CE as found experimentally for (R)-1. Similarly, the sector rule can predict the sign of 1 Lb band CE and thus determine the absolute conﬁguration from the experimental 1 Lb band CE for monosubstituted benzene derivatives with benzylic chirality centers if the sequence of rotatory contributions are known for the benzylic substituents R1 and R2 (Figure 3.1). It must, however, be stressed that the relative order of magnitude, for example, the methyl vs. the hydroxyl group contribution in 1 has been disputed [24]. While it is only the vibronic contribution (orientation and sequence of R1 and R2 ) that determines the 1 Lb band CE of monosubstituted benzene derivatives, with additional achiral ring substituents there is an additional induced rotatory contribution to the 1 Lb CE, which may have the same or opposite sign as that of the vibronic contribution. The magnitude of the induced electronic transition moment is related to the spectroscopic moment of the ring substituent, introduced by Platt [25] and Petruska [8]. Depending on the spectroscopic moment and ring position(s) of the substituent(s), the induced rotatory contribution can reinforce, decrease, or even override the vibronic contribution, whose effects are summarized by a chirality rule [17]. When (R)-α-phenylethyl alcohol [(R)1)] is substituted with a chlorine atom, having a negative spectroscopic moment, the induced rotatory contribution overrides the negative vibronic contribution and results in positive 1 Lb CEs for o- and p-substitution [(R)-6 and (R)-2, respectively], while for msubstitution the negative 1 Lb CE is preserved [(R)-4]. In contrast, when the triﬂuoromethyl group, having a positive spectroscopic moment, is the aromatic substituent, o-, m- and p-substituted derivatives [(R)-3, (R)-5, (R)-7] give equally negative 1 Lb CEs; i .e. the induced rotatory contribution of the triﬂuoromethyl substituent does not overshadow the negative vibronic contribution of the chirality center. It seems that phenylalkylcarbinols and phenylalkylcarbinamines ortho- or para-substituted by an atom or group with positive spectroscopic moment (Cl, Br, CH3 , OH, OMe) show 1 Lb CEs of opposite sign to that of the unsubstituted parent compound. In contrast, derivatives having a substituent with negative spectroscopic moment (CF3 , CN) in the meta position have the same sign of 1 Lb CEs as that of the unsubstituted parent. As a summary, the unambiguous conﬁgurational assignment of benzene derivatives with a benzylic chirality center by semiempirical rules has to meet the following conditions: 1. Benzylic hydrogen is eclipsed with the plane of the benzene ring or else the exact conformation must be known. 2. Reliable priority order for the rotatory contributions of the benzylic R1 and R2 substituents (vibronic contribution to the 1 Lb CE).

75

76

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

3. Clear-cut allocation of the R1 and R2 substituents to sectors. 4. When vibronic and induced contributions have opposite signs, one has to know which is the dominant contribution. Even in the absence of auxochromic substituents on the phenyl ring, the consistency of the sector rule for benzene 1 Lb transition has been recently called into question by different reports having in common the use of high-level quantum-mechanics calculations for predicting ECD spectra. It must be stressed that ECD calculations offer a valuable chance for validating sector rules, the beneﬁt of which the original authors of the rules didn’t have. In fact, provided that a reliable calculation method is chosen, the calculations allow one to observe the chiroptical response of a molecule in a certain speciﬁc geometry (that used as an input). Moreover, such a geometry may be manipulated accessing hypothetical conformations which cannot be physically observed. Pescitelli et al. have reported the ECD spectra of a homologous series of simple chiral aliphatic compounds (R)-PhCH(Me)R [(R)-8-11] with R = Et, nPr, i Pr and tBu, respectively [23]. Compound (R)-11 showed a negative 1 Lb CE in agreement with Smith’s rule; however, the lowest-energy conformation for this derivative had the C–H bond not coplanar with the ring as prescribed by the rule, thus the agreement was fortuitous. The lowest-energy conformations of the remaining compounds do have the prescribed conformation. Low-temperature ECD spectra (at 183 K), which should be dominated in all cases by the respective lowest energy conformer, consist of a series of more intense maxima with positive sign alternated by a series of weaker maxima with negative sign (Figure 3.2). The ﬁrst series is allied with the allowed vibrational progression with spacing 920–1000 cm−1 , and its sign is at odds with that predicted by the 1 Lb CE sector rule. Time-dependent density functional theory (TDDFT) calculations reproduced instead the dominant sign of 1 Lb CE’s, although a more correct treatment would necessarily include vibronic effects [26, 27]. A further exception to the 1 Lb CE sector rule has been reported by Butz et al. [24] concerning (R)-α-phenylethyl alcohol [(R)-1]. According to these authors, the apparent consistency between the negative 1 Lb CE observed for (R)-1 and the sector rule is due to an incorrectly assumed conformation. The correct lowest-energy structure, found by geometry optimizations and gas-phase experiments, has the methyl bond roughly perpendicular to the plane of the phenyl, and the C–H bond is well oustide from the plane. To reconcile such a structure with the observed ECD spectrum, the sector signs for the rule must be reversed. Later, the authors substantiated their conclusions by TDDFT calculations also using a solvent model [28]. In this study, the relevant dihedral angles for 1 (Ph-Cα and Cα-O) were varied systematically and their impact on the sign of calculated 1 Lb CE was ascertained. Moreover, a rigorous theoretical approach was followed by Autschbach and co-workers in their critical evaluation of the same sector rule. The authors generated true nodal surfaces delimiting the sectors by placing a negative charge on a grid set on the top of a benzene ring at a ﬁxed distance and varying its position systematically [29]. Since the benzene ring becomes dissymetrically perturbed by the charge, its transitions acquire non-negligible rotational strengths. In this case too, the signs obtained for the 1 Lb sector rule were opposite to the original ones. The above ﬁndings demonstrate that at least for simple benzenes with chiral substituents (i.e., endowed with third-sphere chirality in Snatzke’s terminology) [30, 31], the sector rule for 1 Lb CE is far from being generally valid and should be used with caution. Since in previous papers [13–15, 17], Smith reviewed the application of sector and helicity rules to benzene derivatives with contiguous chirality center in detail, the present chapter focuses mainly on benzene natural products with condensed carbocyclic and especially heterocyclic rings such as in tetralin, dihydrobenzo[b]furan, chroman, isochroman

77

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

1

1 (a)

×20

×20

(b) 0

0 ×2

293 k 183 k

–1

–1

×2

293 k 183 k

Δε

Δε

–2

–2

nPr

Et

–3

–3 (R)-(1-methylpropyl)-benzene [(R)-8]

(R)-(1-methylbutyl)-benzene [(R)-9]

–4

–4 190 200 210 220 250 260 270 280 λ (nm)

190 200 210 220 250 260 270 280 λ (nm)

1 (c)

0 (d)

×20

0

–1 Δε

–2

×2

293 k 183 k

×20 293 k 173 k

–4 Δε

–2 iPr

–6 tBu

–8

–3

–4

(R)-(1,2,-dimethylpropyl)-benzene [(R)-10] ×.5 190 200 210 220 250 260 270 280 λ (nm)

–10 ×.5 –12

(R)-(1,2,2-trimethylpropyl)benzene [(R)-11]

190 200 210 220 250 260 270 280 λ (nm)

Figure 3.2. Experimental ECD spectra of (R)-PhCH(Me)R derivatives (R)-8-11 at room temperature (solid lines) and at 173–183 K (dashed lines) in hexane or heptane. (Reprinted from reference 23, with permission from Elsevier).

and benzodioxane derivatives. In these derivatives the phenyl chromophore is embedded in a chiral ring, thus they exhibit second-sphere chirality [30, 31]. Some of Smith’s results have been outlined above to underscore the parallelism between the theories describing the two families of benzene derivatives. The major goal of the recent chapter is to provide guidelines for nonspecialists in the determination of absolute conﬁguration of cyclic natural products with fused benzene ring by means of benzene semiempirical helicity rules.

3.2. TETRALIN AND TETRAHYDROISOQUINOLINE DERIVATIVES 3.2.1. Tetralins and Tetrahydroisoquinolines without Aromatic Ring Substituents; P/M Helicity → Positive/Negative 1 Lb CE In terms of chromophoric system, chiral tetralin and 1,2,3,4-tetrahydroisoquinoline derivatives belong to the benzene chromophores with chiral second sphere [30, 31]. In

78

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a)

6 X

P helicity

positive

5a

4 3

X

X2 8a 1 8 ωC5a,C4,C3,X < 0 X: CH2 tetralin X: NH tetrahydroisoquinoline negative M helicity sign of the 1Lb-band CE 7

ωC5a,C4,C3,X > 0

5

(b)

X

Figure 3.3. (a) Snatzke’s helicity rule or correlation between the sign of the second sphere contribution of tetralin (tetrahydroisoquinoline) and the 1 Lb band CE [30, 32]. The arrow indicates the direction of the overall spectroscopic moment. P- and M-helicity refer to the absolute conformation of the nonaromatic ring. (b) Sector rules for the third (fourth) sphere contributions to the 1 Lb band CE in tetraline (tetrahydroisquinoline). The plane of the benzene ring is a nodal plane, and signs refer to upper sectors.

these cases, the achiral benzene chromophore (ﬁrst sphere) is chirally perturbed by the fused chiral ring (second sphere) and the substituents of the heterocyclic ring (third sphere), which gives rise to the observed Cotton effects. In tetralin derivatives having no substituents on the fused aromatic ring, the 1 Lb CE is determined mainly by the absolute conformation [30] of the fused nonaromatic chiral ring adopting usually a half-chair conformation–that is, the P - or M -helicity deﬁned by the ωC5a,C4,C3,X torsional angle, which in turn is directed by the absolute conﬁguration of the fused ring (Figure 3.3). Snatzke and Ho [32] developed a so-called helicity rule for the benzene chromophore of chiral tetralin and tetrahydroisoquinoline derivatives (Figure 3.3) according to which if the benzene ring is not further substituted, P -helicity of the nonaromatic ring leads to a positive CE within the 1 Lb band and, vice versa, M -helicity is manifested in a negative one. Moreover, a sector rule [33–36] was introduced to evaluate the contribution of the third sphere, namely, the presence of substituents attached to the nonaromatic ring. This sector rule is similar to the sector rule of benzene derivatives with benzylic chirality center (Figure 3.1a) except for that an additional plane is added, perpendicular to the benzene ring and coinciding the C2 axis, resulting in 16 sectors (Figure 3.3b). There is a simple relationship between the helicity rule and sector rule of tetralin; the two non-coplanar carbon atoms of tetralin lie in positive sectors with P -helicity affording the same prediction by the two rules. Since it is the chiral sphere nearest to the chromophores that generally determines the sign of the 1 Lb CE, the helicity of the nonaromatic ring has to be considered as the dominant contribution. If the relationship between the helicity of the nonaromatic ring and the sign of 1 Lb band CE is known, the chirality (absolute conformation) of the heterocyclic ring can be deduced from the measured ECD spectrum. Since the relative conﬁguration of the substituents at the chiral centers as well as their equatorial or axial orientation can be obtained from NMR experiments (3 JH,H , 3 JC,H , NOE effects) or X-ray analysis, their absolute conﬁgurations can be also assigned. For instance, in the (2S , 3S )-12 tetraline derivative (Table 3.2), the trans-diequatorial arrangement of the C2 and C3 methyl groups and half-chair conformation of the carbocyclic ring can be determined from NMR experiments (e.g., large value of 3 J2H,3H ); and then according to the helicity rule, the measured positive 1 Lb CE suggests that the nonaromatic ring adopts P -helicity [36]. The combination of these two data allows the determination of the absolute conﬁguration as (2S , 3S )-12. Similarly, the positive 1 Lb CEs of (2R)-13, (1R)-14, and (4aS , 9aS )15 derive from P -helicity of the dominant conformer with half-chair conformation. In

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

order to reduce van der Waals repulsion between the 1-Me and peri 8-H in (1R)-14, the 1-Me group preferably adopts a quasi-axial orientation with a distorted half-chair conformation [36]. The measured positive 1 Lb CE implies P -helicity, which is only feasible with (1R)-14 absolute conﬁguration. With a half-chair conformation of the fused ring B, the benzylic 9-Me group of (4aS , 9S , 9aR)-16 would be in an equatorial orientation, however, C9 moves into the plane of the benzene ring to shift the 9-Me to a quasi-axial position and thus reduces the repulsion with peri 8-H and 1-H, resulting in a twist boat conformation. Since the nonaromatic ring adopts a conformation considerably different from the half-chair one, and other equilibrating conformers may also contribute, the helicity rule fails to determine the right absolute conﬁguration in this case. In contrast, the epimeric (4aS , 9R, 9aR)17 corroborates well the helicity rule, since the fused cyclohexene ring has half-chair conformation with an axial 9-Me as the major conformer. The tetrahydroisoquinoline derivatives 18–21 behave similarly, and M -helicity results in a negative 1 Lb CE and vice versa.

3.2.2. Tetralins with Achiral Ring Substituents Snatzke et al. [31, 39] also showed that achiral substituents of the benzene ring with large spectroscopic moment {e.g., qOMe = +21 [(cm mol)/L]−1/2 } [8] in speciﬁc positions inverted the helicity rule. This inversion was attributed to the change of the direction of the sum spectroscopic moment [8, 25, 40, 41] vector which gives the electric transition moment vector (μ)–namely, the translation of the electron charge during the transition. This effect was called the induced rotatory contribution and was described by the chirality rule (vide supra) for benzene derivatives with contiguous chirality center. In Snatzke’s terminology, the achiral ring substituents can induce the inversion of the original helicity rule, which is the consequence of rotating the electric transition moment by approximately 30◦ . Figure 3.4a shows a polarization diagram of the tetralin chromophore, in which the addition of the spectroscopic moments oriented the electric transition moment along the direction of the C2 axis of the chromophore, which gives a positive 1 Lb -band CE for P -helicity of the nonaromatic ring (helicity rule of unsubstituted tetraline). The same helicity rule is valid for 6,7-dimethoxytetralins (Figure 3.4b), since in the presence of the two methoxy substituents at position 6 and 7, the direction of the electric transition moment does not change. In contrast, when tetralin has only one methoxy or hydroxy group at C6, the sum of the spectroscopic moments rotates the electric transition moment by approximately 30◦ , which leads to a sign inversion as shown in Figure 3.4c. Similarly, the inverse helicity rule holds for 5,7-dimethoxytetralins. A systematic study [31, 39] was carried out on substituted tetralin derivatives to reveal the effect of different substitution patterns on the sign of the 1 Lb band. This study clearly demonstrated that 5,8- and 6,7-disubstituted and 5,6,7-trisubstituted tetralins follow the same helicity rule as the unsubstituted tetralin, while 6-monosubstituted and 5,6- and 5,7-disubstituted tetralins obey the inverse one (Figure 3.5). The tetralin chromophore is found in pharmacologically active chiral synthetic derivatives such as melatoninergic ligands 22 [42] and 23 [43], the 5-HT1A receptor antagonist 24 [44], and 2-aminotetralin-2-carboxylic acids 25a,b [45, 46], the absolute conﬁguration of which could have been determined by tetralin helicity rules (Chart 3.1). In 1-aryltetralin lignan natural products such as burseranin (26) [47], the 1-aryl group has exciton coupled interaction with the fused benzene ring [48], which determines the ECD spectrum (Chart 3.1). The absolute conﬁgurations of 1-aryltetralines are elucidated

79

80

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

TAB L E 3.2. Helicity and 1 Lb band CE of Tetralin and Tetrahydroisoquinoline Derivatives 8 7

R1 8a 1

6

5

5a

4

R1 3

4a

R2 2

6

1 10 R 4

5

7 8

3 R3

R2

R2

9a

9

R3

2

NH

1

H

R3

12 : R1 = H, R2, R3 = Me

15 : R1 = Me, R2 , R3 = H

18 : R1, R2 = H, R3 = Me

13 : R1, R3 = H, R2 = Me

16 : R1, R3 = H, R2 = Me

19 : R1, R3 = H, R2 = Me

1

2

3

1

14 : R = Me, R , R = H

2

3

17 : R , R = H, R = Me

20 : R1 = H, R2, R3 = Me 21 : R1, R3 = Me, R2 = H

Compound

Helicity of the Low-Energy Conformer H

(2S , 3S )-12

Me

1

Lb band CE [λ, nm (ε)a or [θ]b

Reference

272 (+0.191)a

36

264.5 (+0.248) Me H

(2R)-13

(1R)-14

H Me

H Me Me

(4aS , 9aS )-15 H2C

H

(4aS, 9S, 9aR)-16

H 2C

H

272 (+0.161)a 265 (+0.176)

36

272.5 (+0.226)a 264.5 (+0.233)

36

272.5 (+0.303)a 265 (+0.330)

36

272.5 (-0.058), 270 (+0.012)a 265 (−0.33), 257.5 (−0.027)

36

273 (+0.568)a 266 (+0.558)

36

266 (−290)b 273 (−800)b

37 38

270 (+292)b

37

CH2

CH2 H Me

H

(4aS, 9R, 9aR)-17

Me

H2C

H

H

H N

(1S )-18

H Me

H

(3S )-19 Me

N H

81

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

TAB L E 3.2. (Continued ) 1

Helicity of the Low-Energy Conformer

Compound

Lb band CE [λ, nm (ε)a or [θ]b

37

265 (+613)b

37

265

H

Reference

(+718)b

H

Me

(1R, 3S )-20 N H Me

H

H

(1R, 3R)-21

N H a CE b CE

Me

Me

reported as ε. reported as θ .

(a)

–

5

+

+

– –

(b)

4

6

3

7

2

+

+

+

MeO

8 1 P-helicity positive1LbCE

– –

+ P-helicity positive1LbCE

(d)

(c) MeO +

–

MeO

–

+

MeO +

MeO –

–

+

–

MeO –

– P-helicity negative1LbCE

+

+ OMe

OMe P-helicity negative1LbCE

Figure 3.4. Polarization diagram of the 1 Lb band, direction of the overall spectroscopic moment, and helicity rule of (a) tetralin (b) 6,7-dimethoxytetralin (c) 6-methoxytetralin, and (d) 5,7dimethoxytetralin.

from the sign of the relatively intense CE in the 270- to 290-nm region, governed by the exciton coupling of the two 1 Lb transitions, and thus the helicity rule cannot be applied for this type of tetralin derivatives [49, 50].

3.3. BENZENE CHROMOPHORES WITH FUSED HETEROCYCLIC RING In the following, the applicability of benzene helicity rules is to be discussed in O-heterocyclic natural products, in which the fused benzene ring is part of a 2,3dihydrobenzo[b]furan, isochroman, chroman, or 1,4-benzodioxane moiety (Chart 3.2). The correlation between the n –π ∗ CE and the absolute geometry will be also

82

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

R1 R2

6

R3 7

R1 4

5

3

P-helicity positive 1Lb CE

2 8

tetralin: R1,R2,R3,R4= H 5,8-dimethoxytetralin: R1,R4= OMe, R2,R3= H 6,7-dimethoxytetralin: R1,R4= H, R2,R3= OMe 5,6,7-trimethoxytetralin:R1= H, R2,R3,R4= OMe

4 3

P-helicity negative1Lb CE

2

R3 7

1

R4

5

R2 6

8

1

6-methoxytetralin: R1,R3= H, R2= OMe 5,6-dimethoxytetralin: R1,R2= OMe, R3= H 5,7-dimethoxytetralin: R1,R3= OMe R2= H

Figure 3.5. Effect of achiral ring substituents of large spectroscopic moment (e.g., OMe) on the tetralin (tetrahydroisoquinoline) helicity rule.

O NHCOEt MeO

OH

NHC

n-Pr N

MeO

22

23

n-Pr

Me 24

O O O

NH2

O

COOH R

O O burseranin (26)

25a: R = H 25b: R = OH

Chart 3.1. Structures of tetralin derivatives 22–26.

R1

6

7 7a 1 O

5 4

4a 3

2 R2

2,3-dihydrobenzo[b]furan

R1

6

5 5a 4

7 8

8a 1 X

3 O2

R1

R2

X: H2 isochroman X: =O dihydroisocoumarin

R1

7 6

7 6

8 8a 1 O 2 R2 3 5 5a 4 X

X: H2 chroman X: =O chroman-4-one

8 8a 1 O 2 R2 3 5 5a O 4

1,4-benzodioxane

Chart 3.2. Chromophores with fused benzene ring.

addressed in some of the related carbonyl derivatives such as dihydroisocoumarins and chromanones. ECD spectroscopy has been often utilized for the elucidation of absolute conﬁguration of these ﬂavonoids by simply comparing the ECD spectra of similar derivatives without a deeper understanding of the factors that determine their chiroptical properties. We aimed to establish helicity rules for unsubstituted chromophores by the synthesis and

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

ECD study of derivatives with known absolute conﬁgurations to see whether the original or inverse tetralin helicity rule is valid for these chromophores. Since the benzene rings are substituted in most of the natural products containing these chromophores, the effect of aromatic ring substituents on the helicity rule has to be studied as well, which will help natural product chemists to choose appropriate ECD reference compound for the unambiguous determination of absolute conﬁguration of novel natural products. The determination of absolute conﬁguration of natural benzene-fused heterocycles has to follow the protocol outlined below: 1. Determination of the relative conﬁguration and axial /equatorial orientation of the ring substituents by NMR methods (3 J H,H , 3 JC,H , NOE effect), X-ray single crystal diffraction, computational conformational analysis or their combinations. These studies also provide information on the conformation of the fused nonaromatic ring whether it has a half-chair, envelope, or boat conformation. 2. Determination of the absolute conformation (helicity) of the fused heterocyclic ring on the basis of benzene ECD helicity rules from the measured 1 Lb band CE. The helicity rule of the unsubstituted chromophore and the effect of ring substituents with large spectroscopic moment have to be known. The helicity of the heterocyclic ring is governed by the absolute conﬁguration of the chirality centers and the preferred equatorial/axial orientation of the substituents. Large substituents prefer equatorial orientation due to 1,3-diaxial interaction, although benzylic substituents sometimes tend to favor axial position in order to reduce van der Waals repulsion with the peri aromatic hydrogen. 3. For a safe conﬁgurational assignment, a major conformer with known conformations and high population is required that dominates the ECD parameters. 4. By merging the information on the helicity, relative conﬁguration, and axial/equatorial orientation of the substituents, the absolute conﬁguration can be deduced.

3.3.1. Benzodioxane Chromophore; P/M-Helicity → Positive/Negative 1 Lb CE The benzodioxane chromophore occurs in chiral nonracemic natural ﬂavanolignans [51–53] and neolignans [54–57] as well as in synthetic derivatives of pharmacological interest [58–62]. Antus et al. [51] prepared 1,4-benzodioxane steroid derivatives 27a–c of known absolute conﬁguration and helicity, the ECD study of which showed that the same 1 Lb band helicity rule is valid for unsubstituted 1,4-benzodioxanes as for analogous tetralins; the P /M -helicity of the heteroring leads to a positive/negative 1 Lb band CE, respectively (Chart 3.3, Table 3.3). This result also afforded the conﬁgurational assignment of the natural ﬂavanolignan (−)-silandrin and (−)-isosilandrin isolated from Silybum marianum [51, 53]. The unsubstituted 1,4-benzodioxane helicity rule was applied to deduce the absolute conﬁgurations of synthetic glycogen phosphorylase inhibitors 28a–b, having a 1,4-benzodioxane moiety connected to a N -(β-D-glucopyranosyl)amide unit (Chart 3.3) [60]. Their opposite 1 Lb band CE is governed by the helicity of the heteroring (Table 3.3), which in turn is dictated by the equatorial orientation of the C2 substituent and then the absolute conﬁguration of the C2 chirality center. The chirality centers of the sugar unit are manifested only in the transitions of the amide chromophore, which are not in a mirror image fashion, but they do not interfere with the characteristic 1 Lb band CE. The ECD data of (2R, 3S )-29a,b, (S)-30a,b, and

83

84

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

1 H O 2 3 O 4 H 2-H 27a β 27b α 27c β

OH

OH HO HO H 3-H β α α

R2 6

O

R1 7

O

1

OH

O

R2 7

O

R1

2 1

O 2-C 28a (R) 28b (S) OH

6

OH

R1 R2 (S)-30a CH2OAc H (S)-30b H CH2OAc

O

OH

5

O

NHCOMe

4

R2 H (2R,3S)-29a (CH2)2NHCOMe H (CH2)2NHCOMe (2R,3S)-29b

OMe

2 3

O

R1

O

1

O

1 2 3

OH R 7

2

H N

O

8

EtO

6

O 1

OH

OMe R (2S, 3S)-31a (CH2)3OH (2S, 3S)-31b CH=CHCH2OH

2

OH

O (S)-32

Chart 3.3. Structures of benzodioxane derivatives.

TAB L E 3.3. Helicity and ECD Data of Benzodioxane Derivatives cpd. 27ac 27bc 27cc (R)-28a (S )-28b (2R, 3S )-29a (2R, 3S )-29b (S )-30a (S )-30b (2S , 3S )-31a (2S , 3S )-31b (S )-32

Helicity

CE {λ, nm (ε)a or [θ]b}

Reference

M M P M P P P M M M M M

285 (−1.49)a 285 (−1.09)a 284 (+1.59)a 279 (−0.09)a,d 278 (0.04)a,d 280 (+389)b 280 (+400)b 284 (−0.18)a 285 (−0.21)a 299 (−1010)b 282 (+2549)b 315sh (−0.41), 289 (−0.58)a

51 51 51 60 60 56 56 61 61 54 54 e

reported as ε. reported as θ . c With cholestane skeleton. d Absolute conﬁgurations are erroneously shown in reference 60. e Unpublished ECD data of reference 61. a CE b CE

(2S , 3S )-31a, containing ring substituents with small spectroscopic moment in different positions, also corroborates the unsubstituted 1,4-benzodioxane helicity rule. One may expect that similarly to the isochroman chromophore (vide infra), ring substituents (devoid of any chirality center) do not invert the helicity rule, because the two benzodioxane oxygens ﬁx the electric transition moment along the long axis of the chromophore. However, this expectation is contradicted by the example of the neolignan (2S , 3S )-31b chemically correlated to (2S , 3S )-31a, which has a C7 3-hydroxy1-propen-1-yl substituent and for which the positive 1 Lb CE derives from M helicity of its heteroring. Although (S )-32, bearing a conjugated C6 α,β-unsaturated ester moiety,

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

OMe OMe 1

O

7

1

8

O

O

OMe

2 3

2 3

Me

4

(2S,3S)-33 M-helicity negative 1Lb CE

6

5

OMe

O

Me

4

(2R,3R)-34a trans-eusiderin P-helicity negative 1Lb CE

Chart 3.4. Structures of synthetic (2S, 3S)-33 and natural trans-eusiderin [(2R, 3R)-34]. a Absolute conﬁguration is shown as proposed in reference 63.

apparently follows the unsubstituted helicity rule, the application of this rule is prone to error in the presence of such conjugated chromophore. Compound (2S,3S )-33, a synthetic benzodioxane of known absolute conﬁguration, was prepared from optically active 1-phenyl-1,2-epoxypropanes and showed negative 1 Lb CE pointing at M -helicity of its heteroring in agreement with the helicity rule (Chart 3.4) [63]. Its ECD data in the 220 to 250-nm 1 La region were opposite to that of the neolignan trans-eusiderin [(2R,3R)-34], on the basis of which the (2R,3R) absolute conﬁguration was assigned to trans-eusiderin [63]. However, both (2S,3S )-33 and trans-eusiderin showed negative 1 Lb CEs, which suggests that they are homochiral if the ring substituents do not interfere. Since the C5 methoxy and C7 allyl substituents are not expected to invert the helicity rule, the absolute conﬁguration of trans-eusiderin most likely has to be revised to (2S,3S ). The 1 La region is more sensitive to substituent effects and overlapping from other chromophores, which may explain the opposite CEs of (2S,3S )-33 and trans-eusiderin (34) in this region. These examples conﬁrm that the benzodioxane helicity rule can be applied safely for unsubstituted derivatives or compounds in which the fused aromatic ring has alkyl or other substituents of low spectroscopic moments. However, ECD calculations are required whenever conjugated groups such as alkenyl or formyl are attached to the fused aromatic ring.

3.3.2. Isochroman Chromophore; P/M-Helicity → Positive/Negative 1 Lb CE Although the isochroman skeleton is far less common in natural products than the 2,3dihydrobenzo[b]furan one, there are several natural 3-alkylisochromans of remarkable biological activities whose absolute conﬁgurations have not been determined yet. For instance, absolute conﬁgurations of tricyclic derivatives 35a–c [64] and the anticoccidial optically active 3-methylisochroman derivative 36 [65], isolated from Penicillium sp., were not reported, as well as those of the topoisomerase II inhibitor CJ-12,373 (cis-37) [66] and the isochroman toxin trans-38 [67], natural 1,3-disubstituted isochromans with a benzylic hydroxy group (Chart 3.5). Moreover, there are several synthetic optically active isochroman derivatives reported with remarkable pharmacological activities such as selective 5-HT1D agonist [68–70], D1 agonist [71, 72], and D4 antagonist [73], which are promising for the treatment of migraine headache, Parkinson’s disease, and schizophrenia, respectively. It was shown that absolute conﬁgurations of these compounds play a decisive role in their pharmacological activities; the (S ) enantiomers of the 5-HT1D agonist 39a (PNU-109291) [68]

85

86

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

H

R1

O

R2 3

R 35 a b c

R1 OH OH OH

CH3 R2 R3 H H OH H H OH

4

5

HO 6 7

3 Me 2

O 8

n-C7H15

HO O

HOOC

1

5

4

8

O

Me

OH OH trans-38 OH

R2

3

2 1 O

7

Me

OH OH cis-37

OMe 36 R1 6

MeO

HO N

O

N

NH2 HCl

39 R2 R1 a CONHMe OMe b H SO2NH2

(1R,3S)-40 (A68930)

Chart 3.5. Structures of natural (35a-c, 36-38) and synthetic (39a,b and 40) isochromans.

and the selective D4 antagonist sonepiprazole 39b (U-101387) [73] possessed a superior afﬁnity for the binding site compared to the (R) enantiomers, while the (1R, 3S ) enantiomer of 40 (A68930) is almost exclusively responsible for the observed selective D1 agonist activity of the racemate [71]. In spite of the apparent importance of chirality in these derivatives, apart from the calculation of ECD parameters, there is no direct and general method for the conﬁgurational assignment of the isochroman skeleton available which can be used on μg quantity of a noncrystalline derivative. Thus, so far X-ray diffraction [72, 73] and correlations [68, 71] were applied to determine the absolute conﬁgurations of optically active isochromans. In order to establish a relationship between the helicity of the isochroman heteroring and the sign of the 1 Lb band CE and study the effect of ring substituents, rigid (41–43) and ﬂexible (45a–g) isochroman derivatives with known absolute conﬁguration and different ring substitution pattern were prepared and their ECD spectra were recorded (Figure 3.6, Scheme 3.1) [74]. Based on the synthetic steroid derivatives 41 and 42 (Figure 3.6a), the same helicity rule holding for tetralins and tetrahydroisoquinolines could be proposed for the isochroman chromophore having no aromatic substituents: P helicity of the heteroring leads to a positive Cotton effect (CE) in the 1 Lb band, and M -helicity is manifested in a negative

(b)

(a) 12 H H 11 O 12a H H 3 10a 6a 4 10 H H

H 12a

H 1 2

6a H

H 11 O

6a

O

11

12a H

P-helicity ωC–10a,C–11,O–12,C–12a>0

M-helicity ωC–10a,C–11,O–12,C–12a>0

positive1Lb CE

negative1LbCE

λmax [nm] Δε 6a–H 12a–H helicity α β M 274 (–0.05), 272 (+0.02), 270 (–0.08) α α P 272 (+0.33), 269 (+0.24) β β P 273 (–0.16), 268 (–0.08), 266 (–0.11) 1L CEs b

41 42 43

Figure 3.6. (a) Structures of steroid-fused isochroman derivatives 41–43 with the helicity of their heterorings and measured 1 Lb band CEs. (b) Helicity rule for the isochroman chromophore with no substituent on the aromatic ring represented on the example of 41 and 42.

87

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

one (Figure 3.6b) [74]. However, a negative 1 Lb band CE was measured for the 6aβ-H, 12aβ-H derivative 43 whose heteroring should adopt P -helicity, provided that ring A of the cholestane skeleton has chair conformation. This discrepancy was attributed to the contribution of a conformer with a boat conformation in ring A, but it made the former helicity rule ambiguous for the conﬁgurational assignment of natural isochromans. Thus 3-methylisochroman derivatives (+)-S -45a–g with different substitution pattern on the aromatic ring have been synthesized by ring closure of the (+)-(S )-1-arylpropane-2-ols 44a–f (Scheme 3.1), in turn prepared by kinetic resolution (44a–c) with Pseudomonas cepacia or chiral bioreduction (44d,e) [75]. Because the conﬁguration of the chirality center is retained during the oxaPictet–Spengler ring closure of optically active alcohols (+)-(S )-44a–f, the absolute conﬁgurations of isochromans 45a–f can be determined readily, provided that the absolute conﬁgurations of the arylpropanols 44a–f are known. For this purpose, we envisaged the employment of the zinc porphyrin tweezer exciton chirality CD [76] and the Mosher’s NMR method [77], both developed for the conﬁgurational assignment of secondary alcohols [75]. The helicity and ECD data of the isochromans (S )-45a–g and (4aR,10bS )-46 are tabulated in Table 3.4. (S )-45a and (4aR,10bS )-46, bearing no substituents on their aromatic rings, have heterorings of P - and M -helicity (see Scheme 3.1 for deﬁnition and representation), respectively, wherein the C3 methyl group of (S )-45 is oriented equatorially (J3H,4H = 10.9 and 3.1 Hz) while the 4a–H and 10b–H of the trans-annulated (4aR,10bS )-46 have a trans-diaxial conﬁguration [75]. Because the 1 Lb bands of (S )-45 of P -helicity and (4aR,10bS )-46 of M -helicity (Scheme 3.1, Table 3.4) show positive and negative CEs, respectively, it follows that the unsubstituted isochroman chromophore obeys indeed the helicity rule established for unsubstituted chiral tetralins and tetrahydroisoquinolines [32]: P /M -helicity of the heteroring results in positive/negative 1 Lb band CE , respectively. This corroborates well the similar spectroscopic moments of the hydroxymethyl (qCH2 OH = −5), aminomethyl (qCH2 NH2 = −5), and ethyl (qEt = +4.5) groups [8]. Since natural or synthetic isochroman derivatives of pharmacological interest often contain substituents with large spectroscopic moment (O-alkyl, hydroxy) on the fused

R4 R3

H

R4 Me

MeOCH2Cl

5

ZnCl2/Et2O 0°C

OR

R2

2

R2

R1 (+)-(S)-44a-f R2 R3 R1 H H a H OMe b OMe H H c OMe H H OMe OMe d -OCH2 OH e H OMe H f OMe H 45g H

44,45

H

4

R3

8

8a

H Me

Me

R5 H H H H H H H

O

1

2

ωC-8a,C-1,O-2,C-3 >0 P-helicity

4 H 5 3

1

4

O

R1 (+)-(S)-45a-g R4 H H H H H Br H

3

3

O 4a

6

10a 6a

2 10b 1 H 10

46

5

H 10b 7 8 9

O 6 4a

H

ωC-6a,C-6,O-5,C-4a <0 M-helicity

Scheme 3.1. Ring-closure to (S)-isochromans 45a–g and structure of (−)-(4aR,10bS)-46 and their helicities.

88

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

TAB L E 3.4. Helicity and ECD Data for Isochromans 45a-g and (−)-(4aR, 10bS )-46 Measured in Acetonitrile [75] CD: λmax [nm] (ε) Compound

Helicity

(S )-45a

P

(S )-45b

P

(S )-45c

P

(S )-45d

P

(S )-45e

P

(S )-45f (S )-45g

P P

(4aR,10bS )-46

M

1

Lb

272sh (+0.06), 2.68 (+0.11), 261sh (+0.09) 279sh (+0.45), 277 (+0.47), 275sh (+0.45) 277 (+0.29), 270sh (+0.21), 269sh (+0.21), 288sh (+0.65), 282 (+0.71), 271sh (+0.41) 295 (+0.98), 287sh (+0.91), 270sh (+0.27) 284 (+0.24), 278 (+0.23) 284sh (+0.35), 278sh, (+0.40), 275 (+0.40)

1

La

221 (+0.24)

1

B

n.d.

237 (+0.78), 219sh 209.5 (−3.04) (+0.40) 228 (−0.27) 215 (+1.98) 205 (−4.58) 230 (+3.44)

208 (+6.49)

243 (−1.78), 226 (+2.38) 204 (+4.75)

234 (+0.49), 228 (+0.38) 210 (−1.14) 230 (+1.34), 224sh n.d. (+1.06), 218sh (+0.93) 273sh (−0.05), 268 (−0.09), 222 (+0.86) 204 (+3.20) 262sh (−0.08)

aromatic ring, their conﬁgurational assignments by ECD spectroscopy required a systematic study whether the different substitution patterns will affect the unsubstituted helicity rule. In the substituted isochromans (S )-45b–g, the different substitution patterns by the methoxy, methylenedioxy, and bromine ring substituents provide different directions of the sum spectroscopic moment vector which allows us to study the effect of ring substituents and compare it with the corresponding data of the substituted chiral tetralins and tetrahydroisoquinolines. The analysis of ECD data of 45b and 45g (Table 3.4) revealed that in contrast to the tetralin and tetrahydroisoquinoline chromophore, 6,8-dimethoxy or 6-methoxy substitution did not invert the sign of the 1 Lb CE, and the same was found for 45c–f as well. This proves that substituents of large spectroscopic moment in different positions of the benzene moiety does not change the isochroman helicity rule and thus the 1 Lb band can be safely used for the conﬁgurational assignment of isochromans such as 35a–c and 36 (Chart 3.5), substituted on the aromatic ring by groups with large spectroscopic moment (alkoxy, OH, halogen). However, the signs of the CEs in the shorter-wavelength aromatic transitions, namely, 1 La and 1 B region of (S )-45b–g, varied with the position and nature of the substituents, making this region less useful for the determination of conﬁguration. In order to test the effect of axial benzylic substituents on the isochroman chromophore such as in trans-38 (Chart 3.5), the 1-methoxy derivatives (−)-(1R, 3S )-47a,b were prepared by oxidation of (+)-(S )-45a,d with 2,3-dichloro-5,6-dicyano1,4-benzoquinone (DDQ) in methanol (Scheme 3.2). The relative conﬁguration of the C1 methoxy group was studied by measuring three-bond carbon-proton coupling constants of (−)-(1R, 3S )-47b, which proved that the introduction of the methoxy group took place diastereoselectively. Due to the anomeric effect, only the trans product (−)-(1R, 3S )-47b was formed, in which the methoxy group is axially oriented, while

89

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

R1 (S)-45a,d

4

H

Me

6

Me

3

DDQ MeOH

R2 7

8

O

2

8a

3

1

1

O

OMe

R1

R2

a H b OMe

H OMe

1L

bCEs

H

2

(–)-(1R,3S)-47a,b

47

OMe

4

ωC-8a,C-1,O-2,C-3 >0 P-helicity

λmax [nm] Δε

266 (+0.03), 260 (+0.02), 255sh (+0.01) 287 (+0.59), 282 (+0.69), 275sh (+0.063)

Scheme 3.2. Preparation, preferred helicity, and ECD data of (−)-(1R,3S)-47a,b 1,3-disubstituted isochromans.

the heteroring has half-chair conformation [78]. The heteroring of (1R, 3S )-47a,b has P -helicity, and their positive 1 Lb CEs are practically the same as that of (S )-45a,d, which proved that the introduction of an axial benzylic alkoxy group does not change the isochroman helicity rule as long as the conformation or helicity of the heteroring remains the same. This result allows the conﬁgurational assignment of 1-alkoxy- or 1-hydroxyisochromans such as cis-37 and trans-38 from their ECD spectra. On the basis of the isochroman helicity rule, the absolute conﬁguration of pseudoanguillosporin A (48a) and B (49), isolated from the endophytic fungus Pseudoanguillospora sp, could be deduced from their ECD spectra (Chart 3.6 and Figure 3.7) [79]. Since pseudoanguillosporin A (48a) shows a negative 1 Lb band CE [284 nm (ε = −0.4 nm)], its heteroring adopts M helicity (Figure 3.7, right), which implies a (3R) absolute conﬁguration with equatorial C3 substituent. In contrast, the synthetic compound (3S )-45b [75] had a positive 1 Lb band CE and P -helicity. Both (3R)-48a and (3S )-45b showed a positive CE around 240 nm in the 1 La region, which is known to be more sensitive to the effect of achiral substituents of the aromatic ring. The absolute conﬁguration of pseudoanguillosporin B (49) at C3 was deduced as (R), since its CD spectrum was nearly identical to that of 48a. Moreover, the absolute conﬁguration of the C6 chirality center on the side chain of pseudoanguillosporin B, distant from the chromophore, was determined by the Mosher’s NMR method [79]. The interpretation of the ECD spectrum of compounds 48a and 49 (vide infra) is also supported by the calculation of the ECD spectrum of the model compound (R)-48b [79], which reproduced well the pattern of bands discussed above for (R)-48a (Figure 3.8). R4 3

R1

R O

HO

3

O OR R1 48a 48b

n-heptyl CH3

2O

6′

OH

1

2

R2 H H

1′

R3 R4 C-3 H CH3 R H CH3 R

OH (3R,6′R)-49

Chart 3.6. Structures of pseudoanguillosporins A (48a) and B (49), and the ECD model compound (R)-48b.

90

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

O2

0,8

4

(3S)-45b

Δε [cm2 mmol–1]

1 R 3 H ωC-8a,C-1,O,C-2 > 0 P-helicity

0,4 0,0 (3R,6'R)-49

H 3 R

–0,4 1

(3R)-48a

4

–0,8 –1,2

O 2 220

240

260

280

300

ωC-8a,C-1,O,C-2 < 0 M-helicity

λ [nm]

Figure 3.7. Left: Measured ECD spectra of (3R)-48a, (3R, 6 R)-49 and (3S)-45b in acetonitrile. Right: P and M-helicity of the isochromane ring; pseudoanguillosporin A (3R)-48a and B (3R, 6 R)-49 have M-helicity.

2

Δε [cm2 mmol–1]

1 0 Absolute minimum –1 Experimental for (3R)-48a in ACN Calculated with BP86/TZVP on (R)-48b

–2

(Boltzmann-weighted over 4 structures at 300 K)

–3 –4

220

240

260 λ [nm]

280

300

320

Second minimum (+0.10 kcal mol–1)

Figure 3.8. Experimental CD spectrum of pseudoanguillosporin A (48a) compared with the ECD calculated on model compound (R)-48b with TDBP86/TZVP as Boltzmann average over four DFT-optimized structures (B3LYP/6-31G(d)); the two most stable ones are shown on the right.

3.3.3. 2,3-Dihydrobenzo[b]furan Derivatives Unsubstituted 2,3-Dihydrobenzo[b]furan Derivatives; P/M-Helicity → Negative/Positive 1 Lb CE. While a tetralin derivative may be described as a benzene derivative with two alkyl substituents having the same magnitude of spectroscopic moments, this condition is not given for the 2,3-dihydrobenzo[b]furan chromophore, since the spectroscopic moment of the alkoxy moiety is larger than that of the alkyl part based on the spectroscopic moment of methoxy (qOMe = +21) and ethyl group (qEt = +4.5) as determined by Petruska [8]. Therefore, it is expected that the electric transition moment vector (μ) does not lie in

91

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

TAB L E 3.5. Helicity of the Heteroring and 1 Lb Band CEs for Rigid and Flexible 2,3-Dihydrobenzo[b]furan Derivatives Compound

Standard Projection

(−)-50

O C1

1L

b

CE {λ, nm (ε),a [θ]b }

Reference

M

297 (+0.89)a ,c

80, 81

P

289 (−3.30)a ,c

80, 81

P

280 (−2.51)a ,d

81, 82

M

292 (+2.95)a ,e

81–83

M M

288 (+0.72)a ,e 292 (−0.05)a ,c

81, 83 81, 83

P

281 (−0.37)a ,e

84

H

P P

a ,e

285 (−1.61) 300 (+0.46)a ,e

84 84

CH3

M

282 (+2658)b ,e

85, 86

C4 H H

H

(+)-51

Helicity

C1 H

O

C4

(+)-(2S , 3S )-52

H Ph O

CH3 H

(−)-(2R, 3S )-53 (−)-(2R, 3S )-54 (+)-(2R, 3S )-55 (+)-(2S , 3S )-56f f

(+)-(2S , 3S )-57 (+)-(2S , 3S )-58f

H O Ar

CH2

H

OH

H Ph O

CH3

(−)-(2R, 3S )-59f O Ph

H H

CE reported as ε. CE reported as θ . Solvent of ECD measurement is c n-hexane, d ethanol, e methanol, f revised absolute conﬁguration. Note: The standard projections show the deﬁnition of the helicity for the heteroring in the major conformer; M-helicity corresponds to negative ωC7a,O,C 2,C 3 torsional angle.

a b

the direction of the pseudo C2 axis of this chromophore, which can lead to an inversion of the original tetralin helicity rule. In order to obtain a correlation between the stereochemistry of chromophores (helicity of the heteroring) and the sign of the 1 Lb band CE, a stereocontrolled synthesis, conformational and ECD study of rigid [(−)-50, (+)-51] [80] and ﬂexible [52–55 (Chart 3.7)] [82, 83] 2,3-dihydrobenzo[b]furan derivatives with known absolute conﬁguration was performed. The 1 Lb band CD data and the preferred helicity of the ﬁve-membered O-heterocyclic ring in (−)-50, (+)-51, 52 (−)-53, 54 and (+)-55 are tabulated in Table 3.5. A comprehensive NMR study on the conformation of the dihydrofuran ring and ring A of the cholestane skeleton revealed that the heterocyclic ring of the cholestane derivatives (−)-50 and (+)-51 adopt M -and P -helicity, respectively. The known helicity of the heterocyclic rings in (−)-50 and (+)-51 and their measured 1 Lb band CE allowed us to set a helicity rule for the unsubstituted 2,3-dihydro-benzo[b]furan chromophore [80]. P /M -helicity of the heterocyclic ring leads to a negative/positive CE within the 1 Lb band; that is, the inverse form of the tetralin helicity rule is applicable, which is attributed to the large spectroscopic moment of furan O-1.

92

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Substituted 2,3-Dihydro-benzo[b]furans; Effect of Substitution. On the basis of the above helicity rule, the originally reported (2R, 3R) absolute conﬁguration of trans-norneolignans (+)-56 and (+)-57, isolated from Krameria cystisoides [84], had to be revised as (2S,3S ) (Chart 3.7, Table 3.5). Although both (+)-56 and (+)-57 have a 3hydroxyprop-1-yl aromatic substituent at C5, this was not expected to invert the helicity rule due to the small spectroscopic moment of alkyl groups. Since the ECD spectra of (+)-56 and (+)-57, published by Achenbach et al. [84], exhibit a negative CE at 281 and 285 nm, respectively, their heteroring should adopt P helicity. Taking into account that the C2 and C3 substituents are equatorial in both cases (3 J2H3H 8.9 and 9.3 Hz), their absolute conﬁgurations are (2S ,3S ). Because (+)-conocarpan [(+)-58] was chemically correlated with (+)-(2S , 3S )-56, its absolute conﬁguration had to be revised to (+)-(2S , 3S ) as well [80, 81]. However, (+)-conocarpan [(+)-58] has a positive 1 Lb CE at 300 nm (Table 3.5), which suggested that a conjugated C5 1-propen-1-yl substituent inverts the helicity rule. With (+)-(2S , 3S ) conﬁguration, the heteroring of (+)-conocarpan preferably adopts P -helicity to ensure the low-energy quasi -equatorial arrangement of the C2 and C3 substitutents; that is, the positive 1 Lb CE derives from P -helicity of the heterocyclic ring. This observation corroborates well the large spectroscopic moment of the 1-propen-1-yl group (q = +15). Recently, conocarpan (58) possessing a wide range of biological activites [87–89] have attracted the attention of synthetic chemists and enantioselective synthesis of (−)-(2R, 3R)-conocarpan [90, 91], and its enantiomer [(+)-(2S,3S )-58] [86] was reported, which unequivocally conﬁrmed our conﬁgurational assignment by ECD. (−)-Epi -conocarpan [(−)-59], the C2 epimer of (+)-conocarpan [(+)-58] isolated from roots of Piper regnelli [85], was synthesized and converted to (+)-(2S, 3S )-conocarpan [(+)-58] [86], which conﬁrmed its (−)-(2R,3S ) absolute conﬁguration (Chart 3.7, Table 3.5). (−)-(2R,3S )-epi -conocarpan has a positive 1 Lb CE at 282 nm and a heteroring of P -helicity with an equatorial orientation of the C2 aryl group, which does not follow the expected helicity rule. The cis orientation of the C2 and C3 substituents forces the dihydrofuran ring into a nearly planar conformation and thus the third sphere contribution and especially that of the pseudo-axial C3 methyl group most likely overrides the second sphere contribution—that is, the helicity of the heteroring. Flexible unsubstituted 2,3-dihydrobenzo[b]furans (+)-52, (−)-53 and (−)-54 prepared in a stereocontrolled manner also conﬁrmed the validity of the helicity rule

6 5

7 7′a

4

1

H 1 O

4′a

O

2A 34 5

H

H

6 5

H

H

(−)-50

H

(+)-51

O

(+)-(2S,3S)-52

7

A

R2

1

O2 3

B

R3

CH2OH (−)-(2R,3S)-53: R1,R2,R3= H (−)-(2R,3S)-54: R1 = H,R2 = OMe, R3 = OH (+)-(2R,3S)-55: R1 = OMe,R2 = OMe, R3 = OH

R1 OH

HO

R1

7 7a 1 O2 Ph 3 4a 4 CH3

O

CH3

CH3

(+)-(2S,3S)-56: R1 = H (+)-(2S,3S)-57: R1 = OMe

(+)-(2S,3S)-58 (+)-conocarpan

O OH

OH CH3 (–)-(2R,3S)-59 (–)-epi-conocarpan

Chart 3.7. Structures of 2,3-dihydrobenzo[b]furan derivatives for Table 3.5.

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

Ph O

93

CH3

H CH3 H

P-helicity 2,3-trans relative configuration pseudoequatorial Ph and Me groups

O H

H Ph

M-helicity 2,3-trans relative configuration pseudoaxial Ph and Me groups

Figure 3.9. Equlibrating conformers of (+)-(2S, 3S)-52 with the P-helicity conformer as the dominant one.

established for the unsubstituted chromophore based on the steroid derivatives (−)-50 and (+)-51. The P - or M -helicity of the heterocyclic ring is controlled by the equatorial arrangement of the phenyl group at C2 (Figure 3.9), whose contribution to the ECD is less signiﬁcant compared to that of the 2,3-dihydrobenzo[b]furan chromophore, possibly because of increased mobility and distance relatively to the chirality elements. Thus, the substitution pattern of the C2 phenyl (ring B) does not inﬂuence signiﬁcantly the 1 Lb band CE either; (−)-53 and (−)-54 have consistently positive 1 Lb band CE. On the contrary, the substitution of the aromatic ring A at C7 position by a methoxy group (qOMe = +21) [8] changes the sign of the 1 Lb band CE, since the homochiral (−)-54 and (+)-55 show opposite signs for the M -helicity of the heterocyclic ring. The weak negative 1 Lb band CE of (+)-55 proved that a substituent at C7 such as a methoxy group possessing a large spectroscopic moment (q) also reversed the helicity rule; that is, P /M -helicity of the heterocyclic ring leads to positive/negative 1 Lb band CD, respectively. Although chiroptical methods (ECD, ORD, and optical rotations) are extensively used in the conﬁgurational assignment of natural 2,3-dihydrobenzo[b]furan neolignans, the number of publications in which the absolute conﬁguration of a neolignan was determined independently, by X-ray or chemical correlation, and its ECD was also measured, are very limited. Rare examples were presented by Yuen et al. [92], who characterized synthetic neolignans 60, 61, 64, and 65 by ECD and also determined their absolute conﬁgurations unambiguously by X-ray analysis and chemical correlations (Chart 3.8). The negative 1 Lb band CE of (−)-(2R, 3S )-60 conﬁrmed our results regarding the effect of a methoxy group at C7 (Table 3.6), since its heterocyclic ring adopts M -helicity due to the equatorially oriented bulky aryl group at C2. The 7-methoxy group is quite a common substituent in numerous dihydrobenzo[b]furan neolignans, and thus its effect on the sign of the 1 Lb band CE could lead to many erroneous conﬁgurational assignment if not properly taken into account. The ECD data of the homochiral analogue (−)-(2R, 3S )-61 revealed that the introduction of an additional conjugated 3-hydroxy-1-propen-1-yl group at C5 does not induce a further change in the sign of the 1 Lb band CE in the presence of a 7-methoxy group. The ECD data of (−)-(2R, 3S )-60 and (−)-(2R, 3S )-61 also proved that the published absolute conﬁgurations of (−)-62 [93] and (+)-63 [84] are incorrect. In fact, ring B substituents do not inﬂuence the sign of the 1 Lb CE; therefore measured positive 1 Lb band CE for (−)-62 and (+)-63 should stem from P -helicity, which implies (−)-(2S , 3R) and (+)-(2S , 3S ) absolute conﬁguration, respectively [81]. The neolignan licarin B, differing from (+)-63 in the substitution of ring B (3,4-methylenedioxy instead of 4-hydroxy), was used by Achenbach et al. [84] as an ECD reference, but its absolute conﬁguration published by Aiba et al. [97] also has to be revised. In (−)-(2R, 3S )-64 and (−)-(2R, 3S )-65, a C5 α,β-unsatured carbonyl moiety is conjugated with the benzene ring and the resultant chromophore cannot be considered identical with the previous ones. If the intense negative CEs around 330 nm are

94

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

OMe

OMe

OMe

7

O OMe HO

HO (CH2)3

5

O

R1

3

OH

2

CH2OH

CH2OH R

(−)-(2R,3S)-60

1

CH3 2

(−)-(2R,3S)-61: R = OH, R = H (−)−(2S,3R)-62: R1,R2 = OMe (woorenogenin)

OMe

(+)-(2S,3S)-63

OMe

HO 6

O

O R

OMe

OMe

1

O2

CH2OH

OH MeOOC 5

OH HO CH3

OH

(CH2)3 5

(+)-(2S,3S)-66

(−)-(2R,3S)-64: R = CHO (−)-(2R,3S)-65: R = COOMe

O

(2S)-67 (S)-hexahydromarmesin R2

OMe R1O

O

O Ar

HO

OH (2S)-68 wutaiensol

MeO CH3 trans-(2R,3R)-69 R1 = H or allyl R2 = H or allyl

Chart 3.8. Structures of 2,3-dihydrobenzo[b]furan derivatives for Table 3.6. Ar in (2R, 3R)-69 is O-methyl-O, O-methylenepyrogallyl, piperonyl, or tri-O-methylpyrogallyl group.

TAB L E 3.6. Helicity of the Heteroring and 1 Lb Band CEs for 2,3-Dihydrobenzo[b]furan Derivatives Compound (−)-(2R, 3S )-60c (−)-(2R, 3S )-61c (−)-(2S , 3R)-62d (+)-(2S , 3S )-63d (−)-(2R, 3S )-64c (−)-(2R, 3S )-65c (+)-(2S , 3S )-66d (2S )-67c (2S )-68 (2R, 3R)-69d

Helicity M M P P M M P P P M

CE {λ, nm (ε)a or [θ]b} 296 (−1739)b 279 (−11372)b 265 (+1.0)a 295 (+1.36), 270 (+5.30)a 335 (−9862), 279 (+2240)b 376 (+806), 327 (−9884)b 263sh (+2600)b 292 (−2435)b 272 (+3223)b negative CE from ORD

Reference 92 92 93 84 92 92 85 94 95 96

reported as ε. reported as θ . c absolute conﬁguration was determined independently from ECD data. d Revised absolute conﬁguration. a CE b CE

taken into account, these compounds apparently follow the inverse form of the helicity rule but accompanying weaker opposite CEs indicate that the n –π ∗ transitions can make the assignment ambiguous. The neolignan (+)-(2S , 3S )-66 (Chart 3.8, Table 3.6) possesses a C5 methoxycarbonyl group, which has a large negative spectroscopic moment

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

(qCOOMe = −17) [8]. It was shown earlier that a substituent with large spectroscopic moment at C5 such as a 1-propen-1-yl group (q = +15) inverts the helicity rule of the unsubstituted chromophore. Thus a similar behavior may be expected from a C5 methoxycarbonyl substitutent; that is, P-helicity results in a positive 1 Lb CE as shown in Table 3.6. This assumption is supported by the fact that (+)-66 was isolated together with (+)-(2S , 3S )-conocarpan (58) from Piper regnellii [85], which suggests that they are most likely homochiral derivatives. The absolute conﬁguration of (S )-hexahydromarmesin [(2S )-67] was deduced unequivocally by a chemical correlation to (S )-marmesin [94], and the negative 1 L band CE corresponds to P -helicity of its heteroring with pseudoequatorial C2 b substituent. This implies that a C6 hydroxy substituent (qOH = +20) does not invert the helicity rule of the unsubstituted chromophore. The related (2S )-68 neolignan, wutaiensol, has a C5 3 -hydroxy-1-propen-1-yl and a C5 OMe substituent, the presence of which inverts the unsubstituted helicity rule as shown earlier. Thus the published (S ) absolute conﬁguration is in accordance with our ﬁndings, since it shows positive 1 Lb CE and its heteroring has P -helicity [95]. Gottlieb et al. [96] classiﬁed benzofuranoid neolignans into structurally homogeneous groups by constitution and ORD curves and proposed their conﬁgurational assignment. The structure (2R, 3R)-69 represents a group of neolignans with 5,6-dioxygenated benzo[b]furan chromophore. Because we have already shown that the presence of a C5 substituent with large spectroscopic moment inverts the unsubstituted helicity rule, the same is expected for the 5,6-dioxygenated chromophore. Accordingly, negative 1 Lb CEs of 69 derivatives originates from M -helicity and (2R, 3R)-69 absolute conﬁguration. As a consequence, the absolute conﬁgurations assigned to these neolignans may have to be revised as denoted in Chart 3.5. There are numerous publications that used incorrect absolute geometries and ECD results of the presented structures for the conﬁgurational assignment of isolated dihydrobenzo[b]furan neolignans. Whenever the effect of an aromatic substitution pattern on the sign of the 1 Lb CE was not taken into account and an improper ECD reference compound was chosen for comparison, the determination of absolute conﬁguration by ECD data led to incorrect absolute geometry [84, 85, 93, 96–104]. In contrast, there are recent publications that use the dihydrobenzo[b]furan helicity rule [80, 81] properly taking care of the substitution pattern [54, 105–107]. As an example, the C2 and C3 absolute conﬁguration of difengpiol A (70), a neolignan with a nonprecedented C2 cyclohexenediol and a C7 methoxy substituent, was determined as (2S , 3R) on the basis of the positive 1 Lb CE and P -helicity of its heteroring (Chart 3.9). The 9,10-dihydrophenanthrofuran derivative (−)-pleionesin A (71) was isolated from the orchid Pleione yunnanensis together with three related derivatives [108] and the dihydrobenzo[b]furan helicity rule [80, 81] was applied to determine the absolute conﬁguration. However, (−)-pleionesin A (71) contains an inherently chiral biphenyl chromophore with an axial chirality element along the biaryl axis instead of the benzene chromophore of dihydrobenzo[b]furans, and therefore the helicity rule cannot be utilized safely for the conﬁgurational assignment. Similarly, the helicity rule should not be applied whenever there is an additional benzylic chirality center in the dihydrobenzo[b]furan lignan [109, 110] as exempliﬁed by (−)-radulignan (72) [110], in which the contribution of the benzylic chirality center may override that of the dihydrobenzo[b]furan moiety rendering the assignment ambiguous.

95

96

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Our observations in Tables 3.5 and 3.6 can be concluded in the following points and are summarized in Figure 3.10: 1. P /M -helicity of the heterocyclic ring results in negative/positive 1 Lb CE, respectively, if the aromatic ring has no other ring substituent (we refer to this statement as the unsubstituted helicity rule). 2. Substituents with low spectroscopic moment such as an alkyl or 3-hydroxypropan1-yl at C5 and possibly in other positions do not change the unsubstituted helicity rule. 3. A hydroxyl or alkoxy group at C6 does not invert the unsubstituted helicity rule even in the presence of a C4 alkyl group. 4. 2,3-dihydrobenzo[b]furans with C5 substituents having a large spectroscopic moment such as 1-propen-1-yl, CHO, COOMe, OH, OMe follows the inverse helicity rule. 5. A C7 methoxy group in the presence or absence of C5 1-alken-1-yl substituent also causes inversion. 6. Cis-dihydrobenzo[b]furans do not follow the expected helicity rule because the contribution of the pseudoaxial benzylic substituent, belonging to the third sphere, determines the sign of the 1 Lb CE.

3.3.4. Chroman Chromophore; P/M-Helicity → Negative/Positive 1 Lb CE In chroman derivatives, the magnitude of the spectroscopic moment belonging to the alkoxy part of the molecule (qOMe = +21) [8] is approximately ﬁve times larger than that of the alkyl moiety (qEt = 4.5) [8]. Therefore, similarly to the unsubstituted dihydrobenzo[b]furan chromophore, the sum vector is rotated by more than 30◦ and a sign inversion of the tetralin helicity rule is expected (Figure 3.11). On the basis of the spectroscopic moment belonging to the N -methyl group (qNHMe = 27) [8], one can predict that the same rule holds for tetrahydroquinoline chromophore as well. Structures, ECD data, helicity, and references of selected synthetic and natural chroman and tetrahydroquinoline derivatives are tabulated in Chart 3.10 and Table 3.7. The 1 Lb band CEs of rigid synthetic steroid model compounds (−)-73 and 74 and the ﬂexible (−)-(S )-ﬂavan [(S )-75, see conformation in Figure 3.12a], prepared with known absolute

OH OMe HO

7

6 5

1

O 2 1′ 3

4

4′

OH

3′ 2′ CH2OH OH

HO HO

O O O

(+)-(2S,3R,3′S,4′R)-70 difengpiol A 286 nm (+0.49) HO

OMe

OH

MeO

OMe

(−)-pleionesin A (71) 290 nm (+0.53)

Chart 3.9.

O O OH

OH

O

CH2OH OH (−)-radulignan (72) 290 nm (+5.0)

97

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

conﬁguration, conﬁrmed unambiguously the above assumptions. In fact, P/M-helicity of the chroman or tetrahydroquinoline chromophore having no aromatic ring substituents results in a negative/positive 1 Lb band CE, respectively (Chart 3.10, Table 3.7) [81, 111]. The helicity rule is also applicable to the cis-4-hydroxyﬂavan (−)-(2R,4R)-76, in which both the C2 phenyl and C4 hydroxy groups adopt equatorial orientation if the heteroring has half-chair conformation with M -helicity (Figure 12b) affording a positive 1 Lb band CE [81, 112]. Since the chroman chromophore frequently has aromatic ring substituents of large spectroscopic moment (such as hydroxy or methoxy groups) in natural products, the effect of ring substitution on the chroman helicity rule has to be addressed as well. Although the corresponding 5- and 7-hydroxy or -methoxy substituted rigid model compounds have not been synthesized, the effect of substitution on the chiroptical properties can be assessed from the published ECD data of natural chroman derivatives [113]. The comparison of the ECD data of (+)-catechin [(2R, 3S )-77], (−)-rubinetinidol [(2R, 3S )78], (+)-afzelechin [2R, 3S )-79], all of which have P -helicity and negative 1 Lb band CE (Figure 3.12c), with the ECD data of unsubstituted derivatives [(−)-73, (−)-75] of P -helicity proves that neither 5-hydroxy nor 5- and 7-dihydroxy substitutions of the ring A have an inﬂuence on the sign of the 1 Lb CE [81]. The same conclusion can be made

R

4

R O

R6 7 6

5

5

O1

2

3 4 R3

R1

R2

O

1

P-helicity positive 1Lb CE inverse helicity rule

P-helicity negative Lb CE unsubstituted helicity rule −R3,R4,R5,R6 = H

−R4 = 1-propen-1-yl, R3,R5,R6 = H

−R4=

−R6 = OMe, R3,R4,R5 = H or alkyl

3

alkyl or 3-hydroxypropan-1-yl 5

6

−R6 = OMe, R4 = 3-hydroxyprop-1-en-1-yl

R ,R ,R = H −R5

4

3

6

R3,R5 = H

= OH, R = alkyl, R ,R = H

−R4 = CHO or COOMe, R3, R5,R6 = H* −R4,R5 = OH or OMe, R3,R6 = H

Figure 3.10. Dependence of the 2,3-dihydrobenzo[b]furan helicity rule on the substitution pattern of the aromatic ring. *Benzaldehyde or benzoic acid methyl ester chromophore instead of the benzene.

X X chroman: X = O tetrahydroquinoline: X = NH

X μ = Σq

P-helicity negative 1Lb CE

Figure 3.11. Platt polarization diagram of the 1 Lb band for chroman (X = O) and tetrahydroquinoline (X = NH) chromophores having no substituents on the aromatic ring. P-helicity results in a negative 1 Lb band CE. Smaller arrows represent the spectroscopic moment vectors (q), while the longer one represents the electric transition moment vector (μ).

98

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

H1

1 H

2 3

R1

O

4

4

R1

R4

O

OAc

H

H

R2

(+)-(2R,3S)-80: R1,R2,R4 = H, R3 = OMe (−)-(2R,3S)-81: R1,R3 = OMe, R2,R4 = H (+)-(2R,3S)-82: R1,R2,R3,R4 = OMe

OAc

5

R2

4 H

B

R5

OH

R1 (2R,3S)-77: R1,R2,R3,R4 = OH, R5 = H (+)-catechin R4 (2R,3S)-78: R1,R3,R4,R5 = OH, R2 = H (−)-rubinetinidol (2R,3S)-79: R1,R2,R4 = OH, R3,R5 = H (+)-afzelechin

HO 6

R1 R

6

A

1 H O C3 2

(+)-(2S,3S)-83: R1,R2,R4 = H, R3 = OMe (+)-(2S,3S)-84: R1,R3 = OMe, R2,R4 = H (+)-(2S,3S)-85: R1,R2,R3,R4 = OMe

O

HO 7

R1 7

3

OH (−)-(2R,4R)-76

(−)-(S)-flavan (S)-75 R3

H

H

R2

2 3

H4 H (−)−73: X = O 74: X = NH

R4

O

2

X

R2

1 H

O

O

2

(R)-86: R1,R2 = H (S)-87: R1= H,R2 = OMe (S)-88: R1 = Me,R2 = OMe R1

O C 3

4′

8′

O

8

CH3

(−)−(R)-89

A

2

OH H

R2

(2R,4′R,8′R)-90 δ-tocopherol H6

O 2 1 Me R3 (−)-(S)-91: R1,R2,R3 = H (+)-(2R,6S)-93 (+)-(S)-92: R1,R2,R3 = OMe H

B

Chart 3.10. Structures of chroman derivatives 73–93.

on the basis of ECD data of synthetic trans-3-acetoxyﬂavans (2R, 3S )-80–82. Cis-3acetoxyﬂavans (2S , 3S )-83–85 would apparently also follow the helicity rule if their heterorings had half-chair conformation with M -helicity and equatorial C2 aryl group. However, the small coupling constants of 2-H and 3-H [for (2S , 3S )-83, 3 J2,3 = 1.5 Hz, 3 J3,4 = 4.5 and 3.0 Hz) [114] suggest that the heteroring preferably adopts a boat or twist boat conformation with axial or pseudo-axial C2 aryl group. The comparison of the ECD data of (−)-(S )-75 with those published for (R)-7-hydroxy-ﬂavan 86 [115], having M helicity and a positive 1 Lb band CE, revealed that 7-hydroxy substitution does not invert the original rule either [81]. The positive 1 Lb CE of (R)-86 served as a reference to determine the correct absolute conﬁguration of ﬂavans (S )-87 and (S )-88 (Figure 3.12a), although the possible effect of an additional C5 methoxy group was not considered [115]. The ECD data of (−)-(R)-89 [116] and δ-tocopherol [(2R, 4 R, 8 R)-90] [117] corroborate the chroman helicity rule and the case of δ-tocopherol also demonstrates that the presence of a C6 hydroxy substituent does not put a limit to the empirical rule (Figure3.12d,e). Reported ECD data of synthetic [118] and natural [120] isoﬂavans such as those of (−)-(S )-91 and (+)-(S )-92 would suggest that isoﬂavans surprisingly follow an inverse helicity rule with respect to ﬂavans; P /M -helicity of the heteroring gives positive/negative 1 Lb CE [121]. For instance, the heteroring of synthetic derivative (+)(S )-92 has a half-chair conformation of M -helicity with an equatorial C3 aryl group (Figure 3.12f), and it shows an intense negative CE at 288 nm accompanied by weak

99

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

TAB L E 3.7. Helicity and ECD Data of Chroman Derivatives Compound (−)-73c 74c (−)-(S )-75 (−)-(2R, 4R)-76 (+)-(2R, 3S )-77 (−)-(2R, 3S )-78 (+)-(2R, 3S )-79 (+)-(2R, 3S )-80 (−)-(2R, 3S )-81 (+)-(2R, 3S )-82 (+)-(2S , 3S )-83 (+)-(2S , 3S )-84 (+)-(2S , 3S )-85 (R)-86 (S )-87 (S )-88 (−)-(R)-89 (2R, 4 R, 8 R)-90 (−)-(S )-91 (+)-(S )-92 (+)-(2R, 6S )-93

Helicity

CE {λ, nm (ε)a or [θ]b}

References

M M P M P P P P P P

277 (+1.83)a 308 (+3.14)a 276 (−1.01)a 276 (+1.28)a 282 (−0.36)a 284 (−1.50)a 270 (−0.44)a 272 (−6800)b 284 (−10000)b 279 (−5100)b 274 (+6800)b 273 (+2600)b 271 (+1100)b Positive 1 Lb CE Negative 1 Lb CE Negative 1 Lb CE 278 (+0.09), 270 (+0.06)a 298 (−0.33)a 282sh (−0.60), 275 (−0.63)a,f 300 (−78), 288 (−4780), 274 (+17)b 283sh (−0.57), 276 (−0.67)

81, 111 81, 111 81, 111 81, 112 113 113 113 114 114 114 114 114 114 115 115 115 116 117

e e e

M P P M P M M M

d

118 119

CE reported as ε. CE reported as θ . c with cholestane skeleton. d unpublished ECD data. e not determined, conformation of heteroring is different from half-chair. f1 La CE: 228 (−1.39). a b

transitions at higher and shorter wavelength (Table 3.7). This ﬁnding is contradictory to the expectations, since the preferred conformation of the heteroring does not deviate from the half-chair, there is no axial benzylic substituent, and the contribution of the third sphere is not considered signiﬁcant. In order to explain this discrepancy, TDDFT calculations were carried out on the synthetic isoﬂavan (−)-(S )-91, and, for comparison, on ﬂavan (−)-(S )-75. In both cases, DFT-optimized geometries (at B3LYP/6-31G(d) level) were employed as input structures, which showed the heteroring in the expected halfchair conformation with helicity depicted in Figure 3.12a,3.12f. Surprisingly, none of the low-energy transitions of isoﬂavan (S )-91 are of pure 1 Lb character as found for ﬂavan (S )-75. On the contrary, three transitions predicted by B3LYP/TZVP calculations between 247 and 257 nm result from a combination of exciton-coupled excitations centered on the two aromatic rings, plus charge-transfer transitions. Since exciton-coupled interactions are determined by the relative orientation of the electric transition moments, which is in turn are inﬂuenced by the substitution pattern of ring A and B, the ECD of isoﬂavans with different substitutions cannot be compared easily and ECD calculations should be considered for a safe conﬁgurational assignment. On the basis of the above considerations, the determination of absolute conﬁguration for some recently reported isoﬂavan derivatives 94–98 using the inverse chroman helicity rule of isoﬂavans [121] is possibly prone to error (Chart 3.11) [122–125]. This is especially true in the presence of a conjugated double bond ring substituent as in desmodin A [(−)-94] [122] and glabridin (96)

100

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

H

(a) Ar O

H

OH

O Ph

(e)

OH

(g)

H

2

Ar

Me (2R,4′R,8′R)-90 P-helicity negative 1Lb CE

(R)-89 M-helicity positive 1Lb CE

H (2R, 3S)-3-hydroxyflavans 77-79 (2R, 3S)-3-acetoxyflavans 80-82 P-helicity negative 1Lb CE

(f) O R

O

Ar O OR

H H (2R,4R)-76 M-helicity positive 1Lb CE

(S)-flavans 75,87,88 P-helicity negative 1Lb CE (d)

H

(c)

(b)

6

O (S)-isoflavans 91,92 M-helicity negative 1Lb CE

Me O

(2R,6S)-93 M-helicity negative 1Lb CE

Figure 3.12a–g. Preferred low-energy conformations and helicity of chroman derivatives viewed from the direction of the fused aromatic ring.

O

O

O

O

Me

O

O

O

OMe

OH

Me OMe

OH

desmodin A [(−)-94]

OMe

OH

OH

desmodin B [(+)-95]

glabridin (96)

OMe

O

O

OH

HO OH isoflavan-4-ol (97)

(+)-98

Chart 3.11. Structures of isoﬂavans 94–97 and the homoisoﬂavan derivative (+)-98.

[123] or with an additional chiral dihydrobenzo[b]furan ring as in desmodin B [(+)-95] [122]. The bridged tricyclic tetrahydro-2,6-methano-2H -1-benzoxocine derivative (+)(2R, 6S )-93 also follows the inverse helicity rule; that is, its negative 1 Lb CE (Table 3.7) corresponds to a ﬁxed half-chair conformation of M -helicity (Figure 3.12g) [119]. Due to the bridged structure, the C2–C3 and C6–C5 bonds are axially oriented and the contribution of the third sphere presumably overrides that of the second sphere (helicity of the heteroring), thus inverting the helicity rule. As a summary, the same helicity rule (P /M -helicity negative/positive 1 Lb CE) was found for the unsubstituted chroman chromophore as for the unsubstituted dihydrobenzo[b]furan, but, in contrast to the latter, methoxy or hydroxy ring substituents do not change the correlation. The chroman helicity rule can be safely utilized for the conﬁgurational assignment of ﬂavans, cis-4-hydroxyﬂavans, trans-3-oxygenated ﬂavans, and 2-alkylated chromans [81, 121], while the helicity of the heteroring is not decisive for the 1 Lb CE of isoﬂavan and bridged tetrahydro-2,6-methano-2H -1-benzoxocine derivatives.

101

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

3.4. TETRALONE, DIHYDROISOCOUMARIN, AND CHROMAN-4-ONE CHROMOPHORES In this section, the carbonyl derivatives of tetralin, isochroman, and chroman—that is, tetralones, dihydroisocoumarins, and chroman-4-ones—are discussed, in which besides the π –π * transition, the carbonyl n –π * transition is frequently used for determination of absolute conﬁguration.

3.4.1. Tetralone Derivatives Tetralone (3,4-dihydronaphthalen-1(2H )-one) derivatives, containing both a tetraline and acetophenone chromophore, are widespread in nature [126–134], and their n –π * transition of the conjugating carbonyl group above 300 nm may be used for the determination of absolute conﬁguration instead of the 1 Lb band, since its sign is expected to be independent from the substitution pattern of the aromatic ring [112]. The fused carbocyclic ring of tetralones most likely adopts an envelope conformation with C3 out of the plane of the benzene ring [127], the conformation (helicity) of which is expected to determine the sign of the n –π * CE as found also for natural ﬂavanones and isoﬂavanones [121]. ECD data and preferred conformation of the carbocyclic ring of natural tetralones (Chart 3.12) are tabulated in Table 3.8. For tetralones 99–103 and 105, 106 having diverse substitution pattern, the M -helicity of the fused carbocyclic ring with envelope conformation results in positive n –π* CEs, although (2S , 4R)-100 had a very weak negative low-energy CE. In contrast, tetralones 104 and 107–110 show contradictory correlations; P -helicity of the heteroring affords positive n –π * CEs. Although absolute conﬁgurations were determined independently from ECD study by either Mosher’s NMR method or X-ray analysis for 99, 100, 103, 104, 108 and 109, some of these derivatives also show an inconsistent relationship between the helicity of the nonaromatic ring and the sign of the n –π * CE. This ﬁnding suggests that in contrast to the case of ﬂavanones and isoﬂavanones [121], there is no straightforward general correlation between the sign of the n –π * CE and conformation of the nonaromatic ring. A possible explanation can be that conformers other than the represented envelope one as well as the position and nature of the substituents may play a non-negligible role, which renders the conﬁgurational assignment difﬁcult if based only on the helicity of the nonaromatic ring or a simple comparison of the ECD

O

O

OH O OH

CH2 OH

H3C

O

CH3

OH

R1 R2 R1 R2 callianthone A [(2S,4R)-101]: R1= OH, R2= H (2S,4S)-99: R1= H, R2 = OH (2S,4R)-100: R1 = OH, R2 = H callianthone B [(2S,4S)-102]: R1= H, R2= OH O

O H3C

O

O

O OH (2R,4S)-104

hemiculone [(3S,4S,1′R)-103]

OH O

OH O Me

H3C

Me OH (3S,4S)-105

Me OH

OH

pyrolone A [(3S,4S)-106]

Me R1 R2 OH pyrolone B cis-isoshinanolone 1 2 [(2R,4S)-107] (3R,4R)-108: R = OH, R = H trans-isoshinanolone (3R,4S)-109: R1= H, R2= OH

O

Ar

OH

Chart 3.12. Structures of tetralone derivatives for Table 3.3.

(3R)-110 Ar: 3,4-dihydroxyphenyl

102

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

TAB L E 3.8. Conformation of the Carbocylic Ring with Orientation of the Substituents and Their n –π * CEs Low-Energy Conformera Helicityb CE {λ, nm (ε),c [θ],d θ e}

Compound (2S , 4S )-99

f

(2S , 4R)-100f

301 (+272)d

126

OH

M

341 (−285)d 304 (+520)

126

OH

M

324 (+5.29)e

127

H

M

345 (+1.68)e

127

M

301 (+1.5)c

128

P

321 (+4160)d

131

M

333 (+1200), 323 (+1070)d

131

M

338 (+2.4), 316 (+0.8)e

130

P

323 (+2.4)

130

P

331 nm positive CEi

129

P

positive CEi

129

M

320 (−2.2)a

133

H

O H

OH

HO O H

HO O H

CH3

(2S , 4S )-102

M

HO

H

(2S , 4R)-101

Reference

HO O

OH

CH3

(3S , 4S , 1 R)-103f

H

OH O

O

H

(2R, 4S )-104f

H

H

O

OH

Me

(3S , 4S )-105g

H

Me

O

H

OH

(3S , 4S )-106 H

(2R, 4S )-107

H

O

OH

Me

(3R, 4R)-108h

OH O OH

H

H

(3R, 4S )-109h

H O OH

OH

H H

(3R)-110

Ar O

a

Envelope conformer is in equilibrium with the distorted chair but their helicities are identical. Deﬁned by the sign of the ωC5a,C4,C3,C2 torsional angle. c CE reported as ε. d CE reported as θ . e CE reported as θ ε. f Absolute conﬁguration was determined by Mosher’s method. g The reported absolute conﬁguration is probably wrong and should be reassigned as (3R, 4R)-105. h Determination of absolute conﬁguration is based on X-ray diffraction analysis of a related derivative. i From HPLC-CD. b

103

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

spectra. In this case, the conﬁgurational assignment is especially prone to errors and a thorough conformational analysis and ECD calculation cannot be spared. As an example to demonstrate what was stated, we analyzed the conformation and ECD spectra of compounds 100, 104, 105, and 109 with molecular modeling (MMFF conformational search and DFT optimization at B3LYP/6-31G (d) level) and TDDFT ECD calculations. For tetralones 100, 104, and 109, the substituents of the carbocyclic ring adopt a preferred equatorial position in the lowest-energy conformer (populations above 85% at 300 K) which constrains the ring in a more or less ﬁxed envelope conformation with well-deﬁned helicity (Table 3.9). The result of ECD calculations run with TDDFT method, B3LYP/TZVP level, on the three compounds are also shown in Table 3.9. Apparently, ECD calculations reproduce experimental data for 100, 104, and 109, apart from a wavelength shift, and conﬁrm the assigned absolute conﬁguration. However, at least in the case of tetralone (2S , 4R)-100 the assignment based on the n –π *

TAB L E 3.9. Summary of Geometry Optimizationsa and CD Calculationsb for Selected Tetralones Compound (2S , 4R)-100

Lowest-Energy Conformerc (Population)c

Structure OH O

1L

Helicity

n –π * CE λ (sign)

b CE λ (sign)

M

285 (+)

300 (−)

P

317 (+)

261 (−)

M

331 (−)f

273 (−)f

P

298 (+)a

295 (−)a

OH

HO

OH

O H

(86%)d

OH

(2R, 4S )-104

H

O

H

H

O

OH

Me (89%)d

OH

(3S , 4S )-105c

O

H

Me

O

H

OH (85%)e

OH

(3R, 4S )-109

OH

O

H O OH OH a With

OH

H (87%)d

B3LYP/6-31G(d) on MMFF-calculated low-energy minima. B3LYP/TZVP. c Optimizations with B3LYP/6-311+G(d,p) in methanol (IEF-PCM); TDDFT calculations with B3LYP/TZVP in methanol. d At 300 K, using internal energies. e Using free energies; includes two 4-OH rotamers. f Boltzmann-weighted average for four low-energy structures. The result for the absolute lowest-energy minimum was similar. b With

104

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

helicity rule must be considered fortuitous. In fact, the relative position of aromatic 1 Lb and carbonyl n –π * bands varies for the three compounds due to to the presence of OH groups hydrogen-bonded to the carbonyl. On passing from 104 (unbound C = O) to 109 (one C = O· · ·HO bond) and to 100 (two C = O· · ·HO bonds), the n –π * transition experiences the expected hypsochromic effect moving from 317 nm (104) to 298 nm (109) and to 285 nm (100). In this latter case the ﬁrst calculated transition (corresponding to the negative CE observed at 341 nm) turns out to be the 1 Lb and not the n –π * one, while for tetralone 109 the two transitions are almost degenerate (and heavily mixed). Moreover, the calculated rotational strength for the n –π * transition for (2S , 4R)-100 is positive, in contrast with the M -helicity of the carbocylic ring. The conformational situation for tetralone (3S , 4S )-105 is less clear-cut because the 3-methyl and 4-hydroxy substituents are not allowed to occupy simultaneously an equatorial position. On the basis of relative conformational energies [135], the methyl group is expected to dictate the conformation of the ring, showing a stronger preference for the equatorial position than the hydroxyl group. This is conﬁrmed by NMR experiments [131] and by our calculations. In this case, we run B3LYP geometry optimizations using a larger basis set (6-311+G(d , p)) and including a solvent model (IEF-PCM) for methanol, followed by frequency calculations to estimate true free energies. The two lowest-energy conformers show an equatorial 3-methyl group (they differ for the rotation of 4-OH) and amount to an overall population of 85%. They are followed by two other minima (overall 15% population) with axial 3-methyl and equatorial 4-OH group. ECD calculations were run with B3LYP/TZVP including again IEF-PCM for methanol and considering all four minima. Very interestingly, the average TDDFT-calculated ECD spectrum for (3S , 4S )-105 (which is dominated by the lowest-energy structure at long wavelengths) shows negative CEs for both n –π * and 1 Lb , in contrast with the experimental data (compare Tables 3.8 and 3.9). According to this outcome, the reported absolute conﬁguration for (3S , 4S )-105 is wrong and should be reassigned as [(+)-ECD(333)]-(3R, 4R)-105. The above calculation results demonstrate that the observed n –π * CEs for tetralones depend heavily on the substitution pattern of both the aromatic and carbocyclic rings; therefore we discourage the use of the relative helicity rule for conﬁgurational assignments.

3.4.2. Dihydroisocoumarin Chromophore Optically active synthetic isochromans (S )-45a,d,c (Scheme 3.1) were converted to the corresponding dihydroisocoumarins (S )-111a–c by oxidation with Jones reagent [136] or dimethyldioxirane (DMDO) [137] as shown in Scheme 3.3 [75]. The (S )-dihydroisocoumarins 111a–c have very similar ECD patterns; positive π –π * and n –π * transitions at 278–307 nm and 252–268 nm, respectively, followed by a negative and positive band in the high-energy region (Table 3.10). Their ECD TAB L E 3.10. ECD Data for Dihydroisocoumarins (S )-111a–c Compound (S )-111a (S )-111b (S )-111c

CD λmax [nm] (ε) π → π ∗ : 289sh (+1.08), 278sh (+2.00); n → π ∗ : 252 (+4.19); 230 (−4.73), 204 (+13.83). π → π ∗ : 307sh (+1.10), 300 (+1.28), 294sh (+1.20); n → π ∗ : 268 (+7.62); 244 (−4.64), 226 (+10.48), 204 (−6.22). π → π ∗ : 304sh (+2.44), 296 (+2.48); n → π ∗ : 258 (+4.17); 239 (−1.29), 229sh (+0.67), 206 (+11.21).

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

data conﬁrmed that regardless of the substitution pattern of the aromatic ring, the positive n –π * transition of 3-alkyldihydroisocoumarins derives from P -helicity of the heteroring and thus (S ) absolute conﬁguration, in accordance with previous ECD studies on synthetic steroidal dihydroisocoumarin derivatives of 41–43 [74] (Figure 3.6a). The heteroring of (S )-dihydroisocoumarins 111a–c adopts a half-chair or envelope conformation with P -helicity, as deﬁned by the positive ωC-5a,C-4,C-3,O torsional angle, to ensure the favorable equatorial orientation of the C3 methyl group (Scheme 3.3). The dihydroisocoumarin n –π * transition was consistently applied to the conﬁgurational assignment of synthetic derivatives [138] and natural products [139–144] in agreement with the above helicity rule. In order to check the applicability of the dihydroisocoumarin helicity rule for 3,4disubstituted dihydroisocoumarins, the 3,4-cis-dimethyl-dihydroisocoumarin derivative 113 was prepared by catalytic hydrogenation of (S )-(+)-ascochin (112), a natural product isolated from the endophytic fungus Ascochyta sp. (Scheme 3.4), and its chiroptical data were measured and reproduced by calculation [145]. The addition occurred with cis diastereoselectivity due to the inherent (4S ) chirality center (the formyl group was also reduced during the hydrogenation), resulting in (3S , 4S ) absolute conﬁguration for 113. The ECD spectrum of 113 shows a negative CE at 307 nm and a positive one at 267 nm (Figure 3.13). Thus, according to the literature [74, 146], the lactone n –π * CD band should be assigned to the latter band. Interestingly, the synthetic (3S )-3-methyldihydroisocoumarin derivative 111b (Scheme 3.3) has also a positive CE at 268 nm [75], and except for this transition, its ECD curve was almost the mirror image of that of 113 (Figure 3.13). In order to conﬁrm the position of the lactone n –π * ECD transition and hence the semiempirical rule of dihydroisocoumarins, a TDDFT calculation was carried out

(S)-45a Jones reagent (S)-45c (S)-45d or DMDO/ dry acetone

H

H 5a 4 Me 2 3 1 O

R3 R2

R1

1

4

O

O

R2 H OMe H

3

Me

R3 H OMe H

O

2

positive n-π* CE in the 252 – 268 - nm range

H

(S)-111a-c R1 111a H 111b H 111c OMe

3

Me

1

4

O

ωC-5a,C-4,C-3,O >0 P-helicity

Scheme 3.3.

O

11 6 5

HO 7

HO 11 10

10 4

9 3

1 O

HO 6 H2-Pd/C 7

8

OH O (S)-(+)-ascochin (112)

5 4 8

H 9

3

1 O

OH O (3S,4S)-tetrahydroascochin (113)

Me H4 Me

3

1

O

O

2

P-helicity ωC-8a, C-1, O, C-3 > 0

Scheme 3.4. Conversion of (S)-(+)-ascochin (112) to 113. P-helicity of 113 and DFT-calculated most stable conformation.

105

106

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

12 9 n→ π*

Δε and R [10–40 cgs]

6 3

(S)-111b

0

×10 113

–3 –6 –9 –12

200

220

240

260

280

300

320

340

λ (nm)

Figure 3.13. Measured ECD spectra of 113 (solid line) and (S)-111b (dotted line), and TDDFTcomputed rotational strengths R (vertical bars) for the absolute minimum of (3S, 4S)-113 found by DFT (Scheme 3.4).

on the absolute minimum-energy structure computed for 113 [145]. The conformational analysis of 113 resulted in a hydrogen-bonded (O-11 → HO-7) conformer as the most stable structure (1.7 kcal/mol lower DFT-energy than the second minimum), as shown in Scheme 3.4. The TDDFT CD calculation on this conformer demonstrated that the lactone n –π * CD transition of 113 appears with positive CE as the third computed transition from the red at 246 nm, namely as part of the 250 to 270-nm ECD band, overlapped with aromatic π –π * transitions (Figure 3.13). In particular, the most redshifted transition computed at 288 nm is of the 1 Lb -type, and it is apparently responsible for the weak ECD signal above 280 nm. Since 113 has (3S , 4S ) absolute conﬁguration and P -helicity of the heteroring (Scheme 3.4), its computed positive n –π * transition is in accordance with the semiempirical rule. (S )-111b has again P -helicity, also resulting in positive n –π * CE at 268, although all the other corresponding transitions have opposite signs to those of 113, due to the different substitution pattern of the aromatic ring. Since the lactone n –π * transition is only one of the contributors to the 267-nm ECD band among several π –π * transitions, its application for a safe conﬁgurational assignment is endangered in the current case by overlapping transitions. Another example for the combined use of the dihydroisocoumarine helicity rule and TDDFT calculations is offered by natural products phomolactone A (114) and B (115), isolated from Phomopsis sp. (Chart 3.13) [147]. The n –π * CE of 114 at 262 nm is

R1 R2 6

5

5a

4

R4 3

7 3

R

8

O2

8a

OH

1

O

R1 R2 R3 114 115 116 117

OH Cl OH H OH Cl OH H

R4

H n-Pr H n-Pr H Me H Me

n-π* CE 262 (−5.84) 258 (−2.08)

Chart 3.13.

107

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

8 6

Δε [cm2 mmol–1]

4 2 0 –2 Experimental CD of 113 Experimental CD of 114 Calculated CD of 116

–4 –6 200

220

240

260

280

300

320

340

360

λ (nm)

Figure 3.14. Measured (MeCN) ECD spectrum of 114 (solid line) compared with TDDFT-calculated ECD of (R)-116 (gray dotted line) and experimental ECD of 113 (dashed–dotted line).

negative and its high-energy ECD transitions are also opposite to those of (3S , 4S )tetrahydroascochin (113), which suggests that 114 has a heteroring with M helicity and hence (3R) absolute conﬁguration (Figure 3.14). Dihydroisocoumarin 115 showed the same −/+/–/+ ECD pattern from the low-energy to the high-energy region as 114, which allowed its assignment as (3R) as well. TDDFT ECD calculations were employed to conﬁrm the conﬁgurational assignment and support the ECD correlation discussed above. The calculations conﬁrmed that neither the length of the alkyl chain nor its conformation affects the shape of ECD bands allied with the dihydroisocoumarin chromophore. Their sign is entirely determined by the ring A chirality, which is in turn dictated by the absolute conﬁguration of C3. This ﬁnding corroborated the ECD correlation for n –π * CE of dihydroisocoumarin discussed above, and also simpliﬁed the treatment of compounds 114 and 115, instead of which we could consider the methyl analogues 116 and 117 (Chart 3.13). Figure 3.14 shows the TDDFT-calculated ECD spectrum (B3LYP/TZVP) as Boltzmann-weighted average over two DFT-optimized geometries for (R)-116, in a good agreement with the experimental spectrum of 114 below 300 nm, which conﬁrms the absolute conﬁguration established above as (R)-114.

3.4.3. Chroman-4-one Chromophore The chroman-4-one chromophore is found in natural ﬂavanones, 3-hydroxyﬂavanones, 2-alkylchroman-4-ones, and isoﬂavanones exempliﬁed by compounds 118–121 (Figure 3.15). Snatzke established a relationship between the chirality of cyclic aryl ketones and their high-wavelength n –π * CEs [148] which was extended to correlate the helicity of the heteroring and the sign of the n –π* CE in ﬂavanones [149], 3-hydroxyﬂavanones [149], 2-alkylchromanones [150], and isoﬂavanones (Figure 3.15) [121, 151]. According to this rule, P -helicity of the heteroring adopting envelope conformation is manifested in a positive n –π * CE above 300 nm, such as in (S )-ﬂavanone, (2R, 3R)-3-hydroxyﬂavanone, (R)-2-methylchroman-4-one and (R)-isoﬂavanone.

108

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

1

O 4

H

H

1

O

2 3

4

R

O (S)-flavanone [(S)-118]: R = H (2R,3R)-3-hydroxyflavanone [(2R,3R)-119: R = OH

1

Me

O

2 3

4

2 3

O (R)-isoflavanone [(R)-121]

O (R)-2-methylchroman-4-one [(R)-120] H

H R

Ph O

Ph

Me O

O

O

O

O

H

H envelope conformation P-helicity positive n-π* CE

envelope conformation P-helicity positive n-π* CE

envelope conformation P-helicity positive n-π* CE

Figure 3.15. Correlation between the helicity of the heteroring in (S)-ﬂavanone, (2R, 3R)-3hydroxyﬂavanone, (R)-2-methylchroman-4-one, and (R)-isoﬂavanone and the sign of the n–π * CE.

1

O

1

O

O

8a

2 4 5

OR

Me O

(R)-122: R = H (R)-123: R = Me

Br

C

3

3

4 2

O

H ωC8a,O1,C2,C3 > 0 P-helicity positive n→π* negative 1Lb (π→π*)

O

O

O 124

Chart 3.14. Structures of chromanones 122–124 and preferred conformation of the heteroring for (R)-122 and (R)-123.

A recent application of the chromanone n –π * helicity rule is demonstrated by the 2-methyl-chroman-4-one derivatives 122 and 123 isolated from the endophytic fungus Nodulisporium sp. with 6% enantiomeric excess (Chart 3.14) [152]. The separations of their enantiomers were carried out with HPLC using a chiral stationary phase; and then their LC/CD spectra were recorded on-line, which allowed their conﬁgurational assignment on the basis of their long-wavelength n –π * CEs (Figure 3.16). Since the ﬁrst-eluted enantiomers of both 122 and 123 have positive n –π * and negative π –π * CEs around 340 and 310 nm, respectively, their heterorings adopt P -helicity, which implies (R) absolute conﬁguration, provided that the methyl group is equatorial ly oriented [152]. In fact, the X-ray data of the p-bromobenzoate of 122 (compound 124) showed that the chromanone heteroring has an envelope conformation with torsion angles ωC8a,O1,C2,C3 44.9(7)◦ and ωC5,C4a,C4,O − 1.3(7)◦ . The equatorial position of the methyl group is further in agreement with the large coupling constant of J = 12.8 Hz observed for the transdiaxial protons 2-Hax and 3-Hax . It must be stressed that similarly to tetralone derivatives, intramolecular hydrogen bonding of the carbonyl group may cause considerable blue shift of the characteristic n –π * transition, and thus the unambiguous assignment of the n –π *

109

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

9 6

(S)-122

n π* (R)-122

π π* 3 0 Δε

–3 (S)-123

–6 –9 –12

240

270

300

330

360

390

Wavelength (nm)

Figure 3.16. LC/CD spectra of (S)-122 (solid line), (R)-122 (dotted line), and (S)-123 (dashed line) in hexane/isopropanol 9:1.

and π –π * transitions may require ECD measurements in solvents of different polarity or excited-state calculations.

3.5. CONCLUSION The large amount of ECD data compiled on natural and synthetic derivatives containing a fused benzene chromophore makes it often possible to compare the measured ECD data of a new compound with those of analogues with known absolute conﬁguration, allowing a fast determination of absolute conﬁguration. The present chapter aimed to give guidelines for the scope and limitation of these correlations by discussing the safe applications and pitfalls of semiempirical helicity rules on benzene derivatives. As demonstrated by some examples, the application of high-level quantum-mechanics calculations may be extremely useful, especially in ambiguous cases, to explain the background and limitation of these semiempirical rules.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8.

L. Doub, J. M. Vandenbelt, J. Am. Chem. Soc. 1947, 69 , 2714–2723. H. B. Klevens, J. R. Platt, J. Chem. Phys. 1949, 17 , 470–484. J. R. Platt, H. B. Klevens, Chem. Rev . 1947, 41 , 301–310. J. Horwitz, E. H. Strickland, C. Billups, J. Am. Chem. Soc. 1969, 91 , 184–190. H. E. Smith, T. C. Willis, J. Am. Chem. Soc. 1971, 93 , 2282–2290. N. Berova, B. Kurtev, G. Snatzke, Tetrahedron 1983, 39 , 1371–1378. N. Berova, B. Kurtev, G. Snatzke, Croat. Chem. Acta 1987, 60 , 243–262. J. Petruska, J. Chem. Phys. 1961, 34 , 1120–1136.

110

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31.

32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43.

44. 45.

L. P. Fontana, H. E. Smith, J. Org. Chem. 1987, 52 , 3386–3389. S. T. Pickard, H. E. Smith, J. Am. Chem. Soc. 1990, 112 , 5741–5747. L. A. Mitscher, F. Kautz, J. LaPidus, Can. J. Chem. 1969, 47 , 1957–1963. R. J. Lorentzen, J. H. Brewster, H. E. Smith, J. Am. Chem. Soc. 1992, 114 , 2181–2187. H. E. Smith, L. P. Fontana, J. Org. Chem. 1991, 56 , 432–435. H. E. Smith, J. R. Neergaard, J. Am. Chem. Soc. 1996, 118 , 7694–7701. H. E. Smith, J. R. Neergaard, J. Am. Chem. Soc. 1997, 119 , 116–124. D. R. Michals, H. E. Smith, Chirality 1993, 5 , 20–23. H. E. Smith, Chem. Rev . 1998, 98 , 1709–1740. O. Korver, S. Dejong, T. C. Vansoest, Tetrahedron 1976, 32 , 1225–1229. A. O. Patil, W. T. Pennington, I. C. Paul, D. Y. Curtin, C. E. Dykstra, J. Am. Chem. Soc. 1987, 109 , 1529–1535. T. Schaefer, G. H. Penner, R. Sebastian, Can. J. Chem. 1987, 65 , 873–877. U. Berg, T. Liljefors, C. Roussel, J. Sandstrom, Acc. Chem. Res. 1985, 18 , 80–86. J. L. Broeker, R. W. Hoffmann, K. N. Houk, J. Am. Chem. Soc. 1991, 113 , 5006–5017. G. Pescitelli, L. Di Bari, A. M. Caporusso, P. Salvadori, Chirality 2008, 20 , 393–399. P. Butz, G. E. Tranter, J. P. Simons, Phys. Chem. Comm. 2002, 5 , 91–93. J. R. Platt, J. Chem. Phys. 1951, 19 , 263–271. M. Dierksen, S. Grimme, J. Chem. Phys. 2006, 124 . N. Lin, F. Santoro, X. Zhao, A. Rizzo, V. Barone, J. Phys. Chem. A 2008, 112 , 12401–12411. N. A. Macleod, P. Butz, J. P. Simons, G. H. Grant, C. M. Baker, G. E. Tranter, Isr. J. Chem. 2004, 44 , 27–36. M. D. Kundrat, J. Autschbach, J. Chem. Theory Comput. 2009, 5 , 1051–1060. G. Snatzke, Angew. Chem. Int. Ed . 1979, 18 , 363–377. G. Snatzke, M. Kajt´ar, F. Snatzke. Relationship of chiroptical properties with structure. I. Organic compounds. Aromatic chromophores, in Fundamental Aspects and Recent Development in Optical Rotatory Dispersion and Circular Dichroism, F. Ciardelli and P. Salvadori, eds. Hyden & Sons, London, 1973, pp. 148–172. G. Snatzke, P. C. Ho, Tetrahedron 1971, 27 , 3645–3653. N. J. Luche, A. Marquet, G. Snatzke, Tetrahedron 1972, 28 , 1677–1682. E. Dorhege, G. Snatzke, Tetrahedron 1970, 26 , 3059–3067. J. Schulte, G. Snatzke, Chem. Ber. 1989, 122 , 1373–1374. S. Hagishita, K. Kuriyama, Bull. Chem. Soc. Jpn. 1982, 55 , 3216–3224. J. Toda, S. Matsumoto, T. Saitoh, T. Sano, Chem. Pharm. Bull . 2000, 48 , 91–98. J. C. Craig, S. Y. C. Lee, R. P. K. Chan, I. Y. F. Wang, J. Am. Chem. Soc. 1977, 99 , 7996–8002. G. Snatzke, M. Kajt´ar, F. Werner-Zamojska, Tetrahedron 1972, 28 , 281–288. J. Sagiv, Tetrahedron 1977, 33 , 2303–2313. J. Sagiv, Tetrahedron 1977, 33 , 2315–2320. S. Kloubert, M. Mathe-Allainmat, J. Andrieux, S. Sicsic, M. Langlois, Bioorg. Med. Chem. Lett. 1998, 8 , 3325–3330. S. Yous, S. Durieux-Poissonnier, E. Lipka-Belloli, H. Guelzim, C. Bochu, V. Audinot, J. A. Boutin, P. Delagrange, C. Bennejean, P. Renard, D. Lesieur, Bioorg. Med. Chem. 2003, 11 , 753–759. A. C. Trillat, M. Mathe-Allainmat, B. Bremont, I. Malagie, C. Jacquot, A. M. Gardier, M. Langlois, Eur. J. Med. Chem. 1998, 33 , 437–444. A. P´eter, G. T´oth, E. Olajos, F. F¨ul¨op, D. Tourwe, J. Chromatogr. A 1995, 705 , 257–265.

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

46. A. P´eter, G. T¨or¨ok, D. W. Armstrong, G. T´oth, D. Tourwe, J. Chromatogr. A 2000, 904 , 1–15. 47. T. Nakanishi, Y. Inatomi, H. Murata, K. Shigeta, N. Iida, A. Inada, J. Murata, M. A. P. Farrera, M. Iinuma, T. Tanaka, S. Tajima, N. Oku, Chem. Pharm. Bull . 2005, 53 , 229–231. 48. P. B. Hulbert, W. Klyne, P. M. Scopes, J. Chem. Res., Synop. 1981, 27–27. 49. M. Klaes, T. Ellendorff, T. J. Schmidt, Planta Med . 2010, 76 , 719–U724. 50. C. Zhao, A. Nagatsu, K. Hatano, N. Shirai, S. Kato, Y. Ogihara, Chem. Pharm. Bull . 2003, 51 , 255–261. 51. S. Antus, E. Baitz-G´acs, G. Snatzke, T. S. T´oth, Liebigs Ann. Chem. 1991, 633–641. 52. Z. Samu, S. Nyiredy, E. Baitz-G´acs, Z. Varga, T. Kurt´an, Z. Dinya, S. Antus, Chem. Biodivers. 2004, 1 , 1668–1677. 53. I. Szilagi, P. Tetenyi, S. Antus, O. Seligmann, V. M. Chari, M. Seitz, H. Wagner, Planta Med . 1981, 43 , 121–127. 54. T. H. Kim, H. Ito, K. Hayashi, T. Hasegawa, T. Machiguchi, T. Yoshida, Chem. Pharm. Bull . 2005, 53 , 641–644. 55. Y. W. Chin, J. W. Kim, Tetrahedron Letters 2004, 45 , 339–341. 56. N. Noda, S. Kubota, Y. Miyata, K. Miyahara, Chem. Pharm. Bull . 2000, 48 , 1749–1752. 57. M. Z. Xu, W. S. Lee, J. M. Han, H. W. Oh, D. S. Park, G. R. Tian, T. S. Jeong, H. Y. Park, Bioorg. Med. Chem. 2006, 14 , 7826–7834. 58. Z. Varga, I. Seres, E. Nagy, L. Ujhelyi, G. Balla, J. Balla, S. Antus, Phytomedicine 2006, 13 , 85–93. 59. P. M´atyus, K. Horv´ath, Med. Res. Rev . 1997, 17 , 523–535. 60. Z. Czak´o, L. Juh´asz, A. Ken´ez, K. Czifr´ak, L. Soms´ak, T. Docsa, P. Gergely, S. Antus, Bioorg. Med. Chem. 2009, 17 , 6738–6741. 61. K. K´onya, R. Ferenczi, A. Czompa, A. Kiss-Szikszai, T. Kurt´an, S. Antus, Arkivoc 2008, 200–210. 62. A. Czompa, Z. Dinya, S. Antus, Z. Varga, Arch. Pharm. 2000, 333 , 175–180. 63. A. Arnoldi, L. Merlini, J. Chem. Soc. Perkin Trans. 1 1985, 2555–2557. 64. L. Han, X. S. Huang, I. Sattler, U. Moellmann, H. Z. Fu, W. H. Lin, S. Grabley, Planta Med . 2005, 71 , 160–164. 65. N. Tabata, H. Tomoda, Y. Iwai, S. Omura, J. Antibiot. 1995, 48 , 83–84. 66. T. Inagaki, K. Kaneda, Y. Suzuki, H. Hirai, E. Nomura, T. Sakakibara, Y. Yamauchi, L. H. Huang, M. Norcia, L. M. Wondrack, J. A. Sutcliffe, N. Kojima, J. Antibiot. 1998, 51 , 112–116. 67. J. Malmstrom, C. Christophersen, J. C. Frisvad, Phytochemistry 2000, 54 , 301–309. 68. M. D. Ennis, N. B. Ghazal, R. L. Hoffman, M. W. Smith, S. K. Schlachter, C. F. Lawson, W. Bin Im, J. F. Pregenzer, K. A. Svensson, R. A. Lewis, E. D. Hall, D. M. Sutter, L. T. Harris, R. B. McCall, J. Med. Chem. 1998, 41 , 2180–2183. 69. I. Bouchelet, B. Case, A. Olivier, E. Hamel, Br. J. Pharmacol . 2000, 129 , 501–508. 70. A. Bhattacharya, K. W. Schenck, Y. C. Xu, L. Nisenbaum, E. Galbreath, M. L. Cohen, J. Pharmacol. Exp. Ther. 2004, 309 , 825–832. 71. M. P. Deninno, R. Schoenleber, K. E. Asin, R. Mackenzie, J. W. Kebabian, J. Med. Chem. 1990, 33 , 2948–2950. 72. M. P. Deninno, R. J. Perner, H. E. Morton, S. Didomenico, J. Org. Chem. 1992, 57 , 7115–7118. 73. R. E. TenBrink, C. L. Bergh, J. N. Duncan, D. W. Harris, R. M. Huff, R. A. Lahti, C. F. Lawson, B. S. Lutzke, I. J. Martin, S. A. Rees, S. K. Schlachter, J. C. Sih, M. W. Smith, J. Med. Chem. 1996, 39 , 2435–2437. 74. S. Antus, G. Snatzke, I. Steinke, Liebigs Ann. Chem. 1983, 2247–2261.

111

112

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

75. G. Kerti, T. Kurt´an, T. Z. Illy´es, K. E. K¨ov´er, S. S´olyom, G. Pescitelli, N. Fujioka, N. Berova, S. Antus, Eur. J. Org. Chem. 2007, 296–305. 76. T. Kurt´an, N. Nesnas, Y. Q. Li, X. F. Huang, K. Nakanishi, N. Berova, J. Am. Chem. Soc. 2001, 123 , 5962–5973. 77. J. M. Seco, E. Quinoa, R. Riguera, Tetrahedron-Asymmetry 2001, 12 , 2915–2925. 78. K. Inoue, Y. Makino, N. Itoh, Tetrahedron-Asymmetry 2005, 16 , 2539–2549. 79. I. Kock, S. Draeger, B. Schulz, B. Elsasser, T. Kurt´an, A. Ken´ez, S. Antus, G. Pescitelli, P. Salvadori, J. B. Speakman, J. Rheinheimer, K. Krohn, Eur. J. Org. Chem. 2009, 1427–1434. 80. T. Kurt´an, E. Baitz-G´acs, Z. M´ajer, S. Antus, J. Chem. Soc., Perkin Trans. 1 2000, 453–461. 81. S. Antus, T. Kurt´an, L. Juh´asz, L. Kiss, M. Holl´osi, Z. M´ajer, Chirality 2001, 13 , 493–506. 82. L. Juh´asz, L. Szil´agyi, S. Antus, J. Visy, F. Zsila, M. Simonyi, Tetrahedron 2002, 58 , 4261–4265. 83. L. Juh´asz, J. Visy, M. Simonyi, K. Krohn, S. Antus, Tetrahedron: Asymmetry 2002, 13 , 1219–1222. 84. H. Achenbach, J. Gross, X. A. Dominguez, G. Cano, J. V. Star, L. D. Brussolo, G. Munoz, F. Salgado, L. Lopez, Phytochemistry 1987, 26 , 1159–1166. 85. P. J. C. Benevides, P. Sartorelli, M. J. Kato, Phytochemistry 1999, 52 , 339–343. 86. Y. Natori, H. Tsutsui, N. Sato, S. Nakamura, H. Nambu, M. Shiro, S. Hashimoto, J. Org. Chem. 2009, 74 , 4418–4421. 87. D. C. Chauret, C. B. Bernard, J. T. Arnason, T. Durst, H. G. Krishnamurty, P. SanchezVindas, N. Moreno, L. SanRoman, L. Poveda, J. Nat. Prod . 1996, 59 , 152–155. 88. P. S. Luize, T. Ueda-Nakamura, B. P. Dias, D. A. G. Cortez, C. V. Nakamura, Biol. Pharm. Bull . 2006, 29 , 2126–2130. 89. M. P. De Campos, V. Cechinel, R. Z. Da Silva, R. A. Yunes, S. Zacchino, S. Juarez, R. C. Cruz, A. B. Cruz, Biol. Pharm. Bull . 2005, 28 , 1527–1530. 90. D. L. J. Clive, E. J. L. Stoffman, Chem. Commun. 2007, 2151–2153. 91. D. L. J. Clive, E. J. L. Stoffman, Org. Biomol. Chem. 2008, 6 , 1831–1842. 92. M. S. M. Yuen, F. Xue, T. C. W. Mak, H. N. C. Wong, Tetrahedron 1998, 54 , 12429–12444. 93. K. Yoshikawa, H. Kinoshita, Y. Kan, S. Arihara, Chem. Pharm. Bull . 1995, 43 , 578–581. 94. H. Ishii, F. Sekiguchi, T. Ishikawa, Tetrahedron 1981, 37 , 285–290. 95. H. Ishii, T. Ishikawa, I. S. Chen, S. T. Lu, Tetrahedron Lett. 1982, 23 , 4345–4348. 96. O. R. Gottlieb, J. C. Mourao, M. Yoshida, Y. P. Mascarenhas, M. Rodrigues, R. D. Rosenstein, K. Tomita, Phytochemistry 1977, 16 , 1003–1006. 97. C. J. Aiba, R. G. C. Correa, O. R. Gottlieb, Phytochemistry 1973, 12 , 1163–1164. 98. I. R. Nascimento, L. M. X. Lopes, L. B. Davin, N. G. Lewis, Tetrahedron 2000, 56 , 9181–9193. 99. C. J. Aiba, J. B. Fernandes, O. R. Gottlieb, J. G. Soaresmaia, Phytochemistry 1975, 14 , 1597–1604. 100. H. Achenbach, W. Utz, A. Usubillaga, H. A. Rodriguez, Phytochemistry 1991, 30 , 3753–3757. 101. F. Li, X. W. Yang, Helv. Chim. Acta 2007, 90 , 1491–1496. 102. I. R. Nascimento, L. M. X. Lopes, Phytochemistry 1999, 52 , 345–350. 103. H. Achenbach, W. Utz, B. Lozano, E. M. G. Touche, S. Moreno, Phytochemistry 1996, 43 , 1093–1095. 104. L. Duan, H. W. Tao, X. J. Hao, Q. Q. Gu, W. M. Zhu, Planta Med . 2009, 75 , 1241–1245. 105. L. Fang, D. Du, G. Z. Ding, Y. K. Si, S. S. Yu, Y. Liu, W. J. Wang, S. G. Ma, S. Xu, J. Qu, J. M. Wang, Y. X. Liu, J. Nat. Prod . 73 , 818–824.

E L E C T R O N I C C D O F B E N Z E N E A N D O T H E R A R O M AT I C C H R O M O P H O R E S

106. S. Garcia-Munoz, M. Alvarez-Corral, L. Jimenez-Gonzalez, C. Lopez-Sanchez, A. Rosales, M. Munoz-Dorado, I. Rodriguez-Garcia, Tetrahedron 2006, 62 , 12182–12190. 107. S. M. O. Van Dyck, G. L. F. Lemiere, T. H. M. Jonckers, R. Dommisse, L. Pieters, V. Buss, Tetrahedron-Asymmetry 2001, 12 , 785–789. 108. H. Dong, C. Wang, Y. Li, S. Guo, J. Yang, Magn. Reson. Chem. 2010, 48 , 256–260. 109. F. Martin, A. E. Hay, V. R. Q. Condoretty, D. Cressend, M. Reist, M. P. Gupta, P. A. Carrupt, K. Hostettmann, J. Nat. Prod . 2009, 72 , 852–856. 110. J. G. Sena, S. L. Nimmo, H. S. Xavier, J. M. Barbosa, R. H. Cichewicz, J. Nat. Prod . 2009, 72 , 1344–1347. 111. S. Antus, E. Baitz-G´acs, J. Kajt´ar, G. Snatzke, A. L. Tok´es, Liebigs Ann. Chem. 1994, 497–502. 112. G. Snatzke, F. Znatzke, A. L. Tok´es, M. R´akosi, R. Bogn´ar, Tetrahedron 1973, 29 , 909–912. 113. O. Korver, C. K. Wilkins, Tetrahedron 1971, 27 , 5459–5465. 114. H. vanRensburg, P. S. vanHeerden, D. Ferreira, J. Chem. Soc. Perkin Trans. 1 1997, 3415–3421. 115. G. Cardillo, L. Merlini, G. Nasini, P. Salvadori, J. Chem. Soc. (C) 1971, 3967–3970. 116. A. Czompa, T. Kov´acs, S. Antus, J. Het. Chem. 2000, 37 , 991–995. 117. F. Mazzini, G. Pescitelli, L. Di Bari, T. Netscher, P. Salvadori, Chirality 2009, 21 , 35–43. 118. M. Versteeg, B. C. B. Bezuidenhoudt, D. Ferreira, Tetrahedron 1999, 55 , 3365–3376. 119. T. Kurt´an, E. Baitz-G´acs, Z. M´ajer, A. B´enyei, S. Antus, J. Chem. Soc., Perkin Trans. 1 2002, 888–894. 120. L. Verbit, J. W. Clark-Lewis, Tetrahedron 1968, 24 , 5519–5527. 121. D. Slade, D. Ferreira, J. P. J. Marais, Phytochemistry 2005, 66 , 2177–2215. 122. C. Lima, G. Zappia, G. Delle Monache, B. Botta, Eur. J. Org. Chem. 2006, 5445–5448. 123. M. Kim, S. U. Kim, Y. U. Kim, J. Han, Bull. Korean Chem. Soc. 2009, 30 , 415–418. 124. D. Won, B. K. Shin, S. Kang, H. G. Hur, M. Kim, J. Han, Bioorg. Med. Chem. Lett. 2008, 18 , 1952–1957. 125. S. Awale, T. Miyamoto, T. Z. Linn, F. Li, N. N. Win, Y. Tezuka, H. Esumi, S. Kadota, J. Nat. Prod . 2009, 72 , 1631–1636. 126. E. Couche, A. Fkyerat, R. Tabacchi, Helv. Chim. Acta 2003, 86 , 210–221. 127. Y. L. Chen, C. H. Tan, J. J. Tan, S. J. Qu, S. H. Jiang, D. Y. Zhu, Planta Med . 2008, 74 , 1826–1828. 128. D. Laurent, G. Guella, I. Mancini, M. F. Roquebert, F. Farinole, F. Pietra, Tetrahedron 2002, 58 , 9163–9167. 129. G. Bringmann, K. Messer, W. Saeb, E. M. Petes, K. Peters, Phytochemistry 2001, 56 , 387–391. 130. L. Liu, A. L. Li, M. B. Zhao, P. F. Tu, Chem. Biodivers. 2007, 4 , 2932–2937. 131. K. Kagawa, K. Tokura, K. Uchida, H. Kakushi, T. Shike, H. Nakai, Chem. Pharm. Bull . 1992, 40 , 2083–2087. 132. K. Machida, E. Matsuoka, T. Kasahara, M. Kikuchi, Chem. Pharm. Bull . 2005, 53 , 934–937. 133. N. Allouche, B. Morleo, O. Thoison, V. Dumontet, O. Nosjean, F. Gueritte, T. Sevenet, M. Litaudon, Phytochemistry 2008, 69 , 1750–1755. 134. J. Thiem, M. Gerken, G. Snatzke, Liebigs Ann. Chem. 1983, 448–461. 135. S. E. Boiadjiev, D. A. Lightner, J. Am. Chem. Soc. 2000, 122 , 11328–11339. 136. P. S. Steyn, C. W. Holzapfel, Tetrahedron 1967, 23 , 4449–4461. 137. P. Bovicelli, A. Sanetti, R. Bernini, P. Lupattelli, Tetrahedron 1997, 53 , 9755–9760. 138. P. Salvadori, S. Superchi, F. Minutolo, J. Org. Chem. 1996, 61 , 4190–4191.

113

114

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

139. K. Krohn, R. Bahramsari, U. Florke, K. Ludewig, C. KlicheSpory, A. Michel, H. J. Aust, S. Draeger, B. Schulz, S. Antus, Phytochemistry 1997, 45 , 313–320. 140. K. G. Guimaraes, J. Filho, T. R. Dos Mares-Guia, F. C. Braga, Phytochemistry 2008, 69 , 439–444. 141. Y. Asakawa, Y. Tada, T. Hashimoto, Phytochemistry 1994, 37 , 233–235. 142. K. Fujimoto, T. Fujimaki, E. Okuyama, M. Yamazaki, Chem. Pharm. Bull . 1999, 47 , 1426–1432. 143. K. Baba, Y. Q. Xiao, M. Taniguchi, H. Ohishi, M. Kozawa, Phytochemistry 1991, 30 , 4143–4146. 144. M. Taniguchi, M. Yanai, Y. Q. Xiao, T. Kido, K. Baba, Phytochemistry 1996, 42 , 843–846. 145. K. Krohn, I. Kock, B. Els¨asser, U. Fl¨orke, B. Schulz, S. Draeger, G. Pescitelli, S. Antus, T. Kurt´an, Eur. J. Org. Chem. 2007, 1123–1129. 146. K. Krohn, R. Bahramsari, U. Fl¨orke, K. Ludewig, C. KlicheSpory, A. Michel, H. J. Aust, S. Draeger, B. Schulz, S. Antus, Phytochemistry 1997, 45 , 313–320. 147. H. Hussain, N. Akhtar, S. Draeger, B. Schulz, G. Pescitelli, P. Salvadori, S. Antus, T. Kurt´an, K. Krohn, Eur. J. Org. Chem. 2009, 749–756. 148. G. Snatzke, Tetrahedron 1965, 21 , 439–448. 149. W. Gafﬁeld, Tetrahedron 1970, 26 , 4093–4108. 150. W. J. McGahren, G. A. Ellestad, G. O. Morton, M. P. Kunstman, J. Org. Chem. 1972, 37 , 1636–1639. 151. C. Galefﬁ, P. Rasoanaivo, E. Federici, G. Palazzino, M. Nicoletti, B. Rasolondratovo, Phytochemistry 1997, 45 , 189–192. 152. J. Dai, K. Krohn, U. Fl¨orke, S. Draeger, B. Schulz, A. Kiss-Szikszai, S. Antus, T. Kurt´an, T. van Ree, Eur. J. Org. Chem. 2006, 3498–3506.

4 ELECTRONIC CD EXCITON CHIRALITY METHOD: PRINCIPLES AND APPLICATIONS Nobuyuki Harada, Koji Nakanishi, and Nina Berova

4.1. INTRODUCTION AND HISTORICAL OVERVIEW The electronic CD exciton chirality method enables one to determine the absolute conﬁguration (AC) of various chiral compounds in a nonempirical manner without reference to compounds with known AC [1–7]. Namely, if a compound contains two identical chromophores in a chiral position, where each chromophore undergoes an intense π –π ∗ transition, these two electrically allowed π –π ∗ transitions interact with each other to generate the so-called bisignate intense CD Cotton effects (CEs). The sign of the bisignate CEs reﬂects the absolute arrangements of the two chromophores in the molecule. By observing the exciton coupled CD, one can determine the AC of chiral compounds. The exciton CD is very intense and its generation mechanism is simple. Since the exciton chirality rule can be proven by derivation of quantum mechanical equations without numerical calculation as shown below, the CD exciton chirality method (ECM) is classiﬁed as a nonempirical rule. The nature and mechanism of the CD exciton coupling are explained in this chapter. The science of stereochemistry started when Louis Pasteur ﬁrst succeeded in the so-called “optical resolution” of racemic tartaric acid in 1848 [8, 9], and the theory of “tetrahedral carbon atom” was then proposed independently by J. H. van’t Hoff and J. A. Le Bel in 1874 to explain the enantiomeric structures of optically active compounds (Figure 4.1) [9–12]. In 1895, A. Cotton discovered an anomalous dispersion effect in the optical rotation phenomena, which became known as the Cotton effect (CE) [13]. However, because it was not possible to determine the ACs of chiral compounds at that time, E. Fischer proposed a convention where the (D)-AC was arbitrarily assigned to (+)-glyceraldehyde as the standard of chiral organic compounds [14]. Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

115

116

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 4.1. Historical overview of the determination of AC.

Quantum mechanics appeared upon entering the twentieth century; and in 1928, L. Rosenfeld reported the equation expressing the rotational strength R, a parameter governing the sign and magnitude of optical rotation, which is equal to the imaginary part of the scalar product between electric transition moment and magnetic transition moment [15]. In the 1930s, physicists and physical chemists challenged to determine the AC of chiral compounds by the calculation of optical rotation, developing their own theories. For example, W. Kuhn proposed the coupled-oscillator concept based on the classical theory [16]. E. U. Condon, W. J. Kauzmann, and H. Eyring applied the quantum mechanical one-electron theory [17], while J. G. Kirkwood developed the polarizability theory [18]. However, none were sufﬁciently convincing to claim that ACs of chiral compounds could be determined. For example, the coupled-oscillator theory, later extended to the exciton CD theory, had been applied to compounds with regular substituents, but not to compounds with two or more chromophores. The optical rotations of target compounds were small and not suited for determining ACs. The concept of “exciton coupling” was later developed by A. S. Davydov in 1948 for studying the UV spectra of molecular crystals [19]. The history of AC determination of chiral compounds could have been quite different, if physicists and physical chemists had realized in the 1930s that the coupled-oscillator theory would be more suited for the compounds with two identical chromophores and collaborated with organic chemists to synthesize such compounds. That is, it would have been possible to determine the AC of chiral organic compounds prior to Bijvoet’s discovery in X-ray crystallography. Unfortunately, at that time, the concept of the chiral interaction between two identical chromophores had not yet been adopted. It was later that the UV exciton theory was developed by A. S. Davydov. In 1951, the AC of a chiral compound was ﬁrst determined by a totally different method. Namely, J. M. Bijvoet succeeded to determine the L-(2R,3R) AC of (+)-tartaric acid by using the anomalous scattering effect of heavy atoms in X-ray crystallographic diffraction experiments [20]. It was fortunate that the AC arbitrarily selected in Fischer’s convention agreed with the results of the X-ray Bijvoet method. Later in the ﬁeld of chiroptical spectroscopy, the relation between helical structures of biopolymers and chiroptical spectra was rationalized by using the so-called coupledoscillator mechanism. In 1956, W. Mofﬁtt extended the theory to study the OR and CD

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

of proteins [21]. In 1962, I. Tinoco and co-workers applied the theory to the CD of DNAs and oligonucleotides [22]; the same year, J. A. Schellman applied the theory to helical polypeptides [23]. In these studies, the ACs of biopolymers or oligomers were already known through the AC of monomeric units determined by X-ray crystallography. In the ﬁeld of organic stereochemistry, the coupled-oscillator theory was ﬁrst applied in 1962 by S. F. Mason to calycanthine, a natural dimeric alkaloid with C2 symmetry [24]. The CD spectrum of calycanthine shows a positive CE at 259 nm and negative CE at 240 nm in the 1 La transition (252 nm) of the aniline chromophores, from which the AC was determined; the AC was later conﬁrmed by X-ray crystallography [25]. It should be noted that this is the ﬁrst application of the coupled-oscillator mechanism to a chiral organic compound. S. F. Mason and co-workers applied the same method to other chiral compounds, but later it became clear that some compounds were not suited for the coupled oscillator mechanism. For example, the AC of Troger’s base assigned by the coupled-oscillator method [26] was later revised by X-ray crystallography [27]. In 1969, N. Harada and K. Nakanishi reported the dibenzoate chirality rule for determining the AC of chiral glycols [28]. Here the absolute helicity between two benzoate chromophores (i.e., AC of the original glycols) could be unambiguously determined from the bisignate CEs of dibenzoates. This dibenzoate chirality rule was based on the coupled-oscillator mechanism, and it opened a general protocol for determining ACs as the CD exciton chirality method [1, 2]. In the history of the absolute conﬁgurational assignment, there were many controversies, and among them the biggest one was raised in 1972–1973 [29]. Namely, it was claimed that the ACs determined by the X-ray Bijvoet method disagreed with those assigned by the exciton coupling mechanism and that the ACs determined by the Bijvoet method had to be reversed because of an error in the Bijvoet theory. If this was true, all organic chemistry textbooks would have to be revised. In 1973, Y. Saito immediately pointed out that there is no error in the theory of the X-ray Bijvoet method [30], and H. H. Brongersma and P. M. Mul experimentally conﬁrmed the Bijvoet method [31]. In the same year, S. F. Mason reported that the dipole velocity treatment of CD led to the correct AC [32]. In 1976, N. Harada reported the synthesis and CD of an ideal compound connecting the X-ray Bijvoet and CD exciton chirality methods (Section 4.4) [33]. It is now established that both methods lead to the same and correct AC.

4.2. OUTLINE AND PRINCIPLE OF CD EXCITON CHIRALITY METHOD The CD exciton chirality method (ECM) has been successfully applied to a variety of natural products and synthetic chiral compounds to determine their ACs. This method enables one to determine the AC of a chiral compound without any reference, that is, it is a nonempirical method [1–7]. For example, the CD and UV spectra of cholest-5ene-3β,4β-diol bis(p-bromobenzoate) 1 are illustrated in Figure 4.2, where UV spectrum shows an intense band of the allowed π –π ∗ transition at 244.0 nm, which is polarized along the long axis of the p-bromobenzoate chromophore [5]. In this region, the CD shows intense negative ﬁrst and positive second CEs (λext 243.6 nm, ε −30.4: λext 236.2 nm, ε +21.2; A = −51.6) [5]; the CE at longer wavelength is called the ﬁrst CE, while the shorter wavelength extremum the second CE. The exciton CD shows two CEs of similar intensity but opposite signs (Figure 4.2), which are called “bisignate” CEs. The exciton bisignate CD reﬂects the “exciton chirality”—that is, helical sense between two electric transition dipole moments (ETDMs) involved in the excitation:

117

118

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

+40

+20

Br

group i 3β-equatorial

CD

Δε

O

3 O

0

–20

–40 UV

CD 243.6 (–30.4) 236.2 (+21.2) A = –51.6

ε × 10–4

Br

O

H

4 O O

O O

1

O group j 4β-axial

Br

4β-axial group j in rear Br

6 4

UV 244.0 (41,800)

200

300 λ (nm)

250

2

X-ray crystallographic stereoview

3β-equatorial group i in front

0

(a)

(b)

Figure 4.2. (a) CD and UV spectra of cholest-5-ene-3β,4β-diol bis(p-bromobenzoate) 1: UV in 0.3% 1,4-dioxane/EtOH; CD in 10% 1,4-dioxane/EtOH. (b) Negative exciton chirality between long axes of two p-bromobenzoate chromophores: Newman projection and X-ray crystallographic stereoview. (Redrawn from reference 5, with permission.)

(i) If the exciton CD shows negative ﬁrst and positive second CEs, the two ETDMs constitute a counterclockwise screw sense as in the case of bis(p-bromobenzoate) 1 (Figure 4.2). (ii) If the exciton CD shows positive ﬁrst and negative second CEs, the two ETDMs constitute a clockwise screw sense. From this relation, the AC of the target compound, namely, cholest-5-ene-3β,4β-diol, can be determined. This is the electronic CD exciton chirality method [1–7]. In the case of bis(p-bromobenzoate) 1, the π –π ∗ transition at 244 nm is polarized along the long axis of the p-bromobenzoate chromophore, and the exciton chirality corresponds to the helicity between the long axes of two chromophores. As illustrated in the Newman projection, the two long axes constitute anticlockwise screw sense generating negative ﬁrst and positive second CEs. The counterclockwise screw sense between two p-bromobenzoate groups is directly observed in the X-ray crystallographic stereoview in Figure 4.2 [5].

4.2.1. Principles and Nonempirical Nature of Exciton Chirality Method The second example used for explaining the principle of the ECM is 5α-cholestane-2β,3βdiol bis(p-dimethylaminobenzoate) 2 (Figure 4.3). When two identical chromophores i and j , with intense UV π –π ∗ transition (ground state 0 → excited state a), exist in a molecule, two chromophores interact with each other to split the excited state into two energy levels (α and β states), while the ground state (0) remains unsplit [1]. This phenomenon, the exciton coupling or exciton interaction, generates two electronic transitions, from ground state 0 to excited states α and β—that is, transitions 0 → α and 0 → β. The wavefunction, energy, dipole strength, and rotational strength for the α-state

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

α−state:

N

wave function, energy, dipole strength, rotational strength,

O group j O H O

H

N

β−state:

2

O group i

5α-Cholestane-2β,3β-diol bis(p-dimethylaminobenzoate) a

0 group i

β α

a

0 0 total system group j

wave function, energy, dipole strength, rotational strength,

⎯ 2) (fiafj0 – fi0fja) yaa = (1/√ E a = Ea – Vij D a = (1/2)(li0a – lj 0a)2 R a = +(1/2)ps0 Rij • ( li 0a × lj0a) yab = (1/√2) (fiafj0 + fi0fja) Eb = Ea + Vij D b = (1/2)(li 0a + lj 0a)2 R b = –(1/2)ps0 Rij • ( li 0a × lj 0a)

interaction energy, Vij = Rij –3{li0a lj0a – 3Rij–2( li 0aRij) (lj0aRij)} If Vij > 0, the a-state is lower in energy than the b-state. α-state, longer wavelength side => 1st Cotton effect. β-state, shorter wavelength side => 2nd Cotton effect.

Figure 4.3. Theoretical summary of the CD ECM.

and β-state are summarized in Figure 4.3, where Vij is deﬁned as the interaction energy between two electric transition moments μi 0a and μj 0a . If Vij is positive, the α-state corresponds to the transition at longer wavelength, while the β-state corresponds to the transition at shorter wavelength. As shown in Figure 4.3, the rotational strength R α of the α-state is opposite in sign to that of the β-state, R β , but their absolute values are equal. Note that the sign and magnitude of R α and R β are governed by the triple product R ij • (μj 0a × μj 0a ) [1]. These equations were next applied to bis(p-dimethylaminobenzoate) 2 in Figure 4.4, where electric transition moments μi 0a and μj 0a of the benzoate chromophores were assigned as shown. Since vectors μi 0a and μj 0a are set in-phase, the interaction energy Vij becomes positive, and hence the α-state is lower in energy than the β-state. Two vectors μi 0a and μj 0a constitute a clockwise screw, and so the resultant vector (μj 0a × μj 0a ) becomes parallel to the distance vector R ij . Therefore the triple product R ij • (μj 0a × μj 0a ) becomes positive, and R α is positive, while R β is negative. This leads to the CD in Figure 4.4b, where the CE at longer wavelength (1st CE) is positive and that at shorter wavelength (2nd CE) is negative [1]. Figure 4.4c shows the UV and CD spectra of bis(p-dimethylaminobenzoate) 2, which has an intense π –π ∗ transition (λmax 307 nm, ε 54,300) polarized along the long axis of the chromophore. The CD spectrum shows positive 1st and negative 2nd CEs in agreement with the theoretical conclusion: 1st CE, λext 320 nm, ε +61.7 and 2nd one, λext 295 nm, ε −33.2. The amplitude of the exciton CD is deﬁned as A = ε1 − ε2 , where ε1 and ε2 are ε values of 1st and 2nd CEs, respectively. In the case of dibenzoate 2, A = +94.9. From these results, the AC of the original glycol is readily determined. In Figure 4.4a, the in-phase combination of vectors μi 0a and μj 0a was considered. But we can theoretically choose another case of in-phase relation and two cases of outof-phase combination. What will happen in these cases? Due to the self-consistency of the exciton CD theory, the theoretical results agree with that in Figure 4.4b; that is, the exciton CD depends only on the mutual absolute arrangements of two long

119

120

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

N

μi0a

O O H 2

N O

Vector product μi0a × μj0a

O

+40

O H

Parallel

Rij

group i (1) Interaction energy Vij > 0 (2) Parallel, so R α > 0,and R β < 0

H

O H O

O 3

N +20

2

Δε

(a)

OBz-dma OBz-dma

ε × 10–4

μj0a

group j

N

320 (+61.7)

+60

CD

0 307 (54,300)

in EtOH

Positive 1st Cotton –20

295 (–33.2)

λ –40 Negative 2nd Cotton

6 4 2

UV

0 200

250

300

350

λ (nm)

(b)

(c)

Figure 4.4.

Application of the CD ECM to 5α-cholestane-2β,3β-diol bis(p-dimethyl-

aminobenzoate) 2: CD and UV spectra in EtOH. (Redrawn from reference 28g, with permission.)

axes of benzoate chromophores. In application of the CD ECM for AC determination, it is unnecessary to consider the in-phase or out-of-phase combination of ETDMs. The exciton chirality governing the sign and intensity of CEs is deﬁned as shown in Table 4.1 [1]. As discussed, the nonempirical nature of the CD ECM is easily proved, indicating the simplicity of exciton CD mechanism. Further details of the quantum mechanical molecular exciton CD theory are described in the Section 4.3.

TAB L E 4.1. Deﬁnition of Exciton Chirality Qualitative Deﬁnition

Quantitative Deﬁnition

CEs

Positive exciton chirality

R ij • (μioa × μjoa )Vij > 0

Negative exciton chirality

R ij • (μioa × μjoa )Vij < 0

Positive ﬁrst (at longer wavelength) and negative second (at shorter wavelength) Cotton effects Negative ﬁrst (at longer wavelength) and positive second (at shorter wavelength) Cotton effects

Source: Redrawn from reference 28g, with permission.

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

121

The qualitative deﬁnition of exciton chirality is simple: If two ETDMs constitute a clockwise screw sense, CD shows positive ﬁrst and negative second CEs, and vice versa. In general, intense exciton CD CEs are observed at the long axis-polarized transition, and the essential points of the ECM are summarized as follows [1]. 1. If the long axes of two interacting chromophores constitute a clockwise screw sense, the CD shows a positive ﬁrst CE at longer wavelength and a negative second CE at shorter wavelength (Table 4.1 and Figure 4.5). 2. If they constitute a counterclockwise screw sense, a negative ﬁrst CE at longer wavelength and a positive second CE at shorter wavelength result (Table 4.1 and Figure 4.5). Pertinent features of the exciton CD follow. 1. The intensity of the exciton CD (A value) is inversely proportional to the square of the interchromophoric distance Rij , provided that the remaining angular part is the same [1]. A(= ε1 − ε2 ) ∝ Rij −2 2. The A value of exciton split CD is a function of the dihedral angle between two transition moments. In vicinal glycol dibenzoates, the sign of the exciton split CEs remains unchanged from 0◦ to 180◦ . The qualitative deﬁnition shown in Table 4.1 is applicable to a dibenzoate with the dihedral angle of more than 90◦ . The maximum A value is around 70◦ [1]. 3. The A value is proportional to the square of absorption coefﬁcient ε of the chromophore. Therefore, it is advisable to use chromophores undergoing intense π –π ∗ transition. In general, a weak transition along the short axis of chromophores is unsuitable.

Figure 4.5. Exciton coupled CD CEs and UV absorption band. In general, the CD zero-crossing point corresponds to λmax of UV band. (Redrawn from refernece 1, with permission.)

122

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

4. The CEs of α and β states have identical rotational strength of opposite signs, and two CEs are conservative, satisfying the sum rule. Thus the integrated areas of positive and negative CEs are equal.

Rk = 0

5. Rotational strength R should be origin-independent because it is a physically observable quantity. Equations in Figure 4.3 satisfy the origin-independence of rotational strength.

4.3. THEORY OF EXCITON CD SPECTROSCOPY 4.3.1. CD Spectra and Rotational Strength of CE The rotational strength R, a parameter representing the sign and intensity of a CE, is experimentally obtained from the observed CD spectra as shown in Eq. (4.1) [34]. R = 2.296 × 10−39

ε(σ )/σ dσ

(cgs unit)

(4.1)

where σ is wavenumber. The rotational strength R is theoretically formulated by Eq. (4.2) as proposed by Rosenfeld [15]. R = Im{< 0|μ|a > • < a|M |0 >}

(4.2)

where Im denotes the imaginary part of the terms in brackets, < > denotes the integration over conﬁguration space, μ and M are operators of electric and magnetic moment vectors, respectively. The dot stands for scalar product of two vectors, 0 and a are wavefunctions of ground and excited states, respectively. Rotational strength R is thus equal to the imaginary part of the scalar product of electric and magnetic transition moments. A Gaussian distribution approximation of a CD Cotton effect curve leads to Eq. (4.3) [1, 34]. ε(σ ) = εmax exp{−((σ − σo )/σ )2 }

(4.3)

where εmax is the maximum intensity of the CE, σo is the central wavenumber of the CE, and σ is half the band width at 1/e peak height of the Gaussian curve. From eq. (4.1) and (4.3), we obtain √ R = 2.296 × 10−39 π εmax σ/σo

(4.4)

From Eqs. (4.3) and (4.4), the CD curve is formulated as √ ε(σ ) = (σo /(2.296 × 10−39 π σ ))R exp{−((σ − σ0 )/σ )2 }

(4.5)

where σ can be evaluated from observed UV–Vis spectra. Provided that rotational strength R is calculated by Eq. (4.2), a CD spectrum can be reproduced by theoretical calculation [1, 34].

123

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

4.3.2. Molecular Exciton Theory of a Binary System with Two Chromophores According to the quantum mechanical exciton theory, in the exciton coupling system composed of two identical chromophores i and j , exciton wavefunctions are expressed by (4.6) and (4.7), where each chromophore undergoes excitation 0 → a [1]: Ground state :

φi 0 ,

φj 0

(4.6)

Excited state :

φia ,

φja

(4.7)

The Hamiltonian operator of the coupling system is formulated as H = Hi + Hj + Hij

(4.8)

where Hi and Hj are the Hamiltonian of groups i and j , respectively, and Hij is the interaction energy term between two groups i and j . The ground-state wavefunction and energy of a binary system are expressed as ψ0 = φi 0 φj 0

(4.9)

E0 = 0

(4.10)

The singly excited state of the binary system splits into two energy levels, α and β states. For the α state, wave function : Energy :

√ ψaα = (1/ 2){φia φj 0 − φi 0 φja } α

E = Ea − Vij

(4.11) (4.12)

where Vij is the interaction energy between two groups i and j and is approximated by the point dipole approximation: Vij = μi 0a μj 0a Rij−3 {ei • ej − 3(ei • eij )(ej • eij )}

(4.13)

where μioa , μjoa , and Rij are absolute values of vectors μioa , μjoa , and R ij , respectively; ei , ej , and eij are unit vectors of μioa , μjoa , and R ij , respectively. For the terms μioa , μjoa , and R ij , see Eqs. (4.18), (4.19), and (4.31), respectively. For the β state, wave function: Energy:

√ ψaβ = (1/ 2){φia φj 0 + φi 0 φja }

(4.14)

E β = Ea + Vij

(4.15)

These equations indicate that the binary system has two electronic transitions, 0 → α and 0 → β, in UV–Vis spectrum. If Vij > 0, the α state is lower in energy than the β state, and therefore the transition 0 → α locates at longer wavelengths, while the transition 0 → β is at shorter wavelengths. The electric dipole moment operator μ of a whole system is deﬁned as μ=

μi =

er is

(4.16)

124

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

where μi is the electric dipole moment operator of group i , e is the elementary charge, and r is is the distance vector of electron s in group i from the origin. The electric transition moment <0|μ|a>α of the transition 0 → α is formulated as √ (4.17) < 0|μ|a>α = μo aα = ψo μψ aα d τ = (1/ 2)(μioa − μjoa ) where μioa =

φio μi φia d τi

(4.18)

φjo μj φja d τj

(4.19)

μjoa =

Those are electric transition moments of transition 0 → a in groups i and j , respectively. The electric transition moment of the transition 0 → β is similarly expressed as < 0|μ|a>β = μoa β =

√ ψo μψ a β d τ = (1/ 2)(μioa + μjoa )

(4.20)

The magnetic moment operator M of a whole system is formulated as M =

M i = (e/2mc)

r is × p is

(4.21)

where m is the mass of electron, c is the velocity of light, p is is the linear momentum of electron s in group i , and × stands for vector product of two vectors. The magnetic moment operator is further changed as M = (e/2mc)

Ri × pi +

mi

(4.22)

where R i is distance vector of group i from the origin, p i and m i are linear momentum and internal magnetic moment operators of group i , respectively. The magnetic transition moment < a|M |0>α of the excitation 0 → α is calculated as < a|M |0>α = ψaα M ψo d τ √ = (1/ 2){(e/2mc)R i × p iao + m iao − (e/2mc)R j × p jao − m jao } (4.23) where p iao and m iao are linear momentum and internal magnetic moment of group i , respectively. For the linear momentum of a group, the next equation is useful [35]: p oa = −(2π imc/e)σo μoa

(4.24)

where i is the symbol of imaginary, σo is excitation energy expressed in wavenumber units, and μoa is electric transition moment of transition 0 → a. For group i , p iao = (2π imc/e)σo μioa

(4.25)

125

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

Accordingly, √ α = (1/ 2){i π σo R i × μioa − i π σo R j × μjoa + m iao − m jao }

(4.26)

In a similar manner, the magnetic transition moment of excitation 0 → β is calculated as √ β = (1/ 2){i π σo R i × μioa + i π σo R j × μjoa + m iao + m jao } (4.27) The dipole strength D α of a binary system for the excitation 0 → α is expressed as D α = (1/2)(μioa − μjoa )2

(4.28)

D β = (1/2)(μioa + μjoa )2

(4.29)

Similarly,

Rotational strength R α of the α state is derived from Eqs. (4.2), (4.17), and (4.26): R α = Im{< 0|μ|a>α • α } = (1/2)Im{(μioa − μjoa ) • (m iao − m jao )} + (1/2)π σo R ij • (μioa × μjoa )

(4.30)

where R ij is the interchromophoric distance vector from group i to group j and is deﬁned as (4.31) R ij = R j − R i Similarly, β

R β = Im{< 0|μ|a>β • < a|M |0 >} = (1/2)Im{(μioa + μjoa ) • (m iao + m jao )} − (1/2)π σo R ij • (μioa × μjoa )

(4.32)

In the case of π → π ∗ transition of common molecules, internal magnetic transition moments m iao and m jao are negligible. Therefore, rotational strengths are approximated as: (4.33) R α = +(1/2)π σo R ij • (μioa × μjoa ) R β = −(1/2)π σo R ij • (μioa × μjoa )

(4.34)

These equations indicate that the CEs of α and β states have equal intensity but of opposite signs. Thus exciton CD satisﬁes the sum rule. (4.35) RA = R α + R β = 0 The rotational strength is proportional to the triple product of interchromophoric distance and electric transition moments of groups i and j . Therefore, provided that chromophores exhibiting intense π –π ∗ transitions are used, intense exciton CEs are observable. The rotational strengths of exciton CD satisfy the origin-independence as shown in Eqs. (4.31), (4.33), and (4.34).

126

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

4.3.3. Exciton CD of N-mer and Dimer: Quantitative Deﬁnition of Exciton Chirality The exciton theory is applicable to UV–Vis and CD spectra of N -mer having N identical chromophores [1]. When N chromophores undergoing intense π –π ∗ transition (0 → a) interact with one another, the excited state splits into N energy levels. The wavenumber σk of k th excitation is formulated as Cik Cjk Vij (4.36) σk − σo = 2 where coefﬁcients Cik and Cjk are obtained by solving the N th-order secular equation. The rotational strength R k is expressed as R k = −π σo

Cik Cjk R ij • (μioa × μjoa )

(4.37)

For the N -mer, CD curve is formulated as √ ε(σ ) = {σo /(2.296 × 10−39 πσ )} R k exp{−((σ − σk )/σ )2 }

(4.38)

The Taylor expansion of Eq. (4.38) against σk /σ around σo /σ gives the second term of expansion as √ ε(σ ) = {2σo /(2.296 × 10−39 π σ )} exp{−((σ − σo )/σ )2 }{(σ − σo )/σ } × R k {(σk − σo )/σ } (4.39) From Eqs. (4.36), (4.37), and (4.39), the CD equation of N -mer is obtained: √ ε(σ ) = {4 π σo2 /(2.296 × 10−39 σ 2 )}{(σo − σ )/σ } exp{−((σo − σ )/σ )2 } (4.40) × Cik Cjk R ij • (μioa × μjoa ) Cik Cjk Vij It should √ be noted that for α state √ in the case of a binary system, √ the coefﬁcients √ are always 1/ 2 and −1/ 2, and for β state they are 1/ 2 and 1/ 2, for any mutual conﬁguration of two identical chromophores. Therefore, Eq. (4.40) is simpliﬁed as √ ε(σ ) = {2 π σo2 /(2.296 × 10−39 σ 2 )}{(σo − σ )/σ } exp{−((σo − σ )/σ )2 } × R ij • (μioa × μjoa )Vij

(4.41)

This is the exciton CD equation of a binary system. The next term of Eq. (4.41), √ {2 π σo2 /(2.296 × 10−39 σ 2 )}{(σo − σ )/σ } exp{−((σo − σ )/σ )2 } represents an anomalous dispersion curve with positive and negative extrema. The sign and intensity of exciton CD depend on the quadruple term R ij • (μioa × μjoa )Vij . Therefore, the term R ij • (μioa × μjoa )Vij is adopted as the quantitative deﬁnition of exciton chirality. This term is also expressed as R ij • (μioa × μjoa )Vij = Dioa Djoa Rij−2 eij • (ei × ej ){ei • ej − 3(ei • eij )(ej • eij )} (4.42)

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

where Dioa and Djoa are transition dipole strengths of groups i and j , respectively. From this equation, it is suggested to use chromophores undergoing intense π –π ∗ transition in the UV–Vis spectrum for obtaining intense exciton CD. This equation also indicates that the exciton CD amplitude is inversely proportional to the square of the interchromophoric distance Rij .

4.3.4. Theoretical Simulation of Exciton CD As discussed above, the ECM is simple in mechanism, and exciton CD spectra were simulated by various theoretical methods. There were reported examples of simulation by the DeVoe coupled oscillator method [1, 6, 36], the π -electron SCF-CI-DV MO method [1; See Chapter 5, this volume], and more recent ab initio and related MO methods—for example, TDDFT B3LYP/6-32G(d) [37]. For complex CD spectra and molecules, the TD-HF/6-31G(d), TDDFT B3LYP/6-32G(d), and CAM-B3LYP/6-32G(d) methods would be very useful (see theoretical chapters). The simulations are important for conﬁrming the ACs determined by the ECM, and these theoretical methods provide better understanding of the CD generation mechanism.

4.4. THE CONSISTENCY BETWEEN X-RAY BIJVOET AND CD EXCITON CHIRALITY METHODS These methods are based on totally different physical phenomena, but they should give the same AC for a speciﬁc compound. However, it was claimed in 1972 that the ACs determined by X-ray and CD exciton methods disagreed and that the ACs determined by the X-ray Bijvoet method should be revised [29]. This claim was based on the X-ray and CD analyses of compounds (–)-5 and (+)-6 in Figure 4.6, where the CD of the weak 1 Lb transition (∼290 nm) of aniline chromophore polarized along the short axis was analyzed as an exciton couplet. However, this claim was subsequently retracted. Note that the ECM should be applied to an intense UV transition (Section 4.5), but not to a weak UV transition. In 1976, the synthesis and CD spectrum of a chiral cage compound (+)-3 with two anthracene chromophores as an ideal model for ECM were reported [33]; the results unambiguously proved the consistency between X-ray Bijvoet and CD exciton methods (Figure 4.6). Compound (+)-3 was synthesized from diester (+)-4, which was chemically correlated with compounds (−)-5 and (+)-6. The ACs of (−)-5 and (+)-6 had been determined by the Bijvoet method [38]. As expected, compound (+)-3 showed strong exciton-coupled CD CEs at the strong 1 Bb transition of anthracene polarized along the long axis: λext 268.0 nm (ε +931.3), 249.7 (−720.8), A = +1652.1 (Figure 4.6), showing that the strong UV transition gives rise to intense the exciton coupled CD. Since the UV π –π ∗ transition at 267.2 nm is polarized along the long axis of the anthracene chromophore, the exciton split CEs at 268.0 nm and 249.7 nm are generated by the exciton coupling between two ETDMs of anthracene groups. The long axes of the anthracene moieties constitute a clockwise screw sense leading to positive ﬁrst and negative second CEs, and therefore the AC of compound (+)-3 was determined as shown. This result agrees with that determined by the X-ray Bijvoet method (Figure 4.6) [33]. Another controversy regarding the ACs of clerodin (7) and related diterpenes should be noted. In 1962, the AC of clerodin 7 [39] was determined by the X-ray Bijvoet method, which was opposite to the AC shown in Figure 4.7 [40]. Since this AC was believed to

127

128

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 4.6.

CD and UV spectra of (6R,15R)-(+)-6,15-dihydro-6,15-ethanonaphtho-[2,3c]pentaphene 3 in dioxane/EtOH, and chemical correlation between compound 3 and related compounds. (Redrawn from reference 33a, with permission.)

(−)-clerodin 7

3-epicaryoptin derivative bis(p-Cl-benzoate) 9

steroidal model 11

Figure 4.7. Correct ACs of (–)-clerodin 7 and related compound 9, and comparison of exciton CD of compounds 9 and 11.

be correct, clerodin 7 had been used as a reference of AC for newly isolated compounds of the clerodane family for many years. For example, in 1974 the AC of 3-epicaryoptin 8 was determined by comparison of CD and chemical correlation with 7 [41]. At the same time, an unexpected result was reported; the CD ECM was applied to 3-epicaryoptin derivative 3,6-bis(p-Cl-benzoate) 9, but the observed positive couplet was opposite to the negative couplet expected from the AC of 9, which was based on the X ray of 7. To explain the discrepancy, it was postulated that one of the benzoate groups is twisted by an intramolecular H bond, generating a positive exciton chirality, and the result was reported as an exception of the CD ECM [41]. To solve the problem of this unexpected discrepancy, the steroidal model compound 11 was synthesized in 1978 [42], because compounds 9 and 11 have the same relative

129

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

conﬁgurations at key positions (Figure 4.7). The CD spectrum of 11 showed a positive couplet, concluding that the conformation of the benzoate group was not twisted by an H bond. Since 3,6-bis(p-Cl-benzoate) 9 and 11 showed CD couplets of the same sign, 9 should have the same AC as 11. From the exciton CD and molar rotation data, the ACs of clerodane diterpenes 7 and 8 were reversed [42]. The revised ACs were conﬁrmed by X-ray analysis later [43]. Unfortunately, the incorrect AC of clerodin 7 had been used as a reference of the clerodane diterpene family for about 16 years.

4.5. SUITABLE CHROMOPHORES FOR THE CD ECM AND EXAMPLES It is essential to select suitable chromophores for the application of the CD ECM. The following issues should be considered: (i) intense π –π ∗ transition, (ii) unambiguous direction of ETDM, and (iii) symmetrical structure. Figure 4.8 shows typical chromophores useful for the CD ECM, in which arrows show the ETDM direction leading to exciton coupled CD. In general, the long-axis polarized transitions are suitable for exciton CD, because of the larger UV intensity. (a) Para-Substituted Benzoate Chromophores for Glycols The intramolecular CT or 1 La transition (230–310 nm) of para-substituted benzoate chromophores has been used for determining the AC of many glycols [1, 3]. The intramolecular CT transition is polarized along the long axis of the benzoate chromophore, which is almost parallel to the alcoholic C–O bond. Therefore, the AC of

Chromophores for -OH groups: O X O X = H, Br, OMe, NMe2 230–310 nm O

Ph O

X

N

X = OMe, NMe2 O 300–360 nm

Ph N

270 nm fluorescent

M = 2H, Zn, Mg

For -C=C- groups:

O

N

O

O 260 nm fluorescent

280 nm fluorescent

230 nm fluorescent

Ph

O O

Me2N

MeO

N 305 nm

O 260 nm

O 350 nm

O

420 nm fluorescent

For -COOH groups:

O

Me2N

N Ph

O

For -NH2 groups:

O

O

O 235 nm fluorescent

N M

HN

N

N M

Ph N

N

M = 2H, Zn, Mg Ph 420 nm fluorescent

Figure 4.8. Chromophores useful for the CD ECM, where arrows show the direction of ETDM. (Redrawn from reference 5, with permission.)

130

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

the glycol part can be determined from the exciton CD data. In contrast, ortho- and meta-substituted benzoate chromophores are not suited, because their ETDMs are not parallel to the alcoholic C–O bond. (b) Cinnamate, β-Naphthoate, and Other Chromophores for Glycols These chromophores are characterized by their strong absorption at longer wavelengths [44]. (c) Tetraphenyl-Porphyrin-Carboxylic Acid (TPP-COOH) Tetraarylporphyrins and their zinc and copper analogues deserve special attention as chromophores. They possess a very intense, sharp, and narrow Soret band, which is red-shifted to ∼420 nm and has ε ∼450,000–550,000. These porphyrins and metalloporphyrins belong to the most powerful and versatile CD chromophores [45–47]. One of the ﬁrst examples for application of porphyrin in structural studies includes the red-tide toxin brevetoxin, where the CD exciton coupling was observed over a long distance of ˚ [45, 47, 48]. A detailed discussion on the application of porphyrins as CD ∼40–50 A reporter groups as well as an account of the theoretical analysis of porphyrin-porphyrin exciton interactions can be found in references 45–47. The Soret band originates from the two degenerate transitions Bx and By (Figure 4.9), which are perpendicular to each other; theoretically the porphyrin Soret band should be considered as a circular oscillator [46, 47]. However, due to the rotational ﬂexibility around the meso porphyrin 5-C-phenyl junction (librational averaging), the transitions Bx and By can be represented by one effective transition moment along the 5–15 axis (Figure 4.9), and the exciton CD reﬂects the chirality between two effective transition moments. The large ε value and red shift of Soret band above 400 nm where most other chromophores do not absorb make the TPP-COOH extremely useful and versatile chromophore for exciton CD analysis. The typically very intense couplet can be measured with a high S /N ratio at very low concentration; thus it allows reliable studies on a microscale and, as mentioned above, also in cases where a very long-range coupling is involved [45, 48, 49]. Bx

15 N N 5 Ph M N N Effective transition moment

Ph

By X

3 O O

3α,17β

13, M: Zn2+ . UV-Vis: 419 nm ε 550,000 (CH2Cl2) fluorescence: λem 646, Φf 0.10

OR

TPP-COOH

O 17 O

12, M: H, H. UV-Vis: 418 nm ε 440,000 (CH2Cl2) fluorescence: λem 650, Φf 0.12

O

Δε +200 CD

+100

A = +270

0 –100

416 nm (–117) 419 nm (914,000)

–200

419 nm (550,000) monomer

350

1

bis (Zn-TPP) derivation CH2Cl2

UV-Vis

X Rij

14, X = H e = 15,000 Rij = 13.6 Å no coupling

424 nm (+153)

e × 10–6

Ph

400

0 450 λ (nm) 500

15, X = NMe2 e = 28,000 Rij = 13.6 Å A = +21 16, X = TPP e = 440,000 Rij = 24.4 Å A = +193 17, X = Zn-TPP e = 550,000 Rij = 24.4 Å A = +270

Figure 4.9. (Top) UV–Vis, and ﬂuorescence data for TPP-COOH and its derivatives. (Bottom) CD and UV–Vis data of steroidal bis(tetraarylporphyrin) derivatives 16 and 17 together with corresponding benzoate derivatives 14 and 15. (Redrawn from reference 5, with permission.)

131

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

The large steric size of porphyrin chromophores, however, calls for attention when they are introduced at vicinal positions, such as in diols, polyols, diamines, and amino alcohols. The steric repulsion between bulky porphyrins may cause conformational changes upon porphyrin derivatization; therefore additional conformational studies by NMR or molecular modeling are recommended [47]. Figure 4.9 illustrates the striking increase in the A value seen in tetraarylporphyrin (TPP) and its Zn derivative (Zn-TPP) compared to p-dimethylaminobenzoate at the Rij ˚ (Figure 4.9). Other examples for efﬁcient porphyrin–porphyrin CD distance of 24.0 A ˚ can be found [48]. coupling over 40–50 A (d) Benzamide, Phthalimide, and 2,3-Naphthalenedicarboximide Chromophores for Diamines and Amino Alcohols The CD ECM is also applicable to the intramolecular CT band of benzamide groups [1]. The transition is polarized along the long axis of the chromophore. However, in some cases such as N -methyl benzamides, the benzamide moiety exists as a mixture of (E ) and (Z ) isomers, and therefore, the mutual orientation of the ETDMs is uncertain. Thus in these cases, one should be cautious in assigning AC by exciton CD [50]; see Section 4.10 (1). The C2v -symmetrical phthalimide and 2,3-naphthalenedicarboximide chromophores are ideally suited for the ECM application to diamines and amino alcohols, because their long axis-polarized ETDMs are exactly parallel to the amine C–N bond [51, 52]. For example, the 2,3-naphthalenedicarboximide chromophore exhibits an intense 1 Bb transition around 260 nm, which is polarized along the long axis of the chromophore. Figure 4.10 shows the CD and UV spectra of trans-1,2-cyclohexanediyl bis(2,3-naphthalenedicarboximide) (1R,2R)-(−)-18, where the UV 1 Bb transition shows +100 in 10% 1,4-dioxane-EtOH

+50

O –50 Δε

H N O O

CD

–100

H N O (1R,2R)-(–) CD 264.8 (–154.6) 257.5 ( 0.0) 251.9 ( +38.9) UV A = –193.5 UV 259.0 (94,600) 252.2 (97,400)

–150

–200

200

250

300

350

ε x10–4

0

20

O

H

N H N

OO

H 18

(A) NMR: J1-H,2-H = 11.9 Hz

O

O

N

O

15 10

O

N

H O

(B)

5 0

400

λ (nm)

Figure 4.10.

CD and UV spectra of (1R,2R)-(−)-trans-1,2-cyclohexanediyl bis(2,3-naph-

thalenedicarboximide) 18 in 10% 1,4-dioxane/EtOH. (Redrawn from reference 52.)

132

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

two split bands. This split represents the exciton coupling, because the UV spectrum of mono(2,3-naphthalenedicarboximide) shows a single band in this region. The CD showed negative ﬁrst and positive second CEs (λext = 264.8 nm, ε −154.6; λext = 251.9 nm, ε +38.9; A = −193.5). From the negative A value, a counterclockwise helicity between two long axes of 2,3-naphthalenedicarboximide chromophores was assigned to compound (−)-18, leading to the (1R,2R)-AC [52]. Note that compound (−)-18 can take diequatorial conformer (A) and diaxial conformer (B) as shown in Figure 4.10. In general, conformer (A) is considered to be more stable than conformer (B). However, if the electric repulsion between two polar groups is strong, conformer (B) may become more stable. To conﬁrm the conformational preference of (A), the measurement of 1 H NMR coupling constant J1,2 is crucial. However, since these two protons are equivalent, it is not possible to determine J1,2 by regular 1 H NMR spectroscopy. Hence, the 13 C satellite signal method was used to give J1,2 = 11.9 Hz, indicating that conformer (A) was more stable than (B) [53]. Figure 4.11 illustrates a submicroscale chemical protocol developed for the analysis of sphingosines and dihydrosphingosines isolated from new cell lines. First the NH2 group of d-erythro-sphingosine 19 was blocked as a naphthalimide group yielding a derivative 20. Then the OH groups were converted to 2-naphthoate groups yielding derivative 21 that could be sensitively detected by HPLC, mass spec, CD and ﬂuorescence analyses. The relative and ACs were assigned by comparing the observed CD with the standard CD curves of erythro- and threo-sphingosines/dihydro-sphingosines [54]. (e) Chromophores for Carboxylic Acids and Oleﬁn Compounds The chromophores suitable for chiral carboxylic acids are listed in Figure 4.8. The application of the ECM to oleﬁn compounds is unique. The isolated oleﬁn group shows a π –π ∗ transition below 200 nm, and therefore the exciton method is not applicable in a straightforward manner. However, the double bond can be transformed via oleﬁn metathesis into chromophores (see Figure 4.8) that absorb at longer wavelength, so that the entire exciton couplet can be conveniently measured [55]. (f) Compounds with Preexisting Chromophore that Interfere with Exciton CD: Use of Red-Shifted Chromophores When a natural product has a preexisting chromophore that could interfere with observation of exciton CD, chromophores with a longer wavelength UV λmax than the preexisting chromophore can be used to avoid overlap of CEs. Red-shifted chromophores for derivatization of hydroxyl groups are shown in Figure 4.12. For amino groups, the red-shifted

fluorescent fluorescent

NH2 HO

C13H27

OH D-erythro-sphingosine 19 50 μg

O N O

O N O

λmax 258 nm λem 370 nm

C13H27

HO OH

20

C13H27

O O λmax 234 nm λem 360 nm

O

O 21

Figure 4.11. By a selective two-step microscale chemical derivatization procedure, two different types of chromophores are introduced in D-erythro-sphingosine 19.

133

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

Δe 263 (+25) +20

N O O

chrom-1

O chrom-3

N O chrom-2

O

UV λmax 382 nm (ε 27,000)

+10

CD

O

UV λmax 382 nm (e 34,000) UV λmax 410 nm (e 37,000) RO OR

389 (+16) e × 10–4

Me2N

0 –10 –20

22a 412 274 (58,000) (63,000)

22b 8 455 (–25)

O O-Cin H OAc : 22a, R = H : 22b, R = chrom-3 ester

6 4

H

UV 2 200

300

400

500 λ (nm)

0

Figure 4.12. Red-shifted chromophores and application to taxinine. (Redrawn from reference 57, with permission.)

chromophores, such as Schiff bases and cyanine dyes, are also useful for exciton CD analysis, in particular, when it is desirable to avoid possible interference with other electronic transitions present in the substrate [56]. As shown in Figure 4.12, taxinine derivative α-glycol 22a exhibits an intense CE around 263 nm due to the π –π ∗ transition of the highly strained enone group. In the previous application of the ECM, unsubstituted benzoate chromophores were used where a negative exciton couplet was clearly observed despite the overlap with the enone CE [28a]. However later, to avoid the overlap, a red-shifted chromophore (chrom-3) was used for derivatization to yield ester 22b. As expected, the CD of 22b showed a clearer negative exciton couplet indicating a counterclockwise screw sense between the two hydroxyl groups in full agreement with the previous report of the AC [57].

4.6. THE USE OF PREEXISTING CHROMOPHORES IN NATURAL PRODUCTS FOR EXCITON COUPLING Some natural products already have one or two chromophores such as those shown in Figure 4.13, which can be used in exciton CD to determine their ACs.

Figure 4.13. Exciton CD chromophores found in natural products.

134

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a) Substituted Benzene and Polyacene Chromophores for the CD ECM As demonstrated in Section 4.4, the long-axis polarized 1 Bb transition of polyacene chromophores is ideally suited for observing exciton coupled CD, because of large ETDM. The UV data of polyacenes with D2h symmetry are shown in Figure 4.13. In the polyacene systems, there is no ambiguity for determining the long and short axes, and therefore the CD ECM leads to clear-cut conclusions. (b) Conjugated Dienes, Enones, Ene-Esters, Ene-Lactones, and DieneEsters as Exciton CD Chromophores The moieties in Figure 4.13 are useful chromophores for the CD ECM. The transition moment of their π –π ∗ band is almost parallel to the chromophoric long axis. (c) Natural Products with Two Chromophores Showing Exciton CDs The exciton coupling CD mechanism is applicable also to compounds already having two different chromophores, which exhibit long-axis polarized π –π ∗ transitions at different wavelengths—that is, i.e., nondegenerate system. The ACs of some natural products, such as those in Figure 4.14, were established by direct analysis of their CD spectra. In such cases the interaction of at least two preexisting chromophores leads to exciton split CDs. For abscisic acid (23), the opposite AC was once assigned, but it was later revised as shown by several studies. One was the application of exciton CD to the interaction between the enone and diene-carboxylic acid chromophores showing a positive couplet [58]. The case of quassin (24) is unique because of the exciton coupling between two identical preexisting α-methoxy-enone groups [59]. The AC of dendryphiellin F (25) was determined by exciton CDs from the interaction between diene and diene-carboxylate chromophores [60], while that of arnottin II (26) was determined from the dehydrotetralone and phthalide chromophoric interaction [61]. The AC of a derivative of vinblastine 27, a dimeric alkaloid, was originally determined by X-ray crystallography (Figure 4.15) [62]. Experimental and theoretical CD studies were carried out to clarify the CD mechanism of vinblastine and related compounds [63, 64]. The molecule 27 can be considered as containing two moieties, that of cleavamine 27a and that of vindoline 27b, both of which exhibit the intrinsic CD CEs due to their own local chirality [64]. In addition, an exciton-coupled through-space interaction between them takes place (Figure 4.15). To obtain the net exciton CD, the intrinsic CD bands were subtracted from the CD of 27. The resulting “difference CD” appeared as an intense positive couplet around 220 nm (Figure 4.15d), generated by the positive exciton chirality between 1 Bb transition moments of the indole and indoline chromophores at 225 nm and 215 nm, respectively, and by that conﬁrming the (S ) AC at 16’-C.

Figure 4.14. Natural products with two preexisting chromophores showing exciton CD.

135

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

(a)

(c)

27 λ (nm)

(b)

(d)

λ (nm)

Figure 4.15. (a) Vinblastine 27 consists of cleavamine 27a and vindoline 27b. (b) ETDMs of indole and indoline chromophores. (c) The sum CD spectrum (dotted line) = CD(27a) + CD(27b), and CD spectrum of 27. (d) Difference CD = CD(27) − {CD(27a) + CD(27b)}. (Redrawn from reference 64, with permission.)

4.7. SUPRAMOLECULAR APPROACH IN THE ECM: APPLICATION OF PORPHYRIN TWEEZERS The use of tetraarylporphyrins and their metal derivatives as CD chromophores has initiated a new supramolecular approach for the determination of the AC of chiral compounds containing a single stereogenic center and one site for chromophoric derivatization. This group includes various natural products carrying only a single functionality, such as secondary hydroxyl, primary or secondary amino, and carboxyl groups. These compounds are unsuitable for application of the conventional ECM where at least two intramolecularly interacting chromophores are required. The supramolecular approach employs a dimeric zinc porphyrin reagent 31, now commercially available as “Zn-tweezers,” which is capable of forming 1:1 host–guest complexes upon adding a solution of N,N -bidentate conjugate 30, prepared by reacting the chiral substrate 28 with an achiral carrier 29 as shown in Figure 4.16 [65]. The application of the porphyrin tweezers method to (S )-α-(2-naphthyl)ethanol is shown in Figure 4.16b. The formation of 1:1 sandwiched chiral host–guest complex 34

136

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

M N N Zn N N

H chiral XH substrate 28

M L X = O, NH

HO

M L

X

H N

N N Zn N N

NH2

Zn

L H O

:N H

P-1 HH N:

Zn

conjugate 30 (guest)

+ O

O

H

P-2 O

H N carrier 29

O O O O 1:1 host-guest complex 32

Zn porphyrin host 31 “tweezer”

NH2 (a)

+100 M H3C O

P-1

O :N H

P-2

P-1

P-2

H tweezer

M

L

M

L

HH N:

chiral conjugate 33 of (S)-abs. config

+50 Δε 0 –50

host-guest complex 34 –100

preferred conformation

433 nm (+91)

CD A = +170

e × 10–5

L

423 nm (–79)

15 10

422.2

416.6

UV/Vis 400

420

440 λ (nm)

5 0

(b)

Figure 4.16. Porphyrin tweezers method: (a) Preparation of bidentate conjugate 30 from chiral substrate 28 and carrier 29; Formation of a 1:1 host guest complex 32 with Zn-porphyrin tweezers 31. (b) Example of (S)-α-(2-naphthyl)ethanol conjugate 33: two conceivable conformations of complex 34; observed CD spectrum of 34 in methylcyclohexane. (Redrawn from reference 65b, with permission.)

from conjugate 33 proceeds under steric control and leads to positive ﬁrst and negative second CEs in the Soret region. The origin of the intense exciton CEs is due to a preferred conformer with a clockwise interporphyrin twist, where the larger group L (2-naphthyl) protrudes from the binding pockets in order to avoid unfavourable steric interactions. The interporphyrin twist in the complex is thus dictated by the steric orientation of L (2-naphthyl) and M (methyl) at the stereogenic center of the substrate. When there is no ambiguity in the assignment of L and M groups, the sign of the couplet determines the AC at this center. Recently, a more reliable discrimination of preferred interporphyrin helicity of the host–guest complexes by theoretical analysis of both steric and electronic factors involved in stereocontrolled complexation process has become possible. This analysis relied on molecular mechanics calculations by Merck Molecular Force Field (MMFFs) or OPLS2005 approaches coupled to Monte Carlo-based conformational analysis and quantum mechanical treatment of free conjugates [66–68]. The porphyrin tweezers method is now well established and has allowed successful determinations of AC of some natural products, such as isotomenoic acid 35, an irregular diterpene [69], and bovidic acid 37, an 18-carbon hydroxyfuranoid acid [70] (Figure 4.17). More recently, other types of porphyrin-based tweezers have been developed. Structural changes in the tweezers, such as introduction of various substituents at the aryl groups and in the bridge between the two porphyrins, allow for tuning the complexation ability

137

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

O H OH O

s P-2

15

Zn Isotomenoic acid 35

H O N H

5

O O

MeO P-1 NH2 15′ HO Zn 5′

36

O O

O

H H O 6

OH

9

4

O

P-2 O 5

15

Zn Bovidic acid 37

6

H

N H

15′

N Zn H2

5

O O

P-1

O

5′

38

O O

Figure 4.17. Applications of the porphyrin tweezers method to natural products.

of the tweezers and extension of its application to other types of chiral substrates [67, 71, 72].

4.8. APPLICATION OF THE CD ECM: FUNDAMENTAL EXAMPLES As shown, there exist numerous applications of ECM. In the following, some selected examples of fundamental application are explained.

4.8.1. Exciton Coupling Between Polyacene and Related Chromophores Polyacenes (e.g., naphthalene and anthracene) are ideal chromophores for exciton coupling, because the long and short axes are clearly assigned, and the long-axis polarized transition has a large ETDM, yielding intense bisignate CEs. Chiral 1,1 -binaphthyls are typical atropisomers showing exciton CEs. For example, (S )-1,1 -binaphthyl-2,2 -dimethanol 39 shows intense positive ﬁrst and negative second CEs (λext 231.3 nm, ε +342.4: λext 224.3 nm, ε −329.0; A = +671.4) in the longaxis 1 Bb transition (λmax 224.4 nm, ε 132,700) (Figure 4.18). The positive A value leads to an (S ) AC [1]. The exciton CD sign depends on the dihedral angle between two naphthalene planes. From the theoretical viewpoint, the exciton CD of 1,1 -binaphthalene compounds undergoes sign changes around 110–120◦ [73]. Therefore, in the application of the ECM, information of the dihedral angle is necessary. In most cases, the dihedral angle is distributed around 90◦ , which was supported by X-ray crystallography [74] and MO calculations. Chiral 1,1 :4 ,1 -ternaphthalene-2,2 -dimethanol (aS,aS )-(+)-40 with three naphthalene groups in positions of axial chirality is a unique atropisomer [75]. Enantiopure compound (+)-40 was prepared by the CSDP acid method. As shown in Figure 4.18, the CD spectrum of (aS,aS )-(+)-40 shows intense negative ﬁrst and positive second CEs (λext 231.7 nm, ε −333.9: λext 223.4 nm, ε +225.4; A = −559.3) in the long-axis 1 B transition (λ b max 224.0 nm, ε 186,500). In Figure 4.18, chromophores 1 and 2 constitute a counterclockwise screw, and the same negative screw holds for chromophores 2 and 3, because of C2 -symmetry. However, chromophores 1 and 3 have no helicity because the long axes of naphthalene moieties 1 and 3 are almost parallel to each other. Thus the AC of (+)-40 was determined to be (aS,aS ) by the exciton CD method and conﬁrmed by X-ray crystallography of the CSDP ester [76]. The dihedral angle between naphthalene groups ranges over −84.1◦ , −87.2◦ ,

138

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

HOH2C

+300

+200 CH2OH

Δe

+100 Δe 0

CD

(S) HOH2C

CH2OH

–100 224.4 (132,700)

e × 10–4

+200 (S)-[CD(+)231.3]-39

0 CH2OH

chrom 2 –200

15

CH2OH

–200 10 224.3 (–329.0)

–300

5

–400

chrom 3

–400 chrom 1

UV 200

300 λ (nm) 400

0

obsd CD 285.4 ( +21.9) A = –559.3 231.7 (–333.9) 223.4 (+225.4)

e × 10–4

+400 231.3 (+342.4)

(aS,aS)-(+)-40

200

20

(aS,aS)-(+)-40 obsd UV 292.8 ( 24,000) 224.0 (186,500) 250

10

0 300 λ (nm) 350

Figure 4.18. CD and UV spectra of (S)-1,1 -binaphthyl-2,2 -dimethanol 39 (EtOH) and (aS,aS)-(+)-1,1 :4 ,1 -ternaphthalene-2,2 -dimethanol 40 (95% aq. EtOH). (Redrawn from references 1 and 75, respectively, with permission.)

−89.1◦ , and −89.4◦ as determined by the X-ray analysis. That is, three naphthalene chromophores are almost perpendicular to one another. Figure 4.19a shows the CD and UV spectra of (1R,1 S ,2S )-2,2 -spirobi[benz[e] indan]-1,1 -diol diacetate 41 [77]. The exo/endo conﬁguration of two acetate groups was determined by 1 H NMR data showing nonequivalent four methylene protons. Racemic diacetate was enantioseparated by chiral HPLC to give an enantiomer [CD(–)230.2]-41 whose CD and UV spectra are shown in Figure 4.19. In the region of the naphthalene 1 Bb transition (λmax 224.8 nm, ε 172,700), the intense negative ﬁrst and positive second CEs (λext 230.2 nm, ε −961.5 : λext 221.6 nm, ε +567.1; A = −1528.6) led to the (1R,1 S ,2S ) AC [77]. [6,6]-Vespirene 42 is a unique member of chiral 9,9 -spiroﬂuorene compounds, and its chiroptical activity arises from the 9,9 -spirobiﬂuorene system twisted by the side-chain bridge (Figure 4.19b) [78]. The AC of (−)-42 was determined to be R by the coupled oscillator analysis of its CD spectrum [78, 79]. However, the CD spectrum showed a complex pattern deviated from the ideal exciton bisignate CEs, implying that the 9,9 spirobiﬂuorene chromophores of 42 may be strongly strained. Thus, it was questionable whether the CD ECM is applicable to such a strained system in a straightforward manner. To conﬁrm the AC of compound 42, enantiopure compounds 43 and 44 with two anthracene and naphthacene chromophores, respectively, were prepared (Figures 4.20) [80]. These spiroaromatics are more suited than [6,6]-vespirene 42 for the ECM, because of their less strained structures and clear deﬁnition of exciton chirality between two polyacene chromophores. Enantiopure [6,6]-vespirene (−)-42 and compound (R)-(+)-43 were synthesized starting from diester (R)-(+)-45 (Figure 4.20). The UV spectrum of (+)-43 shows an intense 1 Bb transition (λmax 288.6 nm, ε 152,000), which is polarized along the chromophoric long axis. In the 1 Bb transition region, the CD shows intense positive ﬁrst and negative second CEs (λext 300.5 nm, ε +551.0 and λext 278.5 nm, ε −560.7; A value = +1111.7). The present results unambiguously indicate that the long-axis electric transition moments of the two anthracene chromophores constitute a clockwise screw sense (i.e., positive exciton chirality), leading to the (R) AC [80]. In such spiroaromatics, there is the so-called spiro-conjugation—that is homoconjugation between p orbitals surrounding the spiro quaternary center. It is known

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

(a)

(b)

41

λ (nm)

Figure 4.19. (a) CD and UV spectra of (1R,1 S,2S)-2,2 -spirobi[benz[e]indan]-1,1 -diol diacetate 41 in EtOH. (b) Chiral spiroaromatics. (Redrawn from reference 77, with permission.)

(a)

(b)

λ (nm)

Figure 4.20. (a) Synthesis of [6,6]-vespirene (R)-(−)-42 and chiral spiroaromatics (R)-(+)-43 and (R)-(+)-44 starting from diester (R)-(+)-45. (b) CD and UV spectra of (R)-(+)-43 in EtOH. (Redrawn from reference 80, with permission.)

139

140

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(5R,12R)-(–)

(5R,12R)-(–)-46

λ (nm)

Figure 4.21. CD and UV spectra of (5R,12R)-(−)-1,15-diethynyl-5,12-dihydro-5,12[1 ,2 ]benzenonaphthacene 46 in EtOH. (Redrawn from reference 84, with permission.)

that in some cases, such spiro-conjugation makes a dominant contribution to the CD spectra [81]. However, in the case of (+)-43, the spiro-conjugation effect is negligibly small, because the observed CD spectrum shows intense conservative bisignate CEs due to the exciton coupling. Bis(naphthacene) compound (R)-(+)-44 also showed intense positive ﬁrst and negative second CEs leading to the (R)-AC, although the CD curve was complex due to vibronic structures. The AC determination of (R)-(+)-43 and (R)-(+)-44 conﬁrmed the (R)-ACs of dimethyl ester (+)-45 and [6,6]-vespirene (−)-42. The (R)-AC of (−)-42 was later conﬁrmed by X-ray crystallography [83]. (5R,12R)-(−)-1,15-Diethynyl-5,12-dihydro-5,12[1 2 ]benzenonaphthacene (46) is a triptycene derivative having one naphthalene and two phenylacetylene chromophores, with no conformational ﬂexibility due to its cage structure and linear acetylene groups (Figure 4.21). The CD spectrum of (5R,12R)-(−)-46 shows strong exciton CEs (λext 245.5 nm, ε −138.2 : λext 215.0 nm, ε +113.6; A = −251.8) in the UV absorption region (λmax 241 nm, ε 75,000). Clearly the CEs originate from the coupling between the longaxis polarized 1 Bb transition (λmax 220.2 nm, ε 107,300) of naphthalene and the long-axis polarized 1 La transition (λmax 234.2 nm, ε 15,000) of phenylacetylene [84]. The long axes of the naphthalene and one phenylacetylene moiety constitute a negative exciton chirality. Similarly, a negative helicity is found between the naphthalene and the other phenylacetylene. On the other hand, the long axes of two phenylacetylenes are parallel to each other, indicating nil exciton chirality. Thus, the total exciton chirality is negative, leading to the (5R,12R) AC of (−)-46 [84, 85]. This absolute conﬁgurational assignment is in line with the chemical correlation results [84].

4.8.2. Application of the CD ECM to Acyclic 1,2-Glycols and Polyols The CD exciton chirality method has been applied to dibenzoates or bis(2-anthroates) of acyclic 1,2-glycols, which show typical bisignate CEs as exempliﬁed in Figures 4.22

141

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

Δε R

ε x 10–4

273 (+97) +90 OChrom CD +60 Chrom = p-BrBz H OChrom A = +190 or 2-anthroyl +30 47 first CD, (+) 0 second CD, (–) Exciton Chirality: –30 H3C OChrom zero –60 OChrom OChrom H OChrom 258 –90 253 OChrom H R OChrom ChromO 15 (147,000) Chrom = (–93) R H 2-anthroyl H H H H H 10 (S)-48 R H H OChrom [47A] [47B] UV 5 in CH3CN J (trans) = 6.8 ~ 8.4 Hz, [47C] J (gauche) = 3.6 Hz 0 200 250 300 350 λ (nm)

Figure 4.22. Applications of the CD exciton method to acyclic terminal 1,2-glycols. (Redrawn from reference 87, with permission.)

50

λ (nm)

51

λ (nm)

Figure 4.23. Application of the CD ECM to acyclic internal 1,2-glycols with threo-conﬁguration. (Redrawn from reference 86.)

and 4.23 [86, 87]. Acyclic dibenzoates or bis(2-anthroates) can rotate around the bond connecting two benzoates or 2-anthroates, and hence the CD sign depends on the conformational equilibrium. Based on the exciton CD and the conformational analysis by 1 H NMR, the AC of acyclic 1,2-glycols can be determined (Figures 4.22 and 4.23).

142

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

In terminal acyclic 1,2-glycols, diester 47 {bis(p-Br-benzoate) [86] or bis(2anthroate) [87]} can adopt three rotational conformers 47A, 47B, and 47C (Figure 4.22); here conformer 47B is unstable due to two gauche relations among three bulky groups. In contrast, because conformers 47A and 47C have only one gauche relationship between two bulky groups, they are more stable and dominate the equilibrium. The stable conformer 47A has a positive exciton chirality between two chromophores, while in conformer 47C the two chromophores are trans, and hence no exciton chirality is generated. Thus the CD of diester 47 reﬂects the positive exciton chirality of 47A. The 1 H NMR coupling constants {J (trans) = 6.8–8.4 Hz, J (gauche) = 3.6Hz} support this conclusion. The CD of (S )-1,2-propanediol bis(2-anthroate) 48 shows intense exciton CEs, from which the AC was assigned (Figure 4.22) [87]. Similarly, the CD ECM is applicable to internal 1,2-glycols with threo-conﬁguration (Figure 4.23) [86, 87]. For example, diester 49 {bis(p-Br-benzoate) or bis(2-anthroate)} adopts three rotational conformers 49A, 49B, and 49C, in which conformers 49B and 49C are unstable because of three gauche relationships. Conformer 49A with two gauche relations between bulky groups dominates the equilibrium. Conformers 49A and 49B have positive and negative exciton twists between two chromophores, respectively, while in conformer 49C two chromophores are in the transrelationship, and therefore no exciton chirality is generated. The preference of 49A is supported by the large 1 H NMR coupling constant {J (trans) = 6.1–8.7 Hz}. After all, the CD spectrum of diester 49 reﬂects a positive exciton chirality of conformer 49A. For example, (2S,3S )-2,3-butanediol bis(p-Br-benzoate) 50 exhibits bisignate CEs of positive exciton chirality and an 1 H NMR coupling constant {J (trans) = ∼6.1 Hz}, from which the AC can be determined (Figure 4.23) [86]. The above relationship between the AC and exciton CD CEs holds for most internal 1,2-glycols. However, if a glycol has polar or extremely bulky groups (R1 and R2 ), the equilibrium is changed; groups R1 and R2 adopt a trans-relation to diminish the electric repulsive force or steric repulsion, and conformer 49B becomes dominant. The preference of 49B is supported by 1 H NMR coupling constant {J (gauche) = 2.9–4.1 Hz}. The CD spectrum of (2R,3R)-diethyl tartrate bis(p-Br-benzoate) 51 shows a negative exciton couplet reﬂecting the preference of 49B [86], where the ethyl ester groups are trans as conﬁrmed by 1 H NMR coupling constant {J (gauche) = 2.9 Hz}. If the groups R1 and R2 are identical, the 1 H NMR vicinal coupling constant between two methine protons cannot be measured because of the same chemical shifts. In such cases, the 13 C satellite band method is useful to determine the Jvic value [86, 88]. In erythro-1,2-glycols, the determination of AC is more difﬁcult. When the two groups R1 and R2 are identical, the glycol is a meso-isomer and hence achiral. If they are different, the glycol is chiral. The exciton CD CEs of erythro-diester are weak, and they depend on the equilibrium of the rotational conformations. Thus assignment of ACs requires further analysis by other methods. Noteworthy, the AC of 1,3-acyclic polyols can be determined much easier and in more straightforward manner due to a better discrimination between CD of 1,3syn/1,3-anti diols bis(p-Br-benzoates) or bis(p-methoxycinnamates). An acyclic anti 1,3-bis(acylate) adopts a planar zigzag form in its most stable conformer and exhibits a typical CD exciton couplet corresponding to the sign of the screw sense between the two gauche-oriented chromophores, while the syn-analogue also in most stable zigzag conformation has almost parallel p-Br-benzoate or p-methoxycinnamate chromophores that exhibit negligible coupling [89–91].

143

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

4.8.3. Application of the ECM to Compounds Containing a Single Suitable Chromophore: Nondegenerate System The exciton coupled CD is observable even in a nondegenerate system composed of two different chromophores, which undergo different electronic excitations. The ﬁrst example of the nondegenerate exciton CD used for determining the AC of a chiral compound is the case of chromomycin A3 (52), (Figure 4.24). The AC of 52 was determined in the very early stage of the development of the CD ECM [92]. The ﬁrst strategy for determining the AC of 52 was to apply the dibenzoate chirality method to glycol derivative (53). However, benzoylation of glycol 53 yielded monobenzoate 54 due to the steric hindrance. This unexpected result led to a general protocol to apply the CD ECM to nondegenerate systems [92]. The CD spectrum of monobenzoate 54 showed intense bisignate CEs (λext 270 nm, ε −19.9: λext 230 nm, ε +16.8; A = −36.7), which are stronger than those of glycol 53 (λext 270 nm, negative: λext 220 nm, positive) as seen in Figure 4.24 [1, 2]. The CD data of 54 imply that the bisignate CEs originate from the interaction between the long-axis transition of benzoate (λmax 230 nm, ε 14,000) and the long-axis 1 Bb transition of naphthalenoid chromophore (λmax 270 nm, ε 57,200), where the naphthalenoid 1 Bb transition is red-shifted due to the conjugation with a carbonyl group. If this mechanism is true, the following can be expected. If a proper benzoate group is introduced whose long-axis transition is close to the naphthalenoid 270 nm, the exciton CD would be enhanced, because the exciton theory tells us that the exciton coupling is more effective when two transitions are close to each other in energy. Hence the p-methoxybenzoate chromophore (λmax 256 nm, ε 18,000) was chosen to yield monobenzoate 55, the CD of which showed much intense bisignate CEs as expected (λext 271 nm, ε −70.6: λext 250 nm, ε +34.0; A = −104.6) (Figure 4.24). This result not only veriﬁed the exciton mechanism of nondegenerate systems, but also established the absolute conﬁgurational assignment. Since the ﬁrst CE is negative and the second positive, the long axes of two chromophores constitute a counterclockwise screw, leading to the AC as shown in Figure 4.24 [2, 92]. This assignment was later conﬁrmed by chemical correlation [93].

+40

250 (+34.0) 55

R = p-MeO-C6H4CO- HO

H

OMe OH O OH

+20 Δe 0

OH OH O 54

R = C6H5CO-

–20

H

MeO OH 52

MeO

MeO

R1

OMeO

H

OR OH O

O O

53, R = H 54, R = Bz 55, R = p-MeOBz

Me

–40

53

R=H in EtOH

–60

271 (–70.6)

–80 200

300

λ (nm)

MeO MeO H

H O

O H H O O OH O Me Me O Me H

54, R1 = H 55, R1 = OMe

400

Figure 4.24. Preparation of benzoate 54 and p-methoxybenzoate 55 of chromomycin A3 derivative: CD spectra of p-methoxybenzoate, benzoate, and alcohol derivatives. (Redrawn from reference 1, with permission.)

144

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

OR

+40 R=H λext 226 nm, Δe +14.8

+30

H R = Bz λext 235 nm, Δe +34.0 220 nm, Δe –14.3

+20 +10 Δe 0

MeO 56, R = H 57, R = Bz

in EtOH

Figure 4.25. CD spectra of

Me O O

–10

17-dihydroequilenin 3-methyl ether (56) and 17-benzoate (57). (Redrawn from reference 92, with

17

–20

H 300

λ (nm)

MeO

400

57

permission.)

To conﬁrm the nondegenerate exciton method further, the exciton method was applied to equilenin derivative 17-benzoate (57) (Figure 4.25) [2, 92]. The CD of benzoate 57 shows intense bisignate CEs around 230 nm (λext 235 nm, ε +34.0: λext 220 nm, ε −14.3; A = +48.3), while alcohol 56 exhibits a positive CE (λext 226 nm, ε +14.8). It is thus clear that the exciton CD of benzoate 57 originates from the interaction between long-axis 1 Bb transition of naphthalenoid chromophore (λmax 230 nm) and the long-axis transition of benzoate (λmax 230 nm). The positive exciton couplet agrees with the AC as shown in Figure 4.25. These results corroborated the methodology of nondegenerate exciton coupling. The nondegenerate ECM is applicable also to the conjugated diene–benzoate system as exempliﬁed in Figure 4.26, where the CD and UV of 3-methylenecholest-4-ene-6β-ol benzoate (58) is shown (λext 242.0 nm, ε −30.8: λext 225.0 nm, ε +39.1; A = −69.9) [1]. The long-axis polarized π –π ∗ transition of the s-trans diene appears at 234 nm (ε 20,000), which matches the long-axis polarized benzoate transition at 230 nm (ε 15,300). These ETDMs constitute a counterclockwise screw generating negative ﬁrst and positive

(a)

(b)

225.0 (+39.1) +40

58 H

H +20 Δε 0

CD

in EtOH

OBz

H

58

e × 10–4

200

–20

242.0 (–30.8) 235.0 (33,800)

–40

300

OBz-p-Cl 59 Me

O O

2

UV 200

4

λ (nm)

0

H O

H

CD in MeOH λext 247 nm, Δe –24.4 λext 224 nm, Δe +23.0 A = –47.4

H

Figure 4.26. (a) CD and UV spectra of 3-methylenecholest-4-en-6β-ol benzoate (58). (b) CD data of 3-oxocholest-4-en-6β-ol p-chlorobenzoate (59). (Redrawn from reference 1, with permission.)

145

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

second CEs. The exciton coupling between conjugated diene and benzoate chromophores determine the AC of diene–alcohol compounds. The nondegenerate ECM was applied to the conjugated enone–benzoate system, 3-oxocholest-4-en-6β-ol p-chlorobenzoate (59) (λext 247 nm, ε −24.4 : λext 224 nm, ε +23.0; A = −47.4) (Figure 4.26) [59]. The π –π ∗ transition of conjugated enone (λmax 241 nm, ε 16,600) couples with the p-chlorobenzoate transition (λmax 240 nm, ε 21,400) to generate negative ﬁrst and positive second CEs, stemming from the counterclockwise screw sense.

4.8.4. UV λmax Separation of Two Different Chromophores versus Exciton CD As described above, the exciton coupling is the most effective in a degenerate system having two identical chromophores. On the other hand, it is interesting to know how the exciton CD decreases, when the difference between UV λmax values of two chromophores increases. To clarify the effect of UV λmax separation on the exciton CD, we carried out the CD calculation by the exciton theory and also synthesized steroidal model compounds having two different p-substituted benzoate chromophores (Figures 4.27). Figure 4.27a shows the CD calculation result how the exciton CD decreases in intensity, when the separation between two λmax values increases [1]. The important aspect is that even when two chromophores undergo excitations at 230 nm and 310 nm, respectively, their interaction provides observable bisignate CEs, whose signs are governed by the exciton chirality between two ETDMs. The observed CD spectra of steroidal model compounds agree well with the calculated curves. Thus the ECM is applicable to nondegenerate systems, in which two λmax values are much separated. Based on this

223.0 (+15.7)

R=H

230 nm–230 nm

(1)

300 λ (nm) 350

200

238.5 (–15.2) 300 λ (nm)

200

200

250

300

R = Cl 350

230 nm–250 nm

(3) 200

300 230 nm–260 nm

(4) 200

300 230 nm–280 nm

(5) 200

250

200

300

350

229.5 (+14.5)

R = OMe

350 200

230 nm–257 nm 300

250 249.5 (–10.4)

350

350

O O

350

250

230 nm–240 nm

242.3 (–18.2)

O O

232.5 (+11.2)

230 nm–310 nm

(6)

350

227.0 (+16.8)

230 nm–240 nm

(2)

230 nm–230 nm

R = NMe2 R 230 nm–310 nm 350

(a)

200

250

312.0(–3.7)

350

(b)

Figure 4.27. (a) Calculated exciton CD curves, when changing λmax of two chromophores. (b) Observed CD spectra of cholest-5-ene-3β,4β-diol 3-benzoate 4-p-substituted benzoates in EtOH. (Redrawn from reference 1, with permission.)

146

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

mechanism, the CD allylic benzoate method was developed, as will be discussed in the next section.

4.8.5. Allylic Benzoate Exciton Method for Determining the AC of Allylic Alcohols

e × 10–4

0 CD

Δε –5

229.5 (–8.72) –10

(a)

229.0 (+11.45) +10

2 BzO

228.9 (16,400)

H 60

(b)

OBz

BzO CD

+5 Δε 0

61

1

60

200

250 λ (nm) 300 (a)

BzO 0

2

228.5 (13,100)

1 UV

e × 10–4

The chiroptical method for determining the AC of allylic alcohols was ﬁrst developed as the Mills’ rule [94] and the Brewster’s benzoate rule [95], in which the AC was correlated with optical rotation. Later the benzoate sector rule was proposed by Harada and Nakanishi to correlate the AC with CD CEs at 230 nm, where the position of the C=C double bond against the benzoate chromophore was a key factor to govern the CD CE [96]. As an extension of this concept and based on the nondegenerate exciton CD mechanism, the allylic benzoate exciton method was developed for determining the AC of allylic alcohols. Figure 4.28a shows the CD and UV spectra of cholest-4-en-3β-ol benzoate 60, where a negative CE is seen at the long-axis polarized UV band (230 nm) [1, 97]. The mechanism of this CE is interpreted as follows. The C=C double bond undergoes an allowed π –π ∗ transition around 190 nm polarized along its long axis. This transition couples with the long-axis transition of benzoate chromophore at 230 nm; and by this nondegenerate exciton coupling, the negative ﬁrst CE is observed around 230 nm. On the other hand, it is difﬁcult to observe the expected positive second CE, because the π –π ∗ band of C=C double bond locates below 200 nm. Since the long axes of two chromophores constitute a counterclockwise screw sense, the benzoate CE becomes negative as the ﬁrst CE of the non-degenerate exciton coupling. Another stereoisomer, cholest-4-en-3α-ol benzoate 61, exhibits a positive CE around 230 nm, reﬂecting the positive exciton chirality between benzoate and C=C double bond chromophores (Figure 4.28b). Note that the CD intensity of axial allylic benzoate is generally larger than that of equatorial allylic benzoate, as exempliﬁed in Figure 4.28 [1, 97]. Allylic benzoate 60 undergoes a negative CD below 210 nm, which is opposite to the expected positive second CE around 190 nm. A similar phenomenon is observed in the case of allylic benzoate 61. These results may originate from the participations of the

H 61

UV 200

250 λ (nm) 300

0

(b)

Figure 4.28. (a) CD and UV spectra of cholest-4-en-3β-ol benzoate 60 in EtOH. (b) CD and UV spectra of cholest-4-en-3α-ol benzoate 61 in EtOH. (Redrawn from references 97 and 1, respectively, with permission.)

147

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

benzenoid 1 B transition (around 200 nm) of benzene chromophore and the weak π –σ ∗ transition of the C=C double bond chromophore. Despite such complexity, the 230-nm CE reﬂects solely the allylic benzoate chirality, which is useful for determining the AC of allylic alcohols [1, 97]. More recent studies led to a signiﬁcant advance in application of allylic benzoate method to homoallylic alcohols, amines, and other enes by the use of microscale cross metathesis [98, 99]. By applying the microscale cross-metathesis protocol the double bond can be transformed into more suitable for coupling chromophores, such as styrene or p-substituted analogues, so the entire CD couplet can be observed and used for an AC determination. The introduction of ﬂuorescent styrene provides an additional beneﬁt, since the CD can be measured in emission by the more sensitive ﬂuorescent detected CD (FDCD) method [98b, 99].

4.9. RECENT APPLICATIONS OF THE CD EXCITON CHIRALITY METHOD The following are the recent interesting application examples of the CD ECM for determining the ACs of chiral natural and synthetic compounds. (1) ACs of Ciguatoxin and Related Compounds Intense exciton coupled CDs are useful for comparison of CD data with those of a reference compound as exempliﬁed below. The AC of C5 in CTX4A (62), a ciguatoxin precursor, was determined by the use of the exciton coupled CD spectra as summarized in Figure 4.29 [100]. The AB ring fragment (64) of CTX4A was stereoselectively synthesized, and its p-bromobenzoate 65 exhibits the exciton split CEs. It is obvious that these CEs originate from the exciton coupling between conjugated diene in the C5 side chain and C11 p-bromobenzoate, and these two chromophores constitute a clockwise screw sense (Figure 4.29). The CD spectrum of CTX4A tris(p-bromobenzoate) 63 also shows the exciton split CEs, which was interpreted as follows. Although there are six possible interactions among four chromophores in 63, the coupling between 1,3-diene and C11 p-bromobenzoate makes a dominant contribution, because the remaining ones are weak due to the remote distances. This interpretation was conﬁrmed by the CD data of 65. The ACs of CTX4A and ciguatoxin (CTX) were thus determined by comparison of the exciton CD CEs of 63 and 65. The ACs of the CTXs were later conﬁrmed by total syntheses [101].

Me CTX4A (62), R = H 63, R = OBz-Br-p

H 5

O

H

H

A 11 O O H H ORH H

H

Me H OR O H 47 O O O M O H H H H O H Me Me H Data of 63, CD (MeOH) λext 246 nm (Δe +32) 230 nm (Δe –28)

H Me RO O H HO

O O H H

32

H

H

A 5

O H H

O

H O

11

H OR

O

64, R = H 65, R = OBz-Br-p Data of 65, CD (MeOH) λext 242 nm (Δe +25) 225 nm (Δe –14)

Figure 4.29. ACs of ciguatoxin precursor CTX4A (62) and related compounds, along with their CD data.

148

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(2) AC of Urothion The AC of acyclic terminal 1,2-glycol was determined by the ECM as follows. To determine the AC of urothion (66), a yellow pteridine pigment isolated from human urine, the compound was subjected to desulfurization yielding 67, which was converted to tris(p-chlorobenzoate) 68 (Figure 4.30) [102]. An authentic sample of (R)-67 was synthesized starting from d-glucose. Since the [α]D values of diols (S )-67 and (R)-67 were too small to assign their ACs by comparison, tris(p-chlorobenzoates) (S )-68 and (R)-68 were prepared and their CD spectra were compared. The CD spectrum of (R)-68 was opposite in sign to that of (S )-68, thus leading to the (R)-AC of urothion 66. The bisignate CEs at 247 and 228 nm are mainly caused by the exciton coupling between the two benzoate groups in the side chain. According to the ECM applied to acyclic terminal 1,2-glycols (Section 4.8.2), the positive ﬁrst CE leads to the S conﬁguration, which agrees with that obtained by comparison of CD spectra. It should be noted that the intense exciton CEs are thus useful for determining the AC by comparison of chiroptical data. (3) AC of Cephalocyclidin A, a Five-Memberd Ring cis-α-Glycol The ECM was applied to the ﬁve-membered ring 1,2-cis-glycol system. The unprecedented pentacyclic structure of cephalocylidin A (69) was elucidated by X-ray crystallographic, 1 H-NMR, and CD methods [103]. To determine the AC, 2,3-bis(p-methoxycinnamate) (70) was synthesized, which showed bisignate CEs of negative exciton chirality (Figure 4.31). The observed exciton CD is weak, reﬂecting the small dihedral angle of cis-glycol moiety in a ﬁve-membered ring. The AC of compound 69 was determined as shown. It should be advised that in cases with small dihedral angle between two hydroxyl groups, the use of more symmetrical chromophore such as p-dimethylaminobenzoate or p-bromobenzoate would be suitable for obtaining unambiguous results.

O HN H2N

N

H OR

O SCH3

N N

HN

OH

S H

RHN

OR

N N

N

OH

(S)-68, CD (EtOH), λext 247 nm (Δe +8.0) 228 nm (Δe –3.0) UV (EtOH) λmax 241 nm (e 43,600)

(S)-67, R = H (S)-68, R = Bz-p-Cl

urothion (R)-66

Figure 4.30. Urothion and exciton CD data.

O

O

H H 3 OO 2 OHO

OH

HO

2 3

HO O 69

MeO N MeO

70

N O OO

Compound 70, CD (CH3OH) λext 325 nm (Δε –6.6), 283 nm (Δε +5.8)

O O 9 O 1 H

OAcO

HO H

NOEs

Compound 71, CD (CH3CN) λext 270.7 nm (Δe +20.3) 227.8 nm (Δe –18.1)

71

Figure 4.31. ACs of cephalocyclidin A 69 and dihydro-β-agarofuran sequiterpene 71 as determined by the CD ECM.

149

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

(4) AC of Dihydro-β-Agarofuran Sequiterpene Dihydro-β-agarofuran sequiterpene 71 is unique because the natural product itself contains cinnamate and benzoate chromophores at C1 and C9 in an ideal 1–3 transrelationship for exhibiting exciton CEs (Figure 4.31) [104]. The CD spectrum of 71 shows positive ﬁrst and negative second CEs, from which the AC was unambiguously determined as shown. (5) AC of Strevertenes This is an interesting example of the exciton CD due to the interaction between conjugated pentaene and p-dimethylaminobenzoate chromophores [105]. The relative stereostructures of strevertenes A (72) and G (73), antifungal macrolides, were determined by X-ray crystallography of compound 73, where the pentaene moiety adopts all transconﬁguration and almost planar conformation as shown in Figure 4.32. To determine the AC of compound 72, the CD ECM was applied as follows. As the UV λmax of the pentaene chromophore is located around 330 nm, p-dimethylaminobenzoate (λmax = 311 nm) was selected as an exciton coupling partner with a red-shifted absorption suitable for coupling with the pentaene moiety. Thus, methyl ester 74 was benzoylated to give a mixture of mono-benzoates, which was separated by reverse-phase HPLC. All possible six mono-benzoates were isolated, and the esteriﬁcation positions were assigned by 1 H NMR spectra. Of these benzoates, 15-p-dimethylaminobenzoate (75) was suitable for application of the ECM, because the allylic position generates a clear exciton chirality (Figure 4.32). The difference CD spectrum {CD(75) − CD(74)} between 15-benzoate 75 and alcohol 74 is shown in Figure 4.32, where the exciton coupling between pentaene and p-dimethylaminobenzoate chromophores generates intense positive ﬁrst and negative

336 (+37.6)

O

OH

HO

O

R1 OR2

Strevertene A (72): R1 = COOH, R2 = H methyl ester (74): R1 = COOMe, R2 = H derivative (75): R1 = COOMe, R2 = Bz-p-NMe2

C6

C4

C3

C5

Δε

H16 O

C14 H15

C7

+145°

+20 N

0 Diff. CD = CD(75) – CD(74)

in MeOH

C10

C9

354 (+18.8)

O

C11 C8

C12 O7 C29 C13 C1 C14 O9 O2 C16 O1 C20 C18 C22 C15 C24 C27 C17 C19 O8 C25 C23 C21 C31 C26 C28

H3CO2C

O6

O5

O4

O3 C2

+40

–20

303 (–21.4) UV(75) UV(74)

–40

e × 10–5

OH OH OH

320

336 353

1

in MeOH

C30

Strevertene G (73) : R1 = CH2OH, R2 = H 200

250

300 λ (nm)

350

400

Figure 4.32. ACs of strevertenes A (72) and G (73), X-ray crystallographic stereoview of (73), difference CD spectrum {CD(75) − CD(74)}, and UV spectra of (75) and (74). (Redrawn from reference 105, with permission.)

150

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

second CEs. The positive A value indicates that the long axes of both chromophores constitute a clockwise screw sense as shown in Figure 4.32, where the trans-antiperiplanar relation between H-15 and H-16 was assigned by the 1 H NMR coupling constant (J15,16 = 9.2 Hz). The ACs of 15-benzoate 75 and natural products 72 and 73 were thus determined [105]. (6) AC of Phomopsidin The example shows an exciton CD due to the interaction between remote diene-ester and p-nitrobenzoate chromophores. The CD spectrum of phomopsidin 76, a marinederived fungal metabolite, shows only one very weak CE at 266 nm associated with the diene-carboxylic acid chromophore at 6-C with an intense UV absorption at 266 nm (Figure 4.33a) [106]. To determine the AC, alcohol 76 was converted to p-nitrobenzoate 77, which exhibited bisignate CEs due to the exciton coupling between diene-ester and p-nitrobenzoate. From the positive ﬁrst CE, the AC having a clockwise helicity was determined. (7) AC of Spiroxin A, a Bis-Acetophenone Fungal Metabolite In this application, red-shifted chromophores were used to prevent the overlap of CEs. Spiroxin A (78) is a bis-acetophenone with a spiroketal moiety that locks the two conjugated chromophores (Figure 4.33b) [107]. The relative stereochemistry had previously been established by NMR. Spiroxin A 78 exhibits a complex CD in the 200 to 280 nm region, which was difﬁcult to interpret. To determine the AC, two phenolic hydroxyl groups were esteriﬁed with retinoic acid because the UV band of all-trans retinoic acid ester (λmax 356 nm, ε 39,500) is well red-shifted from the absorption of the existing chromophore. However, while spiroxin A bis(retinoate) (79) did not exhibit useful exciton couplet in the retinoic ester region, the difference CD between bis(retinoate) 79 and spiroxin A 78 provided a clear-cut negative exciton couplet that allowed for the AC assignment [107]. (8) AC of Pinellic Acid The CD allylic benzoate method was applied to pinellic acid 80, a long-chain allylic alcohol, to determine the AC as shown in Figure 4.34 [108]. The relative conﬁguration COOR1

(a)

O

(b) 76, R1 = R2 = H 77, R1 = CH3, R2 = p-NO2Bz

ax

OR

spiroxin A 78, R = H CI spiroxin A bis(retinoate) 79, R =

O

O

77

H

CD (MeOH), λext 271 nm (Δε +22.4) λext 244 nm (Δε –7.9) UV (MeOH), λmax 264 nm (ε 35,000)

6 11

eq H R2O

O

O O OR O

Difference CD = CD (79) – CD (78), λext 385 nm (Δε –17.3), 331 (+17.4)

Figure 4.33. (a) CD and UV data of phomopsidin methyl ester p-nitrobenzoate 77. (b) Spiroxin A 78 and CD data.

OR4 12 9 13 R1O 10 OR3 pinellic acid 80, R1 = R2 = R3 = R4 = H O

OR2

81, R1 = CH3, R2 = p-BrBz, R3 & R4 = acetonide

Figure 4.34. Pinellic acid 80 and exciton CD.

81, J9,10 = 7.0 Hz CD (CH3OH), λext 245 nm (Δe +6.9), 221 (+2.13), 209 (+5.7)

151

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

of three chirality centers was determined by NOE experiments of methyl ester-acetonide, from which the syn conﬁguration of the 12-C/13-C vicinal diol was assigned. The acetonide was then converted to p-bromobenzoate 81, the 1 H NMR spectrum of which indicated an antiperiplanar relationship between the 9 and 10 protons (J9,10 = 7.0 Hz). The CD of the allylic p-bromobenzoate showed a positive CE at λext 245 nm (ε +6.97), indicative of the S conﬁguration at 9-C. This predicted the AC of pinellic acid to be either (9S,12S,13S ) or (9S,12R,13R). The remaining question was solved by a stereospeciﬁc synthesis of both diastereomers. The comparison of spectral data with those of the authentic samples indicated that the AC of the natural product was (9S,12S,13S) as shown. (9) AC of Phorboxazole The CD allylic benzoate method was applied to determine the AC of allylic alcohol moiety as follows. The AC of phorboxazole A (82) was determined as shown by total synthesis except for the conﬁguration of 38-C (Figure 4.35) [109]. Although the AC at the 38-C allylic alcohol had originally been assigned as R by application of the Mosher MTPA method, there was an anomaly in the NMR δ data. The threo and erythro model compounds (83 and 84) were synthesized from (S )-malic acid, and hence the ACs of 33C, 35-C, and 37-C were assigned as shown. To determine the AC of 38-C by the allylic exciton method, these alcohols were converted to 2-naphthoate esters (85 and 86). The NMR coupling constants J38,39 = 9.6 Hz for 85 and J38,39 = 9.2 Hz for 86 indicate that these two protons are in trans-relationship in their stable conformations. Threo ester 85 exhibited a negative CE at λext 234 nm (ε −9.2), indicating a negative twist between the naphthoate and the allylic double bond. In contrast, erythro product 86 showed a positive CE at λext 234 nm (ε +15.1) indicating a positive helicity. Thus the ACs of 83 and 84 were determined as shown. The comparison of the NMR coupling constant J37,38 of 82 with those of 85 and 86 led to the AC of natural product 82 as shown in Figure 4.35 [109]. (10) AC of Gymnocin-B The experiments for the AC determination of gymnocin-B 87, a cytotoxic marine natural product with the largest 15-rings polyether skeleton isolated so far, are very revealing. They demonstrate the effectiveness of porphyrin chromophore to serve as chiroptical probe for AC of remote stereogenic centers residing in a very ﬂexible substrate, available only in a few hundred micrograms (Figure 4.36) [99, 110b]. The compound has two hydroxyl groups at 10-C and 37-C, but it was difﬁcult to clearly assign the helicity between the two remote and sterically hindered OH groups, because of the conformational ﬂexibility arising from the presence of seven-membered rings.

OH

O

phorboxazole A 82, J37,38 = 7.9 Hz

N O

OCH3 OCH3 Br

35

39 38

37

H OH

O

OH

O

O

N O

O O

OCH3

83, R = H n-C4H9

39 38

84, R = H

35 37

H OR

O

n-C4H9 OMe

39

38

OCH3 35 37

H O OMe OR

threo-85, R = 2-naphthoyl J37,38 = 7.0 Hz, J38,39 = 9.6 Hz

erythro-86, R = 2-naphthoyl J37,38 = 3.7 Hz, J38,39 = 9.2 Hz

CD (CH3CN) λext 234 nm (Δe –9.18)

CD (CH3CN) λext 235 nm (Δe +15.1)

Figure 4.35. Phorboxazole A 82 and allylic benzoate method.

152

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

OR H H 37 RO H O Me H H H H H O H 10 O O O H I H J H K C D E F G B O O H O L O H A H H O O H O Me Me 5 Me H H H H Me H O M Me N O H O H O O 55 Gymnocin B 87, R = H H 54 O bis(TPP-cinnamate) 88 NHN O NHN R= CD (MeOH)

29°

15 10

10′

20 5

15′ 20′

5′

C5/Ph rotation

10-TPPcin K (ax) J

λext 419 nm, Δe +11 λext 414 nm, Δe –15

37-TPPcin (eq)

28.7 Å

Figure 4.36. Gymnocin B 87, the lowest-energy conformation of its 10,37-bis(TPP-cinnamate) derivative 88 obtained by Monte Carlo/MMFF94s, and CD data of 88. (Redrawn from reference 110b, with permission.)

Triphenylporphyrin-cinnamate chromophores were introduced into 10-OH and 37-OH by acryloylation/cross-metathesis under microscale conditions. The obtained diester 88 showed a clear exciton split CD, positive ﬁrst and negative ˚ apart (Figure 4.36). After extensecond CEs, even though the two porphyrins are ∼30 A sive conformational analysis of the derivative by the MMFF94s/Monte Carlo calculation, the (10S,37S) AC was assigned to the derivative 88. The CD curve of 88 calculated by the DeVoe’s coupled oscillator method for the Boltzmann-weighted conformers agreed well with the observed CD spectrum. The AC of gymnocin B 87 was thus determined as shown [99, 110b]. (11) AC of Axially Chiral Binaphthoquinones The exciton CEs of 1,2 -binaphthyl derivative are well suited for AC determination than the CEs of binaphthoquinone as shown below. The AC of (–)-8 -hydroxyisodiospyrin 89, a naturally occurring bi(naphthoquinone), was determined by the synthesis of the opposite enantiomer (S )-(+)-89 (Figure 4.37) [111]. The AC of a synthetic intermediate was determined by X-ray analysis, and the intermediate was then converted to binaphthalene 90, with its CD showing intense exciton coupled CEs. From the positive sign of the couplet, an S conﬁguration was assigned to 90, which was further converted to (S )-(+)89. The AC of the natural product was determined to be (R)-(−)-89. Thus the ACs of these compounds were conﬁrmed by the CD ECM [111]. The CD spectrum of (S )-90 shows intense exciton CEs, while (S )-(+)-89 exhibits two positive and one negative CEs around 360–260 nm, but their ε values are much smaller than those of 90, and the CD curve deviates from the ideal pattern of the exciton coupling. Thus, to determine the ACs by the exciton method, it is important to select the

O

OMe OMe

OH

(S)-(+)-89 (S)-90 MeO MeO MeO

OMe OMe

225 nm (Δε –60.0)

CD (1,4-dioxane)

O O

CD (CH3CN) λext = 239 nm (Δε +46.7), HO

λext = 356.9 nm (Δε +6.9), O

298.5 nm (Δε +11.2), 263.8 nm (Δε –23.2)

OH

Figure 4.37. ACs and CD data of axially chiral binaphthalene and binaphthoquinone. Numerical CD data were obtained from the spectra reported in reference 111.

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

most appropriate chromophores—that is binaphthalene rather than binaphthoquinone as exempliﬁed in this case. (12) AC of Pre-anthraquinones The example shows an intense exciton coupling between naphthalene–ketone and anthraquinone. The ACs of atropisomeric pigments 91 and 92 were determined by spectroscopic methods including ECM (Figure 4.38) [112]. Atropisomer 91 exhibits intense negative ﬁrst and positive second CEs at 272 nm and 251 nm, respectively, thus leading to an AC with negative helicity between long axes of two aromatic chromophores. The AC at the 3 position was deduced by chemical correlation; the reductive cleavage of 93 yielded (R)-torosachrysone methyl ether with known AC. Pigment 92 exhibits weak bisignate CEs compared to 91, because the two aromatic chromophores are connected by two σ bonds, and hence the dihedral angle between two long axes is close to 180◦ . (13) AC of Spiroleptosphol Figure 4.39 shows an interesting example of triol tribenzoate (97) where the AC has been successfully determined in the presence of three exciton interactions. Spiroleptosphol (93), a γ-methylidene-spirobutanolide, exhibited cytotoxicity against P388 murine leukemia and HeLa human cervix carcinoma (Figure 4.39) [113]. The relative stereostructures of compound 93 and spiroleptosphol C (94) were determined by NMR spectroscopy and/or by X-ray crystallography [114]. Compound 93 was converted to dibenzoate 96, which exhibited typical exciton CEs of negative exciton chirality leading to a counterclockwise helicity between two benzoate chromophores at the 6- and 7-positions (Figure 4.39). The AC of 93 was thus unambiguously determined as shown [114]. Triol 93 was also converted to tribenzoate 97, the CD of which showed intense positive ﬁrst and negative second CEs [114]. These CD data are interpreted as follows: The exciton chirality between 4- and 6-benzoates is clockwise, and that between 4- and 7-benzoates is also clockwise, while that between 4- and 6-benzoates is anticlockwise. The total exciton chirality becomes thus positive and leads to positive ﬁrst and negative second CEs, conﬁrming the AC of triol 93. (14) AC of Leucettamol A, α,ω-Bifunctionalized Sphingolipid The intense exciton CEs are useful for amplifying the chiroptical properties of natural products with very weak optical rotation as exempliﬁed below, where the deconvolution ECM was also useful for AC determination. Leucettamol A (98) was isolated as a marine natural product exhibiting a variety of biological activities. The relative conﬁguration of 2-amino-3-hydroxy end groups was determined to be erythro by NOE experiments of a

OMe OH O

OMe OH O

(–)-92

(+)-91 3'

3'

MeO

CD (MeOH), λext 272 nm (Δε –159.3) 251 nm (Δε +158.8) UV (EtOH) λmax 274 nm (log ε 4.50)

OH O HO

OH O

OMe

MeO

OH O HO

OH O

OMe

Figure 4.38. Pre-anthraquinones and CD data.

CD (CHCl3), λext 286 nm (Δε +14.4) 257 nm (Δε –8.2) UV (EtOH) λmax 278 nm (log ε 4.36)

153

154

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

O

O 4

O HO

12

OH 17 R

6 HO 7

spiroleptosphol (93), R = H spiroleptosphol C (94), R = OH

O RO

4

O

OH

O O

6

RO 7

O (95), R = H (96), R = Bz

4 6

(97)

BzO 7

O

H

O O

6

OBz

O

7 H O

OH R O

CD (MeOH) of (96) λext 238 nm, Δe –18.9 222 nm, Δe +10.9

O O

H

(96), R = side chain

(93)

O BzO

H 6

H4O R O

7 H O

CD (MeOH) of (97) λext 238 nm, Δe +38.4 220 nm, Δe −18.1 204 nm, Δe −6.4

(97), R = side chain

Figure 4.39. Determination of the absolute stereochemistry of spiroleptosphol (93) by the CD ECM. The CD data of (97) were obtained from the reported ﬁgure.

bis-oxazolone derivative. Compound 98 was originally assigned to be racemic because it did not exhibit any measureable optical rotation [115]. To reinvestigate the AC of leucettamol A 98, it was catalytically reduced to give perhydro-derivative 99, which was then converted to N,N ,O,O -tetrabenzoyl derivative 100. The CD spectrum of compound 100 showed bisignate CEs due to the exciton coupling between benzoate/benzamide chromophores (Figure 4.40) [116]. Since the 2,3-N,Odibenzoyl chromophores are remote from 28,29-N,O-dibenzoyl chromophores, the simple additivity of exciton coupled CD is applicable. Erythro-101, selected as a model and prepared from (2S )-alanine, exhibited a negative exciton couplet (Figure 4.40) [117]. The simulated CD curve of (ent-erythro-101 + ent-erythro-101) agreed well with the observed CD of (erythro / erythro)-100 (Figure 4.40), but other combinations—for example (threo-102 + threo-102) and (ent-erythro-101 + threo-102)—disagreed with the observed CD of 100. The ACs of 100 and hence 98 were unequivocally determined to be (2R,3S,28S,29R). Thus leucettamol A is optically active; redetermination of the speciﬁc rotation of 98, averaged over 10 measurements, gave [α]D = −3.8±0.1 (c 4.4,

OH

NH2

NHBz 3

OH leucettamol A (2R,3S,28S,29R)-(–)-98 NHR

OR

99: R = H

100: R = Bz

NH2 OR

OBz erythro-101

CD (MeOH) λext = 235 nm, Δe = –5.6 λext = 220 nm, Δe = +1.6

NHR NHBz

erythro l erythro-100: obsd CD (MeOH), λext = 238 nm, Δe = +10.3, λext = 222 nm, Δe = –2.8 simulated CD based on (ent-erythro-101 + ent-erythro-101): λext = 235 nm, Δe = +11.3, λext = 220 nm, Δe = –3.2

OBz threo-102

CD (MeOH) λext = 237 nm, Δe = +3.0 λext = 221 nm, Δe = –3.5

Figure 4.40. ACs and CD data of leucettamol A (98) and related compounds.

155

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

MeOH). The deconvolution exciton CD method is thus useful for determining the AC of dimeric sphingolipids. (15) AC of 11-Deoxydiaporthein A by the Exciton Allylic Benzoate Method The AC of allylic alcohol was determined by ECM and X-ray analysis as shown below. 11-Deoxydiaporthein A (103) was isolated from a marine fungus [118], and its relative stereostructure was determined by NMR. To determine its AC, compound 103 was converted to benzoate (104), the CD spectrum of which showed a positive CE (λext = 230 nm, ε +5.8), reﬂecting a clockwise helicity between the long axes of benzoate and oleﬁn chromophores (Figure 4.41a). The R conﬁguration at 7-C was determined by the CD allylic benzoate method. The AC determined by CD was conﬁrmed by X-ray crystallography of a single crystal of compound 103 obtained by recrystallization from chloroform. Interestingly, the crystal contained chloroform molecules as crystal solvent, and based on the strong anomalous scattering effect of chlorine atoms, the AC of 103 was established [118]. (16) AC of Cortistatin A This is an example of a natural product whose AC was determined by the exciton coupling of its two preexisting trans-diene and isoquinoline chromophores. Cortistatin A (105), an anti-angiogenic steroidal alkaloid, was isolated from a marine sponge. The relative stereostructure of 105 was elucidated by 2D-NMR (COSY and NOESY) and conﬁrmed by X-ray as shown in Figure 4.41b. Interestingly, compound 105 itself exhibited bisignate CEs of negative exciton chirality, leading to the AC as shown [119]. Compound 105 has a conjugated diene (UV λmax = ∼234 nm) and a C17 isoquinoline (UV λmax = ∼220 nm), which couple to generate exciton CEs. The data indicate that the long axes of the two chromophores constitute a counterclockwise screw in the conformation shown in Figure 4.41, leading to the AC shown. (17) AC of trans-Acenaphthene-1,2-diol The ECM is applicable to glycols with 120◦ dihedral angle as follows. Chiral transacenaphthene-1,2-diol (106) and related compounds were prepared by baker’s yeastmediated reduction, and the AC of (−)-trans-106 was determined to be (1S,2S) by the CD ECM [120]. Diol (−)-106 (97% ee) was converted to bis(p-dimethylaminobenzoate) (+)-107, which exhibited positive ﬁrst and negative second CEs. The negative exciton

OH

H

O 7 OR

OH OH

11-deoxydiaporthein A (103), R = H AC of (103) by X-ray of crystal (103)/CHCI3 (a)

(104), R = Bz CD λext = 230 nm, Δe = +5.8

OH

11 17

HO O N

N

H H N

H

cortistatin A (105) relative structure of (105) by X-ray

H CD of (105) λext 237 nm, Δe = –17 217 nm, Δe = +35

17S

(b)

Figure 4.41. (a) AC of 11-deoxydiaporthein A (103) as determined by the CD allylic benzoate and X-ray methods. (b) AC of cortistatin A (105) as determined by X-ray and ECM methods.

156

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

CD couplet led to the (1S,2S) AC as shown in Figure 4.42a. When reporting the CD data, the intensity should be provided in ε units, not in raw θ /m◦ units, otherwise it is impossible to judge the intensities of the observed CEs. (18) AC of Kolokoside A Aglycone The ECM is useful for the AC determination of cyclic 1,2-glycols as exempliﬁed below. The AC of kolokoside A (108), a triterpenoid glycoside, was determined by the CD ECM (Figure 4.42b) [121]. The aglycone was esteriﬁed to give 2,3-bis(pdimethylaminobenzoate) (109), whose CD spectrum showed negative ﬁrst and positive second. Thus the (2R,3R) AC of the aglycone was determined. (19) AC of Dinemasone B This is an example of a cyclic trans-1,2-glycol. The AC of dinemasone B (110), a bioactive metabolite from fungi, was determined by the CD ECM [122]. Glycol 110 was converted to bis(p-bromobenzoate) (111), which exhibited a negative exciton CD. The AC of 110 was thus assigned as shown in Figure 4.42c. (20) AC of Oligonaphthalenes The following are unique chiral oligonaphthalenes whose ACs were determined by porphyrin–porphyrin coupling at very long distances (Figure 4.43) [123]. These chiral oligomers were synthesized from chiral (S )-binaphthyl derivative by repeating the oxidative coupling giving the diastereomeric products—that is, 4-mers, 8-mers, and 16-mers. For example, the coupling reaction of (S,S,S )-4-mer yielded diastereomeric (S,S,S,S,S,S,S )-8-mer and (S,S,S,R,S,S,S )-8-mer. To determine their ACs, the CD ECM using tetraphenylporphyrin (TPP) carboxylic acid was applied, because the Soret band (λmax = 420 nm) of TPP-carboxylic ester is far separated from the naphthalenoid UV bands. It has been previously reported that the exciton CD method using TPP-esters was applicable to remote hydroxyl groups, where the interchromophoric distance R ranges ˚ [47, 48]. around 50 A Two terminal phenol groups of diastereomers were esteriﬁed to yield bis(TPP-ester) 112 and 113. When applying the exciton method to these systems, it is important to elucidate the dihedral angle between naphthalene moieties. The X-ray crystallographic

N

N

N

H

HOOC O

O

O

H

O

O O

O

O O

(1S,2S)-(+)-trans-107, 97%ee CD (MeOH)

N

H

CD (CH3CN)

(a)

OH O

O O O

109

Kolokoside A aglycone bis(p-dimethylaminobenzoate)

λext 329 nm, postive λext 308 nm, negative

O

O

Br 111

Br Dinemasone B, bis(p-bromobenzoate) CD (CH3CN)

λext 320 nm, Δe = –22 λext 296 nm, Δe = +13

λext 252 nm, Δe = –45.3 λext 235 nm, Δe = +22.9

(b)

(c)

Figure 4.42. Application of ECM and ACs: (a) (1S,2S)-(+)-trans-acenaphthene-1,2-diol; (b) kolokoside A aglycone; (c) dinemasone B.

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

(S,S,S,S,S,S,S,S,S,S,S,S,S,S,S)-114,

(S,S,S,S,S,S,S,S,S,S,S,S,S,S,S)-114, UV (CHCI3) λmax 420.0 nm, e = 756,000 CD (CHCI3) λext 434 nm, Δe = –2.8 418 nm, Δe = +2.7

n = 7, R = 66 Å HN N

H N

N O

O

OO

OO

OO O

O

(S,S,S,S,S,S,S)-112, n = 3 UV (CHCI3) λmax 420.0 nm, e = 725,000 CD (CHCI3) λext 428 nm, Δe = –10.1 418 nm, Δe = +12.6

n (S,S,S,S,S,S,S)-112, n = 3 O

O

O

O O

OO

O

OO

O

O O

OO

O

O

O

(S,S,S,R,S,S,S)-113, n = 3 UV (CHCI3) λmax 420.0 nm, e = 755,000 CD (CHCI3) λext 428 nm, Δe = +10.7 418 nm, Δe = –11.1

Figure 4.43. Exciton CD data and the AC of oligonaphthalenes.

analyses of some derivatives and the CONFLEX-MM2 calculation indicated the average dihedral angle to be around 89.4◦ ∼90.0◦ , that is a right angle. Thus, it is predicted that if bis(TPP-ester) 8-mer takes an (S,S,S,S,S,S,S ) conﬁguration, the dihedral angle between two terminal TPP-ester groups becomes −90◦ . That is, it is calculated 90◦ × 7 = (360◦ × 2) − 90◦ . Thus a negative exciton couplet is predicted. If bis(TPP-ester) 8-mer takes an (S,S,S,R,S,S,S ) conﬁguration, the dihedral angle between two terminal TPP-ester groups becomes +90◦ , because 90◦ × 6 − 90◦ = (360◦ ) + 90◦ . That is a positive exciton couplet is predicted. As listed in Figure 4.43, bis(TPP-ester) 8-mer 112 exhibited negative ﬁrst and positive second CEs, while bis(TPP-ester) 8-mer 113 exhibited an opposite couplet. Thus the observed CD curves are mirror images of each other. Thus, an (S,S,S,S,S,S,S ) conﬁguration was assigned to 112 and an (S,S,S,R,S,S,S ) conﬁguration was assigned to 113. The ECM method is applicable to a more remote diol diester—for example bis(TPP-ester) 16-mer 114, where two chromophores are separated at ∼66 ˚ but still gave an observable exciton CD. From the sign of the ﬁrst CE, an A, (S,S,S,S,S,S,S,S,S,S,S,S,S,S,S ) AC was assigned to 16-mer 114 (Figure 4.43) [123]. (21) AC Assignment of Acetylene Alcohols by the CD ECM The combination of Sonogashira reaction and ECM enables the AC determination of chiral terminal acetylene alcohols as shown below. Enantiopure acyclic acetylene alcohols are employed as chiral synthons for bioactive compound syntheses. It is generally difﬁcult to determine the ACs of acyclic acetylene alcohols, but the 1 H NMR anisotropy method using MαNP acid ester has been developed for enantioresolution of racemic alcohols and simultaneous determination of AC. The CD ECM is also useful for determining ACs of acetylene alcohols. A method for determining the ACs of terminal acetylene alcohols was developed from the exciton CD due to the interaction between p-methoxyphenylacetylene (λmax 252 nm) and p-methoxybenzoate (λmax 257 nm) [124]. For example, acetylene alcohol (+)115 was converted by a Sonogashira reaction to p-methoxyphenylacetylene alcohol (−)116, which was esteriﬁed giving benzoate (−)-117 (Figure 4.44). The CD spectrum of (−)-117 showed exciton bisignate CEs (Figure 4.44). The negative A-value led to an (R)-AC for (−)-117, consistent with the result from the 1 H NMR anisotropy method. This method is thus useful for determining the ACs of terminal acetylene alcohols [124].

157

158

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

+20 CD

H OH

(R)-(–)-117

–10 –20

obsd CD 266.0 (–27.9) 246.2 (+21.0)

–30

A = –48.9 UV

e x 10–4

+10 Δe 0

300

(R)-(–)-116

CH3O

OCH3 252 nm

obsd UV 258.2 (41,500)

250

(R)-(+)-115

4

H O

O

(R)-(–)-117

2 CH3O

in EtOH 200

H OH

257 nm

350

Figure 4.44. A scheme for the AC determination of terminal acetylene alcohol by combination of the Sonogashira reaction and the CD ECM: CD and UV spectra of 1-(4-methoxyphenyl)1-dodecyn-3-ol 4-methoxybenzoate (R)-(−)-117 in EtOH. (Redrawn from reference 124, with permission.)

(22) ACs and Exciton CD of Unique 1,3-Diethynylallene Compounds The ACs of allene compounds can be determined by the ECM as exempliﬁed below. Allene compounds are unique chiral synthons devoid of chirality centers, and they have been employed for the syntheses of various chiral compounds. Allene (M )-(−)-118 is one of such compounds, the AC of which was determined as follows. Compound (−)-118 was converted to 1,3-bis{(4-dimethylaminophenyl)ethynyl}allene (−)-119, with typical intense exciton split CEs (Figure 4.45) [37, 125]. The negative A value indicated that two identical chromophores of (4-dimethylaminophenyl)ethynyl oleﬁn constitute a counterclockwise sense, and thus an (M )-AC was assigned to allene (−)-119. It should be noted that the dihedral angle between two (4-dimethylaminophenyl)ethynyl groups is 90◦ due to the allene skeleton, which satisﬁes the requirement of an ideal exciton CD mechanism. The CD spectral curves of (M )-(−)-119 as calculated by TDDFT and π -electron SCF-CI-DV MO methods were in good agreement with the observed spectrum conﬁrming the above assignment by the ECM [37]. These absolute conﬁgurational assignments were consistent also with the X-ray analysis and chemical correlation results [125].

H i-Pr3Si (M)-(–)-118

N N

(M)-(–)-119 (e.r. > 91:9) CD (hexane) λext 328 nm, Δε = –77.9 287 nm, Δε = +37.2

(M)-(–)-119

Figure 4.45. Exciton CD of 1,3-diethynylallene derivative.

159

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

4.10. EXAMPLES REQUIRING CAUTION AND THEORETICAL ANALYSIS OF EXCITON CD The following are examples, where cautious considerations of CD were necessary. That is, the observed CD spectra could not be interpreted by the exciton coupling mechanism in a straightforward manner. (1) AC of Antitumor Antibiotic AT2433-A1 with a Secondary Amino Group To determine the AC of antitumor antibiotic AT2433-A1 (120), amino sugar bis(p-Brbenzoyl) derivative (121∗ ) was derived from the natural product (Figure 4.46). Its CD spectrum showed a negative couplet leading to the AC as shown [126]. However, it was later pointed out that this assignment was wrong as explained below [50]. To clarify the reasons for the wrong assignment, the authentic samples 121 and 122 were synthesized from a starting material with known AC. Surprisingly, the CD of 121 showed a weak positive exciton couplet, while that of 122 showed a strong negative one. The 1 H NMR of 121, a benzamide derivative of the secondary amine, indicated the presence of (Z ) and (E ) amide isomers, where the exciton chirality is negative and positive, respectively. Since the CD contributions cancel in some extent, the sign and intensity of observed CD are governed by those of the prevailing (E ) amide. On the other hand, the CD of the primary amine derivative 122 reﬂects straight the AC because of its (Z ) conformation. Therefore, when the ECM is applied to secondary amines, the analysis of (E ) and (Z ) conformations is critical. The AC of the secondary amine in natural product 120 was conﬁrmed by the total synthesis of a related compound [50]. (2) Anomalous CD CEs of 1,1 -Biphenanthryl Compounds Enantiopure 2,2 -dimethoxy-1,1 -biphenanthryl (aR)-(+)-123 (Figure 4.47) was synthesized by oxidative coupling of 2-phenanthrol, followed by enantioresolution, where its AC was determined by the axial chirality recognition method [127]. It was expected that

CH3 N O

O

O

CI OH

O O HN CH3 OH

p-BrBzN CH3

N

N H O

OH OCH3

OCH3

O

p-BrBzO p-BrBzN

p-BrBzN Me

OBz-p-Br

R

121*(AC originally assigned) CD (CH3CN), λext 249 nm (negative CD) 219 nm (positive CD)

AT2433-A1 (120)

O H Ar N O Ar Me O H

Ar O H OCH3

121, (Z )-amide negative chirality

O Ar

H

N O Me O H

O

Ar H Ar

OCH3

121, (E)-amide positive chirality

O H N O H O H

OCH3

OBz-p-Br

121, R = Me CD (CH3CN), λext 251 nm (Δe +5.5) 219 nm (Δe –4.4)

O H OCH3

122, (Z )-amide negative chirality

122, R = H CD (CH3CN), λext 252 nm (Δe –33.4) 234 nm (Δe +12.6)

Figure 4.46. Application of the ECM to secondary amine. Numerical CD data were obtained from the spectra reported in reference 50.

160

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

+200

(aR)-(+)-123 +100 CD Δe

(aR)-(+)-125

OCH3

OCH3

(aR)-(–)-124

OCH3

OCH3

e x 10–5

0

(aR)-(–)-124

(aR)-(+)-125

–100 2 OCH3 –200

200

OCH3

UV

250

1

300

350 λ (nm)

0

(aR)-(+)-123

(aR)-126

Figure 4.47. Chemical correlation of 1,1 -biphenanthryl derivative (aR)-(+)-123 and related compounds, along with their CD and UV spectra. (Redrawn from reference 127, with permission.)

the CD spectrum of (aR)-(+)-123 should show a negative exciton couplet around the 1 Bb transition of phenanthrene chromophore, because the two long axes of phenanthrene groups constitute an anticlockwise screw sense as illustrated in Figure 4.47. However, compound (aR)-(+)-123 showed positive and negative CEs at 274 nm 258 nm, respectively, disagreeing with the expectation. The anomalous CD results led to the reinvestigation the ACs of (+)-123 and related compounds. Starting from (aR)-(+)-1,1 -binaphthyl-2,2 -diol, compound (aR)-(−)-124 was synthesized, and it was easy to convert (aR)-(−)-124 to the target compounds (aR)-(+)-123 and (aR)-(+)-125 with a binaphthyl chromophore (Figure 4.47). The CD spectrum of (aR)-(+)-125 showed intense exciton CEs (λext 239.8 nm, ε −197.1; λext 227.8 nm, ε +133.3; A = −330.4) reﬂecting a negative helicity between two long axes of naphthalene moieties, which conﬁrmed the (aR)-AC of (+)-125. Compound (aR)(−)-124 also exhibited exciton CEs (λext 249.6 nm, ε −112.6; λext 236.4 nm, ε +60.5; A = −173.1), but the A value decreased to about half. That is, the conjugation of the naphthalene chromophore with a double bond diminished the exciton CD intensity, because the corresponding ETDM deviated from the long axis of the naphthalene chromophore [127]. In contrast, compound (aR)-(+)-123 exhibited positive and negative CEs (λext 274.2 nm, ε +38.5; λext 258.0 nm, ε −52.5; A = +91.0) (Figure 4.47). These results conﬁrmed that the CEs of 1,1 -biphenanthryl derivative (+)-123 are opposite in sign to those of 1,1 -binaphthyl derivative (+)-125 despite the same ACs. Furthermore, the CD shape of (+)-123 is complex, and the A value is about 1/4 compared to that of (aR)-(+)125. Additionally, more intense negative and positive CEs were observed around 220 nm. To gain insight into the anomalous CD behavior of compound (aR)-(+)-123, the CD and UV spectra of 1,1 -biphenanthryl (aR)-126 were calculated by the π -electron SCF-CI-DV MO method (for this MO method, see Chapter 5, this volume). The UV

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

calculation of phenanthrene chromophore revealed its complexity due to the presence of another electronic transition polarized along the short axis around 255 nm in addition to the intense long axis-polarized 1 Bb transition around 260 nm. Therefore, in (aR)-126, four ETDMs interact with one another, giving rise to positive ﬁrst and negative second CEs. The simulated CD and UV spectra agreed well with the observed; the (aR)-AC of (+)-123 was conﬁrmed also by the MO calculation [127]. It should be emphasized that the CD ECM itself is correct, but the electronic transitions of compounds (aR)-(+)-123 and (aR)-126 are complex, and therefore the simple and qualitative application is not valid for these compounds. In general, symmetrical chromophores (e.g., linear polyacenes such as naphthalene, anthracene, etc.) are more suitable for the CD ECM than the less symmetrical ones such as nonlinear condensed aromatics (e.g., phenanthrene).

4.11. CONCLUSION As discussed above, the CD ECM is useful for determining the ACs of various chiral compounds. The exciton coupling between two or more chromophores generates exciton split and intense bisignate CEs that reﬂect the helicity between ETDMs (positive or negative exciton couplet). The AC of the compound can be unambiguously determined from the sign of the couplet. In general, the exciton couplet CDs due to through-space chromophoric interaction are much stronger than the CEs of isolated chromophores, such as those due to ketone n → π ∗ , benzenoid π → π ∗ , and conjugated diene or enone π → π ∗ transitions. Their unique bisignate shapes facilitate the recognition of an exciton couplet. The CD ECM is readily proved by the quantum mechanical exciton theory as described, and therefore it is classiﬁed as a nonempirical method. Thus the ACs of chiral compounds can be determined by the exciton CD method without any reference compound. It was established that both X-ray Bijvoet and CD exciton chirality methods give the same AC, although assignments are based on totally different phenomena. However, the unambiguous determination of AC by ECM requires a very careful selection of the appropriate chromophores. It is critical that this selection takes into account not only the structural and conformational features of the chiral substrate, but also other basic requirements of this method. We hope that this chapter clariﬁes the main aspects of the CD exciton chirality method and provides useful guidelines for its application in stereochemical analysis.

ACKNOWLEDGMENTS The authors thank the co-workers in these studies, whose names are listed in references, and Dr. George A. Ellestad, Department of Chemistry, Columbia University, for his valuable suggestions. We are very grateful to JASCO Co., for their continuing instrumental and technical support.

REFERENCES 1. N. Harada, K. Nakanishi, Circular Dichroic Spectroscopy—Exciton Coupling in Organic Stereochemistry, University Science Books, Mill Valley, CA, and Oxford University Press, Oxford, 1983. 2. N. Harada, K. Nakanishi, Acc. Chem. Res. 1972, 5 , 257–263.

161

162

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

3. N. Berova, K. Nakanishi, Exciton chirality method: Principles and applications, in Circular Dichroism: Principles and Applications, K. Nakanishi, N. Berova, and R. Woody, eds., John Wiley & Sons, New York, 2000, pp. 337–382. 4. N. Berova, N. Harada, K. Nakanishi, Electronic spectroscopy: Exciton coupling, theory and applications, in Encyclopedia of Spectroscopy & Spectrometry, J. Lindon, G. Tranter, and J. Holmes, eds., Academic Press, London, 2000, pp. 470–488. 5. N. Berova, G. A. Ellestad, N. Harada, Modern methods in natural products chemistry: characterization by circular dichroism spectroscopy, in Comprehensive Natural Products Chemistry II, L. Mander and H. W. Liu, eds., Elsevier, Oxford, 2010, Chapter 9, pp. 91–146. 6. S. Superchi, E. Giorgio, C. Rosini, Chirality 2004, 16 , 422–451. 7. N. Berova, L. Di Bari, G. Pescitelli, Chem. Soc. Rev . 2007, 36 , 914–931. 8. (a) L. M. Pasteur, C. R. Acad. Sci. Paris 1848, 26 , 535–538. (b) L. M. Pasteur, Anal. Chim. Phys. 1848, 24 , 442–4459. 9. See also: (a) H. D. Flack, Acta Cryst. 2009, A65 , 371–389. (b) Z. S. Derewenda, Acta Cryst. 2008, A64 , 246–258. 10. J. H. van’t Hoff, Arch. Neerl. Sci. Exactes Nat. 1874, 9 , 445–454. 11. J. A. Le Bel, Bull. Soc. Chim. Fr. 1874, [2], 22 , 337–347. 12. See also E. L. Eliel, S. H. Wilen, Stereochemistry of Organic Compounds, John Wiley & Sons, New York, 1994, pp. 1–10. 13. (a) A. Cotton, Compt. Rend . 1895, 120 , 989–991. (b) A. Cotton, Ann. Chim. Phys. 1896, 8 , 347. 14. E. Fischer, Ber. Dtsch. Chem. Ges. 1891, 24 , 2683–2687. 15. V. L. Rosenfeld, Z. Physik 1928, 52 , 161–174. 16. (a) W. Kuhn, Z. Physik. Chem. 1929, B 4, 14–36. (b) W. Kuhn, Trans. Faraday Soc. 1930, 26 , 293–308. 17. (a) E. U. Condon, W. Alter, H. Eyring, J. Chem. Phys. 1937, 5 , 753–775. (b) E. U. Condon, Rev. Modern Phys. 1937, 9 , 432–457. (c) W. J. Kauzman, J. E. Walter, H. Eyring, Chem. Rev . 1940, 26 , 339–407. 18. (a) J. G. Kirkwood, J. Chem. Phys. 1937, 5 , 479–491. (b) W. W. Wood, W. Fickett, J. G. Kirkwood, J. Chem. Phys. 1952, 20 , 561–568. 19. (a) A. S. Davydov, Zhur. Eksptl. i Teoret. Fiz . 1948, 18, 210–218. (b) A. S. Davydov, Theory of Molecular Excitons, translated by M. Kasha, M. Oppenheimer, Jr., McGraw-Hill, New York, 1962. 20. (a) J. M. Bijvoet, A. F. Peerdeman, A. J. van Bommel, Nature 1951, 168 , 271–272. (b) J. M. Bijvoet, A. F. Peerdeman, A. J. van Bommel, Proc. K. Ned. Acad. Wet. B 1951, 54 , 16–19. (c) A. J. van Bommel, Proc. K. Ned. Acad. Wet. B 1953, 56 , 268–271. (d) J. Trommel, J. M. Bijvoet, Acta Crystallogr. 1954, 7 , 703–709. (e) J. M. Bijvoet, A. F. Peerdeman, Acta Crystallogr. 1956, 9 , 1012–1015. 21. (a) W. Mofﬁtt, J. Chem. Phys. 1956, 25 , 467–478. (b) W. Mofﬁtt, D. D. Fitts, J. G. Kirkwood, Proc. Natl. Acad. Sci. USA 1957, 43 , 723–730. 22. (a) I. Tinoco, Jr., Advan. Chem. Phys. 1962, 4 , 113–160. (b) I. Tinoco, Jr., R. W. Woody, D. F. Bradley, J. Chem. Phys. 1963, 38 , 1317–1325. (c) I. Tinoco, Jr., Radiat. Res. 1963, 20 , 133–139. (d) I. Tinoco, Jr., C. A. Bush, Biopolym. Symp. 1964, 1 , 235–250. (e) W. C. Johnson, I. Tinoco, Jr., Biopolymers 1969, 8 , 715–731. 23. (a) J. A. Schellman, P. Oriel, J. Chem. Phys. 1962, 37 , 2114–2124. (b) P. M. Bayley, E. B. Nielsen, J. A. Schellman, J. Phys. Chem. 1969, 73 , 228–243. 24. (a) S. F. Mason, Proc. Chem. Soc. 1962, 362–363. (b) S. F. Mason, G. W. Vane, J. Chem. Soc. B 1966, 370–374. 25. A. F. Beecham, A. C. Harley, A. McL. Mathieson, J. A. Lamberton, Nature 1973, 244 , 30–32.

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

26. S. F. Mason, G. W. Vane, K. Schoﬁeld, R. J. Wells, J. S. Whitehurst, J. Chem. Soc. B 1967, 553–556. 27. S. H. Wilen, J. Z. Qi, P. G. Williard, J. Org. Chem. 1991, 56 , 485–487. 28. (a) N. Harada, K. Nakanishi, J. Am. Chem. Soc. 1969, 91 , 3989–3991. (b) M. Koreeda, N. Harada, K. Nakanishi, Chem. Commun. 1969, 548–549. (c) N. Harada, K. Nakanishi, Chem. Commun. 1970, 310–311. (d) N. Harada, H. Sato, K. Nakanishi, Chem. Commun. 1970, 1691–1693. (e) N. Harada, S. Suzuki, H. Uda, K. Nakanishi, J. Am. Chem. Soc. 1971, 93 , 5577–5579. (f) S.-M. L. Chen, N. Harada, K. Nakanishi, J. Am. Chem. Soc. 1974, 96 , 7352–7354. (g) N. Harada, S.-M. L. Chen, K. Nakanishi, J. Am. Chem. Soc. 1975, 97 , 5345–5352. 29. (a) J. Tanaka, F. Ogura, H. Kuritani, M. Nakagawa, Chimia 1972, 26 , 471–473. (b) J. Tanaka, K. Ozeki-Minakata, F. Ogura, M. Nakagawa, Nature (London), Phys. Sci . 1973, 241 , 22–23. (c) J. Tanaka, K. Ozeki-Minakata, F. Ogura, M. Nakagawa, Spectrochim. Acta, Part A 1973, 29 , 897–924. 30. Y. Saito, Kagaku-to-Kogyo (Japan) 1973, 26 , 128–134. 31. H. H. Brongersma, P. M. Mul, Chem. Phys. Lett. 1973, 19 , 217–220. 32. S. F. Mason, J. Chem. Soc., Chem. Commun. 1973, 239–241. 33. (a) N. Harada, Y. Takuma, H. Uda, J. Am. Chem. Soc. 1976, 98 , 5408–5409. (b) N. Harada, Y. Takuma, H. Uda, Bull. Chem. Soc. Jpn. 1977, 50 , 2033–2038. (c) N. Harada, Y. Takuma, H. Uda, Bull. Chem. Soc. Jpn. 1978, 51 , 265–270. (d) N. Harada, Y. Takuma, H. Uda, J. Am. Chem. Soc. 1978, 100 , 4029–4036. 34. P. J. Stephens, N. Harada, Chirality 2010, 22 , 229–233. 35. D. Bohm, Quantum Theory, Prentice-Hall, New York, 1951, p. 427. 36. E. Giorgio, K. Tanaka, W-D. Ding, G. Krishnamurthy, K. Pitts, G. A. Ellestad, C. Rosini, N. Berova, Bioorg. Med. Chem. 2005, 13 , 5072–5079. 37. J. L. Alonso-Gomez, A. G. Petrovic, N. Harada, P. Rivera-Fuentes, N. Berova, F. Diederich, Chem. Eur. J . 2009, 15 , 8396–8400. 38. (a) N. Sakabe, K. Sakabe, K. Ozeki-Minakata, J. Tanaka, Acta Crystallogr. Sect. B 1972, 28 , 3441–3446. (b) J. Tanaka, C. Katayama, F. Ogura, H. Tatemitsu, M. Nakagawa, Chem. Commun. 1973, 21–22. 39. D. H. R. Barton, H. T. Cheung, A. D. Cross, L. M. Jackman, M. Martin-Smith, J. Chem. Soc. 1961, 5061–5073. 40. I. C. Paul, G. A. Sim, T. A. Hamor, J. M. Robertson, J. Chem. Soc. 1962, 4133–4145. 41. S. Hosozawa, N. Kato, K. Munakata, Tetrahedron Lett. 1974, 15 , 3753–3756. 42. N. Harada, H. Uda, J. Am. Chem. Soc. 1978, 100 , 8022–8024. 43. D. Rogers, G. G. Unal, D. J. Williams, S. V. Ley, G. A. Sim, B. S. Joshi, K. R. Ravindranath, J . Chem. Soc., Chem. Commun. 1979, 97–99. 44. (a) N. Zhao, P. Zhou, N. Berova, K. Nakanishi, Chirality 1995, 7 , 636–651. (b) D. Rele, N. Zhao, K. Nakanishi, N. Berova, Tetrahedron 1996, 52 , 2759–2776. (c) N. Zhao, N. Berova, K. Nakanishi, M. Rohmer, P. Mougenot, U. J. Jurgens, Tetrahedron 1996, 52 , 2777–2788. 45. X. Huang, K. Nakanishi, N. Berova, Chirality 2000, 12 , 237–255. 46. G. Pescitelli, S. Gabriel, Y. Wang, J. Fleischhauer, R. W. Woody, N. Berova, J. Am. Chem. Soc. 2003, 125 , 7613–7628. 47. S. Matile, N. Berova, K. Nakanishi, J. Fleischhauer, R. W. Woody, J. Am. Chem. Soc. 1996, 118 , 5198–5206. 48. S. Matile, N. Berova, K. Nakanishi, Chemistry and Biology 1996, 3 , 379–392. 49. A. Mammana, G. Pescitelli, T. Asakawa, S. Jockusch, A. G. Petrovic, R. R. Monaco, R. Purrello, N. J. Turro, K. Nakanishi, G. Ellestad, M. Balaz, N. Berova, Chem. Eur. J . 2009, 15 , 11853–11866.

163

164

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

50. J. D. Chisholm, J. Golik, B. Krishnan, J. A. Matson, D. L. Van Vranken, J. Am. Chem. Soc. 1999, 121 , 3801–3802. 51. (a) F. Kazmierczak, K. Gawronska, U. Rychlewska, J. Gawronski, Tetrahedron: Asymmetry 1994, 5 , 527–530. (b) J. Gawronski, F. Kazmierczak, K. Gawronska, P. Skowronek, J. Waluk, J. Marczyk, Tetrahedron 1996, 52 , 13201–13214. (c) J. Gawronski, M. D. Rozwadowska, F. Kazmierczak, Polish J. Chem. 1994, 68 , 2279–2288. 52. M. Watanabe, T. Abe, Y. Fujii, N. Harada, Enantiomer 1998, 3 , 197–202. 53. N. Harada, H.-Y. Li, N. Koumura, T. Abe, M. Watanabe, M. Hagiwara, Enantiomer 1997, 2 , 349–352. 54. (a) A. Kawamura, N. Berova, V. Dirsch, A. Mangoni, K. Nakanishi, G. Schwartz, A. Bielawska, Y. Hannun, I. Kitagawa, Bioorg. Med. Chem. 1996, 4 , 1035–1043. (b) V. Dirsch, J. Frederico, N. Zhao, G. Cai, Y. Chen, S. Vunnam, J. Odingo, H. Pu, K. Nakanishi, N. Berova, D. Liotta, A. Bielawska, Y. Hannum, Tetrahedron Lett. 1995, 36 , 4959–4962. 55. K. Tanaka, G. Pescitelli, L. Di Bari, T. L. Xiao, K. Nakanishi, D. W. Armstrong, N. Berova, Org. Biomol. Chem. 2004, 2 , 48–58. 56. D. Gargiulo, N. Ikemoto, J. Odingo, N. Bozhkova, T. Iwashita, N. Berova, K. Nakanishi, J. Am. Chem. Soc. 1994, 116 , 3760–3767. 57. G. Cai, N. Bozhkova, J. Odingo, N. Berova, K. Nakanishi, J. Am. Chem. Soc. 1993, 115 , 7192–7198. 58. N. Harada, J. Am. Chem. Soc. 1973, 95 , 240–242. 59. M. Koreeda, N. Harada, K. Nakanishi, J. Am. Chem. Soc. 1974, 96 , 266–268. 60. A. Guerriro, M. D’Ambrosio, V. Cuomo, F. Vanzanella, F. Pietra, Helvetica Chim. Acta, 1989, 72 , 438–446. 61. T. Ishikawa, M. Murota, T. Watanabe, T. Harayama, H. Ishii, Tetrahedron Lett. 1995, 36 , 4269–4272. 62. J. W. Moncrief, W. N. Lipscomb, J. Am. Chem. Soc. 1965, 87 , 4963–4964. 63. J. P. Kutney, D. E. Gregonis, R. Imhof, I. Itoh, E. Jahngen, A. I. Scott, W. K. Chan, J. Am. Chem. Soc. 1975, 97 , 5013–5015. 64. C. A. Parish, J-G. Dong, W. G. Bornmann, J. Chang, K. Nakanishi, N. Berova, Tetrahedron 1998, 54 , 15739–15758. 65. (a) X. Huang, B. H. Rickman, B. Bohran, N. Berova, K. Nakanishi, J. Am. Chem. Soc. 1998, 120 , 6185–6186. (b) T. Kurt´an, N. Nesnas, Y.-Q. Li, X. Huang, K. Nakanishi, N. Berova, J. Am. Chem. Soc. 2001, 123 , 5962–5973. (c) T. Kurt´an, N. Nesnas, F. E. Koehn, Y.-Q. Li, K. Nakanishi, N. Berova, J. Am. Chem. 2001, 123 , 5974–5982. 66. X. Huang, N. Fujioka, G. Pescitelli, F. E. Koehn, T. R. Williamson, K. Nakanishi, N. Berova, J. Am. Chem. Soc. 2002, 124 , 10320–10335. 67. N. Berova, G. Pescitelli, A. G. Petrovic, G. Proni, Chem. Commun. 2009, 5958–5980. 68. (a) A. G. Petrovic, Y. Chen. G. Pescitelli, N. Berova, G. Proni, Chirality, 2010, 22 , 129–139. (b) Y. Chen, A. G. Petrovic, M. Roje, G. Pescitelli, Y. Yang, M. Kayser, N. Berova, G. Proni, Chirality, 2010, 22 , 140–152. 69. J. W. Van Klink, S.-H. Baek, A. J. Barlow, H. Ishii, K. Nakanishi, N. Berova, N. B. Perry, R. T. Weavers, Chirality 2004, 16 , 549–558. 70. H. Ishii, S. Krane, Y. Itagaki, N. Berova, K. Nakanishi, P. J. Weldon, J. Natural Products 2004, 67, 1426–1430. 71. V. V. Borovkov, J. M. Lintuluoto, Y. Inoue, J. Am. Chem. Soc. 2001, 123 , 2979–2989. 72. (a) Q. Yang, C. Olmsted, B. Borhan, Org. Lett. 2002, 4 , 3423–4326. (b) X. Li, M. Tanasova, C. Vasileiou, B. Borhan, J. Am. Chem. Soc. 2008, 130 , 1885–1893. (c) X. Li, B. Borhan, J. Am. Chem. Soc. 2008, 130 , 16126–16127. (d) M. Tanasova, C. Vasileiou, O. O. Olumolade, B. Borhan, Chirality 2009, 21 , 374–382. (e) M. Tanasova, Q. yang, C. C. Olmsted, C. Vasileiou, X. Li, M. Anyika, B. Borhan, Eur. J. Org. Chem. 2009, 4242–4253.

E L E C T R O N I C C D E X C I T O N C H I R A L I T Y M E T H O D : P R I N C I P L E S A N D A P P L I C AT I O N S

73. (a) I. Hanazaki, H. Akimoto, J. Am. Chem. Soc. 1972, 94 , 4102–4106. (b) S. F. Mason, R. H. Seal, D. R. Roberts, Tetrahedron 1974, 30 , 1671–1682. (c) L. Di Bari, G. Pescitelli, P. Salvadori, J. Am. Chem. Soc. 1999, 121 , 7998–8004. 74. K. Harata, J. Tanaka, Bull. Chem. Soc. Jpn. 1973, 46 , 2747–2751. 75. N. Harada, N. Hiyoshi, V. P. Vassilev, T. Hayashi, Chirality 1997, 9 , 623–625. 76. N. Harada, V. P. Vassilev, N. Hiyoshi, Enantiomer 1997, 2 , 123–126. 77. (a) N. Harada, J. Iwabuchi, Y. Yokota, H. Uda, Y. Okamoto, H. Yuki, Y. Kawada, J. Chem. Soc., Perkin Trans. 1 1985, 1845–1848. (b) N. Harada, J. Iwabuchi, Y. Yokota, H. Uda, Croatica Chim. Acta 1989, 62 , 267–278. 78. (a) G. Hass, V. Prelog, Helv. Chim. Acta 1969, 52 , 1202–1218. (b) G. Hass, P. B. Hulbert, W. Klyne, V. Prelog, G. Snatzke, Helv. Chim. Acta 1971, 54 , 491–509. 79. J. Sagiv, A. Yogev, Y. Mazur, J. Am. Chem. Soc. 1977, 99 , 6861–6869. 80. N. Harada, H. Ono, T. Nishiwaki, H. Uda, J. Chem. Soc., Chem. Commun. 1991, 1753–1755. 81. K. Shingu, H. Kuritani, A. Kato, S. Imajo, Tetrahedron Lett. 1980, 21 , 3997–4000. 82. H. Ono, N. Harada, Enantiomer 1998, 3 , 245–250. 83. N. Harada, H. Ono, Enantiomer 1996, 1 , 115–118. 84. N. Harada, Y. Tamai, Y. Takuma, H. Uda, J. Am. Chem. Soc. 1980, 102 , 501–506. 85. N. Harada, Y. Tamai, H. Uda, J. Am. Chem. Soc. 1980, 102 , 506–511. 86. N. Harada, A. Saito, H. Ono, S. Murai, H.-Y Li, J. Gawronski, K. Gawronska, T. Sugioka, H. Uda, Enantiomer 1996, 1 , 119–138. 87. I. Akritopoulou-Zanze, K. Nakanishi, H. Stepowska, B. Grzeszczyk, A. Zamojski, N. Berova, Chirality 1997, 9 , 699–712. 88. N. Harada, H.-Y. Li, N. Koumura, T. Abe, M. Watanabe, M. Hagiwara, Enantiomer 1997, 2 , 349–352. 89. N. Harada, A. Saito, H. Ono, J. Gawronski, K. Gawronska, T. Sugioka, H. Uda, T. Kuriki, J. Am. Chem. Soc. 1991, 113 , 3842–3850. 90. P. Zhou, N. Zhao, D. Rele, N. Berova, K. Nakanishi, K. J. Am. Chem. Soc. 1993, 115 , 9313–9314. 91. (a) D. Rele, N. Zhao, K. Nakanishi, N. Berova, Tetrahedron 1996, 52 , 2759–2776. (b) N. Zhao, N. Berova, K. Nakanishi, M. Rohmer, P. Mougenot, U. J. J¨urgens, Tetrahedron 1996, 52 , 2777–2788. 92. N. Harada, K. Nakanishi, S. Tatsuoka, J. Am. Chem. Soc. 1969, 91 , 5896–5898. 93. Y. A. Berlin, M. N. Kolosov, L. A. Piotrovich, Tetrahedron Lett. 1970, 1329–1331. 94. J. A. Mills, J. Chem. Soc. 1952, 4976–4985. 95. J. H. Brewster, Tetrahedron 1961, 13 , 106–122. 96. (a) N. Harada, M. Ohashi, K. Nakanishi, J. Am. Chem. Soc. 1968, 90 , 7349–7351. (b) N. Harada, K. Nakanishi, J. Am. Chem. Soc. 1968, 90 , 7351–7352. 97. (a) N. Harada, J. Iwabuchi, Y. Yokota, H. Uda, K. Nakanishi, J. Am. Chem. Soc. 1981, 103 , 5590–5591. (b) N. Harada, Y. Yokota, J. Iwabuchi, H. Uda, M. Ochi, J. Chem. Soc., Chem. Commun. 1984, 1220–1221. 98. (a) H.-U. Humpf, N. Berova, K. Nakanishi, M. B. Jarstfer, C. Dale Poulter, J. Org. Chem. 1995, 60 , 3539–3542. (b) K. Tanaka, G. Pescitelli, K. Nakanishi, N. Berova, Monatsh. Chem. 2005, 136 , 367–395. 99. K. Tanaka, K. Nakanishi, N. Berova, J. Am. Chem. Soc. 2003, 125 , 10802–10803. 100. M. Satake, A. Morohashi, H. Oguri, T. Oishi, M. Hirama, N. Harada, T. Yasumoto, J. Am. Chem. Soc. 1997, 119 , 11325–11326. 101. M. Hirama, T. Oishi, H. Uehara, M. Inoue, M. Maruyama, H. Oguri, M. Satake, Science, 2001, 294 , 1904–1907.

165

166

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

102. A. Sakurai, H. Horibe, N. Kuboyama, Y. Hashimoto, Y. Okumura, J. Biochem. 1995, 118 , 552–554. 103. J. Kobayashi, M. Yoshinaga, N. Yoshida, M. Shiro, H. Morita, J. Org. Chem. 2002, 67 , 2283–2286. 104. M. L. Kenney, F. Cort´es-Selva, J. M. Perez-Victoria, I. A. Jim´enez, A. G. Gonzalez, O, M. Mu˜noz, F. Gamarro, S. Castanys, A. G. Ravelo. J. Med. Chem. 2001, 44 , 4668–4676. 105. J. Guo, G. Schlingmann, G. T. Carter, K. Nakanishi, N. Berova, Chirality, 2000, 12 , 43–51. 106. H. Kobayashi, S. Meguro, T. Yoshimoto, M. Namikoshi, Tetrahedron 2003, 59 , 455–459. 107. T. Wang, O. Shirota, K. Nakanishi, N. Berova, L. A. McDonald, L. R. Barbieri, G. Carter, Can. J. Chem. 2001, 79 , 1786–1791. 108. T. Sunazuka, T. Shirahata, K. Yoshida, D. Yamamoto, Y. Harigaya, T. Nagai, H. Kiyohara, H. Yamada, I. Kuwajima, S, Omura, Tetrahedron Lett. 2002, 43 , 1265–1268. 109. T. F. Molinski, L. J. Brzezinski, J. W. Leahy, Tetrahedron: Asymmetry 2002, 13 , 1013–1016. 110. (a) K. Tanaka, G. Pescitelli, L. Dibari, T. L. Xiao, K. Nakanishi, D. W. Armstrong, N. Berova, Org. Biomol. Chem. 2004, 2 , 48–58. (b) K. Tanaka, Y. Itagaki, M. Satake, H. Naoki, T. Yasumoto, K. Nakanishi, N. Berova, J. Am. Chem. Soc. 2005, 127 , 9561–9570. 111. R. W. Baker, S. Liu, M. V. Sargent, Aust. J. Chem. 1998, 51 , 255–266. 112. M. S. Buchanan, M. Gill, P. Millar, S. Phonh-Axa, E. Raudies, J. Yu, J. Chem. Soc., Perkin Trans. 1 1999, 795–801. 113. M. Hashimoto, T. Tsushima, T. Murakami, M. Nomiya, N. Takada, K. Tanaka, Bioorg. Med. Chem. Lett. 2008, 18 , 4228–4231. 114. T. Murakami, T. Tsushima, N. Takada, K. Tanaka, K. Nihei, T. Miura, M. Hashimoto, Bioorg. Med. Chem. 2009, 17 , 492–495. 115. F. H. Kong, D. J. Faulkner, J. Org. Chem. 1993, 58 , 970–971. 116. D. S. Dalisay, S. Tsukamoto, T. Molinski, J. Natl. Prod . 2009, 72 , 353–359. 117. G. M. Nicholas, T. F. Molinski, J. Am. Chem. Soc. 2000, 122 , 4011–4019. 118. S. Yoshida, K. Kito, T. Ooi, K. Kanoh, Y. Shizuri, T. Kusumi, Chem. Lett. 2007, 36 , 1386–1387. 119. S. Aoki, Y. Watanabe, M. Sanagawa, A. Setiawan, N. Kotoku, M. Kobayashi, J. Am. Chem. Soc. 2006, 128 , 3148–3149. 120. L. Wang, X. Wang, J. Cui, W. Ren, N. Meng, J. Wang, X. Qian, Tetrahedron: Asymmetry 2010, 21 , 825–830. 121. S. T. Deyrup, J. B. Gloer, K. O’Donnell, D. T. Wicklow, J. Nat. Prod . 2007, 70 , 378–382. 122. K. Krohn, M. H. Sohrab, T. van Ree, S. Draeger, B. Schulz, S. Antus, T. Kurtan, Eur. J. Org. Chem. 2008, 5638–5646. 123. (a) K. Tsubaki, K. Takaishi, H. Tanaka, M. Miura, T. Kawabata, Org. Lett. 2006, 8 , 2587–2590. (b) K. Tsubaki, Org. Biomol. Chem. 2007, 5 , 2179–2188. 124. J. Naito, Y. Yamamoto, M. Akagi, S. Sekiguchi, M. Watanabe, N. Harada, Monatsh. Chem. 2005, 136 , 411–445. 125. J. L. Alonso-Gomez, P. Schanen, P. Rivera-Fuentes, P. Seiler, F. Diederich, Chem. Eur. J . 2008, 14 , 10564–10568. 126. J. Golik, T. W. Doyle, B. Krishnan, G. Dubay, J. A. Matson, J. Antibiot. 1989, 42 , 1784–1789. 127. T. Hattori, K. Sakurai, N. Koike, S. Miyano, H. Goto, F. Ishiya, N. Harada, J. Am. Chem. Soc. 1998, 120 , 9086–9087.

5 CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS: THEORETICAL DETERMINATION OF THE ABSOLUTE STEREOCHEMISTRY AND EXPERIMENTAL VERIFICATION Nobuyuki Harada and Shunsuke Kuwahara

5.1. INTRODUCTION Electronic circular dichroism (ECD) is very useful for characterization of chiral organic compounds with π -electron chromophores. That is, ECD enables one to determine the absolute conﬁgurations (ACs) of various natural products and chiral synthetic compounds by the use of appropriate CD methods, exempliﬁed by the CD exciton chirality method [1–5]. In general, chiral compounds containing a chiral conjugated π -electron chormophore such as conjugated diene and enone exhibit medium CD Cotton effects. These chromophores contained in chiral compounds naturally adopt twisted conformations falling in the category of the inherently dissymmetrical chromophore. The observed CD Cotton effects are generally governed by the helicity of the diene or enone moiety, from which the absolute conﬁguration can be determined. In some cases, however, the allylic group makes a contribution to the CD, and hence the observed CD does not agree with the chromophore helicity. On the other hand, compounds containing further extended and conjugated π -electron chromophores exhibit much more intense CD Cotton effects, as will be discussed in this chapter. In these cases, the CD Cotton effects are mostly governed by the helicity or twisted structure of the π -electron chromophore itself. Therefore, the CD spectra of these systems can be calculated by the π -electron approximation such as the π -electron self-consistent-ﬁeld / conﬁguration interaction / dipole velocity molecular orbital (SCFCI-DV MO) method [6–8]. In fact, we have determined the ACs of various natural products and chiral synthetic compounds by the use of the π -electron SCF-CI-DV MO method. In this chapter, the principle and applications of this method are explained in detail. Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

167

168

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

To conﬁrm the ACs as determined by the theoretical method, chiral model compounds and/or natural products themselves were synthesized in enantiopure forms. In such cases, the CSDP acid method and/or MαNP acid method are very useful for enantioresolving racemic compounds into enantiopure derivatives, and the methods simultaneously enable one to determine the ACs by X-ray crystallography and/or by 1 H NMR anisotropy [9–12]. The CD spectra of the synthesized model compounds were compared with those of the compounds in question. By these experimental studies, the theoretical method was established to lead to the correct absolute conﬁgurational assignments. Besides the theoretical method, the combination of CD spectroscopy and X-ray crystallographic analysis with an internal reference is also reliable for determining the ACs of various chiral compounds. This method has been applied to various natural products, chiral spiro compounds, light-powered chiral molecular motors, chiral C60 fullerene bis-adducts, and so on which led to the unambiguous assignment of ACs as described in this chapter. In the case of light-powered chiral molecular motors, the motor rotation mechanism and dynamics were also clariﬁed by CD spectroscopy together with 1 H NMR spectroscopy. CD spectroscopy is thus useful not only for studying chiral stereochemistry, but also for the static and dynamic behavior of natural products and chiral synthetic functional compounds. In this chapter are described the research results carried out mostly by the authors’ group.

5.2. THEORETICAL CALCULATION OF CD AND UV SPECTRA BY THE π -ELECTRON SCF-CI-DV MO METHOD The CD and UV spectra of an extended π -electron system can be calculated by the π electron SCF-CI-DV MO method, where the rotational strength Rba and dipole strength Dba are expressed as follows [6]. Rba = 2(ψa |∇|ψb )(ψa |r × ∇|ψb )μB 2 /(π σba )

(5.1)

Dba = 2(ψa |∇|ψb ) μB /(π σba )

(5.2)

2

2

2

where ∇ is the del operator, r is a distance vector, μB is the Bohr magneton, and σba is the excitation wavenumber of the transition a → b. The z -axis components of the electric and magnetic transition moments are formulated, respectively, as [6] (ψa |∇|ψb )z =

(Cra Csb − Csa Crb )<∇rs >cos Zrs

bonds

(ψa |r × ∇|ψb )z =

(Cra Csb − Csa Crb )<∇rs >(Xrs cos Yrs − Yrs cos Xrs )

(5.3) (5.4)

bonds

cos Zrs = (Zr − Zs )/Rrs

(5.5)

Xrs = (Xr + Xs )/2

(5.6)

where Cra is the coefﬁcient of atomic orbital r in the wavefunction ψa ; <∇rs > is the expectation value of a dipole velocity ∇rs , which is directed along the bond rs in the direction r → s; Xr , Yr , and Zr are the x , y, and z coordinates of an atom r, respectively; and Rrs is the interatomic distance between atoms r and s. The x and y components of the electric and magnetic transition moments can be similarly calculated.

169

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

In the π -electron SCF-CI-DV MO method, the following standard values of atomic orbital parameters are proposed: for sp2 carbon, Z (C) = 1.0, W (C) = −11.16 eV, ˚ = −2.32 eV, <∇> (C–C, 1.388 A) ˚ = 4.70 × (rr|rr)(C) = 11.13 eV, β(C–C, 1.388 A) 107 cm−1 ; for ether oxygen, Z (O) = 2.0, W (O) = −33.00 eV, (rr|rr)(O) = 21.53 eV, β(C–O) = −2.00 eV, <∇> (C–O) = 6.00 × 107 cm−1 ; for sp 2 nitrogen, Z (N) = 1.0, W (N) = −14.12 eV, (rr|rr)(N) = 12.34 eV, β(C–N) = −2.32 or −2.55 eV, <∇> (C– N) = 4.70 or 5.17 × 107 cm−1 [1]. The electric repulsion integral (rr|ss) was approximated by the Nishimoto–Mataga equation. The resonance integral β and del value <∇> were calculated by the use of following equations, respectively [1]: ˚ β(1.388 A) ˚ cos θ β = [S /S (1.388 A)]

(5.7)

˚ ˚ <∇> = [<∇>(empirical, 1.388 A)/<∇>(theoretical, 1.388 A)] × <∇>(theoretical) cos θ

(5.8)

where θ is a dihedral angle. The overlap integral S and <∇>(theoretical) were calculated on the basis of the Slater orbitals. The conﬁguration interactions between all singly excited states were included. The curves of the component CD and UV bands were approximated by the Gaussian distribution [13], ε(σ ) =

εk exp[−{(σ − σk )/σ }2 ]

(5.9)

k

ε(σ ) =

εk exp[−{(σ − σk )/σ }2 ]

(5.10)

k

where σ is half the bandwidth at 1/e peak height. The σ value of 2500 cm−1 was used as a standard value [1].

5.3. SOME ESTABLISHED EXAMPLES OF THE π -ELECTRON SCF-CI-DV MO METHOD The following are some examples of the application of the π -electron SCF-CI-DV MO method applied to various chiral natural products and synthetic chiral compounds with extended π -electron chromophores. These examples were already explained in reference 3, and hence the summary of the results (i.e., comparison of observed and calculated CD and UV–Vis spectra, absolute stereostructures, and experimental veriﬁcation by Xray crystallography and/or by synthesis) are brieﬂy described. The exciton CD studies of chiral spiroaromatics of 9,9 -spirobiﬂuorene skeleton are described in Chapter 4 of this volume.

5.3.1. Absolute Conﬁguration of (+)-1,8a-Dihydro-3,8-dimethylazulene Chiroptically active 1,8a-dihydro-3,8-dimethylazulene (+)-(1) was isolated from the cell culture of the liverwort Calypogeia granulata Inoue (Figure 5.1) [14]. The labile intermediate 1 with a unique 1,8a-dihydroazulene skeleton shows very intense chiroptical activity, [α]D + 1164 and intense CD Cotton effects as shown in Figure 5.1, suggesting

170

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Δe 314.0 (+19.7) +20

CD

0

H e × 10–4

–20

–40 235.2 (–47.4)

(8aS)-(+)

3

227.5 (25,600) –60 Obsd in hexane

2

UV 1 308.5 (5,400)

200

Figure 5.1. CD and UV spectra of naturally occurring (8aS)-(+)-1,8a-dihydro-3,8dimethylazulene (1) in hexane. (Redrawn from

0

300 λ (nm)

reference 15, with permission.)

Δe 219 (+46.2) H +40

(8aR)

CD

e × 10–4

+20

0

–20

219 (27,300) 313 (–13.9)

Calcd

UV

3

2

313 (9,900) 1

Figure 5.2. CD and UV curves of (8aR)-1,8a-dihydroazulene 2 calculated by the 200

300 λ (nm)

0

π -electron SCF-CI-DV MO method. (Redrawn from reference 15, with permission.)

171

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

a strongly twisted conjugated tetraene system. Therefore, it is reasonable to consider that the chiroptical activity of 1 is mainly due to the twist of the π -electron chromophore. To determine the AC of 1 theoretically, we carried out the calculation of the CD curve of 1,8a-dihydroazulene (2) on the basis of the π -electron framework approximation, using the SCF-CI-DV MO method, where its AC was arbitrarily chosen to be (8aR) (Figure 5.2) [15]. The theoretically calculated CD and UV curves agree well with the observed CD and UV spectra except for the sign of the CD ε values (compare Figures 5.1 and 5.2). That is, the observed CD curve of compound 1 is almost a mirror image of the curve calculated for the model compound (8aR)-2. Accordingly, the AC of the labile biosynthetic

Br

OCH3

O

OCH3

O O

O

H

(8aS)-(+)-4

(1S,8aS)-(+)-3

CH3O (1S,3aR,4S,7R,8aS)-(+)-5 X-ray

Scheme 5.1. A synthesis of the model compound (1S,8aS)-(+)-3.

+20

321.0 (+5.7) CD

0 Δe –20

–40

221.3 (–24.5)

3

e × 10–4

OCH3

(1S,8aS)-(+) –60

223.2 (23,700) 2 Obsd in EtOH

UV 324.3 (6,000)

1

Figure 5.3. CD and UV spectra of (1S,8aS)-(+)200

300 λ (nm)

0

1,8a-dihydro-1-methoxy-8a-methylazulene (3) in EtOH. (Redrawn from reference 15, with permission.)

172

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

intermediate (+)-1 was theoretically determined to be (8aS ). This conclusion was proved experimentally by the synthesis of model compounds, as described in the following. As a chiral model compound, (1S,8aS)-(+)-(3) was selected because the angular position 8a is blocked by a methyl group, and hence it resists the oxidation to azulene. The model compound was synthesized (Scheme 5.1) [15, 16], starting from the enantiopure Wieland–Miescher ketone (S )-(+)-(4) [17] via an intermediate bromide (+)-5, the AC of which was conﬁrmed by the Bijvoet method in X-ray crystallography [15, 16]. The CD and UV spectra of (1S,8aS)-(+)-3 are shown in Figure 5.3; the CD curve of (1S,8aS )-(+)-3 is quite similar, in both sign and shape of Cotton effects, to that of dihydroazulene (+)-1. Therefore, it was proved experimentally that the natural product (+)-1 has 8aS absolute conﬁguration. Thus the present results verify the theoretical determination of the absolute conﬁguration of (+)-1 discussed above.

5.3.2. Circular Dichroism and Absolute Stereochemistry of Chiral Troponoid Spiro Compounds The SCF-CI-DV MO method has been successfully applied to a chiral troponoid spiro compound (6) as follows (Figure 5.4) [18]. Racemic spiroacetal (±)-6 could be enantioseparated by chiral HPLC of (+)poly(triphenylmethylmethacrylate). In the HPLC, the ﬁrst-eluted fraction gave an enantiomer (−)-6, [α]D −4700, which shows the CD and UV spectra as illustrated in Figure 5.4. To determine the absolute stereochemistry of (−)-6, we calculated the

Δe

Obsd OO

+100

287 (+80.4)

N

+50

N

(S)-(–) e × 10–4

CD

0

–50

285 (23,500)

3 398 (–45.3)

–100

2

MeOH UV

378 (7,900) 1

Figure 5.4. CD and UV spectra of troponoid 300

400 λ (nm)

500

spiro compound (S)-(−)-6 in MeOH. (Redrawn from reference 18, with permission.)

173

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

CD spectra of (S )-6 by applying the π -electron SCF-CI-DV MO method, where the absolute conﬁguration was arbitrarily chosen as S . The calculated (calcd) CD and UV spectra shown in Figure 5.5 are in a good agreement with the observed (obsd) spectra. Accordingly, the absolute stereochemistry of (−)-6 was theoretically determined to be S . The present conclusion is in line with the X-ray crystallographic results of a related compound [19].

5.3.3. Absolute Stereochemistry of the Halenaquinol Family Marine Natural Products The π -electron SCF-CI-DV MO method for calculation of CD spectra was next applied to the determination of the AC of the compounds of the halenaquinol family (Chart 5.1) [20–23]. To determine the absolute conﬁguration of halenaquinol (+)-7, we ﬁrst planned to apply the CD exciton chirality method [21] to the interaction between the naphthalene and benzoate chromophores. During the synthetic studies of a pertinent benzoate derivative 8, we obtained a naphthalene–diene compound (−)-9 (Chart 5.1) [24], which surprisingly exhibited much stronger CD Cotton effects than other halenaquinol derivatives did (Figure 5.6). The result clearly indicates that the major part of the CD Cotton effects originates from the π -electron chromophore composed of the naphthalene–diene moiety, which is twisted by the angular methyl group at the 12b position. Therefore, the twisted

Δe

Calcd OO

+100

289 (+76.3)

N

N

+50

(S)

e × 10–4

CD

0

–50

3

284 (23,500)

–100

2 UV

394 (–106.9) 1

Figure 5.5. CD and UV curves of spiro 300

400 λ (nm)

500

troponoid compound (S)-6 calculated by the π -electron SCF-CI-DV MO method. (Redrawn from reference 18, with permission.)

174

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

HO

1

2

O

12

10

OR

CH3O

OSi

CH3O

4 7

HO

4

OCH3 O

O O

3

CH3O

OBz

CH3O

(12bS)-(+)-7

O CH3O

CH3O

8

(–)-9

7

6

O

(12bS)-10

Chart 5.1. Halenaquinol (7) and related compounds.

naphthalene–diene moiety is an ideal system for the determination of the absolute stereochemistry by applying the π -electron SCF-CI-DV MO method. As a model compound for the theoretical calculation of CD spectra, we adopted the molecule (12bS )-10, which has the essential part of the π -electron system of naphthalenediene compound 9. The absolute conﬁguration of 10 was arbitrarily chosen as 12bS for the calculation, and the molecular geometry of the model compound was calculated by molecular mechanics [25]. The theoretical calculation of the CD and UV spectra of (12bS )-10 by the π electron SCF-CI-DV MO method afforded the curves illustrated in Figure 5.7 [24]. The

229 (+40.9) +40

OSi

CH3O

OCH3 O CH3O

+20 Δe

(12bS)-(–) 338 (+6.4)

CD

e × 10–4

0

Obsd in MeOH –20

218 (42,000) 301 (–23.3)

UV

4

324 (27,000)

2

Figure 5.6. CD and UV spectra of 0 200

300

400 λ (nm)

helanaquinol trans-methoxy diene derivative (3R,4R,12bS)-(−)-(9) in MeOH. (Redrawn from reference 24b, with permission.)

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

+40

223 (+35.5)

175

CH3O

O CH3O

+20

(12bS )

CD

Δe

378 (+3.3)

248 (–5.7)

–20

e × 10–4

0

Calcd

219 (40,300)

322 (–22.4)

4

349 (29,900) UV 2

Figure 5.7. CD and UV curves of the model compound 200

300

(12bS)-10 calculated by the π -electron SCF-CI-DV MO method. (Redrawn from reference 24b, with permission.)

0

400 λ (nm)

+100

CH3O

+50

CH3O

O (12bS )

D × 1036 cgs unit

CD

0

Calcd

R × 1040 cgs unit

–50

30 UV 20 10

200

300

400 λ (nm)

0

Figure 5.8. Calculated rotational and dipole strengths of the model compound (12bS)-10. (Redrawn from reference 24b, with permission.)

176

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

theoretically simulated CD curve is in good agreement with the observed curve of (−)-9. It is thus evident that the basic pattern of the CD and UV spectral curves, including the sign, position, intensity, and shape of the bands, was well reproduced by the calculation. Since the absolute conﬁguration of the model compound 10 is set as 12bS, the comparison of the present calculated and observed CD data leads to the unambiguous determination that the naphthalene–diene compound 9 has the 12bS absolute conﬁguration. Accordingly, the absolute stereochemistry of halenaquinol (+)-7 was theoretically determined to be 12bS . To clarify the applicability of the present theoretical method to such a complex system, we analyzed the composition of the apparent CD and UV bands [24b]. As shown in Figure 5.8, there are nine major electronic transitions that contribute to the CD and UV bands. The ﬁrst and second transitions with weak positive rotational strength at 374.5 and 351.6 nm, respectively, generate the weak positive Cotton effect at 378 nm (Figure 5.8). Furthermore, the third transition with an intense negative rotational strength at 324.4 nm results in the negative Cotton effect at 322 nm, and the sixth transition with a strong positive rotational strength contributes mainly to the intense positive Cotton effect at 223 nm. The correspondence between the component transitions and the apparent CD is thus clear. Therefore, the present analysis makes the absolute conﬁgurational determination of the halenaquinol compounds more reliable. Theories and theoretically obtained results should be proved experimentally. We succeeded in the ﬁrst total synthesis of (+)-halenaquinol 7 and related natural products starting from (8aR)-(−)-Wieland–Miescher ketone [26–29]. The CD and UV spectra of the compounds synthesized were, of course, identical to those of the natural products. By these total syntheses of halenaquinol family compounds, we have proved, in an excellent way, that their absolute conﬁgurations theoretically determined were correct [26].

5.3.4. Atropisomerism of Natural Products: CD and Absolute Stereochemistry of the Biﬂavone, 4,4 ,7,7 -Tetra-O-methylcupressﬂavone Atropisomers are chiral compounds devoid of a chirality center. Those compounds are unique because the rotation about a single bond connecting two bulky moieties is sterically hindered, and hence their rotational conformers are sufﬁciently stable to be resolved into enantiomers. A natural product of biﬂavone, 4,4 ,7,7 tetra-O-methylcupressuﬂavone (11), is one of such atropisomers (Figure 5.9). The CD spectrum of biﬂavone (−)-11 shows strong bisignate Cotton effects of positive ﬁrst and negative second signs at 400–300 nm (Figure 5.9). These Cotton effects look like an exciton split CD; therefore, one may assign the positive exciton chirality—speciﬁcally, clockwise screw sense of P -helicity—to this atropisomer. In fact, during the stereochemical studies of this ﬂavone, the aS absolute stereochemistry (or P -helicity) had initially been assigned to (−)-11. However, such a careless application of the exciton chirality method leads to the erroneous assignment of the AC as

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

+50

177

Obsd 362.0 (+25.6)

267.5 (+21.3) Δe

CD

e × 10–4

0

OH O

CH3O

O

–50

OCH3 10 OCH3

326.2 (–54.4) O

CH3O

225.8 (51,800)

UV OH O (aR)-(–)

273.0 (41,400)

5 324.2 (40,900)

in EtOH

200

300

0

400 λ (nm)

+50

Calcd

Figure 5.9. CD and UV spectra of 4,4 ,7,7 -tetra-O-methylcupressuﬂavone (aR)-(−)-11 in EtOH. (Redrawn from reference 30, with permission.)

359.7 (+28.6)

263.2 (+21.7) Δe

CD

OH O

CH3O –50

e × 10–4

0

O

317.5 (–45.0)

OCH3 10 OCH3

226.8 (78,300)

CH3O 322.6 (66,200)

O

OH O (aR)

5

UV

Figure 5.10. CD and UV curves of 0 200

300

400 λ (nm)

(aR)-4,4 ,7,7 -tetra-O-methylcupressuﬂavone 11 calculated by the π -electron SCF-CI-DV MO method. (Redrawn from reference 30, with permission.)

178

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

described below [30]. We calculated the CD and UV spectra of biﬂavone (−)-11 by the π -electron SCF-CI-DV MO method and came to the conclusion that the correct absolute stereochemistry of (−)-11 is aR (or M -helicity). The absolute stereochemistry of a model compound for the calculation was arbitrarily chosen as (aR). The structure of atropisomer (aR)-11 was calculated by molecular mechanics (MMP2) to generate the stable conformation, where the calculated dihedral angle between two ﬂavone moieties was 91◦ . The CD and UV spectra of (aR)-11 were calculated by the π -electron SCF-CI-DV MO method (Figure 5.10). The calculated CD and UV curves are in excellent agreement with the observed curves, including sign, intensity, and position of bands. Based on these results, the absolute stereochemistry of biﬂavone (−)-11 was unambiguously determined as aR (or M helicity) [30]. There are two nonempirical methods for determining the AC of chiral compounds; one is the X-ray Bijvoet method, and the other is the theoretical CD method including the exciton chirality method. These two methods are based on totally different phenomena, but they should come to the same AC for the same compound. In the case of the biﬂavone (−)-11, there had been considerable confusion in the stereochemical studies. After publication of our assignment of (aR) absolute conﬁguration based on theoretical CD studies [30], the opposite AC, namely (aS ), by X-ray crystallography was reported [31]. Which determination is more reliable? Most people may support the determination by X-ray analysis. However, we were conﬁdent about our assignment by the theoretical CD calculation, because of the nonempirical nature of the method. To solve such a problem, we proved the absolute stereostructure by the total synthesis of the natural enantiomer. We designed a synthetic route, where the absolute stereochemistry of an intermediate was determined by X-ray crystallography, and we succeeded in the total synthesis of the natural atropisomer (−)-11 [32, 33]. The CD and UV spectra of the synthetic sample (aR)-(−)-11 were identical to those of the natural sample. Therefore, it was concluded that the absolute stereochemistry of natural biﬂavone (−)-11 is (aR). Thus our theoretical determination of the absolute stereochemistry of biﬂavone 11 by the π -electron SCF-CI-DV MO method was thus proved experimentally by the total synthesis of natural atropisomer 11.

5.4. CD SPECTRA AND ABSOLUTE STEREOSTRUCTURES OF UNIQUE CHIRAL OLEFINS: DISCOVERY AND DEVELOPMENT OF LIGHT-POWERED MOLECULAR MOTORS There are various kinds of chiral compounds devoid of chirality centers. Those compounds cannot take a planar structure because of strong steric hindrance. Some examples of these compounds are chiral oleﬁns, (E )-1,1 ,2,2 ,3,3 ,4,4 -octahydro-4,4 biphenanthrylidene (12) and its (Z )-isomer (13) as shown in Chart 5.2. Oleﬁns 12 and 13 can exist as chiral compounds, and they have been actually resolved into enantiomers by chiral HPLC using a chiral stationary phase [34]. The CD spectra of these oleﬁns showed intense Cotton effects reﬂecting their twisted π -electron chromophores. However, their absolute conﬁgurations have remained undetermined. To solve this problem, we carried out the theoretical calculation of their CD spectra by the π -electron SCF-CI-DV MO method, synthesis of enantiopure compounds, and the experimental determination of their absolute conﬁgurations as follows [35].

179

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

8

8 8

8

10

5

10

5

10

1

1 3

5

1

1

3

[CD(+)239.0](M,M)-(E)-12

10

5

[CD(–)239.0](P,P)-(E)-12

3

3

[CD(+)238.1](M,M)-(Z)-13

[CD(–)238.1](P,P)-(Z)-13

Chart 5.2. Absolute stereochemistry of unique chiral oleﬁns, (E)-1,1 ,2,2 ,3,3 ,4,4 -octahydro-4,4 biphenanthrylidene (12) and its (Z)-isomer (13).

5.4.1. Synthesis of Enantiopure Chiral Oleﬁns 12 and 13, and Their CD Spectra During our calculation of CD and UV spectra by the π -electron SCF-CI-DV MO method, we realized that the reported CD ε values [34] are too small compared to the calculated values. To obtain reliable CD and UV data, we synthesized the enantiopure target compounds. According to the reported procedure [34], the racemic oleﬁns, trans-(±)-12 and cis-(±)-13 were synthesized as shown in Scheme 5.2. The relative stereostructures of trans-oleﬁn 12 and cis-oleﬁn 13 were determined by 1 H NMR spectroscopy and then conﬁrmed by the X-ray crystallographic analysis of trans-oleﬁn 12 [35] as illustrated in Figure 5.11. The molecular framework of this compound is thus nonplanar and strongly twisted, supporting that this molecule can take a chiral form. Next the enantioseparation of racemic trans-oleﬁn (±)-12 by chiral HPLC was attempted. We found that hydrocarbon 12 could be completely resolved into enantiomers using a chiral stationary phase of (+)-poly(triphenylmethylmetacrylate) under the reverse phase condition using MeOH as eluent and a column temperature of 3◦ C. To remove a small amount of the polymer of the chiral stationary phase, which was present as a contaminant, the fraction of each enantiomer was puriﬁed by HPLC (ODS-C18 , MeOH). From the ﬁrst-eluted fraction, enantiopure oleﬁn [CD(+)239.0]-(E )-12 was obtained, and its 1 H NMR spectrum was identical to of racemate (±)-12. The CD and UV spectra of the ﬁrst-eluted trans-oleﬁn [CD(+)239.0]-(E )-12 are shown in Figure 5.12, where the UV spectrum shows a broad band at 329.8 nm, which

8 8 10

5

1 O 14

10

5

+ 1

3

3 (±)-(Z)-13 (±)-(E)-12

Scheme 5.2. Preparation of racemic oleﬁns, (E)-1,1 ,2,2 ,3,3 ,4,4 -octahydro-4,4 -biphenanthrylidene (12) and its (Z)-isomer (13).

180

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 5.11. ORTEP drawing of racemic trans-oleﬁn, (E)-1,1 ,2,2 ,3,3 ,4,4 -octahydro-4,4 -biphenanthrylidene (±)-12. (Redrawn from reference 35, with permission.)

+100 A = +211.5

239.0 (+58.2)

Δe

e × 10–4

0

CD –100 214.2 (−153.3) 216.2 (82,800)

10 (M,M)-(E)

–200 Obsd in MeOH

5

UV

200

250

300 λ (nm)

350

400

0

Figure 5.12. CD and UV spectra of the ﬁrst-eluted trans-oleﬁn [CD(+)239.0]-(E)-12 in MeOH. (Redrawn from reference 35, with permission.)

181

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

may be assigned to the 1 La transition of naphthalene chromophore. In the corresponding region, the CD spectrum shows a broad positive Cotton effect of medium intensity (λext 331.8 nm, ε +26.0). On the other hand, in the 1 Bb transition region, the UV spectrum shows an intense broad band (λext 232.2 nm, ε 61,800 and λext 216.2 nm, ε 82,800), and the CD spectrum shows intense positive and negative Cotton effects (λext 239.0 nm, ε +58.2 and λext 214.2 nm, ε −153.3): The amplitude A value between the peak and trough is +211.5. Such intense CD Cotton effects clearly indicate that the π -electron system of trans-12 is strongly twisted. The enantioseparation of cis-oleﬁn (±)-13 was next examined, and we found that the reverse-phase HPLC used for trans-oleﬁn (±)-12 was not useful. Instead, the chiral HPLC using (+)-poly(triphenylmethylmetacrylate) and hexane as eluent was effective; cis-oleﬁn (±)-13 was partially separated into enantiomers at 3◦ C. To obtain the enantiopure oleﬁn, the ﬁrst-eluted fraction was recycled ﬁve times. Since we found the unexpected thermal racemization of cis-oleﬁn 13 at room temperature, as will be discussed later, the CD and UV spectra were immediately measured after HPLC separation. Figure 5.13 shows the CD and UV spectra of the ﬁrst-eluted cis-oleﬁn [CD(+)238.1]-(Z )-13 in hexane. The ﬁrst-eluted cis-oleﬁn [CD(+)238.1]-(Z )-13 exhibits a broad UV band at 301.9 nm (ε 11,300) at the 1 La transition of naphthalene chromophore. In the corresponding region, the CD spectrum shows a broad weak negative Cotton effect

238.1 (+189.7)

+200

A = +429.0

CD

+100 Δe

e × 10–4

0

–100

–200 223.5 (–239.3)

10

(M,M)-(Z)

223.0 (73,700) –300 Obsd in hexane

5

UV

Figure 5.13. CD and UV spectra of the 200

250

300 λ (nm)

350

400

0

ﬁrst-eluted cis-oleﬁn [CD(+)238.1]-(Z)-13 in hexane. (Redrawn from reference 35, with permission.)

182

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(λext 339.0 nm, ε −12.3). In the 1 Bb transition region, the UV spectrum shows an intense broad band (λext 223.0 nm, ε 73,700), while the CD spectrum shows very intense positive and negative Cotton effects (λext 238.1 nm, ε +189.7 and λext 223.5 nm, ε −239.3): The amplitude A value between the peak and trough is +429.0. The intense CD data thus indicate that the π -electron system of cis-13 is also strongly twisted.

5.4.2. Absolute Stereochemistry of Chiral Oleﬁns trans-12 and cis-13 as Determined by the Calculation of CD and UV Spectra Using the SCF-CI-DV MO Method To determine the absolute stereochemistry of chiral oleﬁns 12 and 13, we next calculated the CD and UV spectra by the π -electron SCF-CI-DV MO method. As a model compound for the calculation, the (M,M)-(E )-enantiomer 12 was arbitrarily chosen, and the atomic coordinates were obtained by the MOPAC AM1 calculation. The calculated CD and UV spectra of trans-oleﬁn (M,M)-(E )-12 are shown in Figure 5.14 [35]. As seen in ﬁgures 5.12 and 5.14, the CD and UV spectra of trans-oleﬁn 12 were reproduced well by the calculation. In the 1 La transition around 300 nm, the positive CD band was obtained by calculation, which agreed with the observed CD, although its intensity was smaller than the observed one. In the 1 Bb transition around 200–250 nm, +100 240.4 (+87.9)

0 Δe –100

e × 10–4

CD

–200 219.3 (–256.0)

10 223.2 (94,400) –300

(M,M)-(E)

Calcd 5

UV A = +343.9

Figure 5.14. CD and UV spectral curves of 200

250

300 λ (nm)

350

400

0

trans-oleﬁn (M,M)-(E)-12 calculated by the π -electron SCF-CI-DV MO method. (Redrawn from reference 35 with permission.)

183

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

the intense positive and negative CD Cotton effects were obtained by calculation. The sign and shape of Cotton effects agreed well with those of observed spectrum, although the calculated intensity was stronger. Therefore, the absolute stereochemistry of the ﬁrsteluted trans-oleﬁn [CD(+)239.0]-(E )-12 was clearly determined to be (M,M) by the theoretical calculation. The ﬁrst-eluted enantiomer is designated as [CD(+)239.0]-(M,M)(E )-12 [35]. The CD and UV spectra of cis-oleﬁn (M,M)-(Z )-13 were similarly calculated by the π -electron SCF-CI-DV MO method as shown in Figure 5.15. When comparing with Figure 5.13, it was clear that the CD and UV spectra of cis-oleﬁn 13 were also wellreproduced by the calculation. In the 1 La transition around 300–370 nm, the negative CD band was obtained by calculation, which agreed with the observed CD, although its intensity was again larger. In the 1 Bb transition around 200–250 nm, the intense positive and negative CD Cotton effects were obtained by calculation. The sign and shape of the Cotton effects agreed well with those of the observed spectrum, although the calculated intensity in this case was weaker. Therefore, the absolute stereochemistry of the ﬁrst-eluted cis-oleﬁn [CD(+)238.1]-(Z )-13 was clearly determined to be (M,M) by theoretical calculation. Thus the absolute stereochemistry of the ﬁrst-eluted enantiomer is designated as [CD(+)238.1]-(M,M)-(Z )-13 [35].

+100 232.6 (+76.7)

0

Δe

10 CD e × 10–4

–100

215.5 (–158.0) 211.9 (57.600) –200

(M,M)-(Z)

5

Calcd A = +234.7

UV

Figure 5.15. CD and UV spectral curves of 200

250

300 λ (nm)

350

400

0

cis-oleﬁn (M,M)-(Z)-13 calculated by the π -electron SCF-CI-DV MO method. (Redrawn from reference 35, with permission.)

184

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

5.4.3. Unexpected Thermal Racemization of cis-Oleﬁn 13 During the studies discussed above, we observed that chiral cis-oleﬁn 13 underwent an unexpected thermal racemization at room temperature. We had ﬁrst considered that if one of these chiral oleﬁns undergoes the racemization, it must be trans-oleﬁn 12, because of less steric hindrance. In cis-oleﬁn 13, two naphthalene moieties overlap with each other as seen in the X-ray stereostructure (Figure 5.16), which generates a severe steric hindrance, and therefore it is difﬁcult to image the racemization. However, it was conﬁrmed that cis-oleﬁn 13 really undergoes thermal racemization, which was monitored by CD spectrum as illustrated in Figure 5.17. The thermal racemization of cis-oleﬁn 13 was also measured by the magnetization transfer experiment of 1 H NMR spectroscopy. On the other hand, it was clariﬁed that trans-oleﬁn 12 does not undergo racemization at room temperature. To obtain the enantiopure cis-oleﬁn 13 and to measure its CD spectrum, we carried out the HPLC separation at lower temperature. That is, the chiral HPLC column was cooled at −30◦ C during the enantioseparation, and the CD spectrum was measured at −50◦ C as illustrated in Figure 5.18, where the observed CD intensity was corrected for volume contraction. The CD intensity of cis-13 in Figure 5.18 is larger than that in Figure 5.13. It is thus important to measure the CD spectrum of the enantiopure sample. The reaction mechanism of the thermal racemization of cis-oleﬁn 13 was later clariﬁed by theoretical calculation [37], where the (M, P)-(Z ) isomer was included as an intermediate. This mechanism indicates that two naphthalene moieties slip by each other. This is the critical reaction step in the light-powered molecular motor discussed below.

Figure 5.16. ORTEP drawing of racemic cis-oleﬁn, (Z)-1,1 ,2,2 ,3,3 ,4,4 -octahydro-4,4 biphenanthrylidene (±)-13. (Redrawn from reference 36, with permission.)

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

185

+200 0h

Racemization of

1h

+100

2h Δe

(M,M)-cis

3h 4h 5h

in hexane at room temp.

0 % 100 UV λmax 222.8 nm –100

CD CD λext 238.1 nm

50

t1/2 = 1.2 h –200

0 200

250

Figure 5.17. Decrease of CD intensity of cis-oleﬁn 0

300 λ (nm)

2

4

6 h 350

[CD(+)238.1]-(M,M)-(Z)-13 in hexane due to the thermal racemization at room temperature. (Redrawn from reference 36 with permission.)

5.4.4. Experimental Determination of Absolute Stereochemistry of trans-Oleﬁn 12 and cis-Oleﬁn 13: Use of the Internal Reference of Absolute Conﬁguration in X-Ray Analysis The absolute stereostructures of trans- and cis-oleﬁns were theoretically determined by the calculation of CD and UV spectra using the π -electron SCF-CI-DV MO method as described above. But the authors believe that the theoretically determined results have to be proved experimentally. That is, the problem is now how to prove the absolute stereochemistry of 12 and 13 experimentally. To solve this problem, we adopted the following strategy. At ﬁrst we thought to synthesize derivatives containing a heavy atom like Br or S and to carry out X-ray crystallography for determining the absolute conﬁguration using the X-ray Bijvoet method. However, all attempts of the synthesis were unsuccessful. It was then decided to introduce an internal reference of the absolute conﬁguration—that is a methyl group in a chiral position, as shown in compounds 15 and 16 (Chart 5.3). As the starting material for the synthesis, racemic cis-alcohol (±)-17 was selected, and it was enantioresolved by the CSDP acid (camphor-sulfonyl-dichloro-phthalic acid) method [9–12] as shown in Scheme 5.3. Cis-alcohol (±)-17 was esteriﬁed with CSDP

186

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

239.6 (+222.2)

+200 Δe

CD

+100

282.5 (+11.9)

e × 10–4

0 338.0 (–14.0) –100 256.8 (–80.1) 10 –200 (M,M)-cis

224.0 (–281.3)

–50.0°C

–300

222.8 (71.900) in hexane 5 UV 301.9 (11,300)

300

200

400

Figure 5.18. CD and UV spectra of cis-oleﬁn

0

[CD(+)238.1]-(M,M)-(Z)-13 in hexane at −50◦ C. (Redrawn from reference 36, with permission.)

λ (nm)

8 8

H CH3

10

5

10

5

1 1

3

H3C

[CD(–)237.2]-(3R,3'R)(P,P)-(E)-15

H

3

H CH3

H3C

H

[CD(–)238.0]-(3R,3′R)(P,P)-(Z)-16

Chart 5.3. Dimethyl-substituted chiral oleﬁns 15 and 16 useful for determining absolute conﬁgurations.

acid (1S,2R,4R)-(−)-18 to yield a diastereomeric mixtures of esters, which was easily separated by HPLC on silica gel. The second-eluted ester (−)-19b was obtained as a solid, which was recrystallized from EtOAc giving large prisms suitable for X-ray crystallography. The AC of the second-eluted ester (−)-19b was unambiguously determined to be (3S, 4S ) by the heavy atom effect of Cl and S atoms and also by the use of the camphorsultam moiety as an internal reference of the AC (Figure 5.19).

187

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

Cl

Cl Cl

Cl Cl N S O

+ OH CH3

O

N S O O

COOH

(1S,2R,4R)-(–)-18

+ O

O

H3C

O

(±)-17

OO

Cl N S O

O

O H3C

O

(3S,4S)-(–)-19b, X-ray

(3R,4R)-(+)-19a

Scheme 5.3. Enantioresolution of alcohol (±)-17 by the CSDP acid method.

Figure 5.19. ORTEP drawing of the second-eluted CSDP ester (3S,4S)(−)-19b. (Redrawn from reference 38, with permission.)

The AC of the ﬁrst-eluted CSDP ester (+)-19a was, therefore, assigned as (3R,4R). The LiAlH4 reduction of the ﬁrst-eluted CSDP ester (3R,4R)-(+)-19a yielded enantiopure cis-alcohol (3R,4R)-(+)-17, which was then oxidized to give enantiopure ketone (3R)-(−)-20 (Scheme 5.4). McMurry coupling reaction of ketone (3R)-(−)-20 yielded the desired dimethyl trans-oleﬁn (−)-15, which was puriﬁed by repeated HPLC under normal-phase and reverse phase conditions affording enantiopure sample, [CD(−)237.2](3R,3 R)-(P, P)-(−)-15 ([α]D −446.2). The large negative optical rotation value of the product (−)-15 indicates that no racemization occurred during the McMurry reaction.

Cl Cl

S O

N O

OO H3C

O

H CH3

HO

O CH3

(3R,4R)-(+)-17 (3R,4R)-(+)-19a

CH3 (3R)-(–)-20

H3C H [CD(–)237.2]-(3R,3′R)(P,P)-(E)-15

Scheme 5.4. Synthesis of dimethyl trans-oleﬁn [CD(–)237.2]-(3R,3 R)-(P, P)-(−)-15.

188

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The relative and absolute stereochemistry of dimethyl trans-oleﬁn (−)-15 was determined by X-ray crystallography as follows. Recrystallization of the product (−)-15 from MeOH afforded prismatic crystals, one of which was subjected to X-ray crystallographic analysis, and the stereostructure was determined as shown in Figure 5.20 [38]. Since compound (−)-15 is just a hydrocarbon containing no heavy atom, the AC of (−)-15 could not be determined by the X-ray analysis. However, it has two methyl groups at chiral positions, and therefore the (3R,3 R) conﬁguration was used as an internal reference of the AC. The absolute helicity of the chiral oleﬁn part was thus determined as (P, P ) from the ORTEP drawing in Figure 5.20. The CD and UV spectra of dimethyl trans-oleﬁn (3R,3 R)-(P, P)-(E )-(−)-15 were next measured as shown in Figure 5.21, where the CD spectrum shows a negative Cotton effect at 237.2 nm. Hence the enantiomer is designated as [CD(–)237.2]-(3R,3 R)-(P, P)(E )-(−)-15. The UV spectrum of (−)-15 is similar to that of trans-oleﬁn [CD(+)239.0]-(M,M)(E )-12. On the other hand, the CD spectra of [CD(−)237.2]-(3R,3 R)-(P, P)-(E )-(−)-15 is also similar to that of [CD(+)239.0]-(M,M)-(E )-12 in position, shape, and absolute intensity, but opposite in sign (compare Figures 5.12 and 5.21). These results clearly indicate that the dimethyl groups in (−)-15 do not change the molecular conformation too much, and that the absolute helicity of the chiral oleﬁn part in [CD(+)239.0]-(E )-12 is (M,M). The absolute stereochemistry of trans-oleﬁn [CD(+)239.0]-(M,M)-(E )-12 as determined previously by the theoretical calculation of CD spectrum was thus conﬁrmed in an experimental manner. To synthesize dimethyl cis-oleﬁn (3R,3 R)-(Z )-16, enantiopure dimethyl trans-oleﬁn [CD(–)237.2]-(3R,3 R)-(P, P)-(E )-(−)-15 was irradiated by a high-pressure mercury lamp using a Pyrex glass ﬁlter to yield dimethyl cis-oleﬁn 16 (Scheme 5.5), the CD spectrum of which showed an intense negative Cotton effect at 238.0 nm. Therefore, the enantiomer was designated as [CD(−)238.0]-(3R,3 R)-(Z )-16. The helical sense of the naphthalene–double bond–naphthalene moiety of [CD(−)238.0]-(3R,3 R)-(Z )-16 was ﬁrst studied by 1 H NMR spectroscopy and it was ﬁnally determined by X-ray crystallography as follows. We ﬁrst obtained single crystals of racemate (±)-16, which were subjected to X-ray analysis, affording the ORTEP drawing as shown in Figure 5.22.

Figure 5.20. ORTEP drawing of dimethyl trans-oleﬁn [CD(–)237.2](3R, 3 R)-(P, P)-(E)-(−)-15. (Redrawn from reference 38, with permission.)

189

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

217.8 (+148.5) +150 H CH3 +100 CD

H3 C H

+50 (3R,3′R)-(P,P)-(E)

Δe

e × 10–4

0

–50 10 237.2 (–92.5) A = –241.0 –100 218.4 (85,700) 5 UV

Obsd in MeOH

Figure 5.21. CD and UV spectra of dimethyl 200

250

300

350

400

0

trans-oleﬁn [CD(–)237.2]-(3R,3 R)-(P, P)-(E)-(−)-15 in MeOH. (Redrawn from reference 38, with permission.)

λ (nm)

H CH3 hν

H3C H

[CD(–)237.2]-(3R,3′R)(P,P)-(E)-15

H

CH3

H 3C H

[CD(–)238.0]-(3R,3′R)(P,P)-(Z)-16

Scheme 5.5. Synthesis of dimethyl cis-oleﬁn [CD(–)238.0]-(3R,3 R)-(P, P)-(Z)-16.

From the ORTEP drawing, the relative stereochemistry of (±)-16 was determined as (3R ∗ ,3 R ∗ )-(P ∗ , P ∗ )-(Z ). Since the 1 H NMR spectrum of chiral oleﬁn [CD(−)238.0](3R,3 R)-(Z )-16 was identical with that of racemate (3R ∗ ,3 R ∗ )-(P ∗ , P ∗ )-(Z )-(±)-16, the absolute helicity of chiral oleﬁn was determined as (P, P ). The absolute stereochemistry of dimethyl cis-oleﬁn [CD(−)238.0]-(3R,3 R)-(P, P)-(Z )-16 was thus unambiguously determined. Later, we obtained single crystals of chiral dimethyl cis-oleﬁn [CD(–)238.0](3R,3 R)-(P, P)-(Z )-16, and the same AC was determined from the X-ray crystallographic analysis [39].

190

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 5.22. ORTEP drawing of dimethyl cis-oleﬁn (3R∗ ,3 R∗ )-(P ∗ , P ∗ )-(Z)-(±)-16. (Redrawn from reference 38, with permission.)

+400 223.4 (+334.0)

+200

CD

e × 10–4

H3 C H

H CH3

Δe

(3R,3′R)-(P,P)-(Z) 0

10

238.0 (–226.9)

–200

A = –560.9 5

222.4 (76,500)

UV

–400

Obsd in hexane

Figure 5.23. CD and UV spectra of dimethyl 200

250

300 λ (nm)

350

400

0

cis-oleﬁn [CD(–)238.0]-(3R,3 R)-(P,P)-(Z)-16 in hexane. (Redrawn from reference 38, with permission.)

191

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

The CD and UV spectra of [CD(–)238.0]-(3R,3 R)-(P, P)-(Z )-16 is shown in Figure 5.23. As in the case of trans-oleﬁn 12 and dimethyl trans-oleﬁn 15, the CD and UV spectra of dimethyl cis-oleﬁn [CD(−)238.0]-(3R,3 R)-(P, P)-(Z )-16 are very similar to those of cis-oleﬁn [CD(+)238.1]-(M,M)-(Z )-13 in position, shape, and absolute intensity, but opposite in sign (compare Figures 5.13 and 5.23). Therefore, the absolute stereochemistry of cis-oleﬁn [CD(+)238.1]-(M,M)-(Z )-13 previously determined by the theoretical calculation of CD was conﬁrmed experimentally. The X-ray crystallographic method using an internal reference of absolute conﬁguration is thus very useful for the absolute conﬁgurational assignment of various chiral compounds. It is interesting that the CD spectrum of dimethyl cis-oleﬁn [CD(–)238.0]-(3R,3 R)(P, P)-(Z )-16 did not show a decrease in intensity at room temperature, unlike the case of cis-oleﬁn [CD(+)238.1]-(M,M)-(Z )-13. That is, dimethyl cis-oleﬁn 16 does not racemize at all, because the two methyl groups block the racemization process.

5.4.5. Unique Photo- and Thermo-chemical Behavior of Chiral Dimethyl Oleﬁn: First Discovery and Development of a Light-Powered Chiral Molecular Motor Further studies of the photo- and thermo-chemistry of chiral dimethyl oleﬁns 15 and 16 led to the ﬁrst discovery of a light-powered chiral molecular motor as described below. During the photochemical studies of chiral dimethyl trans-oleﬁn (−)-15, we found the formation of another product, yellow-colored dimethyl trans-oleﬁn (+)-21 as shown in Scheme 5.6, although oleﬁns (−)-15 and (+)-16 are colorless. Later it was clariﬁed that the product (+)-21 was not directly formed from (−)-15 but from dimethyl cis-oleﬁn (+)-16, and this photochemical step is reversible [40]. The structure of yellow-colored dimethyl trans-oleﬁn (+)-21 was ﬁrst studied by 1 H NMR spectrum, where two methyl groups appeared at δ 0.31 ppm, implying that the high ﬁeld shift is due to the anisotropy effect by a neighboring naphthalene ring. These NMR data suggested the structure of (3R,3 R)-(M,M)-(E )-21, in which two methyl groups are placed at the equatorial position. The relative stereochemistry of oleﬁn 21 was conﬁrmed by X-ray crystallography of racemate (±)-21 as shown in Figure 5.24, where the two equatorial methyl groups are in contact with the naphthalene rings, causing a strong steric hindrance between methyl and naphthalene moieties. This effect makes oleﬁn (+)-21 unstable, and also affects the π electron framework to change its color. This is the major reason why oleﬁn 21 is yellow. Figure 5.25 shows the CD and UV spectra of the yellow dimethyl trans-oleﬁn (3R,3 R)-(M,M)-(E )-(+)-21 together with those of colorless dimethyl trans-oleﬁn

H CH3

H CH3

hn

hν

hn H 3C H (3R,3′R)-(P,P)-(E)-(–)-15 with two axial methyl groups

H CH3

H 3C H

(3R,3′R)-(P,P)-(Z)-(+)-16 with two axial methyl groups

H3C

H

(3R,3′R)-(M,M)-(E)-(+)-21 with two equatorial methyl groups

Scheme 5.6. Photochemical interconversion between chiral dimethyl oleﬁns (−)-15, (+)-16, and (+)-21.

192

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 5.24. ORTEP drawing of yellow colored dimethyl trans-oleﬁn (3R∗ ,3 R∗ )-(M∗,M∗ )-(E)-(±)-21. (Redrawn from reference 40, with permission.)

+200 H CH3

+100

CD

H3C

H

(3R,3′R)-(M,M)-(E)-(+)

e × 10–4

Δe 0

H CH3

–100

10

H3C H UV

5

–200

Figure 5.25. CD and UV spectra of

(3R,3′R)-(P,P)-(E)-(–)

200

300

400 λ (nm)

0

yellow-colored unstable dimethyl trans-oleﬁn (3R,3 R)-(M,M)-(E)-(+)-21 in EtOH together with those of colorless stable (3R,3 R)-(P, P)-(E)-(−)-15 in EtOH. (Redrawn from reference 40, with permission.)

193

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

(3R,3 R)-(P, P)-(E )-(−)-15. It should be noted that in the UV–Vis spectrum of (3R,3 R)-(M,M)-(E )-(+)-21, the 1 La transition of the naphthalene chromophore is largely red-shifted at 320–420 nm, causing the yellow color. In the corresponding region, the CD spectrum shows a broad positive Cotton effect reﬂecting the inversed (M,M) helicity, while oleﬁn (3R,3 R)-(P, P)-(E )-(−)-15 exhibits a negative Cotton effect in the 1 La transition region. In the 1 Bb transition region around 200–270 nm, the UV spectrum shows two absorption bands. On the other hand, the CD spectrum shows intense but complex Cotton effects as shown in Figure 5.25, reﬂecting the strongly twisted π -electron system. The absolute stereochemistry of the unique third isomer (3R,3 R)-(M,M)-(E )-(+)-21 was unambiguously determined. It should be noted that the internal reference method of absolute conﬁguration is thus applicable to the X-ray analysis of racemic compounds. As shown in Scheme 5.6, the photochemical step between stable dimethyl cis-oleﬁn (+)-16 and unstable dimethyl trans-oleﬁn (+)-21 was reversible, as expected for oleﬁn compounds. However, it was surprising that the step between stable dimethyl trans-oleﬁn (−)-15 and stable dimethyl cis-oleﬁn (+)-16 was irreversible. That is, the photochemical conversion from stable dimethyl trans-oleﬁn (−)-15 to stable dimethyl cis-oleﬁn (+)-16 occurred, but the reverse reaction did not proceed. Why? To solve this problem, we postulated the reaction scheme as shown in Scheme 5.7, where trans-oleﬁn (−)-15, cis-oleﬁn (+)-16, and trans-oleﬁn (+)-21 were renamed trans-isomer (−)-22a, cis-isomer (+)-22c, and trans-isomer (+)-22d, respectively. As discussed above, it was observed that the photochemical isomerization between stable cis-isomer (+)-22c and unstable trans-isomer (+)-22d was reversible, but unstable transisomer (+)-22d underwent thermal isomerization to stable trans-isomer (−)-22a. It was assumed that the unstable cis-isomer (3R,3 R)-(M,M)-(Z )-22b must exist as a primary

hn

Δ hn

H

H

CH3

CH3

H3C

H

(3R,3′R)-(M,M)-(Z)-22b

H3C

H

H Δ

CH3

H3C H

(3R,3′R)-(P,P)-(Z)-(+)-22c

(3R,3′R)-(P,P)-(E)-(–)-22a

hn H

CH3 hn

CH3 H

(3R,3′R)-(M,M)-(E)-(+)-22d

Scheme 5.7. The cyclic reaction scheme of photo- and thermo-chemical conversions of unique dimethyl oleﬁns.

194

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

product of the photochemical conversion of (−)-22a, and the reverse reaction from 22b to (−)-22a may occur. However, if the isomer 22b is extremely unstable, the formed isomer 22b immediately and irreversibly converts to stable cis-isomer (+)-22c. Therefore, the total reaction from (−)-22a to (+)-22c becomes irreversible in agreement with the observed results [41]. In fact, we succeeded in detecting the unstable isomer (±)-22b in the photochemical reaction of (±)-22a at −60◦ C by 1 H NMR spectroscopy (Figure 5.26) [41]. Later, we realized that this system makes a unidirectional molecular motor rotation around the central double bond [42]. That is, when looking at the molecule from the left side, the naphthalene moiety on the left rotates counterclockwise against the naphthalene moiety on the right-upper side. The photochemical step makes the rotation in both directions, but the thermal step rotates only counterclockwise. The reaction 22a → 22b → 22c → 22d thus makes the 360◦ rotation in the counterclockwise direction, and the motor returns to the starting place 22a. Therefore, the molecular motor can make a continuous rotation under photoirradiation and heating, where the direction of the motor rotation is governed by the molecular chirality. The photochemical energy is thus converted to the mechanical rotation of the molecule, and this was the discovery of the ﬁrst light-powered molecular motor [42].

5.5. A NEW MODEL OF LIGHT-POWERED CHIRAL MOLECULAR MOTOR WITH HIGHER SPEED OF ROTATION The chiral oleﬁns shown in Scheme 5.7 ideally satisfy the requirements of molecular motor, but its rotation was not fast, because the fourth rotation step (i.e., thermal reaction Aromatic Part

ppm

8.25

(3R*,3R*)-(P*,P*)-trans stable

8.00

7.75

7.50

7.25

7.00

6.75

7.75

7.50

7.25

7.00

6.75

irradiation at –60.0 °C after 1 day

ppm

8.25

8.00

NMR detection of unstable dimethyl cis-oleﬁn in CD2 Cl2 : ∗, unstable cis-oleﬁn (3R∗,3 R∗ )-(M∗ ,M∗ )-(Z)-22b; ◦, stable trans-oleﬁn (3R∗ ,3 R∗ )-(P ∗ , P ∗ )-(E)-22a. (Redrawn from reference

Figure 5.26. 41.)

1H

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

195

22d → 22a), needs higher temperature and hence is slow, because of the severe steric hindrance between methyl group and naphthalene moiety. To make a faster molecular motor, we improved the structure as follows. A new model of chiral molecular motor 23 with higher speed of rotation was designed as shown in Chart 5.4, where the six-membered rings in molecular motor 22 were replaced by ﬁve-membered rings to diminish the steric hindrance between methyl group and naphthalene moiety [43, 44].

5.5.1. Synthesis, CD Spectra, X-ray Structure, and Absolute Stereochemistry To synthesize the new chiral molecular motor, we adopted the strategy shown in Scheme 5.8. Racemic cis-alcohol (±)-24 was esteriﬁed with CSDP acid, giving a diastereomeric mixture of esters, which was easily separated by HPLC on silica gel [43]. One of the advantages of the CSDP acid method is that CSDP esters tend to give single crystals suitable for X-ray crystallography with high probability [9–12]. In fact, the second-eluted CSDP ester (−)-cis-25b was obtained as single crystals by recrystallization from hexane/EtOAc. The single crystal was subjected to X-ray analysis, and the absolute stereostructure was determined by the use of CSDP acid moiety as an internal reference and also by the heavy atom effect (Figure 5.27). That is, the (1S,2S ) absolute conﬁguration was assigned to (−)-cis-25b. The absolute conﬁguration of the ﬁrst-eluted ester was naturally determined as (1R,2R). To recover alcohol, the second-eluted CSDP ester (1S,2S )-(−)-cis-25b was hydrolyzed with KOH/MeOH yielding enantiopure cis-alcohol (1S,2S )-(+)-24. Cis-alcohol (1S,2S )-(+)-24 was next oxidized to yield enantiopure ketone (S )-(+)26. The product was then subjected to the McMurry coupling reaction with TiCl3 and LiAlH4 in THF giving dimethyl trans-oleﬁn (2S,2 S)-(E )-(−)-23a (colorless prisms, yield 21%) and dimethyl cis-oleﬁn (2S,2 S)-(Z )-(−)-23c (pale yellow prisms, yield 5%) as shown in Scheme 5.8. The relative stereochemistry of racemate (±)-23a was determined to be (2S ∗ ,2 S ∗ )∗ (M ,M ∗ )-(E ) by X-ray crystallography as shown in Figure 5.28. Since the 1 H NMR spectrum of chiral trans-oleﬁn (−)-23a was identical to that of racemate (2S ∗ ,2 S ∗ )-(M ∗,M ∗ )(E )-(±)-23a, the absolute stereochemistry of (−)-23a was unambiguously determined to be (2S,2 S)-(M,M)-(E ).

8

CH3

H

5

2 1

1

2

H CH3

5 8

[CD(–)257.8]-(2S,2′S)(M,M)-(E)-23a

CH3

H

H CH3

[CD(–)270.0]-(2S,2′S)(M,M)-(Z)-23c

Chart 5.4. A new model of chiral molecular motor with higher speed of rotation.

196

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Cl

Cl

Cl

Cl

H3C

N

N

+

S

OH

O

S O

COOH

O

O

O (1S,2R,4R)-(–)-18

(±)-24

O

+ O

O

H3C (1R,2R)-(–)-cis-25a

Cl Cl N S

O

O

O

O

(1S,2S)-(–)-cis-25b

O

HO

H3C

CH3

(1S,2S)-(+)-cis-24 (1S,2S)-(–)-cis-25b, X-ray

H

CH3 +

O

CH3

H

H

H

CH3

CH3

CH3

(S)-(+)-26 [CD(–)257.8]-(2S,2′S)(M,M)-(E)-(–)-23a

[CD(–)270.0]-(2S,2′S)(M,M)-(Z)-(–)-23c

Scheme 5.8. Synthesis of chiral dimethyl oleﬁns (−)-23a and (−)-23c.

Figure 5.27. ORTEP drawing of the second-eluted CSDP ester (1S,2S)-(−)-cis-25b. (Redrawn from reference 43, with permission.)

197

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

Figure 5.28. ORTEP drawing of dimethyl trans-oleﬁn (2S∗ ,2 S∗ )-(M∗ ,M∗ )-(E)-(±)-23a. (Redrawn from reference 43, with permission.)

It was surprising to see the CD spectrum of (−)-23a, which showed several complex negative Cotton effects in the 1 Bb transition region at 200–270 nm (Figure 5.29). The CD pattern is much different from that of previous chiral six-membered oleﬁns. On the other hand, in the 1 La transition region at 300–400 nm, a broad positive Cotton effect was observed. Since the CD spectrum showed a negative Cotton effect at 257.8 nm, the enantiomer was designated as [CD(–)257.8]-(2S,2 S)-(M,M)-(E )-(−)-23a.

obsd CD 365.8 ( +16.7) 349.6 ( +18.2) 295.8 ( –10.5) 257.8 (–140.0) 247.8 ( –57.1) 226.6 (–100.2) 214.2 (–109.6)

200 CD Δe

H

CH3

e × 10–4

0

–100 CH3 –200

H

[CD(–)257.8]-(2S,2′S)(M,M)-(E) obsd UV 367.6 (25,600) 352.0 (25,900) 243.6 (39,200) 216.4 (87,700)

UV

200

10

in MeOH

300

400 λ (nm)

5

Figure 5.29. CD and UV spectra of stable dimethyl trans-oleﬁn

0 500

[CD(–)257.8]-(2S,2 S)-(M,M)-(E)-(−)-23a in MeOH. (Redrawn from reference [43], with permission.)

198

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The stereochemistry of the other oleﬁn (−)-23c was determined to be (2S,2 S)(M,M)-(Z )-(−)-23c by 1 H NMR analysis and also by the fact that the UV irradiation of trans-oleﬁn (−)-23a yielded cis-oleﬁn (−)-23c. The CD and UV spectra of cis-oleﬁn (−)-23c are shown in Figure 5.30, where in the 1 La transition region at 350–400 nm a broad negative Cotton effect was observed. On the other hand, in the 1 Bb transition region at 200–290 nm, strong negative and positive Cotton effects were obtained. Since the CD Cotton at 270 nm was negative, this enantiomer was fully designated as [CD(–)270.0](2S,2 S)-(M,M)-(Z )-(−)-23c. For the cis-oleﬁn [CD(–)269.8]-23c, the absolute stereochemistry had been previously assigned by the Feringa group as (2R,2 R)-(P, P)-(Z )-23c by comparison with the CD spectrum of the six-membered ring compound [CD(–)238.0]-(3R,3 R)-(P, P)-(Z )-16 in Figure 5.23. They considered that the pattern of negative (270.0 nm)/positive (232.0 nm) bands of [CD(–)269.8]-23c was similar to that of the negative (238.0 nm)/positive (223.4 nm) bands of six-membered cis-oleﬁn 16 [45]. Since their assignment was thus opposite to ours, it was concluded that such comparison of CD spectra led to the erroneous assignment [43].

5.5.2. Light-Powered Chiral Molecular Motor with Higher Speed of Rotation: Isolation or 1 H NMR Detection of the Unstable Motor Rotation Isomers and Their CD Spectra The new chiral oleﬁns with ﬁve-membered ring systems worked as a light-powered chiral molecular motor with higher speed of rotation as shown in Scheme 5.9. To clarify the

100

obsd CD 379.2 ( –6.5) 304.2 ( –21.5) 291.8 ( –23.9) 270.0 (–159.6) 232.0 (+116.2) 223.8 ( –21.0) 215.4 ( +42.7) 210.2 ( +30.5)

CD

Δe

e × 10–4

0

–100

CH3 –200

in MeOH

UV

200

H

H

CH3

[CD(–)270.0]-(2S,2′S)(M,M)-(Z) obsd UV 369.6 (15,300) 330.4 ( 7,000) 305.8 ( 8,200) 294.4 ( 6,500) 254.2 (28,200) 221.4 (72,100)

300

400 λ (nm)

10

5

Figure 5.30. CD and UV spectra of stable 0 500

dimethyl cis-oleﬁn [CD(−) 270.0]-(2S,2 S)(M,M)-(Z)-(−)-23c in MeOH. (Redrawn from reference 43, with permission.)

199

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

hn

Δ hn

H

CH3

CH3

CH3

H

H

CH3

[CD(+)279.2]-(2S,2′S)(P,P)-(Z)-23b

H CH3

Δ hn

[CD(–)257.8]-(2S,2′S)(M,M)-(E)-(–)-23a

H

CH3

H

H CH3

[CD(–)270.0]-(2S,2′S)(M,M)-(Z)-(–)-23c hn

H

Scheme 5.9. Rotation scheme of a

CH3

light-powered chiral molecular motor of ﬁve-membered ring type rotating with a higher speed.

[CD(+)269.0]-(2S,2′S)(P,P)-(E)-23d

motor rotation mechanism, we ﬁrst tried to isolate the unstable cis-oleﬁn 23b as follows [44]. A solution of the stable trans-oleﬁn 23a in CH2 Cl2 was irradiated with UV light at 312 nm at −78◦ C, and the reaction mixture was subjected to HPLC (ODS, MeOH) at −40◦ C. The desired unstable cis-oleﬁn 23b was obtained as a yellow powder of the second-eluted fraction, and the structure of 23b was determined by 1 H NMR spectra measured at −30◦ C. To avoid the thermal isomerization, the CD spectrum of the unstable cis-oleﬁn (2S,2 S)-(P, P)-(Z )-23b was measured in MeOH at −32.5◦ C as illustrated in Figure 5.31, where very complex and intense Cotton effects were observed in the 1 Bb transition region (200–300 nm), reﬂecting the strongly twisted π -electron structure. In addition, a weak

Δε +100 CD +50

H H CH3 CH3 [CD(+)279.2]-(2S,2′S)(P,P)-(Z)

0

–50 in MeOH at –32.5°C –100

–150 200

300

obsd CD 400.6 ( +8.8) 279.2 (+100.6) 261.0 ( –71.3) 233.0 (–144.1) 221.6 ( +70.9) 400

λ (nm)

Figure 5.31. CD spectrum of the unstable 500

cis-oleﬁn [CD(+)279.2]-(2S,2 S)-(P, P)-(Z)-23b in MeOH at −32.5◦ C. (Redrawn from reference 44, with permission.)

200

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

positive broad Cotton effect appeared at 350–450 nm, and the strong red-shift of the electronic transition caused the yellow color of this unstable motor rotation isomer 23b. Since the unstable cis-oleﬁn (2S,2 S)-(P, P)-(Z )-23b shows an intense positive CD band at 279.2 nm, its AC was fully designated as [CD(+)279.2]-(2S,2 S)-(P, P)-(Z )-23b. We also attempted to isolate the other unstable motor rotation isomer 23d in a similar manner. A solution of the stable cis-oleﬁn (−)-23c in MeOH was irradiated with UV light at 330 nm at −78◦ C for 18 s. The photochemical reaction was monitored by CD spectroscopy performed at −62◦ C to ﬁnd that the reaction had reached a photoequilibrium state between 23c and 23d. The equilibrium ratio was determined to be 23d/23c = 93 : 7 by 1 H NMR spectra measured at −60◦ C. The extremely unstable trans-oleﬁn (2S,2 S)-(P, P)-(E )-23d shows intense and complex CD Cotton effects in the 1 Bb transition region (200–290 nm), reﬂecting the strongly twisted π -electron structure, while at 300–450 nm a negative broad CD band was observed (Figure 5.32). Since the unstable trans-oleﬁn 23d shows a positive CD band at 269.0 nm, its AC was designated as [CD(+)269.0]-(2S,2 S)-(P, P)-(E )-23d.

5.5.3. Light-Powered Chiral Molecular Motor with Higher Speed of Rotation: Dynamics of Motor Rotation Studied by CD Spectroscopy To clarify the motor rotation mechanism, the photochemical and thermal reactions of oleﬁn 23 were studied by CD and 1 H NMR spectroscopy as follows [44]. (1) The Photochemical First Motor Rotation Step. A solution of stable transoleﬁn (2S,2 S)-(M,M)-(E )-(−)-23a in MeOH was irradiated with UV light at 312 nm at −78◦ C, and the change was monitored by CD measured at −25◦ C (Figure 5.33). The photochemical reaction was thus very fast and reached photoequilibrium after 20.5 s of irradiation. The 23b/23a ratio at the ﬁnal stage was 94:6. The reverse reaction 23b → 23a was similarly studied by CD spectroscopy, where a solution of unstable cis-oleﬁn (2S,2 S)-(P, P)-(Z )-23b in MeOH was irradiated with

+150 Δε H

+100 CD CH3

+50

H

[CD(+)269.0]-(2S,2′S)(P,P)-(E)

0

obsd CD 386.0 ( –35.0) 269.0 ( +52.4) 255.4 ( +41.4) 240.0 ( –7.9) 230.4 (+137.1) 220.2 ( –30.2) 214.8 ( +67.7)

–50 in MeOH at –62.0 °C –100

200

CH3

300

400 λ (nm)

500

Figure 5.32. CD spectrum of the unstable trans-oleﬁn [CD(+)269.0]-(2S,2 S)-(P, P)-(E)-23d (93%) in MeOH at −62.0◦ C; the sample contained some of the stable cis-oleﬁn (2S,2 S)-(M,M)-(Z)-23c (7%); the ratio was determined by 1 H NMR spectroscopy. (Redrawn from reference 44, with permission.)

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

Δε 233.0 nm

201

in MeOH –25 °C

+100

7

+50 2

1

0

1 2

1, 0.00 sec 2, 2.62 sec 3, 5.09 sec 4, 7.06 sec 5, 10.37 sec 6, 20.50 sec 7, unstable

7

–50

–100 1

7 220

240

260

280

300

Figure 5.33. CD spectral change due to the photoisomerization of the stable trans-oleﬁn (2S,2 S)-(M,M)-(E)-23a into the unstable cis-oleﬁn (2S,2 S)-(P, P)-(Z)-23b on UV irradiation in MeOH at 312 nm at −78◦ C, as monitored by CD at

320

λ (nm)

−25◦ C. (Redrawn from reference 44, with permission.)

visible light (430 nm) at −78◦ C. After 65 sec irradiation, the reaction 23b → 23a was complete. (2) The Thermal Second Motor Rotation Step. The thermal reaction 23b → 23c was also monitored by CD spectroscopy, where a solution of unstable cis-oleﬁn (2S,2 S)-(P, P)-(Z )-23b in MeOH was kept at 14.8◦ C and CD spectra were measured at intervals of 1 h (Figure 5.34). Based on the Arrhenius and Eyring plots, the dynamics data of the thermal molecular motor rotation step 23b → 23c were determined. The data obtained by CD spectroscopy agreed well with those by 1 H NMR spectroscopy.

Δε 13

+100

in MeOH 14.8 °C

1

+50 0 13

–50 –100

1, 2, 3, 4, 5, 6, 7,

2 1

–150

233.8 nm 220

240

260

280 λ (nm)

0.0 h 1.0 h 2.0 h 3.0 h 4.0 h 5.0 h 6.0 h 300

8, 7.0 h 9, 8.0 h 10, 9.0 h 11, 10.0 h 12, 11.0 h 13, stable cis 320

Figure 5.34. CD spectral change due to the thermal isomerization of the unstable cis-oleﬁn (2S,2 S)-(P, P)-(Z)-23b into the stable cis-oleﬁn (2S,2 S)-(M,M)-(Z)-23c at 14.8◦ C. (Redrawn from reference 44, with permission.)

202

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(3) The Photochemical Third Motor Rotation Step. A solution of stable cisoleﬁn (2S,2 S)-(M,M)-(Z )-(−)-23c in MeOH was irradiated with UV light at 330 nm at −78◦ C giving unstable trans-oleﬁn (2S,2 S)-(P, P)-(E )-23d, and the change was monitored by CD at −60◦ C (Figure 5.35). The photochemical reaction was also very fast and reached photoequilibrium after 23.0 s of irradiation. The 23d/23c ratio at the ﬁnal stage was 93:7. The low-temperature CD spectra (−60◦ C) are very useful for studying the photodynamics of extremely unstable motor rotation isomer. (4) The Thermal Fourth Motor Rotation Step. The thermal reaction 23d → 23a was also monitored by CD spectroscopy as follows; CD spectra of the solution 23d/23c (93:7) in MeOH was measured at −19.1◦ C at intervals of 1 h (Figure 5.36). The thermal step 23d → 23a occurred even at subzero temperatures like −19.1◦ C, indicating that the fourth motor rotation is much faster than that of six-membered ring motor described above. This ﬁve-membered ring molecular motor thus rotates very fast as predicted. From the Arrhenius and Eyring plots, the dynamics data of the thermal fourth molecular motor rotation step 23d → 23a were determined. The molecular motor rotation dynamics data are summarized in Table 5.1. (i) The ﬁrst photochemical rotation step 23a → 23b is much faster under the conditions of CD measurement than under the conditions of 1 H NMR measurement, because of the photoirradiation efﬁciency. That is, the speed of the photochemical rotation step essentially depends on the photoirradiation conditions. As the motor can rotate backward (e.g., 23b → 23a), the direction of motor rotation has to be controlled by choosing the optimal wavelength. Thus, the motor rotates forward on UV irradiation at 312 nm, reaching photoequilibrium in 20 s with the ratio 23b/23a = 94 : 6. (ii) The second step of motor rotation 23b → 23c is thermally controlled, and the dynamics data obtained by 1 H NMR and CD methods are similar. (iii) The third photochemical rotation step 23c → 23d was again faster under the conditions of CD measurement, where the motor rotates forward on UV irra-

Δε in MeOH

–60 °C

+100 1 5

+50 0 5

3

–50 –100 1

–150

1, 0.00 sec 2, 0.69 sec 3, 1.86 sec 4, 4.22 sec 5, 13.78 sec

270.0 nm 220

240

260

280 λ (nm)

300

320

Figure 5.35. CD spectral change due to the photoisomerization of the stable cis-oleﬁn (2S,2 S)-(M,M)-(Z)-23c into the unstable trans-oleﬁn (2S,2 S)-(P, P)-(E)-23d on UV irradiation in MeOH at 330 nm at −78◦ C, as monitored by CD at −60◦ C. (Redrawn from reference 44, with permission.)

203

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

Δε in MeOH

–19.1 °C

+50

1

0

1, 2, 3, 4, 5, 6, 7, 8,

–50

8

–100

258.2 nm 220

240

260

280

0.0 h 1.0 h 2.0 h 3.0 h 4.0 h 5.0 h 6.0 h stable trans 300

320

λ (nm)

Figure 5.36. CD spectral change due to the thermal isomerization of the unstable trans-oleﬁn (2S,2 S)-(P, P)-(E)-23d into the stable trans-oleﬁn (2S,2 S)-(M,M)-(E)-23a at −19.1◦ C. (Redrawn from reference 44, with permission.)

TAB L E 5.1. The Molecular Motor Rotation Dynamics Data as Monitored by 1 H NMR and CD Methods Motor Rotation Forward rotationa 23a → 23b, hν, 312 nmb Backward rotationa 23b → 23a, hν, 430 nmb Forward rotationc 23b → 23c, thermal Forward rotationa 23c → 23d, hν, 330 nmb Backward rotationa 23d → 23c, hν, 430 nmb Forward rotationc 23d → 23a, thermal

a

1

H NMR 400 MHz in CD2 Cl2 15 min 23b/23a = 91 : 9 4 min 23b/23a = 0 : 100 Ea = 20.7 H = = 20.1 S = = −6.17 R = 0.999, 0.999 12 min 23d/23c = 96 : 4 60 min 23d/23c = 25 : 75 Ea = 17.1 H = = 16.5 S = = −9.23 R = 0.999, 0.999

CD in MeOH 20.5 s 23b/23a = 94 : 6 90 s 23b/23a = 0 : 100 Ea = 21.4 H = = 20.8 S = = −6.30 R = 0.999, 0.999 23 s 23d/23c = 93 : 7 90 s 23d/23c = 20 : 80 Ea = 16.8 H = = 16.3 S = = −12.6 R = 0.975, 0.974

Time to photoequilibrium. Ratio at photoequilibrium. c Activation energy Ea and activation enthalpy H = in kcal mol−1 unit: activation entropy S = in cal K−1 mol−1 unit; the value of correlation coefﬁcient R for Arrhenius and Eyring plots, respectively. Source: Adapted from reference [44], with permission. b

204

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

diation at 330 nm, reaching the photoequilibrium state in 23 s with the ratio 23d/23c = 93 : 7. (iv) The fourth thermal rotation step 23d → 23a was similarly monitored by 1 H NMR and CD methods, but in this case the CD study was difﬁcult because of the instability of compound 23d. However, it should be noted that both methods gave similar values of kinetic parameters as seen in Table 5.1. The activation energy of the fourth thermal step was much lower than that found for the motor of sixmembered ring type, and hence the new motor rotates much faster than the old one.

5.5.4. Continuous Rotation of the New Light-Powered Chiral Molecular Motor Studied by CD Spectroscopy Continuous rotation experiments were carried out as follows: (i) For the ﬁrst rotation step, a solution of the enantiopure stable trans-oleﬁn 23a in n-pentanol was irradiated with UV light at 312 nm at −78◦ C for 30 s; (ii) for the second rotation step, the solution was heated at 120◦ C for 20 s; (iii) for the third rotation step, the solution was irradiated with UV light at 330 nm at −78◦ C for 30 s; (iv) for the fourth rotation step, the solution was heated at 120◦ C for 10 s. After each operation, the CD spectrum of the solution was measured at −50◦ C, and the CD intensity at 275 nm was plotted. One cycle of operations (namely, 360◦ rotation of motor) took 90 s under these conditions. Figure 5.37 shows the CD plot for 10 cycles of rotation, indicating that this new molecular motor rotates continuously and is durable for such operations.

5.6. ABSOLUTE CONFIGURATION OF CHIRAL C60 -FULLERENE CIS-3 BISADDUCTS DETERMINED BY X-RAY CRYSTALLOGRAPHY AND CD SPECTROSCOPY The fullerene C60 is a symmetrical and achiral molecule. However, an addition reaction at two chiral positions of the C60 skeleton (e.g. a cis-3 addition) makes the π -electron 1

2

3

4

5

6

7

8

9

10 cycles

+60 +40 +20 Δε 0 –20 –40 –60 –80 –100

Figure 5.37. Continuous rotation of the 0

200

400

600

Time (sec)

800

new molecular motor 23 as monitored by CD. (Redrawn from reference 44, with permission.)

205

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

system in fullerene chiral. Synthetic and absolute conﬁgurational studies of various chiral fullerene cis-3 bisadducts have been carried out by many research groups as listed in Chart 5.5 [46]. Compound [CD(–)288]-27 was synthesized as the ﬁrst chiral cis-3 bisadduct by using a chiral tether, and its AC was tentatively assigned as (R, R, f,s A) {≡ (R, R, f C ), old nomenclature} on the basis of MM2/Monte Carlo calculations, because it was assumed that the product formed should be the most energetically stable diastereomer. Therefore, the stereochemistry of the most stable diastereomer calculated was assigned as shown in Chart 5.5 [47]. In a similar manner, the AC of compound [CD(+)281]-28 was determined as shown by assuming that the product formed should have the most energetically stable stereostructure. Thus the diastereomeric structure (S,S,f,s A) was assigned as the most stable one by molecular mechanics calculation [48]. In addition, the CD spectrum of compound (S,S,f,s A)-28 was calculated by the π electron SCF-CI-DV MO method to compare with the observed CD spectrum [49]. By comparison of the data, the AC of bisadduct [CD(+)281]-28 was determined to be (S,S,f,s A). Later, the same AC was assigned by applying the CD exciton chirality method to a related compound to conﬁrm the previous assignment of 28 [50].

O

O

O

H

H

H

O O

O

(R,R,f,sA)-[CD(–)288]-27* energy calculation (1996)

O

B

O

O

B

H

CH3O O

O

H3C CH3 H H

O

O

O

OCH3

O

O

O

EtO

OEt

(S,S,f,sA)-[CD(+)281]-28 energy calculation (1997) CD calculation (1998) Exciton CD (2000)

(S,S,f,sC)-[CD(+)284]-29* energy calculation (1997–8) revised (2003)

H3CO H H OCH3

H3CO H H OCH3

O

(f,sC)-[CD(−)287]-30 comparison of CD (1999)

O O

O

O O

O

O

O O

O

O

O

O

O

EtO

OEt

EtO

OEt

(S,S,f,sA)-[CD(+)281]-31a 1H NMR analysis (2002)

(S,S,f,sC)-[CD(−)281]-31b 1H NMR analysis (2002)

Chart 5.5. Previously reported chiral C60 -fullerene cis-3 adducts together with their CD data and absolute conﬁgurations, which are designated by the new systematic nomenclature (f,s C and f,s A). Our studies described here suggested that the ACs of compounds 27 and 29 should be revised.

206

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The AC of cis-3 bisadduct [CD(+)284]-29 was tentatively assigned as shown in Chart 5.5 by calculation of the stable diastereomeric structures and their energy levels, because the tether part was synthesized from (2S,3S )-butanediol, and the stereostructure (S,S,f,s C ) was calculated to be the most stable isomer [51]. Compound [CD(–)287]-30 has no chirality center in the tether moiety, and therefore its AC was determined to be (f,s C ) by comparison of CD spectrum with that of compound [CD(+)281]-(S,S,f,s A)-28 [52]. That is, the CD spectra of [CD(–)287]-30 and [CD(+)281]-(S,S,f,s A)-28 were opposite to each other. Regarding the ACs of these chiral fullerene derivatives, a serious problem was raised as follows. Among compounds 27–30, their tether moieties are different from each other, but the remaining C60 chiral chromophores with the cis-3 bisadduct pattern are the same or mirror images of each other. As listed in Table 5.2, however, the CD spectrum of (R,R,f,s A)-27 is opposite to that of (S,S,f,s A)-28, although they have the same (f,s A) AC of the fullerene moiety. On the other hand, compounds (S,S,f,s A)-28 and (S,S,f,s C )-29 have the opposite ACs in the fullerene part, but they exhibited similar CD spectra as seen in Table 5.2. These results strongly indicated that some absolute conﬁgurational assignments were wrong. To solve this problem, a different strategy was taken as follows. In the previous syntheses, only one diastereomer of two possible products was isolated and it was assumed to be the most stable diastereomer. This strategy brought some ambiguity in the determination of AC. To overcome these difﬁculties, we selected a tether with a more ﬂexible conformation to yield two possible disatereomeric products in the synthesis. That is, we thought that it was relatively easy to determine the relative stereochemistry by comparing the 1 H NMR data of two diastereomers [53]. We were able to synthesize the two possible diastereomers [CD(+)281]-31a and [CD(–)281]-31b starting from (2S,3S )-(−)-2,3-butanediol, and these diastereomers were separable by HPLC on silica gel (Chart 5.5). Their CD spectra were almost mirror images of one another, reﬂecting the opposite chirality of the π -electron system in the two C60 skeletons. Careful analysis of the 1 H NMR data (i.e., chemical shift and coupling constant) led to the absolute conﬁgurational assignments, (S,S,f,s A)-[CD(+)281]-31a and (S,S,f,s C )-[CD(–)281]-31b [53]; these results thus conﬁrmed the AC of (S,S,f,s A)[CD(+)281]-28 reported previously by the Diederich and our groups [48–50]. Although the new assignment based on the 1 H NMR analyses of two diastereomers was more reliable than the previous ones using only one diastereomer, we wanted to conﬁrm our results by X-ray crystallography.

TAB L E 5.2. Reported CD Data and ACs of Chiral Fullerene cis-3 Bisadducts Compound

27

28

29

30

AC CD λ (nm) ε CD λ (nm) ε References

(R,R,f,s A)a 720.0 +4.69c 288.0 −75.0c 47

(S,S,f,s A) 706.0 −37.0 281.0 +89.0 48–50

(S,S,f,s C )b 737.0 −36.3 284.0 +60.6 51

(f,s C ) 726.0 +26.7 287.0 −73.2 52

a The

revision of AC was later suggested; see the discussion below. was later revised [51]. c Values are 1/1000 of the ε values reported in reference 47, which are too large. b AC

207

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

Although it was known that most fullerene derivatives were obtained as amorphous solids, we attempted to get single crystals of various chiral cis-3 bisadducts, and we ﬁnally succeeded in obtaining single crystals of diethyl ester [CD(+)280]-32, which was previously synthesized by the Diederich group [50]. As seen in Scheme 5.10, bisadduct [CD(+)280]-32 was prepared starting from (2R,3R)-(−)-2,3-butanediol, the moiety of which would be useful as the internal reference of absolute conﬁguration in X-ray crystallography. The product [CD(+)280]-32 was recrystallized from chloroform/hexane (1:1), giving extremely thin, red plate crystals with a thickness of 1–2 μm, which were too thin for conventional X-ray diffractometers [54]. Therefore, the X-ray diffraction experiment was carried out with extremely strong synchrotron radiation at the SPring-8 in Hyogo ˚ space group P 21 (#4), R = 0.180. Although (Japan): X ray, 22.00 keV, λ = 0.5633 A, the ﬁnal R value remained large, the AC of bisadduct [CD(+)280]-32 was unambiguously determined as (f,s A) by using the (2R,3R) absolute conﬁguration of the tether moiety as an internal reference (Figure 5.38). Thus the use of the internal reference method in X-ray crystallographic analysis is very useful for unambiguous determination of AC [54].

H3C H H CH3 HO

H3C H H CH3

O

O

EtO

O

O

H3C H H CH3 O O O O

O OEt

O

OH

(2R,3R)-(–)2,3-butanediol

O OEt EtO

O

(2R,3R)-tether

Scheme 5.10. Synthesis of chiral cis-3 bisadduct (R,R,f,sA)-[CD(+)280]-32

(R,R,f,s A)-[CD(+)280]-32.

a

Figure 5.38. Absolute stereo-structure of the C60 fullerene cis-3 bisadduct (R,R,f,s A)[CD(+)280]-32 (top) and projection along b-axis

o

c Projection along b-axis

(bottom). (Redrawn from reference 54, with permission.) (See insert for color representation of the ﬁgure.)

208

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The CD and UV–Vis spectra of bisadduct (R,R,f,s A)-32 are shown in Figure 5.39, where the CD spectrum shows an intense positive Cotton effect (ε +93.7) at 280 nm. In addition, compound (R,R,f,s A)-32 exhibited an anomalously large optical rotation [α]28 D +3950 (c = 0.0214, CHCl3 ). Therefore, its AC was designated as (R,R,f,s A)-[CD(+)280]-(+)-32 [54]. Based on these X-ray and CD results, the ACs of C60 fullerene cis-3 bisadducts were rationalized as follows: cis-3 derivatives showing a positive CD band around 280 nm should have the (f,s A) absolute conﬁguration, while cis-3 compounds showing a negative CD around 280 nm should have the (f,s C ) absolute conﬁguration. Therefore, our previous assignments of (S,S,f,s A)-[CD(+)281]-31a and (S,S,f,s C )-[CD(−)281]-31b were corroborated by this study. The assignment of (S,S,f,s A)-[CD(+)281]-28 by Diederich and co-workers was also conﬁrmed. However, it was concluded that the AC of [CD(−)288]27 should be revised to be (R,R,f,s C ) [54]. In a similar manner, revision of the AC was suggested so that cis-3 bisadduct [CD(+)284]-29 should have the (S,S,f,s A) absolute conﬁguration. We found unique phenomena in the CD and UV–Vis spectra as follows. Compound (R,R,f,s A)-[CD(+)280]-(+)-32 exhibits a very weak absorption band at 706.2 nm (ε = 338); this band is due to the forbidden π –π ∗ transition as a result of the small ε value. In the corresponding region, the CD spectrum shows an intense negative Cotton effect at 701.5 nm (ε −36.3). The curve of g value (g = ε/ε) was calculated as illustrated in Figure 5.39, where the maximum value found was g = −0.110 at 697.0 nm [54]. This g value is much larger than that of n–π ∗ forbidden transition of chiral ketones and that of

+0.10

+100

280.0 (+93.7) CD +0.05

+50

701.5 (–36.3) 0 Δe/e

Δe 0

–50

–100

–150

ε × 10–4

10

5

220

obsd CD g-value 701.5 (–36.3) 639.5 (–10.0) 697.0 (–0.110) 593.0 (+5.6) 486.5 (+21.5) g-value 424.5 (–20.1) 697.0 (–0.110) 412.0 (+3.0) 398.0 (–13.3) 635.5 (–0.0191) 378.5 (–32.4) 593.0 (+0.00752) 339.0 (+75.8) 490.0 (+0.0130) 313.0 (+37.5) 425.5 (–0.00717) 280.0 (+93.7) 380.5 (–0.00425) 263.0 (–151.2) 343.0 (–0.00298) 232.0 (–47.9) 706.2 (338) obsd UV UV 706.2 (338) 643.6 (517) × 100 316.5 (37100) 254.5 (106600) 300

400

500 λ (nm)

600

700

–0.05

–0.1

–0.15

Figure 5.39. CD and UV–Vis spectra and g-value curve of cis-3 bisadduct (R,R,f,s A)[CD(+)280]-(+)-32 in ClCH2 CH2 Cl: CD and

800

UV–Vis, top curve; g value (g = ε/ε), bottom curve. (Redrawn from reference 54, with permission.)

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

209

the π –π ∗ allowed transition of other compounds. For example, the n –π ∗ transition of (R)-(+)-3-methylcyclohexanone showed g = +0.03 at 298 nm; π –π ∗ , (+)-hexahelicene, g = +0.007 at 325 nm [55]; π –π ∗ , exciton coupling CD of (2S,3S )-butanediyl bis(4bromobenzoate), g = +0.00035 at 252 nm. A similar large g value was also observed for bisadduct (S,S,f,s A)-[CD(+)281]-31a: g = −0.126 at 696 nm [54]. Why does this transition of chiral cis-3 fullerene make such a large g value? The phenomenon could be interpreted as follows: The π –π ∗ transition is electronically forbidden as discussed above, but is magnetically allowed, because molecular orbitals (MOs) involved are spherical, reﬂecting the spherical shape of fullerene skeleton. Therefore, the transition contains a lot of angular momenta, thus yielding a large magnetic moment and generating an intense CD Cotton effect. The g-value thus becomes large in this transition [54].

5.7. ABSOLUTE STEREOCHEMISTRY AND CD SPECTRA OF AN ALLENO-ACETYLENIC MACROCYCLE The alleno-acetylenic tetrameric macrocycle (P, P, P, P)-(−)-33 is a unique chiral compound devoid of any chirality center, in which each allene moiety takes a P helicity, and hence the compound takes a D4 symmetric structure as illustrated in Figure 5.40 [56, 57]. Chiral macrocycle (P, P, P, P)-(−)-33 was ﬁrst synthesized by Diederich and coworkers starting from tert-alcohol 34 as shown in Scheme 5.11 [56], where racemic alcohol (±)-34 was resolved as camphanate esters. The absolute conﬁguration of alcohol (R)-(−)-34 was determined by X-ray crystallographic analysis of its camphanate ester, where the camphanate group was used as an internal reference of the absolute conﬁguration [58]. The other enantiomer (S )-34 was converted to chiral allene derivative (P )-(+)-35 via a rearrangement reaction. The obtained chiral alleno-acetylene (P )-(+)35 was then dimerized, followed by deprotection, to yield a chiral dimer (P, P)-(+)-36, [α]D 20 +506 (c 1, CHCl3 ), the absolute conﬁguration of which was determined by X-ray crystallographic analysis of bis[Si(iso-Pr)3 ] derivative 37 of the other (M,M)-enantiomer series, where the anomalous dispersion effect of silicon atoms was used [59]. The further dimerization–cyclization of bis(acetylene) (P, P)-(+)-36 furnished the target chiral macrocycle (P, P, P, P)-(−)-33, [α]D 20 −770 (c 1, CHCl3 ), the relative crown conﬁguration of which was determined as shown by X-ray analysis of racemate (±)-33 (Scheme 5.11) [57]. The absolute stereochemistry of macrocycle (P, P, P, P)-(−)-33 was thus unambiguously determined by X-ray crystallography.

Figure 5.40. A novel unique chiral alleno-actylenic (P,P,P,P)-(–)

macrocycle (P, P, P, P)-(−)-33 and its crown stereostructure. (Redrawn from reference 56, with permission.)

210

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

HO

OH

H H (S)-34

(P)-(+)-35

abs. config. by X-ray of camphanate ester of (R)-34

R

R

(P,P)-(+)-36: R = H abs. config. by X-ray of Si(iso-Pr)3 derivative of (M,M)-series

(P,P,P,P)-(–)-33 relat. config. by X-ray of (+)-33 –

Scheme 5.11. Synthesis of macrocycle (P,P,P,P)-(−)-33 and absolute conﬁgurational assignment by X-ray crystallographic analyses of related derivatives.

It is interesting that macrocycle (P, P, P, P)-(−)-33 shows extremely intense CD Cotton effects as illustrated in Figure 5.41; for example, the positive CD band at 253 nm has an intensity of ε = +790, which is about 100 times larger than that of monomer (P )-(+)-35 and ≈ 8 times larger than that of dimer (P, P)-(+)-36. The CD spectrum of macrocycle (P, P, P, P)-(−)-33 is thus unique and a good example for CD studies.

+800 CD

(P,P,P,P)-(–) e × 10–5

+400 Δe 0 –400

3

–800 2

UV

1

Figure 5.41. CD and UV–Vis spectra of enantiopure macrocycle (P, P, P, P)-(−)-33 (thick

200

300 λ (nm)

0 400

line) and CD spectra of (M, M, M, M)-(+)-33 (thin line) in hexane. (Redrawn from reference 56, with permission.)

211

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

The UV spectrum of (P,P,P,P)-(−)-33 shows three maxima of medium intensity around 275–375 nm, which were assigned to the vibronic structure of conjugated acetylene groups because of the vibronic interval of ≈ 2100 cm−1 (Figure 5.41). The three UV bands around 275–375 nm were thus assigned due to a single π –π ∗ transition of medium intensity, but not due to three different π –π ∗ transitions. On the other hand, a very intense UV band is seen at 245 nm (ε 210,000), which was assigned to an allowed π –π ∗ transition [56]. In the region of 270–350 nm, the CD spectrum of (P, P, P, P)-(−)-33 shows three negative extrema corresponding to the UV vibronic structure, while a very intense positive Cotton effect is observed in the region of 230–270 nm (Figure 5.41). The g values (g = ε/ε) of these Cotton effects were calculated to give interesting data as follows: g = approximately −0.007 to −0.008 for the Cotton effects in the region of 270–350 nm; g = approximately +0.004 for the intense Cotton effect at 253 nm. These data clearly indicate that the CD bands in the region of 270–350 nm have a large magnetic transition dipole moment contribution, and the transition is magnetically allowed. On the other hand, the CD band at 253 nm has a large electric transition dipole moment contribution, indicating the character of electric allowed transition [56]. To give insight into the CD mechanism of macrocycle (P, P, P, P)-(−)-33 and also to determine its absolute stereochemistry in a theoretical manner, the CD and UV spectra were calculated by the ZINDO MO method. As illustrated in Figure 5.42, the basic pattern of the CD curve including the sign, amplitude, and position of Cotton effects, but not the vibronic structures around 270–350 nm, was reproduced well by the ZINDO calculation. The (P,P,P,P ) absolute stereochemistry of macrocycle (−)-33 was thus determined by the MO calculation of the ZINDO level, and the theoretical absolute conﬁgurational assignment was consistent with the experimental one [56]. The calculation results show that there are three major electronic transitions as shown in Figure 5.42. The negative CD bands at 270–350 nm are due to the S1 transition, while the positive CD band at 253 nm is due to two degenerate S2 and S3 transitions. The S1 transition is very unique, since its large magnetic transition dipole moment (MTDM) is perpendicular to the ring plane of the macrocycle and is oriented upward, as seen in the

MTDM

+800 +600 +400

R × 1037 cgs unit

Δe

S2,S3 ETDM

+2.0

transition S1

+200

+1.0

0 –200

0

Figure 5.42. Comparison of the observed CD

–1.0

spectrum of (−)-33 (thick line) with the CD curve of (P, P, P, P)-33 (thin line) calculated by the ZINDO MO method, where the graphic inserted

obsd CD –400 S1

–600 200

–2.0

calcd CD

300 λ (nm)

–3.0 400

above shows the electric transition dipole moment (ETDM) and magnetic transition dipole moment (MTDM) of the transition S1 in the geometry of the compound. (Redrawn from reference 56, with permission.)

212

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

graphic inserted in Figure 5.42. This large MTDM is generated by the cyclic movement of an electron along the macrocycle ring during the electronic transition. At the same time, the S1 transition generates a small electric transition dipole moment (ETDM), which is also perpendicular to the ring plane, but is oriented downward. Therefore, MTDM and ETDM are antiparallel to each other and take nonzero values, generating an intense negative CD band at the S1 transition. The ETDM and MTDM of the S2 transition are placed in the macrocycle plane and oriented parallel to each other generating a positive CD band (Figure 5.42). Both ETDM and MTDM take nonzero values, which is the main reason for the intense CD at the S2 transition. The S3 transition has a similar character to that of the S2 transition, generating an intense positive CD. The ETDMs of S2 and S3 transitions are perpendicular to each other reﬂecting the D4 symmetric structure of macrocycle 33. This is the reason that the S2 and S3 transitions are degenerate. The mechanism of intense CD of macrocycle (P, P, P, P)-(−)-33 was thus clariﬁed even by the MO calculation of ZINDO level [56]. A similar result was obtained by the more advanced MO method, that is, the Coulombattenuated hybrid exchange-correlation functional (CAM-B3LYP), supporting the CD mechanism discussed above [57].

5.8. CONCLUSION As discussed above, the ACs of various chiral compounds with extended π -electron system, including natural products and synthetic compounds, have been theoretically determined on the basis of the calculation of their CD spectra by the π -electron SCF-CIDV MO method. The ACs theoretically assigned were proved experimentally by X-ray crystallographic analyses using an internal reference of AC, and/or by the total synthesis of natural enantiomers. The combination of X-ray crystallography and CD spectroscopy is thus very reliable for determining ACs. In addition, in the case of the light-powered chiral molecular motors, CD spectroscopy is useful for clarifying the motor rotation mechanism and dynamics. Therefore, the CD methodology discussed here is a promising and powerful tool for determining the ACs of various chiral compounds with an extended and twisted π -electron chromophore.

ACKNOWLEDGMENTS The authors sincerely thank the co-workers of the studies described here for their contributions, whose names are listed in references, and Dr. George A. Ellestad, Department of Chemistry, Columbia University, for his valuable suggestions.

REFERENCES 1. N. Harada, K. Nakanishi, Circular Dichroic Spectroscopy—Exciton Coupling in Organic Stereochemistry, University Science Books, Mill Valley, CA, and Oxford University Press, Oxford, 1983. 2. N. Harada, Circular dichroism spectroscopy and absolute stereochemistry of biologically active compounds, in The Biology—Chemistry Interface, R. Cooper and J. K. Snyder, eds., Marcel Dekker, Inc., New York, 1999, Chapter 6, pp. 139–190.

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

3. N. Harada, Circular dichroism of twisted π -electron systems: theoretical determination of the absolute stereochemistry of natural products and chiral synthetic organic compounds, in Circular Dichroism: Principles and Applications, K. Nakanishi, N. Berova, and R. Woody, eds., John Wiley & Sons, New York, 2000, pp. 431–457. 4. N. Berova, N. Harada, K. Nakanishi, Electronic Spectroscopy: Exciton coupling, theory and applications, in Encyclopedia of Spectroscopy & Spectrometry, J. Lindon, G. Tranter, and J. Holmes, eds., Academic Press, London, 2000, pp. 470–488. 5. N. Berova, G. A. Ellestad, N. Harada, Modern methods in natural products chemistry: characterization by circular dichroism spectroscopy, in Comprehensive Natural Products Chemistry II , L. Mander and H. W. Liu, eds., Elsevier, Oxford, 2010, Chapter 9, pp. 91–146. 6. (a) C. Kemp, S. F. Mason, Tetrahedron 1966, 22 , 629–635. (b) A. Brown, C. Kemp, S. F. Mason, J. Chem. Soc. A 1971, 751–755. c) W. S. Brickell, A. Brown, C. Kemp, S. F. Mason, J. Chem. Soc. A 1971, 756–760. 7. C. Rossini, C. Bertucci, P. Salvadori, M. Zandomeneghi, J. Am. Chem. Soc. 1985, 107 , 17–19. 8. J. Roschster, U. Berg, M. Pierrot, J. Sandstrom, J. Am. Chem. Soc. 1987, 109 , 492–507. 9. N. Harada, Chiral auxiliaries powerful for both enantiomer resolution and determination of absolute conﬁguration by X-ray crystallography, in Topics in Stereochemistry, Vol. 25, S. E. Denmark, J. S. Siegel, eds., John Wiley & Sons, Hoboken, NJ, 2006, Chapter 6, pp. 177–203. 10. N. Harada, Powerful chiral molecular tools for preparation of enantiopure alcohols and simultaneous determination of their absolute conﬁgurations by X-ray crystallography and/or 1 H NMR anisotropy methods, in Chirality in Drug Research, Ed. E. Francotte, W. Lindner, WILEYVCH, Weinheim, Germany, 2006, Chapter 9, pp. 283–321. 11. N. Harada, M. Watanabe, S. Kuwahara, Novel chiral derivatizing agents powerful for enantioresolution and determination of absolute stereochemistry by X-ray crystallographic and 1 H NMR anisotropy methods, in Chiral Analysis, K. W. Busch and M. A. Busch, eds., Elsevier Science Publishers, Amsterdam, 2006, Chapter 18, pp. 661–691. 12. N. Harada, Chirality 2008, 20 , 691–723. 13. P. J. Stephens, N. Harada, Chirality 2010, 22 , 229–233. 14. R. Takeda, K. Katoh, J. Am. Chem. Soc. 1983, 105 , 4056–4058. 15. N. Harada, J. Kohori, H. Uda, K. Nakanishi, R. Takeda, J. Am. Chem. Soc. 1985, 107 , 423–428. 16. N. Harada, J. Kohori, H. Uda, K. Toriumi, J. Org. Chem. 1989, 54 , 1820–1826. 17. N. Harada, T. Sugioka, H. Uda, T. Kuriki, Synthesis 1990, 53–56. 18. N. Harada, H. Uda, T. Nozoe, Y. Okamoto, H. Wakabayashi, S. Ishikawa, J. Am. Chem. Soc. 1987, 109 , 1661–1665. 19. Y. Okamoto, S. Honda, H. Yuki, H. Nakamura, Y. Iitaka, T. Nozoe, Chemistry Lett. 1984, 1149–1152. 20. D. M. Roll, P. J. Scheuer, G. K. Matsumoto, J. Clardy, J. Am. Chem. Soc. 1983, 105 , 6177–6178. 21. M. Kobayashi, N. Shimizu, Y. Kyogoku, I. Kitagawa, Chem. Pharm. Bull . 1985, 33 , 1305–1308. 22. H. Nakamura, J. Kobayashi, M. Kobayashi, Y. Ohizumi, Y. Hirata, Chem. Lett. 1985, 713–716. 23. F. J. Schmitz, S. J. Bloor, J. Org. Chem. 1988, 53 , 3922–3925. 24. (a) M. Kobayashi, N. Shimizu, I. Kitagawa, Y. Kyogoku, N. Harada, H. Uda, Tetrahedron Lett. 1985, 26 , 3833–3836. (b) N. Harada, H. Uda, M. Kobayashi, N. Shimizu, I. Kitagawa, J. Am. Chem. Soc. 1989, 111 , 5668–5674. 25. N. L. Allinger, J. Am. Chem. Soc. 1977, 99 , 8127–8134; N. L. Allinger, Y. H. Yuh, QCPE Program No. 395, 1980. 26. N. Harada, T. Sugioka, Y. Ando, H. Uda, T. Kuriki, J. Am. Chem. Soc. 1988, 110 , 8483–8487.

213

214

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

27. N. Harada, T. Sugioka, H. Uda, T. Kuriki, J. Org. Chem. 1990, 55 , 3158–3163. 28. N. Harada, T. Sugioka, H. Uda, T. Kuriki, M. Kobayashi, I. Kitagawa, J. Org. Chem. 1994, 59 , 6606–6613. 29. N. Harada, T. Sugioka, T. Soutome, N. Hiyoshi, H. Uda, T. Kuriki, Tetrahedron Asym. 1995, 6 , 375–376. 30. N. Harada, H. Ono, H. Uda, M. Parveen, N. U-P. Khan, B. Achari, P. K. Dutta, J . Am. Chem. Soc. 1992, 114 , 7687–7692. 31. F.-J. Zhang, G.-Q. Lin, Q.-C. Huang, J. Org. Chem. 1995, 60 , 6427–6430; Additions and corrections in J. Org. Chem. 1996, 61 , 5700. Later, the authors changed to the (aR) absolute conﬁguration; G.-Q. Lin, M. Zhong, Tetrahedron Lett. 1997, 38 , 1087–1990. 32. N. Harada, H.-Y. Li, T. Nehira, M. Hagiwara, Enantiomer 1997, 2 , 353–358. 33. H.-Y. Li, T. Nehira, M. Hagiwara, N. Harada, J. Org. Chem. 1997, 62 , 7222–7227. 34. B. Feringa, H. Wynberg, J. Am. Chem. Soc. 1977, 99 , 602–603. 35. N. Harada, A. Saito, N. Koumura, H. Uda, B. de Lange, W. F. Jager, H. Wynberg, B. L. Feringa, J. Am. Chem. Soc. 1997, 119 , 7241–7248. 36. N. Harada, A. Saito, N. Koumura, D. C. Roe, W. F. Jager, R. W. J. Zijlstra, B. de Lange, B. L. Feringa, J . Am. Chem. Soc. 1997, 119 , 7249–7255. 37. R. W. J. Zijlstra, W. F. Jager, B. de Lange, P. T. van Duijnen, B. L. Feringa, H. Goto, A. Saito, N. Koumura, N. Harada, J. Org. Chem. 1999, 64 , 1667–1674. 38. N. Harada, N. Koumura, and B. L. Feringa, J. Am. Chem. Soc. 1997, 119 , 7256–7264. 39. N. Koumura, N. Harada, Enantiomer 1998, 3 , 251–253. 40. N. Koumura, N. Harada, Chem Lett. 1998, 1151–1152. 41. N. Koumura, Ph.D. Thesis, Tohoku University, March 1999. 42. N. Koumura, R. W. J. Zijlstra, R. A. van Delden, N. Harada, B. L. Feringa, Nature 1999, 401 , 152–155. 43. T. Fujita, S. Kuwahara, N. Harada, Eur. J. Org. Chem. 2005, 4533–4543. 44. S. Kuwahara, T. Fujita, N. Harada, Eur. J. Org. Chem. 2005, 4544–4556. 45. M. K. J. ter Wiel, R. A. van delden, A. Meetsma, B. L. Feringa, J. Am. Chem. Soc. 2003, 125 , 15076–15086. 46. For the nomenclature of regioisomeric fullerene derivatives, see: (a) A. Hirsch, I. Lamparth, H. R. Karfunkel, Angew. Chem. 1994, 106 , 453–455; Angew. Chem. Int. Ed. Engl . 1994, 33 , 437–438; (b) for the nomenclature of chiral fullerene derivatives, see: C. Thilgen, A. Herrmann, F. Diederich, Helv. Chim. Acta 1997, 80 , 183–199; (c) for a new and systematic nomenclature for fullerenes (IUPAC Recommendations 2002), see W. H. Powell, F. Cozzi, G. P. Moss, C. Thilgen, R. J.-R. Hwu, A. Yerin, Pure Appl. Chem. 2002, 74 , 629–695. 47. E. Nakamura, H. Isobe, H. Tokuyama, M. Sawamura, Chem. Commun. 1996, 1747–1748. 48. J.-F. Nierengarten, T. Habicher, R. Kessinger, F. Cardullo, F. Diederich, V. Gramlich, J.-P. Gisselbrecht, C. Boudon, M. Gross, Helv. Chim. Acta 1997, 80 , 2238–2276. 49. H. Goto, N. Harada, J. Crassous, F. Diederich, J. Chem. Soc. Perkin Trans. 2 1998, 1719–1723. 50. R. Kessigner, C. Thilgen, T. Mordasini, F. Diederich, Helv. Chim. Acta 2000, 83 , 3069–3096. 51. (a) M. Taki, S. Sugita, Y. Nakamura, E. Kawashima, E. Yashima, Y. Okamoto, J. Nishimura, J. Am. Chem. Soc. 1997, 119 , 926–932; (b) N. Taki, Y. Nakamura, H. Uehara, M. Sato, J. Nishimura, Enantiomer 1998, 3 , 231–239; (c) revised assignment: Y. Nakamura, K. O-kawa, J. Nishimura, Bull. Chem. Soc. Jpn. 2003, 76 , 865–882. 52. T. Ishi-i, K. Nakashima, S. Shinkai, A. Ikeda, J. Org. Chem. 1999, 64 , 984–990. 53. K. Yoshida, S. Osawa, K. Monde, M. Watanabe, N. Harada, Enantiomer 2002, 7 , 23–32. 54. S. Kuwahara, K. Obata, K. Yoshida, T. Matsumoto, N. Harada, N. Yasuda, Y. Ozawa, K. Toriumi, Angew. Chem. Int. Ed ., 2005, 44 , 2262–2265.

CD SPECTRA OF CHIRAL EXTENDED π-ELECTRON COMPOUNDS

55. S. F. Mason, Molecular Optical Activity and the Chiral Discrimination, Cambridge University Press, Cambridge, 1982. 56. J. L. Alonso-G´omez, P. Rivera-Fuentes, N. Harada, N. Berova, F. Diederich, Angew. Chem. Int. Ed . 2009, 48 , 5545–5548. 57. P. Rivera-Fuentes, J. L. Alonso-G´omez, A. G. Petrovic, P. Seiler, F. Santoro, N. Harada, N. Berova, H. S. Rzepa, F. Diederich, Chem. Eur. J . 2010, 16 , 9796–9807. 58. M. K. J. ter Wiel, S. Odermatt, P. Schanen, P. Seiler, F. Diederich, Eur. J. Org. Chem. 2007, 3449–3462. 59. J. L. Alonso-G´omez, P. Schanen, P. Rivera-Fuentes, P. Seiler, F. Diederich, Chem. Eur. J . 2008, 14 , 10564–10568.

215

6 ASSIGNMENT OF THE ABSOLUTE CONFIGURATIONS OF NATURAL PRODUCTS BY MEANS OF SOLID-STATE ELECTRONIC CIRCULAR DICHROISM AND QUANTUM MECHANICAL CALCULATIONS ´ and Karsten Krohn Gennaro Pescitelli, Tibor Kurtan,

6.1. INTRODUCTION Natural products represent a rich source of therapeutically useful compounds. About 70% of the drugs marketed in the 1982–2007 period was more or less directly derived from natural products, especially in the ﬁeld of anticancer drugs [1]. At the same time, the portion of chiral drugs patented as a single enantiomer has increased sharply in recent years [2], mostly as a consequence of the regulatory prescriptions concerning the stereochemical characterization of new drugs [3]. Isolation and identiﬁcation of new chemical compounds from natural sources, including elucidation of their absolute stereochemistry, has therefore important consequences in many disciplines. It is hard to believe that a rather fundamental aspect of chemistry—that is, the assignment of absolute conﬁguration (AC)—was established only in 1951 with the Xray diffraction experiment of Bijvoet et al. [4] on NaRb (+)-tartrate. The observation of anomalous dispersion has remained unsurpassed as the most reliable means for assigning absolute conﬁgurations, though with its well-known limitations. In the last decade, a valid alternative has been offered by quantum mechanical (QM) calculations of chiroptical properties such as electronic circular dichroism (ECD), vibrational CD (VCD), and Raman optical activity (ROA). Thanks to the development of computer technology, QM calculations have become accessible at a higher level of sophistication and their reliability has increased signiﬁcantly. The computational approach for assigning ACs is based on the comparison of experimental and calculated ECD, VCD, or ROA spectra and has certain advantages over X-ray structure elucidation, including its practicability and wider scope. Its main limitations are the appearance of signiﬁcant CD or ROA bands and (at some extent) the molecular size. As far as electronic CD (ECD) is concerned, the Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

217

218

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

compound must show a signiﬁcant ECD spectrum in the commonly accessible spectral range (above 185 nm). This is, however, not uncommon with natural products, most of which contain at least a conjugated moiety leading to distinct ECD spectra. Every molecular calculation is necessarily size-dependent, and demanding computational methods may be restricted to small systems. However, ECD spectra of increasingly larger and complex molecular assemblies have been treated by high-level methods in recent years [5]. The most serious drawback of the computational approach is related to the reliability of the input structure, an issue that will be extensively discussed in the following sections because it represents the main reason for the development of the solid-state ECD/TDDFT (time-dependent density functional theory) approach described in the present chapter.

6.2. ASSIGNING ABSOLUTE CONFIGURATIONS THROUGH ECD CALCULATIONS 6.2.1. Conformation and Conﬁguration Whenever a molecular property is predicted by theory and used for comparison with the experiment, it is crucial to employ a correct input structure. “Correct” means it must represent as faithfully as possible the true structure (or structures) responsible for the observed property. In particular, chiroptical data such as ECD, VCD, and ROA spectra are extremely sensitive to the overall molecular geometry in terms of both conformation and absolute conﬁguration. Conﬁgurational and conformational elements are often strictly intertwined, and chiroptical approaches usually determine only one of them in the knowledge of the other [6, 7]. There are spectacular examples in the literature where two slightly different conformations of a compound (with ﬁxed AC) led to almost mirrorimage computed ECD spectra: In other words, regarding their chiroptical parameters, two conformers behaved as two enantiomers [8]. Generally speaking, any solution CD spectrum amounts to the sum of contributions from all populated conformations, and the set of input structures to be considered in the calculation must be representative of the whole conformational ensemble [6]. As a consequence, when ECD calculations are applied to deduce absolute conﬁgurations, they must rely on an independently established conformational picture, which is gained through the use of other spectroscopic and/or theoretical means. Modern NMR techniques play a major role in deducing solution conformations, in particular when they provide data sensitive to the three-dimensional structure such as NOE effects and scalar J -couplings [9, 10]. In challenging cases represented by very ﬂexible compounds, assistance by molecular modelling may be indispensable. A conformational analysis is normally started with a rapid computational procedure, based on Monte Carlo or molecular dynamics approaches at a low level of theory, such as molecular mechanics (MM) [11]. These conformational search routines provide a set of structures that are further optimized at a higher level of theory, usually density functional theory (DFT) or other ab initio methods. Geometry optimizations provide structures with respective (internal) energies, thereafter used to estimate the population of each calculated conformer. Calculated geometries must be checked against NMR data by considering H–H distances versus observable NOEs, as well as H/H and C/H dihedrals versus measured J -couplings [10]; when necessary, 13 C chemical shifts may be calculated and compared with the experimental set [9]. After a reliable set of input structures has been generated with an initial arbitrary AC, ECD calculations must be run on all minima within a certain energy threshold (2–3

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

kcal/mol)—that is, with signiﬁcant population (say, >3–5%) at the working temperature (normally, 298 or 300 K). Thereafter, a weighted average of all computed ECD spectra is estimated according to the Boltzmann weights at the working temperature. Reliable populations can be obtained by single-point calculations at a higher calculation level than that used for geometry optimizations, for example, using larger basis sets. Furthermore, internal energies should be corrected with zero-point vibrational and entropy terms to afford true free energies. Finally, the weighted average ECD is compared with the experimental ECD spectrum: If the agreement is good with either the calculated ECD (for the initially assumed AC) or its mirror image (i.e., for the opposite AC), the conﬁguration may be assigned. The outlined procedure, ﬁrst developed by Bringmann and co-workers [12], has been used frequently in recent years to assign the AC of natural products, in particular by employing QM semiempirical [7, 13] and TDDFT calculations [14, 15]. Several approximations are often used, the most drastic of which are the neglect of solvent and vibronic effects on geometries and calculated ECD. Treatment of solute–solvent interactions, in both geometry optimization and CD calculation steps, is possible by adopting a suitable solvent model (which increases the computational time) [16]. Inclusion of vibronic effects is a rather complicated process and remains still restricted to some models or small molecules [17]. Provided that the calculation method employed for ECD calculations is efﬁcient, the crucial point of the above procedure lies in the generation of the input structure(s). In fact, the conformational analysis may be both computationally demanding and prone to inaccuracy. The major sources of error lie in the prediction of relative energies, as well as in the possible overlooking of one or more signiﬁcant conformers. Finally, since the CD calculation must be run on each calculated conformer, ﬂexible molecules may represent very difﬁcult cases to handle.

6.2.2. The Origin of the Solid-State ECD/Computational Approach X-ray analysis of crystalline natural products offers a twofold advantage to surmount the difﬁculty just described in the generation of the input structure for ECD calculations. First, in the crystals, the molecular conformation is ﬁxed and univocal (unless polymorphs occur); second, its structure can be determined with high accuracy by diffraction experiments. In the course of a screening for novel bioactive compounds from natural sources, Krohn and co-workers investigated a Phomopsis sp. and isolated (+)-phomoxanthone A (1, Figure 6.1) [18]. The relative conﬁguration of the chirality centers and axial chirality element was established by X-ray single-crystal analysis (Figure 6.1), while the ECD spectrum served for the assignment of the absolute conﬁguration. Atropisomeric biaryl compounds like 1 show intense ECD spectra dominated by the exciton coupling (see Section 6.2.3) between the two aromatic chromophores, and the exciton-coupled spectra are easily correlated with the absolute stereochemistry [19]. Following the computational procedure outlined in the previous section, the authors ﬁrst performed a conformational analysis with MMFF and AM1 methods [20], which afforded eight low-energy conformers within 11 kJ/mol with arbitrarily assumed (aS , 5R, 6R, 10aR, 5 R, 6 R, 10a R)-(+)-1 absolute conﬁguration. They were used as input for ECD calculations with a semiempirical method using the BDZDO/MCDSPD program package [21]. The calculated spectra were then weighted with their respective Boltzmann populations and summed to obtain a weighted-average ECD spectrum, which reproduced the experimental solution spectrum. Surprisingly, however, when the X-ray geometry was used as input for

219

220

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

01

02

OH

O

03

OH

C7 C6

OAc OAc H3C

O 4a

4

4’

C15

C2 C1

C33

C18

013 C32

C9

4a’ O

AcO

C5 C4

C8

07

C3

CH3

C16 C10

OAc

C19

C14 06 C11 04

C17

C12

C20

C31

016

C21

C34

C38

05

O

OH

C22

C24

C30 014

C26

C36

C29

C35 C28

C13

OH

C37

012

08

015

C27

C23

C25

Phomoxanthone A (1)

011 010 09

80 20

40

0 Δε (M–1cm–1)

Calculated CD spectra Weighted average using AM1-optimized structures On X-ray geometry

60

20 –20 Experimental CD spectra Solid-state (KCl disc)

–40

Solution (4:1 CH3OH/CH2CI2) Solution (CH2CI2)

Figure 6.1.

260

300

340

–20 –40

–60 220

0

380

–60

220

260

300

λ (nm)

λ (nm)

(a)

(b)

340

380

(a) Experimental ECD spectra of (aS, 5R, 6R, 10aR, 5 R, 6 R,10a R)-(+)-phomoxanthone

(1) in solution (CH2 Cl2 and methanol/CH2 Cl2 4:1) and in the solid state (KCl disc). (b) Calculated Boltzmann-weighted average ECD spectrum over eight low-energy AM1 structures, as well as calculated ECD spectrum using the X-ray geometry (shown on the top). All calculations run with the BDZDO method.

BDZDO calculations, the resultant spectrum showed a better agreement with the experimental one. Clearly, the comparison between calculated and experimental properties is more justiﬁed when they refer to the same geometry. Therefore, the ECD spectrum was also recorded in the solid state using the KCl pellet technique (described in Section 6.3.2) and, as expected, the match between this latter spectrum and that calculated using the X-ray geometry was improved. In this way the idea was born to use X-ray coordinates as input data for ECD calculations for future characterization of chiral nonracemic natural products. It was immediately clear that the major advantage of the method would lie in skipping the conformational analysis step and its related uncertainties, with consequent time-saving and increased reliability.

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

6.2.3. The Choice of the Computational Method After its ﬁrst application to phomoxanthone A, the major improvement of the approach based on solid-state ECD has consisted in a different choice of the standard computational method employed, which was switched from the semiempirical BDZDO method to TDDFT. Some speciﬁc applications to large molecules and/or molecules containing multiple aromatic chromophores have been dealt with alternative computational methods such as ZINDO or even the semiclassical coupled oscillator DeVoe’s one. In this section, we will brieﬂy discuss the choice of the computational method for solid-state ECD calculations [22]. The reader is also referred to the many chapters of Volume 1 especially devoted to the topic of simulations of ECD spectra [23]. ECD calculations with high-level quantum mechanical (QM) methods are nowadays fully practicable for moderately large organic molecules and metal complexes [24–27]. Reliable ECD predictions require theoretical methods that take electron correlation into account and use of large basis sets. Among the many possible ab initio methods employed for excited states calculations, time-dependent density functional theory (TDDFT) has emerged in recent years as one of those leading to the best accuracy/cost compromise [26, 28] (it must be noted that many DFT functionals contain adjusted parameters and are not strictly “ab initio”). Although DFT functionals are usually designed to reproduce thermochemical data [11, 29], hybrid functionals such as B3LYP, BH&HLYP, and PBE0 [30, 31] predict with high accuracy transition energies and rotational strengths. Several commonly employed DFT functionals have well-recognized drawbacks, especially a poor description of some loosely localized states, such as charge-transfer, diffuse, and Rydberg states, which may be alleviated using some of the new long-range functionals [32–34]. TDDFT arises from a perturbative approach to DFT, and therefore it is intrinsically more accurate in predicting low-lying excited states [35]. Apart from these issues, the scope of TDDFT calculations is practically unlimited. Thus, when we had to choose a general and reliable method for the calculation of solid-state ECD spectra, the choice of TDDFT was almost obvious, and now we refer to this approach as the solid-state ECD/TDDFT approach [22]. Semiempirical quantum mechanical methods rely on strong simpliﬁcations, which include ignoring core electrons and neglecting differential overlap (NDO) [36]. They are much faster than ab initio methods and may be helpful when dealing with complex molecules and simple supramolecular aggregates. Various approximation schemes have been purposely developed for spectroscopic simulations, such as CNDO/S (complete NDO) and ZINDO/S (Zerner’s intermediate NDO). They have been parameterized in order to describe aromatic and heteroaromatic chromophores and, for ZINDO, some transition metals. Both methods reproduce well ECD spectra dominated by “strong” mechanisms of optical activity [37], such as the exciton coupling between aromatic rings, like in biaryls, and the inherent chirality of twisted π -electron systems [7, 38]. The accuracy of semiempirical methods in predicting high-lying electric-dipole forbidden transitions is comparatively much poorer [39]. Another problem with semiempirical methods is related to the two gauge formulations (length and velocity) used for calculating rotational strengths. Unless using gauge-independent methods, dipole-length (DL) and dipolevelocity (DV) values for rotational strengths are always different [40]. The difference is related to basis set completeness; therefore using, for example, TDDFT with double- or triple-ζ quality basis sets including polarization functions leads to substantially equivalent DL and DV rotational strengths. By contrast, semiempirical methods often employ minimal basis sets and very large and unpredictable discrepancies may be obtained. Finally, there are cases where the ECD pattern is dominated by the so-called exciton coupling mechanism [19, 41, 42]. It arises when two or more chromophores, allied

221

222

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

with strong electric-dipole-allowed transitions, are close in the space and arranged in a skewed fashion with respect to each other. In this condition the coupling between the two transition dipoles generates a strong bisignate ECD signal, called an exciton couplet. Analysis of exciton-coupled ECD spectra is often straightforward, as witnessed by the countless applications of the exciton chirality method in the AC assignments of organic compounds, including natural products [6, 19, 41, 42]. On a quantitative ground, excitoncoupled ECD spectra may be simulated by some techniques that may be grouped under the name of semiclassical methods, such as DeVoe’s approach. For a comprehensive description of this method, we refer the reader to the original papers [43] and to recent reviews and applications [27, 44]. Provided that the chromophores under considerations are known and characterized, DeVoe-type calculations are extremely fast and applicable to systems containing even dozens of chromophores. Therefore, it would be a method of choice to estimate the ECD of large supramolecular aggregates in the coupled-oscillator approximation.

6.3. THE SOLID STATE ECD/TDDFT APPROACH: METHODOLOGY AND SCOPE 6.3.1. Principle The essence of the procedure for assigning absolute conﬁgurations by means of the solid-state CD/TDDFT method consists of measuring the solid-state ECD spectrum of a microcrystalline sample and comparing it with the spectrum calculated using the Xray geometry as input structure [22]. The main advantage of the solid-state approach, with respect to similar procedures based on solution ECD calculations, is that it does not require a conformational analysis and therefore it is computationally fast and avoids the difﬁculties connected to conformational searches and geometry optimizations [6b]. Moreover, the experimental and calculated ECD spectra refer to the very same geometry, implying a good agreement between theory and experiment and a reliable assignment. It is of vital importance that an artifact-free experimental solid-state ECD spectrum has to be measured (see Section 6.3.2). Furthermore, the solid-state ECD spectrum must be devoid of bands intrinsic to the solid state—for example, arising from intermolecular interactions in the crystals [45, 46]. A step-by-step formulation of the solid-state CD/TDDFT approach is the following: 1. Isolation of the natural product and determination of its constitution and relative conﬁguration [22]. 2. Growing crystals for X-ray analysis. 3. X-ray single-crystal diffraction analysis. 4. Measurement of microcrystalline solid-state ECD spectrum as KCl pellet, as well as of solution ECD spectra in one or more solvents. 5. Generation of the input geometry for ECD calculations (with initial arbitrary AC), by optimizing hydrogen atoms of the X-ray structure. 6. Calculation of rotational strengths with the TDDFT method, possibly employing various combinations of functionals and basis sets. 7. Generation of the ECD spectrum as a sum of Gaussians. 8. Comparison of the experimental solid-state (and solution) ECD spectrum with the TDDFT-calculated one.

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

The reason for recording both solid-state and solution ECD spectra is that their comparison is useful to reveal the presence of unexpected bands in the solid-state ECD spectrum. Bands much stronger than anticipated, or bands appearing in unexpected positions of the spectrum, may be due to measurement artifacts (Section 6.3.2) or be allied to solid-state intrinsic ECD effects (Section 6.5). In point 5, it is stated that the input geometry for ECD calculations is generated from the X-ray one upon optimization of the hydrogen atoms. Hydrogen atoms are often not accurately located using X-ray data because of low scattering power, distorted electron density, and large librations. In the process of structure reﬁnement, hydrogen atoms are generated in positions determined by distance and angle constraints relative to the heavy atoms they are attached to. The SHELXL routine for structure reﬁnement [47] adopts a “riding” model based on default C–H distances that, for some reason, are much shorter than real ones and would introduce unwanted errors in the calculated ECD spectra. Thus a preliminary DFT optimization of hydrogen atoms is run with B3LYP/6-31G(d), leaving all other atoms frozen, which usually produces geometries whose Y–H bonds (Y = C,O,N) are longer by 10–15% than the input ones. Although librations of carbon, oxygen, and nitrogen atoms are usually much smaller than for hydrogen, their effect may also lead to a slight underestimation of, for example, C–C, C–O, and C–N bond lengths by X-ray diffraction measurements [48]. In our experience, the discrepancies between Xray geometries and DFT-optimized ones (using the X-ray geometries as input structures and keeping all dihedral angles frozen) are very small, because C–C, C–O, and C–N bond lengths may vary within ±2%. Possible exceptions are compounds with a compact polycyclic structure, for example, containing fused or spiro-linked ﬁve-membered rings, where many bond lengths are strictly correlated to each other (see compounds 4 and 15 in Chart 6.1) [49, 50]. In these cases, we observed that relieving the bond lengths by means of DFT geometry optimizations was beneﬁcial for a better agreement between experimental and computed solid-state ECD spectra, but never decisive for assigning the absolute conﬁguration. Concerning point 6, it is clear that the choice of a proper DFT functional and basis set is crucial for the success of the method. One of the advantages of the solid-state approach is that the time saved in the generation of the input structure may be devoted to testing different combinations of functionals and basis sets to look for the best agreement with the experimental spectrum. According to our experience on several compounds with diverse structures, the three hybrid functionals B3LYP [29], PBE0 [31], and BH&HLYP [30], with increasing fraction of exact (Hartree–Fock) exchange, can be used to handle most common situations and often result in calculated ECD spectra in very good agreement with experimental ones. Usually, the spectra calculated with B3LYP and PBE0 are quite similar to each other, and they differ from BH&HLYP for a systematic wavelength shift (the direction depends on the nature of transitions involved). To improve the predictions of energies and transitions dipoles allied with charge-transfer, diffuse, and Rydberg states, use of the Coulomb-attenuated version of B3LYP, known as CAMB3LYP, is advisable [34]. In our hands, however, BH&HLYP has demonstrated to be capable of solving the same kinds of problems efﬁciently. It is remarkable that this “halfand-half” functional is included in the most recent versions of computational packages “for backward-compatibility only,” but its usefulness is indisputable [51]. As for the basis sets, use of the largest possible basis set is recommended for many reasons, some of which have been outlined above [26, 40]. On a practical ground, a split-valence basis set including a reasonable set of polarization functions and a minimum set of diffuse functions would be sufﬁcient in most cases. Our favorite in the ﬁeld is Ahlrich’s triple-ζ

223

224

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

basis set TZVP [52], whose set of ﬂexible polarization functions somehow counterbalances the lack of diffuse functions. When inclusion of these latter seemed necessary, we used the well-known Dunning’s correlation-consistent aug-cc-pVDZ set [53], or the less popular ADZP [54], or else aug-TZVP, obtained by augmenting TZVP with the set of most diffuse functions taken from aug-cc-pVDZ. In almost all cases analyzed thus far, however, use of, for example, aug-TZVP versus TVZP improved the general agreement between experimental and calculated spectra, without being decisive for the conﬁgurational assignment. In point 7, the set of calculated rotational strengths as a function of frequency (in other words, a stick plot) is converted in a more handy ECD spectrum by applying a broadening or band-shape. To each rotational strength is associated a band-shape function, usually of Gaussian or Lorentzian type, with intensity proportional to the absolute rotational strength value. A sum of all bands is then evaluated to generate a full ECD spectrum. Such a procedure requires the selection of a band-width parameter σ that is normally established on an empirical ground, selecting the value providing the best ﬁt with the experimental spectrum in the more relevant spectral region(s). Expressing the rotational strengths in 10−40 cgs units and ε in the common M−1 · cm−1 units (where M is molarity, mol · L−1 ), the ECD spectrum calculated as sums of Gaussians in the wavenumbers (˜ν in cm−1 ) domain is [55] Ri v˜i v˜ − v˜i 2 exp − . (6.1) ε(˜ν ) = 0.0247 σi σi i

If computed transition wavelengths are systematically shifted with respect to experimental ones, a wavelength correction may be applied to better compare computed and experimental spectra. In the so-called UV correction, one looks for the match between computed and experimental UV–vis spectra and then applies the same shift to ECD ones [7]. The comparison between the experimental solid-state ECD spectrum and the TDDFT-calculated one (point 8) is clearly the decisive step. The advantage of using full predictions of ECD spectra with respect to other kinds of treatments such as the exciton chirality approach is that the former provide a full ECD spectrum extending on a more or less wide wavelength range, rather than focusing on a single spectral feature. When the experimental ECD spectrum is well-structured—that is, when it contains several distinct bands—this offers the chance for a critical evaluation of the calculation results going beyond a mere comparison of band signs. This “quality check” may be very important to assess the reliability of the calculation method employed and may point, if necessary, to the need of testing further ones.

6.3.2. Solid-State CD Measurements Solid-state ECD spectra can be measured by different techniques described in several reviews [22, 45, 46, 56, 57]. The microcrystalline pellet or disc method is the most frequently used technique and was adopted as the standard in our solid-state ECD/TDDFT protocol. The crystalline sample is mixed and powdered with a suitable matrix such as KBr, KCl, or CsI, and the microcrystalline powder is pressed to produce a translucent glassy disc. KBr can be used for ECD measurements only above 220 nm, because of its UV cutoff. The use of KCl is less common in the literature, despite the fact that KCl can be used down to 180 nm and its handling is not different from KBr. The shortwavelength range extension may be crucial for the solid-state ECD analysis of natural

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

products containing only weak chromophores with high-lying transitions, such as alkene, ester, or anhydride groups (see compounds 7, 13, and 14 in Chart 6.1). Detailed procedures of pellet preparation are reported in several articles [58, 59]. In our solid-state ECD measurements of natural products, the disc is prepared by grinding and mixing ≈ 180–250 mg of KCl (≥99.999% Fluka, preheated at 100◦ C) and 30–250 μg of sample (depending on the chromophore) with the aid of a Perkin–Elmer vibrating mill equipped with a stainless steel ball for 5 min. The mixture is then pressed under vacuum at 10 tons with a Perkin–Elmer press for 5–10 min to provide a translucent disc. To decrease diffused reﬂections at grain boundaries, the sample and the matrix must be ﬁnely powdered and mixed to provide homogeneous distribution. Elaborated grinding is necessary because the intensity of the scattered light is proportional to 6th power of the particle diameter. The lowest possible sample amount providing acceptable ECD spectra is generally used to decrease the effect of absorption ﬂattening [60] and assure linearity with sample concentration [58]. Because KCl is hygroscopic, measurements are carried out right after the preparation of the disk by mounting it on a rotatable holder placed as close as possible to the detector. Normally, solid-state ECD data are reported as ellipticity φ in mdeg units. However, they can be normalized to ε when necessary, provided that the approximate dimensions of the pellet are known. It is well established that solid-state ECD measurements are easily contaminated by artifacts due to linear dichroism (LD) and linear birefringence (LB) allied with macroscopic sample anisotropies in the sample [56, 61, 62]. These artifacts have both rotation-dependent and rotation-independent contributions, whose presence may be ascertained by, respectively, sample rotation around the incident-light axis (or z axis) and ﬂip (180◦ rotation) around the vertical y axis [61, 62]. However, averaging the various spectra obtained by z -rotation and y-ﬂip is not sufﬁcient for obtaining artifact-free ECD spectra. As clearly pointed out by Kuroda and Shindo, the only proper way to extract true ECD data from a raw spectrum with LB and LD contributions consists of recording ECD, LB, and LD signals simultaneously and applying a speciﬁc protocol based on Mueller matrix formalism [56, 61]. Recording ECD, LB, and LD spectra is possible with a dedicated chiroptical spectrophotometer (Jasco J-800KCM) [61]. Using commercially available ECD spectrophotometers, at least LD data can be recorded, and simultaneous ECD/LD measurement should always be performed whenever possible on solid samples. In our protocol, various ECD spectra are usually recorded upon stepwise 90◦ rotations around the z axis as well as 180◦ ﬂip around the y axis, to exclude the presence of both rotation-dependent and -independent contributions from macroscopic anisotropies. In most cases, the spectra showed negligible changes with rotation or ﬂip [22]. When some inﬂuence of rotation around z axis was noticed, it consisted of a periodic baseline shift leading to minor differences between spectra (Figure 6.2). Solution spectra in various solvents are always recorded to be compared with solid-state ones to further exclude the presence of spectral artifacts in solid-state spectra (Figure 6.2).

6.3.3. Applicability The most important prerequisite of the solid-state ECD/TDDFT approach is the availability of X-ray geometries from single-crystal diffraction. One may question why to measure and calculate an ECD spectrum once the X-ray geometry is known. This is because the conﬁgurational assignment by X-ray anomalous scattering effect is subjected to important limitations that do not apply to the solid-state ECD method. When crystals belonging to noncentrosymmetric space groups exhibit anomalous scattering, the differences observed in Friedel pairs intensities may be used to extract

225

226

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

60

60

40

20

0

–20

–40

0° 90° 180° 270°

40

φ (mdeg)

Δε (M–1cm–1)

Solid-state ECD spectra (KCI disc), disc rotation

Solution ECD spectrum (CH3CN)

20

0

–20

200

225

250 275 λ (nm)

300

325

350

–40

200

225

250

275

300

325

350

λ (nm)

Figure 6.2. Comparison between solution (left) and solid-state ECD spectra (right) measured on KCl pellets upon rotations around the incident light z axis. The sample is bis(4-bromobenzoate) of palmarumycin M1 (29b) discussed in Section 6.6.

phase information and assign absolute conﬁgurations. Such a difference is, however, rather small, and it is measurable only in the presence of one or more atoms whose absorption edge is close to the X-ray wavelength. The scattering factor responsible for the magnitude of the anomalous dispersion effect is in fact both atom- and wavelength-dependent, and it is almost negligible for light elements (up to O). Although a single F or even O atom may sufﬁce for measuring the anomalous dispersion effect, excellent data collection should be performed with well-deﬁned single crystals at low temperatures, recording at least a full set of Friedel pairs and using CuKα radiation. A statistical survey of Cambridge Structural Database (CSD) [63] clearly reveals the scope of the anomalous dispersion method to assign absolute conﬁgurations. The version of the CSD updated to March 2010 contains ≈ 8000 compounds in chiral space groups ﬂagged with “absolute conﬁguration.” However, only 586 consist of H, C, O, and N atoms, and 51 consist of H, C, O, N, F, amounting overall to less than 8% on the total. Even in the presence of a heavy scatterer, assignment of the absolute structure by X-ray analysis is not trivial. Two literature surveys that were reported around 25 years apart [64] revealed “many unsatisfactory features in the original publications,” including incorrect space groups and insufﬁcient Friedel coverage leading to great uncertainty in some of reported conﬁgurations. It must also be stressed that a large majority of natural products contains only H, C, O, and N atoms. Of the around 226,500 entries comprised in the Dictionary of Natural Products as of June 2010 [65], only 15,500 (less than 7%) contain heteroatoms other than O and N. In summary, there is much space available for application of the solid-state ECD method for assigning absolute conﬁgurations of natural compounds. A major issue associated with solid-state ECD spectra is that intermolecular interactions between molecules closely packed in the crystals may give rise to non-negligible contributions to the spectrum [45, 46]. Any phenomenon of this kind cannot be predicted by a calculation run on a single molecule. The impact of crystalline intermolecular effects on the solid-state ECD/TDDFT method will be discussed in Section 6.5.

227

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

6.4. THE SOLID-STATE ECD/TDDFT APPROACH: EXAMPLES OF APPLICATIONS The solid-state ECD/TDDFT approach has been applied thus far to 17 natural products (Chart 6.1), secondary metabolites of fungi, often endowed with biological activities such as antibiotic, antifungal, and anti-inﬂammatory [15, 49, 50, 66–76]. These compounds exhibit a great structural variability and contain different chromophores, fully or partially unsaturated rings, ﬂexible saturated rings or chains, and centers of chirality in diverse O

O

O HO

OH

OH COOCH3

H3C

HO

O O

Globosuxanthone A (2)

O OHO

OCH3 OH

O

H O

O

O O Microsphaeropsone A (8)

O

3,16-Diketoaphidicolan (6)

O H Sinularolide B (5)

OH

O

H

O

O HO

HO

O

O HOH2C

O

HO

O Massarilactone E (4)

OH O Ascochin (3)

O

OH O

Viburspiran (7)

O

OH

H

O

O

O

OH

O

HO

H

HO O

O

MeO

O

OH

O Curvulone A (9)

O

O

O

α,β-Dehydrocurvularin (10) O

OCH3

OH

O

OH

OH

Fusidilactone B (14)

O Tetrahydropyrenopherol (13)

H3C H3CO O RO

OH

O O

N Macropodumine B (12)

H

O

HO

O

OH

O

H

O

OH

Blennolide A (11)

H

H3C

O

OH

O

O

CH3

O

H O O

Papyracillic acid A (R = H, 15a) and methyl acetal (R = CH3, 15b)

H3CO

O

Hypothemycin (16)

O OHOH

O

O 1β,10β-Epoxydesacetoxymatricarin (17)

Chart 6.1. Natural products whose AC has been assigned by means of the solid-state ECD/TDDFT approach.

228

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

spatial relationships. In all cases, the absolute conﬁguration could be established thanks to a good agreement between the experimental and calculated solid-state ECD spectra. In the following, some relevant cases will be discussed in detail to exemplify the adopted procedure and stress its scopes and limitations [22]. After phomoxanthone (1) described above, the ﬁrst case investigated with the solidstate ECD/TDDFT approach was globosuxanthone A (2, Chart 6.1) extracted from Microdiplodia sp. and obtained in a crystalline form suitable for X-ray analysis [68]. Globosuxanthone A is a dihydroxanthenone with potent antitumor activity, also isolated from Chaetomium globosum [77]. ECD spectra of 2 were recorded in solution (methanol and a tertiary solvent mixture) and in the solid state as a KCl pellet (Figure 6.3). The three spectra were roughly consistent, but they displayed some differences probably related to the conformational freedom around the ester and diol moieties. Most importantly, however, the solid-state ECD spectrum did not show any extra band with respect to the solution ones and both the signs and relative intensities of the major bands were preserved. This observation pointed to the essentially intramolecular origin of the solidstate ECD spectrum, meaning that intermolecular couplings in the crystal lattice did not give rise to apparent contributions. Starting from the X-ray geometry of 2 and assuming (1R,2R) conﬁguration, hydrogen atoms were re-optimized with DFT using B3LYP/631G(d). As a result, C–H and O–H bond lengths increased by 15% on the average, demonstrating the necessity of this preliminary optimization. TDDFT and ZINDO calculations were tested in ECD calculations, employing in the ﬁrst case various hybrid functionals (B3LYP, PBE0, and BH&HLYP) and TZVP basis set. Both B3LYP/TZVP and PBE0/TZVP calculations led to a good agreement with the experimental spectrum above 250 nm in terms of position, sign, and (to some extent) relative intensity of bands, including a couple of shoulders in the 300- to 330-nm range (Figure 6.3). The BH&HLYP functional reproduced the sign sequence but with a less satisfying general agreement. On the contrary, ZINDO performed very poorly, which was not unexpected in view of the

Δε (M–1cm–1), φ (mdeg), and R (10–39 cgs)

8 6

Experimental CD in CH3OH solution

4

Experimental solid-state CD (KCl disc)

O OH COOCH3

OH O HO Globosuxanthone A (2)

2 0

C3

O2 C6

–2

C2 C1

C5

C7

Calculated CD on X-ray geometry Calculated R

–4

C4

C12 C10 C11

C8

–6

O5

C13 C14

C9

O1 07 C15

04

–8

O3

200

250

300 λ (nm)

350

O6

400

Figure 6.3. Experimental ECD spectra of (1R,2R)-(−)-globosuxanthone A (2) in methanol solution and in the solid state (KCl disc), as well as ECD spectrum calculated with B3LYP/TZVP on the solid state geometry (shown on the right). Vertical bars represent calculated rotational strengths.

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

molecular and electronic structure. In fact, the most plausible mechanism of optical activity responsible for the moderately intense ECD spectrum of 2 is the second-sphere chirality [78] provided by the chiral ring on the conjugated chromophore (which includes the planar diene). Based on the results obtained with TDDFT calculations, the absolute conﬁguration of globosuxanthone A was assigned as (1R, 2R)-(−)-2 [68]. The conclusion was reached that TDDFT, with a proper choice of the functional and a sufﬁciently ﬂexible basis set, would grant in subsequent cases a wide applicability with a reasonable computational time. In particular, since the conformational analysis step is skipped and the ECD calculation needs to be run on a single structure, various functionals (and possibly basis sets) may be tested to look for the best agreement with the experiment in terms of overall spectral appearance, which can be taken as an indication of the reliability of the conﬁgurational assignment. Essentially the same procedure was followed for successive compounds 3–17 (Chart 6.1) isolated from natural sources, the absolute conﬁgurations of which were assigned by means of the solid-state ECD/TDDFT approach. These included (4S )-(+)ascochin (3), isolated from the fungus Ascochyta sp. from the plant Meliotus dentatus, exhibiting antifungal and algicidal activity [69]; (4S , 5R, 6S , 7R, 10R)(+)-massarilactone E (4), isolated from Coniothyrium sp. associated with Artimisia maritima [49]; (1R, 2R, 3R, 12S , 13R)-(−)-sinularolide B (5), a cytotoxic compound isolated from the soft coral Lobophytum crassum [71]; (4R, 5R, 8S , 9S , 10S , 12R)-(−)-3,16-diketoaphidicolan (6), extracted from the endophytic fungus Phoma sp. isolated from Aizoon canariense [76]; and (1R, 2S , 4R, 5R, 6R, 7R, 8S )-(+)-viburspiran (7), a new octadride member of maleic anhydride natural products, extracted from the endophytic fungus Cryptsporiopsis sp., isolated from Viburnum tinus [74]. A somewhat peculiar case is represented by (1R, 2R)-microsphaeropsone A (8), the ﬁrst natural compound with dihydrooxepino[2,3-b]chromen-6-one skeleton, isolated from Microsphaeropsis sp. from the bush Lycium intricatum. Calculated ECD spectra using the solid-state geometry for 8 showed a great variation with the functional employed, a situation occurring sometimes with complicated chromophores which calls for special caution in the interpretation of results. Although at least BH&HLYP/TZVP calculations reproduced the experimental solid-state ECD spectrum satisfactorily, the assignment was substantiated by means of vibrational CD [79], following the recommendation to use diverse chiroptical techniques to assign the absolute stereochemistry in ambiguous situations [80]. The VCD spectrum of 8 was measured in solution (CDCl3 ) and compared with the Boltzmann-averaged calculated spectrum using B3LYP/6-31G(d) on three DFT-optimized low-energy structures, including SCRF-PCM solvent model for chloroform [73]. More recently, a series of curvularin derivatives with 12-membered lactone ring skeleton belonging to the polyketide family was extracted from the fungus Curvularia sp. 6540, isolated from the marine red algae Gracilaria folifera. All compounds showed antibacterial, antifungal, and antialgal activity, and they included two new compounds, (10S ,15R)-(−)-curvulone A (9, Chart 6.1) and (11R,15R)-(−)-curvulone B, and two known curvularins with unusual (15R) conﬁguration, namely (11R,15R)11-hydroxycurvularin and (10E ,15R)-(+)-10,11-dehydrocurvularin (10) [75]. For compounds 9 and 10, crystals suitable for X-ray analysis were obtained and thus the solid-state ECD/TDDFT approach was applied. In both cases, the absolute conﬁguration assignment was conﬁrmed by independent methods (chemical correlation and X-ray analysis) [75, 81]. Compound 10 was especially interesting because it crystallized in two different forms, depending on the solvent employed (CDCl3 versus wet CH3 OH/CH2 Cl2 ). The

229

230

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

two X-ray geometries showed large discrepancies in the conformation of the ﬂexible lactone ring, and, interestingly enough, the ECD spectra calculated with TDDFT (B3LYP/TZVP) on the two X-ray structures were also very different (Figure 6.4). The solid-state ECD spectrum measured on the crystals obtained from CDCl3 (type A) was well-reproduced by the TDDFT-calculated spectrum using the respective geometry. The above results suggested that the solid-state ECD/TDDFT approach may be well employed for crystalline materials exhibiting polymorphism, provided that the crystal form used for the solid-state ECD measurement is known and the corresponding molecular structure is determined [75]. Another large collection of natural compounds was obtained from Blennoria sp., an endophytic fungus from the succulent Carpobrotus edulis, which led to eight compounds related to the known secalonic acids [72, 82]. Besides secalonic acid B (18, Chart 6.2), a powerful fungicide and algaecide, a series of new monomeric and a mixed dimer derivatives were found, named blennolides (Chart 6.2). These compounds represent the long-awaited monomeric units of the dimeric secalonic acids, in particular the isomeric hemisecalonic acids B and E (or blennolides A and B, 11 and 19). Their rearrangement products 20–22 (blennolides D–F) are structurally unique new natural products, where a highly substituted γ -lactone moiety is linked to a dihydrobenzopyranone. In blennolide G (23), the usual ergochrome monomer is linked to the deoxy analogue of rearranged monomer 20, extending the secalonic acid family with a novel heterodimer. The structure and the relative stereochemistry of blennolide A (11) were conﬁrmed by single-crystal X-ray analysis, while the absolute conﬁguration was determined as

OH

Δε (M–1cm–1) and φ (mdeg)

25 Experimental solid-state CD (KCl disc, type A crystals)

20 10

HO

Calculated CD on X-ray geometry (type A)

H CH3 α,β-Dehydrocurvularin (10) O

Calculated CD on X-ray geometry (type B)

5

O

O

A

0 B –5 200

250

300 350 λ (nm)

400

450

Figure 6.4. Left: Experimental solid-state ECD spectrum of (15R)-(+)-10,11-dehydrocurvularin (10) recorded using crystals isolated from CDCl3 (type A), as well as ECD spectra calculated with B3LYP/TZVP on the X-ray structures of (15R)-10 for the crystals isolated from CDCl3 (type A) and a mixture of wet CH3 OH and CH2 Cl2 (type B). Right: Overlapped solid-state X-ray geometries of (15R)-10 for type A and B crystals (dark and light structures, respectively).

231

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

O

HO

OMe O

OH

OH

O

O

OH

OH

OH

OH

O H

2

OH

O

O

MeO

O

OH

MeO

Secalonic acid B (18)

OH

O

OH

O

O

O O H

O

MeO

Blennolide B (19)

OH

O

O

O

O

Blennolide D (20) x OMe O OH

H O O

Blennolide E (21)

O

OH

OH

O

9

O

H

OH

2

MeO

OH 9

O

O

MeO

O O

Blennolide F (22)

O

O

OH O

Blennolide G (23)

MeO

O

OH

Chart 6.2. Blennolides extracted from Blennoria sp. [72, 82]. Blennolide A (11) is shown in Chart 6.1.

(5S ,6S ,10aR)-(+)-11 after application of the solid-state CD/TDDFT approach. The situation for blennolides D and E (20 and 21) was more challenging because their Xray structures were not available. The overall stereochemistry of 20 and 21 could be established only by means of a combination of several spectroscopic (NOESY, heteronuclear 3 JC,H couplings, ECD) and computational techniques (MM conformational searches, DFT geometry optimizations, TDDFT calculations). In particular, molecular modeling results were essential to rationalize observed NOEs around the rotatable C2–C9 bond [72]. The family of blennolides offered a striking evidence of the advantage provided by the X-ray crystal structure for establishing both relative and absolute stereochemistry. A further conﬁrmation of this latter point is provided by the two analogue macropodumines B and C (12 and 24, Figure 6.5 and Figure 6.6, respectively), belonging to the family of Daphniphyllum alkaloids, a group of fused-heterocyclic fungal metabolites with signiﬁcant medicinal properties. Macropodumine B (12) shows an almost unique structural feature for a natural compound: It is in fact a zwitterion containing a rare cyclopentadienyl anion and an iminio counterion (Figure 6.5). Compounds 12 and 24 were isolated from the Chinese medicinal plant D. macropodum, and the solid-state structure of 12 was determined by X-ray analysis [83]. The absolute conﬁguration of macropodumines B and C was established by comparison of their ECD spectra with TDDFT-calculated ones, using the solid-state protocol for macropodumine B (12) and the corresponding solution protocol for macropodumine C (24). In practice, the two compounds offered the option for a direct comparison between the two methods. For macropodumine B (12), the solid-state ECD/TDDFT approach led to a calculated CD spectrum (B3LYP/TZVP) in very good agreement with the experimental solid-state one (KCl disk, Figure 6.5). Other functional/basis set combinations (using BH&HLYP and ADZP) did not improve the observed agreement. After a moderate computational effort (<4 days calculation time), the absolute conﬁguration could be established as (2R,5S ,6S ,18S )-(−)-12. For macropodumine C (24), an MM-based conformational search followed by DFT geometry optimizations (B3LYP/6-31G(d)) revealed the presence of several minima due to the rotation of CH2 CH2 OH and COOCH3 substituents, also accompanied by ﬂuctuations in the ring system. The ﬁrst four low-energy DFT minima, with internal

232

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

40

Δε (M–1cm–1), φ (mdeg), and R (10–39 cgs)

O

O

OCH3

H

30

OH

20 10

N

0

Macropodumine B (12)

–10 –20

Calculated CD on X-ray geometry

x 1.5

–30

Experimental solid-state CD (KCl disc)

–40

Calculated R

–50 200

250

300

350

400

450

500

λ (nm)

Figure 6.5. Left: Experimental ECD spectrum of (2R,5S,6S,18S)-(−)-macropodumine B (12) in the solid state (KCl disc) and TDDFT-calculated CD spectrum on the X-ray geometry with B3LYP/TZVP; vertical bars represent computed rotational strengths. Right: Electrostatic potential surface computed with B3LYP/6-31G(d) for 12. The dark and light regions indicate positive and negative charge density, respectively.

O OH

Δε (M–1cm–1) and R (10–39 cgs)

30 20

OCH3

HO H

x5

OH O

N

10

Macropodumine C (24) 0 Weighted calc. CD on DFT geometries Experimental CD in CH3CN solution

–10 –20

Calculated R (absolute minimum)

–30 200

250

300

350 λ (nm)

400

450

500

Figure 6.6. Left: Experimental ECD spectrum of (2S,4R,5S,6S,18S)-(−)-macropodumine C (24) in acetonitrile solution and Boltzmann-weighted average of B3LYP/TZVP-calculated ECD spectra on four low-energy DFT structures; vertical bars represent computed rotational strengths on the lowest energy minimum. Right: Overlap between the ﬁrst four lowest-energy DFT structures of 24 computed with B3LYP/6-31G(d).

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

energies within 1.3 kcal/mol and representing 72% overall population, were considered for calculations. TDDFT-calculated ECD spectra were quite different for the various minima, an undesirable situation calling for a very careful estimation of relative populations. The weighted average of calculated ECD spectra for 24 showed a sufﬁciently good agreement with the spectrum measured in CH3 CN solution in the low-energy region, while some discrepancy was obtained at high energies (Figure 6.6). Overall, the absolute conﬁguration could be established as (2S ,4R,5S ,6S ,18R)-(−)-24 after a considerable computational effort (∼17 days). Possible improvements of the result could be achieved by (a) considering further low-energy minima in the calculations, (b) evaluating relative populations at a higher level and/or from free energies, and (c) including a solvent model in energy and ECD calculations. Each of these possibilities would increase the overall computational time at a large extent. For the two similar compounds just discussed the solid-state CD/TDDFT led to a safer conﬁgurational assignment with a much shorter computational time (about 1:4.5 ratio). This observation cannot be generalized, although it must be observed that macropodumines 12 and 24 are relatively rigid, and for more ﬂexible compounds exhibiting the presence of several low-energy minima the advantage of employing the solid-state procedure will be greater. An example of a very ﬂexible compound is offered by tetrahydropyrenopherol (13), a 16-membered dilactone extracted from Phoma sp., isolated from Fagonia cretica, showing anthelmintic and fungicidal activity [70]. Compound 13 contains two identical lactone chromophores giving rise to observable n –π ∗ transitions around 210 nm. The outcome of ECD spectra measurement was surprising because for (−)-13 the sign of the n –π ∗ band was positive in acetonitrile and methanol solution and negative in the solid state (KCl pellet, Figure 6.7). The most obvious explanation for the observed discrepancy lies in the extreme ﬂexibility of the 16-membered ring. In fact, a conformational analysis with MMFF and AM1 methods revealed the existence of at least 60 minima within 3 kcal/mol (AM1 energy). Thus, an attempt to reproduce the solution ECD spectra would be hampered by the necessity of considering so many input structures. X-ray analysis revealed a single structure with quasi-C2 symmetric rectangular shape (Figure 6.7). Use of the solidstate ECD/TDDFT approach led to (4S ,7R,4 S ,7 R)-13 conﬁgurational assignment after B3LYP/TZVP calculations (use of larger basis sets such as aug-cc-pVDZ and aug-TZVP conﬁrmed the same result). Interestingly enough, the sign of calculated n –π ∗ band for the X-ray geometry was in agreement with a sector rule proposed by Snatzke for lactones [84]. Using a signiﬁcant “chirogenic” fragment 25 (Figure 6.8) and varying systematically the relevant dihedral angles around the ester moiety, the sector rule was further veriﬁed and the shape of nodal surfaces was slightly modiﬁed [70]. The signiﬁcance and further examples of such kind of analysis, used to substantiate and rejuvenate sector and helicity rules, are discussed elsewhere in the present volume (Chapters 2 and 3). A second compound presenting a ﬂexible skeleton was fusidilactone B (14), belonging to a family of polycyclic lactones extracted from endophytic Fusidium sp., isolated from Mentha arvensis. Fusidilactone B (14, Chart 6.1) had already been isolated in 2002 but could not be crystallized at that time [85]. The relative conﬁguration of the cyclic core was elucidated by NMR, but it was not possible to assign the relative conﬁguration of the chirality centers of the side chain. When, later, suitable single crystals could be obtained, X-ray analysis revealed the overall relative conﬁguration and made it possible to apply the solid-state ECD/TDDFT approach for determination of the absolute conﬁguration as (2S ,3R,4S ,4aR,7S ,7aS ,3 R,5 S )-14 [15].

233

234

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Δε (M–1cm–1) φ (mdeg), and R (10–39 cgs)

O

2

O

CH3CN solution CH3OH solution

Experimental CD

OH

solid-state (KCI disc)

HO O

0 O Tetrahydropyrenopherol (13) 04

–2 C10

–4

Calculated CD on X-ray geometry Calculated R

–6

C9

05

C8

C7

C11

C6

C12

C5 03 C4

06 C13 C3

190

200

210

220

230

240

λ (nm)

250

260

C14 C15

C1

C16 01

02

C2

Figure 6.7. Experimental ECD spectra of (4S, 7R, 4 S, 7 R)-(−)-tetrahydropyrenophorol (13) measured in acetonitrile and methanol solution and in the solid state (KCl disc), as well as ECD spectrum calculated with B3LYP/TZVP using the solid-state geometry shown on the right. Vertical bars represent rotational strengths.

H

CH3

H

O

O

O

O

O

O

Figure 6.8. Left: Modiﬁed Snatzke’s CH3

H

Upper sector

Lower sector

CH3 25

sector ‘‘comet’’ rule for esters. The signs show sector contributions to the n–π ∗ transition [84]. Right: Structure of the chirogenic fragment 25.

A different situation where the solid-state approach demonstrated its potential for conﬁgurational assignment is the presence of equilibrating isomers in solution. Papyracillic acid A (15a) was extracted from the endophytic fungus Microsphaeropsis sp., isolated from Larix decidua, with its methyl acetal 15b (Scheme 6.1) [50]. The peculiarity in the skeleton of papyracillic acid A (15a) is the presence of a hemiacetal functionality at C7 whose ether-type oxygen is further involved in a spiro-ketal moiety at C4. Two consecutive ring openings may therefore occur easily in solution (Scheme 6.1). In fact, 1 H-NMR spectra of 15a in chloroform showed four sets of signals in approximate 4:2:1:1 ratio due to the coexistence of isomers (possibly including epimers). The observation was in keeping with a previous report by Sterner and co-workers [86] describing isolation of papyracillic acid A from a different source. In both cases, NMR data did not allow a deﬁnite structural assignment. Luckily however, crystals of papyracillic acid A (15a) were obtained allowing the determination of both the correct relative conﬁguration and the absolute conﬁguration, by means of the solid-state ECD/TDDFT protocol. In this

235

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

H3C HO

H3CO

H3C

O 7 4

O O

H3C

O

O

H3C

H3C

H3C H3CO O O

H3CO HO O

HO

H3C

H3CO O

H3C

H3CO O 7 4 O

O

15b (revised structure)

15a H3C HO H3C

H3CO

H3C H3CO HO O

O O

O

H3C

O

O

H3C H3CO O H3CO 7 4 O HC 3

O

previously proposed structure

Scheme 6.1. Possible solution equilibria of papyracillic acid A (15a), and proposed versus revised structure for its methyl acetal (15b).

case too, the advantage of considering the solid-state protocol with respect to the solution one is evident: Reproducing the solution ECD spectrum of 15a is hampered by the uncertainty about the nature of the species equilibrating in solution. In the solid state, a single isomer in a single conformation is present which needs to be considered for ECD calculations. The solid-state ECD/TDDFT approach was applied to both (4S , 6S , 7R)(−)-papyracillic acid A (15a) and its methyl acetal (4S , 6S , 7R)-(−)-15b. Interestingly, the structure of the latter proposed by Sterner and co-workers [86] appeared incorrect in the relative conﬁguration at C4 and was revised as shown in Scheme 6.1.

6.5. THE SOLID-STATE ECD/TDDFT APPROACH: THE ROLE OF CRYSTALLINE INTERMOLECULAR COUPLINGS In the course of application of the solid-state ECD/TDDFT approach, it was noticed that in the large majority of cases, intermolecular interactions in the crystal lattice had negligible or small effects on the observed solid-state ECD spectra. In principle, intermolecular interactions between individual molecules closely packed in the crystals are symmetryallowed for noncentrosymmetric crystal classes and should generate nonvanishing ECD signals [45, 46, 87], particularly in the presence of strong chromophores giving rise to exciton-coupling interactions. However, the solid-state ECD spectra measured for almost all compounds depicted in Chart 6.1 did not display any signal characteristic of solid-state intermolecular exciton couplings. This latter conclusion was supported by two observations. First, solid-state and solution ECD spectra were recorded for each sample and they were always consistent, in the sense that no new signal or signal with strongly altered intensity appeared in the solid-state ECD spectrum with respect to the solution state. Second, experimental solid-state ECD spectra were in all cases well-reproduced by TDDFT calculations run on single molecules with X-ray input geometry, which is possible only if intermolecular couplings are negligible. Only in two of 17 cases considered thus far have we had some hint about the occurrence of ECD features possibly allied to intercrystalline effects. The ﬁrst case was hypothemycin (16), a well-known 14-membered β-resorcylic acid macrolactone endowed with remarkable biological activity [88], which was extracted from fungus Phoma sp., isolated from Senecio kleinii [69]. ECD spectra of 16 recorded in solution and solid state showed some differences in the regions between 290–320 nm and 220–240 nm (indicated by arrows in Figure 6.9). However, an MM-based conformational analysis, followed by

236

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

30

OH Experimental CD in CH3CN solution

Δε (M–1cm–1) and φ (mdeg)

20

O

Experimental solid-state CD (KCl disc) Calculated CD on X-ray geometry

10

CH3

O

H3CO

OH OH Hypothemycin (16) O

O

0

–10

–20

–30 200

X-ray geometry 250

300

350

400

Lowest-energy AM1 structure

λ (nm)

Figure 6.9. Left: Experimental CD spectra of (1R,2R,4S,5S,10S)-(+)-hypothemycin (16) in acetonitrile solution and in the solid state as KCl disc, as well as ECD spectrum calculated with B3LYP/TZVP using the solid-state geometry; arrows indicate the most discrepant regions between solution and solid-state spectra. Right: Comparison between the solid-state geometry of 16 and AM1-computed lowest-energy structure.

AM1 geometry optimizations, revealed a strong preference for a structure similar to the X-ray one (Figure 6.9, right). Unexpectedly, we observed that theoretical ECD spectra, calculated with TDDFT using either B3LYP or PBE0 functionals and TZVP basis set on the solid-state geometry, were more similar to solution ECD spectrum than to solid-state experimental ECD spectrum. Analysis of TDDFT results revealed that the strongest π –π ∗ transitions of the aromatic chromophore (a 2-hydroxy-4-methoxybenzoate) occurred exactly in discrepant regions around 220 and 300 nm. If intermolecular exciton interactions do occur in the solid state, they are likely to manifest especially in those spectral regions. However, the overall effects of such interactions were still quite limited (see note at the end of the chapter). The absolute conﬁguration of (1R, 2R, 4S , 5S , 10S )-(+)hypothemycin was already known [88], and application of the solid-state ECD/TDDFT protocol led to the same assignment. Hypothemycin served as a test case for compounds endowed with strong aromatic chromophores possibly leading to intercrystalline exciton couplings [69]. A similar conclusion was reached for (1R, 5R, 6S , 7S , 10S , 11S )-(+)-1β, 10βepoxydesacetoxymatricarin (17), a sesquiterpenoid isolated from wild safﬂower (Carthamus oxyacantha) [67]. In this case, the solid-state ECD/TDDFT approach was employed for a true conﬁgurational assignment. ECD spectra were recorded for 17 both in solution (acetonitrile) and in the solid state and they were similar, but in the solid-state spectrum a couple of new bands appeared while another band acquired a clear vibronic structure. This latter effect was justiﬁed with the conformational stiffening typical of the solid state, while the appearance of the new bands was attributed to intermolecular couplings in the crystals [67]. In general, the impact of intermolecular couplings in the crystals on the solid-state ECD/TDDFT method seems very modest for all compounds analyzed thus far. Three different hypotheses may justify this observation:

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

(a) A fortuitous arrangement of the chromophores in the crystals makes their exciton coupling interaction weak, since their transition dipoles are aligned either parallel or collinear or exactly perpendicular. (b) The simultaneous couplings among dozens of molecules in the crystals may cancel each other and thus result in a negligible overall effect. (c) The existence of some shielding effect in the crystals, which suppresses the exciton interactions. The phenomenon described in the latter point would be similar to what occurs when excitation energy transfer or charge transfer takes place in condensed media. It has been demonstrated that the presence of an environment embedding two chromophores in close contact to each other may suppress their dipolar interaction, thus screening their exciton interaction and quenching energy and charge transfer [89]. An environment depicted by means of a continuum anisotropic dielectric may efﬁciently simulate a crystal packing. The contribution of such an embedding effect to solid-state ECD spectra of crystalline materials has to be still proved. The ﬁrst two points can in principle be demonstrated by calculating the ECD spectra of molecular composites representative of crystal clusters and comparing them to that of single, “isolated” molecules. The obvious limitation of such an approach is the computational cost, making it impracticable even for a small number of constituent molecules if a high-level QM method like TDDFT is employed. Alternatively, one may turn to using approximate methods such as semiempirical calculations and even semiclassical methods such as the coupled-oscillator DeVoe’s method, described in Section 6.2.3. At the present time, only two literature reports have described full calculations of ECD spectra of small crystal clusters using the above approaches (but see note at the end of the chapter). The ﬁrst one concerns two isomeric naphthylethylidene ketals of α-Lrhamnopyranoside (26 and 27, Figure 6.10), which were synthesized and obtained as single crystals amenable to X-ray analysis [90]. The two isomers showed interesting differences in their solution ECD spectra versus the solid-state ones, all of which are dominated by the exciton coupling between the two aromatic chromophores, namely naphthalene and 1,4-dimethoxybenzene (DMB). For the 2-naphthyl compound 26, ECD spectra measured in acetonitrile solution and in the solid state as KCl pellet were similar to each other and showed a prominent band at 225 nm allied to the naphthalene 1 Bb transition, with negative sign for (1 R)-26 (Figure 6.10). On the contrary, for the 1-naphthyl compound 27, the solid-state ECD spectrum (KCl pellet) was quite different and relatively more intense than that in acetonitrile solution, and for (1 R)-27 it was dominated by a strong negative couplet with zero point at 228 nm (Figure 6.10) [90]. Accordingly, this was attributed to the exciton coupling between naphthalene rings belonging to distinct molecules packed in the crystal. To conﬁrm such interpretation, ECD calculations were run with the DeVoe approach [43] including three transitions for the naphthalene (1 Ba , short-axis polarized, at 226 nm; 1 Bb , long-axis, 235 nm; 1 La , short-axis, 300 nm) and three transitions for DMB chromophore (1 Ba , long-axis, at 207 nm; 1 Bb , short-axis, 208 nm; 1 La , long-axis, 256 nm). DeVoe calculations were ﬁrst run on isolated molecules of 26 and 27 using both the solid-state geometry and MMFF-optimized structures, and for each compound they led to consistent results (Figure 6.11). This ﬁnding demonstrated that the difference observed for 27 between solution and solid-state ECD spectra could not be justiﬁed on the basis of different conformations assumed in the two states [90].

237

238

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

OMe

OMe

O

O

O

AcO O

O

AcO O

O CH3

CH3 1-naphthylethylidene ketal (27)

2-naphthylethylidene ketal (26) 20

Δε (M–1cm–1) and φ (mdeg)

O

40

10

20

Experimental CD (27) CH3CN solution solid-state (KCI disc)

x0.5

0 0 x 10

–10

x10

–20 –20 Experimental CD (26)

–40

–40

CH3CN solution solid-state (KCI disc)

–30

200

220

240 260 λ (nm)

300

280

–60

200

220

240 260 λ (nm)

280

300

Figure 6.10. Experimental ECD spectra of naphthylethylidene ketals of α-L-rhamnopyranoside 26 (left) and 27 (right) in acetonitrile solution and in the solid state (KCl disc).

20

10 DeVoe calculated CD (26) 5 Δε (M–1cm–1)

15

on MMFF structures (weighted average) on the X-ray geometry

10 5 0

0

–5

DeVoe calculated CD (27) on MMFF structures (weighted average) on the X-ray geometry

–10

–5

–15 –10

200

220

240 260 λ (nm)

280

300

320

–20

200

220

240 260 λ (nm)

280

300

320

Figure 6.11. ECD spectra calculated with DeVoe method for naphthylethylidene ketals of α-Lrhamnopyranoside 26 (left) and 27 (right) using either MMFF-optimized structures or the X-ray geometry. For MMFF structures, the Boltzmann-weighted average spectrum over four low-energy conformers at 300 K is shown.

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

(a)

(b)

Figure 6.12. Portions of crystal lattices of 26 (a) and 27 (b) showing the different relative arrangements of neighboring molecules in the crystals.

Examination of the crystal lattices revealed profound differences in the arrangement between neighboring molecules for the two compounds. Both compounds were isolated as orthorhombic crystals belonging to the P 21 21 21 space group; however, they showed different polarity. Molecules of isomer 26 assume a head-to-head and tail-to-tail arrangement in the crystal cell, so that the corresponding chromophores of neighboring molecules are aligned exactly parallel (or, otherwise said, stacked) to each other (Figure 6.12a). In this situation, the exciton coupling between the stacking chromophores is zero, because the projection angles between their transition dipoles (e.g., two naphthalene 1 Bb transition moments) is 0◦ or 180◦ . In contrast, molecules of isomer 27 have a head-to-tail arrangement in the crystal cell, and neighboring chromophores are not aligned or stacked but rather tilted relatively to one another (Figure 6.12b). The naphthalene ring of each molecule of 27 is surrounded by three other naphthalene rings and two DMB rings at close distance. For all of them, the projection angles between the long axes of the aromatic chromophores deviate considerably from 0◦ or 180◦ . Therefore, a strong exciton coupling may occur between electronic transitions of neighboring molecules in the crystals of 27 [90]. The above analysis was substantiated by DeVoe calculations run on a small cluster system composed of the two chromophores from a probe molecule surrounded by a limited number of chromophores belonging to neighboring molecules in the crystal lattice. In the case of 27, the intermolecular couplings between close naphthalene rings were found to generate a strong negative exciton couplet in the 1 Bb region, similar to what was seen in the solid-state ECD spectrum of 27 (Figure 6.10), whose intermolecular origin was thus demonstrated. The study represented the ﬁrst example of a semiquantitative investigation of intermolecular exciton couplings in the solid state [90]. Stimulated by the work of Bringmann et al. [8] discussed below, we have studied again the problem of compounds 26 and 27 by using ZINDO calculations [91]. First, the ECD spectrum was calculated on isolated molecules of 26 and 27 with both ZINDO and TDDFT using CAM-B3LYP and a relatively small basis set (SVP) of the Ahlrich’s series [92]. Because the two sets of spectra were consistent, we concluded that ZINDO was accurate enough to reproduce the ECD spectra of these naphthylethylidene ketals. Then, ZINDO calculations were applied to aggregates extracted from the crystal lattice on a systematic basis. For both lattices, a central probe molecule appears surrounded by 26 close neighbors in a cubic-like fashion. All possible 26 dimers were built and their ECD spectra calculated by ZINDO. For 27, there were two dimers giving exceptionally strong

239

240

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

ECD spectra dominated by a negative exciton couplet in the 1 Bb region with amplitude (peak-to-trough distance in ε units) A = −460 M−1 cm−1 . Several other dimers gave moderately intense ECD spectra with A within ±100 M−1 cm−1 . For 26, the strongest ECD spectrum associated with a dimer showed a couplet with A = −60 M−1 cm−1 ; all other dimers led to couplets with A below ±50 M−1 cm−1 . The eight dimers leading to the stronger calculated ECD spectra for both 26 and 27 were collected together to build a 9-mer including the probe molecule. Not surprisingly, the ECD spectrum calculated with ZINDO for the 9-mer aggregate of 27 showed a strong negative exciton couplet in the 1 Bb region with amplitude around −2000 M−1 cm−1 (Figure 6.13). The aggregate of 26 led instead to a positive and much weaker couplet with amplitude +400 M−1 cm−1 (Figure 6.13) [91]. Although the ZINDO-calculated spectra for the aggregates still fail to reproduce correctly the experimental spectra recorded in the solid state, they reinforce the hypothesis that intermolecular exciton couplings in the crystals play a decisive role in determining the solid-state ECD of 27 but not that of 26. Recently, Bringmann, Kuroda, and co-workers have reported a detailed study aimed to rationalize the ECD spectra of the well-known alkaloid (1R,3R,7P )-(−)-dioncophylline A (28), recorded in solution and in the solid state. In particular, the authors investigated the impact of crystalline inter-molecular interactions on the solid-state ECD spectrum, by means of semiempirical calculations run with ZINDO and CNDO/S methods on small and medium-size aggregates built from the crystal lattice [8]. Dioncophylline A possesses elements of both central and axial chirality and has (1R,3R,7P )-(−) absolute conﬁguration, and, as it happens frequently for biaryls, the chiroptical properties depend strictly on a combination of conﬁgurational and conformational factors. The most important structural parameter is the biaryl dihedral angle (A–B–C–D in Figure 6.14), which amounts to 115◦ in the X-ray structure of 28. In a ﬁrst step, the authors veriﬁed that semiempirical

1500

200

ZINDO calculated CD (27)

150

monomer 9-mer

1000 Δε (M–1cm–1)

100 50

x5

500

0 0

–50

x5 ZINDO calculated CD (26) monomer 9-mer

–100 –150 –200

200

220

240

260 λ (nm) (a)

280

300

–500

320

–1000

200

220

240

260 280 λ (nm)

300

320

(b)

Figure 6.13. ECD spectra calculated with ZINDO/S-CI for naphthylethylidene ketals of α-Lrhamnopyranoside 26 (a) and 27 (b) using the X-ray geometry. As ‘‘monomer,’’ a single isolated molecule has been considered. The ‘‘9-mer’’ is made of a probe molecule surrounded by eight neighbors as found in the crystal lattice, chosen as those giving rise to the strongest exciton coupling with the probe (evaluated with ZINDO, see text).

241

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

Δε (M–1cm–1) and φ (mdeg)

5

0 CH3 CH3 B C D

Experimental solid-state CD (KCI disc)

–5

OH H3CO H3CO

CD calculated on a single molecule (X-ray geometry) –10

–15 200

NH A

CH3

Dioncophylline A (28)

CD calculated for a crystal cluster

225

250

275 λ (nm)

300

325

350

Figure 6.14. Experimental solid-state ECD spectrum (KBr disc) of (1R, 3R, 7P)-(−)-dioncophylline A (28), as well as ECD spectra calculated with CNDO/S-CI on a single molecule (X-ray geometry) geometry and for a cluster of 16 molecules as found in the crystal lattice. (Reprinted from reference 8, with permission from Elsevier.)

QM methods did correctly reproduce the ECD spectrum of 28 by comparing results from different kinds of calculations (TDDFT, DFT/MRCI [24], ZINDO, and CNDO/S) and the experimental spectra in solution (ethanol) and in the solid state (KBr pellet, Figure 6.14). The experimental spectra in the two states showed substantial similarity, indicating a related conformational situation. However, intermolecular couplings in the crystals were expected to be especially strong for this molecule because of its aromatic chromophores (naphthalene and isoquinoline) arranged almost orthogonally. To evaluate intercrystalline couplings, several dimers were built from the crystal lattice by considering all possible independent couples composed of a central probe molecule and the surrounding ones. Some of the dimers led to calculated ECD spectra not far from that of the “isolated” probe molecule, but some others led to striking differences due to strong and efﬁcient intermolecular exciton couplings. However, when the theoretical curves obtained for all dimers where summed to each other and averaged, the resulting spectrum was again very similar to that of the isolated probe molecule. The same result was obtained by calculating the ECD spectrum for an aggregate composed of 16 molecules representing a large crystal portion (Figure 6.14). Therefore, the calculations demonstrated that intermolecular exciton couplings in the crystals do inﬂuence the chiroptical response; however, the various couplings tend to cancel each other so that their net effect is negligible, and the overall solid-state ECD spectrum turns out to be dominated by single-molecule effects [8]. To draw a conclusion from the two latter studies, it can be inferred that the two hypotheses made above to justify the apparent weakness of solid-state intrinsic effects are both viable. In some cases, like α-l-rhamnopyranoside ketal 26, the molecular arrangement in the crystal makes the exciton coupling between chromophores belonging to distinct molecules inherently weak. In some other cases, like dioncophylline A (28), some speciﬁc pairwise intercrystalline couplings may be strong, but they tend to cancel each

242

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

other and sum to a negligible overall effect. Further studies are needed to check how frequently these two situations occur and how general are the conclusions reached thus far.

6.6. OTHER PREDICTIONS OF SOLID-STATE ECD SPECTRA The solid-state ECD/TDDFT approach described in previous sections has been purposely developed for analysis of natural products and seems in fact especially well suited to this kind of compounds for two reasons. First, natural products frequently have rather complicated structures, containing multiple elements of chirality, which may be difﬁcult to elucidate by NMR. Therefore, the necessary efforts for obtaining crystals suitable for X-ray analysis are better justiﬁed. Second, natural compounds are often available in very small amounts, preventing further derivatization to introduce chromophores suitable for other ECD options such as the exciton chirality approach. Obviously, solid-state ECD calculations can be perfectly applied to synthetic compounds as well. The present section reviews the latest literature reports describing the use of solid-state ECD spectra, interpreted by various means, as a source of stereochemical information of organic compounds; a more extensive survey may be found in a recent review [22]. It must be stressed that the solid-state ECD technique has been employed in the past especially in context of inorganic chemistry, notably by Kuroda and co-workers. It is only relatively recently that the work of these authors, to which we draw the attention of the readers interested in inorganic chemistry applications [45, 56, 93], has inspired us and others for focused applications in the ﬁeld of organic stereochemistry. A semisynthetic compound whose solid-state ECD was employed to support an absolute conﬁguration assignment is the bis(4-bromobenzoate) of palmarumycin M1 (29b, Chart 6.3). Palmarumycin M1 (29a) was extracted from Microsphaeropsis sp., an endophytic fungus of Larix decidua [50]. The absolute conﬁguration of 29a was established as (4S , 4aS , 5S , 8R, 8aS )-(+) by X-ray analysis of its derivative 29b. Solution and solid-state ECD spectra of 29b were dominated by the exciton coupling between the bromobenzoate chromophores (the spectra are shown in Figure 6.2 and consist of a positive exciton couplet in the 210- to 270-nm region). The exciton chirality rule was applied on the X-ray structure of 29b and conﬁrmed the same absolute conﬁguration [50]. With respect to the more usual application of the exciton chirality method, which focuses on solution ECD spectra, consideration of the solid state has the expected advantage that only a single conformation needs to be taken into account. Since exciton-coupled ECD spectra depend sensibly on the molecular conformation, in terms of reciprocal arrangement between chromophores, the conformation adopted by the bis-chromophoric compound must be known with certainty [6, 19, 41]. In fact, apparent sign reversal or unexpected low intensity of exciton couplets have been reported as a consequence of unexpected conformations [6b, 94]. The comparison between the spectra of structurally similar compounds, one of which has already established stereochemistry, is the oldest way to gain stereochemical information from ECD spectra. An example of empirical correlation based on solidstate ECD spectra has been reported by Zhao and co-workers [95] concerning a series of diastereomeric pentacoordinate spirophosphoranes derived from α-amino acids (30, Chart 6.3). Some compounds in the series were crystallized and their absolute stereochemistry assigned. Solid-state ECD spectra of all compounds were also measured as KBr pellets and correlated with the conﬁguration at the phosphorous atoms. Thus, the stereochemistry of noncrystalline derivatives was inferred by comparison of their ECD spectra with those of the already characterized analogues [95].

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

R1O

H

OR1

O R2

H O

O

OH

O HN P H HN O

O N

R2

N

O 30 (R2 =

32

OAc R3 R3

OAc

R3 R3

O

O N

R4 R5

AcO 31a R3 = OAC 31b R3 = H

O H H

H

AcO Se

O

R6

Se

O

Bn

iPr,iBu)

29a R1 = H (Palmarumycin M1) 29b R1 = p-BrBz

O

243

33

Chart 6.3. Compounds 29–33, whose solid-state ECD spectra have been used for stereochemical investigations.

A second example of empirical correlation is described by Kamigata and co-workers [96] about diphenyl dichalcogenides Ph–Y–Y–Ph (Y = S, Se,Te), which crystallize as optically active conglomerates exhibiting non-negligible solid-state ECD spectra. The latter were correlated with the helicity deﬁned by the chiral dichalcogenide moiety [96]. For disulﬁdes, in particular, the sign of the n –σ ∗ band between 220 and 300 nm is related to the helicity of the C–S–S–C moiety by a quadrant rule [78, 97]. More recently, Kurt´an et al. [98] have reported the solution and solid-state ECD spectra of a series of diglycosyl disulﬁdes and diselenides to conﬁrm the applicability of the same rule for diselenides. In particular, TDDFT calculations were run on compound 31b, a truncated model of peracetylated diglucosyl diselenide (31a, Chart 6.3), whose X-ray structure was available. High-level QM calculations, including TDDFT and RI-CC2 method [99], were employed by Inoue and co-workers [100] to interpret the solid-state ECD spectrum of a nicotinamide derivative 32 (Chart 6.3). The authors were interested in investigating the effects of cation–π interactions in determining the conformation of two charged (pyridine N -methylated) derivatives of 32. As usual, the solid-state study was useful to restrict to a single conformer of 32 rather than dealing with an assembly of solution conformers. Stecko et al. [101] have used a mixed-type approach making use of solution and solid-state ECD spectra in connection with TDDFT calculations to rationalize the ECD spectra of a series of cycloadducts 33 derived from cyclic nitrones and 2(5H )-furanones. The authors intended to reconsider the applicability of a semiempirical ECD rule for lactones due to Legrand and Bocourt [102]. A parallel comparison was run between ECD spectra in solution versus the solid state, and between the conformation found in solution, studied by means of MM+ conformational searches and DFT geometry optimizations, versus the X-ray solid-state structure for crystalline compounds [101].

6.7. CONCLUSIONS AND PERSPECTIVES A simple inspection of the table of contents of the two volumes of the present treatise reveals immediately the current tremendous importance of quantum mechanics

244

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

calculations in the ﬁeld of chiroptical spectroscopy. Also instructive is a comparison with what can be considered the ﬁrst version of the treatise published in 2000. At that time, a chapter concerning theoretical approaches to electronic optical activity reported that TDDFT calculations had not yet been employed for optical activity or ECD calculations [103]. The great advance seen in the past 10 years is beyond doubt. It is also easy to anticipate that further technological development will make full ECD (or VCD or ROA) calculations suitable for much larger molecules, up to biopolymers. On augmenting the size and complexity of the molecular systems under investigation, the molecular ﬂexibility will also unavoidably increase. It is expected that fast and reliable tools for conformational analysis will be additionally improved in the future, taking also into account the role played by the solvent in a progressively more efﬁcient way. Nonetheless, the advantages offered by the solid state will remain valuable for the reasons discussed above, particularly in Section 6.2.1. Chiroptical spectra arise from a complicated combination of conformational and conﬁgurational factors, and while this represents the major advantage of chiroptical spectroscopies as a source of complete stereochemical information, it may be a drawback in many circumstances when conformational ambiguity should be avoided. This is especially true for the assignment of absolute conﬁgurations, as demonstrated by the solid-state ECD/TDDFT approach discussed in the present chapter, as well as when theoretical aspects are concerned, such as the interpretation of chiroptical spectra. In all these cases, the solid state offers the unique possibility of referring to a situation where a single and well-determined conformation occurs. In addition, a deep understanding of chirogenesis phenomena occurring in the solid state is still lacking and will certainly be the subject of future investigations necessarily based on solid-state chiroptical spectra. Note: The solid-state ECD/TDDFT protocol has been recently applied to determine the AC of (−)-massarigenin A, a bicyclic spiro compound extracted from Microsphaeropsis sp., an endophyte isolated from Arbutus unedo [104]. The impact of solid-state hydrogen bonds on solid-state ECD spectra has been investigated for massarigenin A [104] and hypothemycin (16) [105]. It has been concluded that speciﬁc interactions, such as hydrogen bonds, and through-space exciton couplings may both affect solid-state ECD spectra and be responsible for effects of comparable magnitude.

REFERENCES 1. D. J. Newman, M.C. Cragg, J. Nat. Prod . 2007, 70 , 461–477. 2. H. Caner, E. Groner, L. Levy, I. Agranat, Drug Discov. Today 2004, 9 , 105–110. 3. (a) EMEA, Investigation of chiral active substances, Report 3CC29a, 1993, 381–391. (b) C. Sahajwalla, Regulatory considerations in drug development of stereoisomers, in Chirality in Drug Design and Development, I. K. Reddy and R. Mehvar, eds., Marcel Dekker, New York, 2004, pp. 419–432. (c) C. P. Miller, J. W. Ullrich, Chirality 2008, 20 , 762–770. 4. J. M. Bijvoet, A. F. Peerdman, A. J. van Bommel, Nature 1951, 168 , 271–272. 5. (a) Y. Ren, W. Ke, Y. Li, L. Feng, J. Wan, X. Xu, J. Phys. Chem. B 2009, 113 , 10055–10058. (b) M.-R. Goldsmith, C. B. George, G. Zuber, R. Naaman, D. H. Waldeck, P. Wipf, D. N. Beratan, Phys. Chem. Chem. Phys. 2006, 8 , 63–67. 6. (a) N. Berova, L. Di Bari, G. Pescitelli, Chem. Soc. Rev . 2007, 36 , 914–931. (b) G. Pescitelli, L. Di Bari, N. Berova, Chem. Soc. Rev . 2011, 40 , 4603–4625. 7. G. Bringmann, T. A. M. Gulder, M. Reichert, T. Gulder, Chirality 2008, 20 , 628–642.

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

8. G. Bringmann, K. Maksimenka, T. Bruhn, M. Reichert, T. Harada, R. Kuroda, Tetrahedron 2009, 65 , 5720–5728. 9. G. Bifulco, P. Dambruoso, L. Gomez-Paloma, R. Riccio, Chem. Rev . 2007, 107 , 3744–3779. 10. R. H. Contreras, J. E. Peralta, Prog. Nucl. Magn. Reson. Spectrosc. 2000, 37 , 321–425. 11. (a) C. J. Cramer, Essentials of Computational Chemistry, 2nd ed., John Wiley & Sons, Chichester, 2004. (b) F. Jensen, Introduction to Computational Chemistry, 2nd ed., John Wiley & Sons, Chichester, 2007. 12. (a) G. Bringmann, K.-P. Gulden, H. Busse, J. Fleischhauer, B. Kramer, E. Zobel, Tetrahedron 1993, 49 , 3305–3312. (b) G. Bringmann, S. Busemann, The quantum chemical calculation of CD spectra: the absolute conﬁguration of chiral compounds from natural or synthetic origin, in Natural Product Analysis, M. H. P. Schreier, H.-U. Humpf, and W. Schwab, eds., Vieweg, Braunschweig, 1998, pp. 195–211. 13. (a) G. Bringmann, T. Bruhn, K. Maksimenka, Y. Hemberger, Eur. J. Org. Chem. 2009, 2717–2727. (b) J.-M. Gao, J.-C. Qin, G. Pescitelli, S. Di Pietro, Y.-T. Ma, A.-L. Zhang, Org. Biomol. Chem. 2010, 8 , 3543–3551. (c) C.-N. He, Y. Peng, L.-J. Xu, Z.-A. Liu, J. Gu, A.-G. Zhong, P.-G. Xiao, Chem. Pharm. Bull . 2010, 58 , 843–847. 14. (a) V. Bertin, H. Hussain, S. F. Kouam, E. Dongo, G. Pescitelli, P. Salvadori, T. Kurt´an, K. Krohn, Nat. Prod. Commun. 2008, 3 , 215–218. (b) G. Bringmann, M. Reichert, Y. Hemberger, Tetrahedron 2008, 64 , 515–521. (c) T. Grkovic, Y. Ding, X.-C. Li, V. L. Webb, D. Ferreira, B. R. Copp, J. Org. Chem. 2008, 73 , 9133–9136. (d) N. Cachet, G. GentaJouve, E. L. Regalado, R. Mokrini, P. Amade, G. Culioli, O. P. Thomas, J. Nat. Prod . 2009, 72 , 1612–1615. (e) N. Cachet, E. L. Regalado, G. Genta-Jouve, M. Mehiri, P. Amade, O. P. Thomas, Steroids 2009, 74 , 746–750. (f) S. Ding, L. Jia, A. Durandin, C. Crean, A. Kolbanovskiy, V. Shaﬁrovich, S. Broyde, N. E. Geacintov, Chem. Res. Toxicol . 2009, 22 , 1189–1193. (g) Y. Ding, X.-C. Li, D. Ferreira, J. Nat. Prod . 2009, 72 , 327–335. (h) H. Hussain, N. Akhtar, S. Draeger, B. Schulz, G. Pescitelli, P. Salvadori, S. Antus, T. Kurt´an, K. Krohn, Eur. J. Org. Chem. 2009, 749–756. (i) I. Kock, S. Draeger, B. Schulz, ´ Ken´ez, S. Antus, G. Pescitelli, P. Salvadori, J.-B. Speakman, B. Els¨asser, T. Kurt´an, A. J. Rheinheimer, K. Krohn, Eur. J. Org. Chem. 2009, 1427–1434. (j) Q. Zhu, C.-P. Tang, C.-Q. Ke, X.-Q. Li, J. Liu, L.-S. Gan, H.-C. Weiss, E.-R. Gesing, Y. Ye, J. Nat. Prod . 2009, 73 , 40–44. (k) M. Masullo, C. Bassarello, G. Bifulco, S. Piacente, Tetrahedron 2010, 66 , 139–145. (l) M. Woznica, P. Kowalska, R. Lysek, M. Masnyk, M. Gorecki, M. Kwit, F. Furche, J. Frelek, Curr. Org. Chem. 2010, 14 , 1022–1036. (m) G. Pescitelli, S. Di Pietro, C. Cardellicchio, M. A. M. Capozzi, L. Di Bari, J. Org. Chem. 2010, 75 , 1143–1154. (n) K. Matsumoto, K. Miki, T. Inagaki, T. Nehira, G. Pescitelli, Y. Hirao, H. Kurata, T. Kawase, T. Kubo, Chirality 2011, 23 , 543–548. (o) J. H. Cardellina, R. C. Vieira, V. Eccard, J. Skerry, V. Montgomery, Y. Campbell, V. Roxas-Duncan, W. Leister, C. A. LeClair, D. J. Maloney, D. Padula, G. Pescitelli, I. Khavrutskii, X. Hu, A. Wallqvist, L. A. Smith, ACS Med. Chem. Lett. 2011, 2 , 396–401. (p) I. Ahmed, H. Hussain, B. Schulz, S. Draeger, D. Padula, G. Pescitelli, T. van Ree, K. Krohn, Eur. J. Org. Chem. 2011, 2011 , 2867–2873. (q) I. N. Siddiqui, A. Zahoor, H. Hussain, I. Ahmed, V. U. Ahmad, D. Padula, S. Draeger, B. Schulz, K. Meier, M. Steinert, T. Kurt´an, U. Fl¨orke, G. Pescitelli, K. Krohn, J. Nat. Prod. 2011, 74 , 365–373. 15. S. Qin, K. Krohn, U. Fl¨orke, B. Schulz, S. Draeger, G. Pescitelli, P. Salvadori, S. Antus, T. Kurt´an, Eur. J. Org. Chem. 2009, 3279–3284. 16. Continuum Solvation Models in Chemical Physics: From Theory to Applications, B. Mennucci and R. Cammi, eds., John Wiley & Sons, Chichester, 2007. 17. (a) M. Dierksen, S. Grimme, J. Chem. Phys. 2006, 124 , 174301/1–174301/12; (b) N. Lin, F. Santoro, X. Zhao, A. Rizzo, V. Barone, J. Phys. Chem. A 2008, 112 , 12401–12411. 18. B. Els¨asser, K. Krohn, U. Fl¨orke, N. Root, H.-J. Aust, S. Draeger, B. Schulz, S. Antus, T. Kurt´an, Eur. J. Org. Chem. 2005, 4563–4570.

245

246

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

19. N. Harada, K. Nakanishi, Circular Dichroic Spectroscopy—Exciton Coupling in Organic Stereochemistry, University Science Books, Mill Valley, CA, 1983. 20. (a) T. A. Halgren, J. Comput. Chem. 1996, 17 , 490–519. (b) M. J. S. Dewar, E. G. Zoebisch, E. F. Healy, J. J. P. Stewart, J. Am. Chem. Soc. 1985, 107 , 3902–3909. 21. J. W. Downing, Programm package BDZDO/MCDSPD, Department of Chemistry and Biochemistry, University of Colorado, Boulder, USA; modiﬁed by J. Fleischhauer, W. Schleker, and B. Kramer; ported to LinuX by K.-P. Gulden. 22. G. Pescitelli, T. Kurt´an, U. Fl¨orke, K. Krohn, Chirality 2009, 21 , E181–E201. 23. See Chapters 20–23 in Volume 1. 24. C. Diedrich, S. Grimme, J. Phys. Chem. A 2003, 107 , 2524–2539. 25. T. D. Crawford, Theor. Chem. Acc. 2006, 115 , 227–245. 26. (a) J. Autschbach, Chirality 2009, 21 , E116–E152. (b) J. Autschbach, L. Nitsch-Velasquez, M. Rudolph, Top. Curr. Chem. 2011, 298 , 1–98. 27. L. Di Bari, G. Pescitelli, Electronic Circular Dichroism, in Computational Spectroscopy—Methods, Experiments and Applications, J. Grunenberg, ed., Wiley-VCH, Weinheim, 2010, pp. 241–277. 28. P. Elliott, K. Burke, F. Furche, Excited states from time-dependent density functional theory, in Reviews in Computational Chemistry, K. B. Lipkowitz and T. R. Cundari, eds., John Wiley & Sons, New York, 2008, pp. 91–166. 29. A. D. Becke, J. Chem. Phys. 1993, 98 , 5648–5652. 30. (a) A. D. Becke, Phys. Rev. A 1988, 38 , 3098–3100. (b) A. D. Becke, J. Chem. Phys. 1993, 98 , 1372–1377. 31. C. Adamo, V. Barone, J. Chem. Phys. 1999, 110 , 6158–6170. 32. L. Goerigk, S. Grimme, J. Phys. Chem. A 2009, 113 , 767–776. 33. (a) D. Jacquemin, E. A. Perpete, G. E. Scuseria, I. Cioﬁni, C. Adamo, J. Chem. Theory Comput. 2008, 4 , 123–135. (b) Y. Zhao, D. G. Truhlar, Acc. Chem. Res. 2008, 41 , 157–167. 34. D. Shcherbin, K. Ruud, Chem. Phys. 2008, 349 , 234–243. 35. M. E. Casida, C. Jamorski, K. C. Casida, D. R. Salahub, J. Chem. Phys. 1998, 108 , 4439–4449. 36. M. C. Zerner, Semiempirical molecular orbital methods, in Reviews in Computational Chemistry, K. B. Lipkowitz and T. R. Cundari, eds., John Wiley & Sons, New York, 1991, pp. 313–365. 37. S. F. Mason, Molecular Optical Activity and the Chiral Discrimination, Cambridge University Press, Cambridge, 1982. 38. (a) S. G. Telfer, N. Tajima, R. Kuroda, J. Am. Chem. Soc. 2004, 126 , 1408–1418. (b) S. G. Telfer, N. Tajima, R. Kuroda, M. Cantuel, C. Piguet, Inorg. Chem. 2004, 43 , 5302–5310. (c) N. Harada, Circular dichroism of twisted π -electron systems: theoretical determination of the absolute stereochemistry of natural products and chiral synthetic organic compounds, in Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, pp. 431–457. 39. J. Fabian, L. A. Diaz, G. Seifert, T. Niehaus, Theochem 2002, 594 , 41–53. 40. M. Pecul, K. Ruud, T. Helgaker, Chem. Phys. Lett. 2004, 338 , 110–119. 41. N. Berova, K. Nakanishi, Principles and applications of exciton chirality method, in Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, pp. 337–382. 42. See also Chapter 4 in the present volume. 43. (a) H. DeVoe, J. Chem. Phys. 1964, 41 , 393–400. (b) H. DeVoe, J. Chem. Phys. 1965, 43 , 3199–208. 44. (a) C. Rosini, M. Zandomeneghi, P. Salvadori, Tetrahedron Asymm. 1993, 4 , 545–54. (b)

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

45.

46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58.

59. 60. 61. 62. 63.

64. 65. 66.

67. 68. 69. 70.

L. Di Bari, G. Pescitelli, G. Reginato, P. Salvadori, Chirality 2001, 13 , 548–555. (c) G. Pescitelli, S. Gabriel, Y. Wang, J. Fleischhauer, R. W. Woody, N. Berova, J. Am. Chem. Soc. 2003, 125 , 7613–7628. (d) S. Superchi, E. Giorgio, C. Rosini, Chirality 2004, 16 , 422–451. R. Kuroda, Solid-state CD: Application to inorganic and organic chemistry, in Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, pp. 159–184. R. Kuroda, T. Honma, Chirality 2000, 12 , 269–277. G. M. Sheldrick, Acta Crystallogr. 2008, A64 , 112–122. W. R. Busing, H. A. Levy, Acta Crystallogr. 1964, 17 , 142–146. K. Krohn, Zia-Ullah, H. Hussain, U. Fl¨orke, B. Schulz, S. Draeger, G. Pescitelli, P. Salvadori, S. Antus, T. Kurt´an, Chirality 2007, 19 , 464–470. J. Dai, K. Krohn, B. Els¨asser, U. Fl¨orke, S. Draeger, B. Schulz, G. Pescitelli, P. Salvadori, S. Antus, T. Kurt´an, Eur. J. Org. Chem. 2007, 4845–4854. G. Pescitelli, N. Sreerama, P. Salvadori, K. Nakanishi, N. Berova, R. W. Woody, J. Am. Chem. Soc. 2008, 130 , 6170–6181. A. Schafer, C. Huber, R. Ahlrichs, J. Chem. Phys. 1994, 100 , 5829–5835. R. A. Kendall, T. H. Dunning, Jr., R. J. Harrison, J. Chem. Phys. 1992, 96 , 6796–806. A. Canal Neto, F. E. Jorge, Chirality 2007, 19 , 67–73. P. J. Stephens, N. Harada, Chirality 2009, 22 , 229–233. R. Kuroda, Circular dichroism in the solid state, in Chiral Photochemistry, Y. Inoue and V. Ramamurthy, eds., Marcel Dekker, New York, 2004, pp. 385–413. E. Castiglioni, P. Biscarini, S. Abbate, Chirality 2009, 21 , E28–E36. (a) B. Bosnich, J. M. Harrowﬁeld, J. Am. Chem. Soc. 1972, 94 , 3425–3438. (b) L. Ding, L. Lin, C. Liu, H. Li, A. Qin, Y. Liu, L. Song, H. Zhang, B. Z. Tang, New J. Chem. 2011, 35 , 1781–1786. (a) R. Kuroda, Y. Saito, Bull. Chem. Soc. Jpn. 1976, 49 , 433–436. (b) F. Formaggio, M. Crisma, C. Toniolo, J. Kamphuis, Biopolymers 1996, 38 , 301–304. E. Castiglioni, S. Abbate, G. Longhi, R. Gangemi, Chirality 2007, 19 , 491–496. R. Kuroda, T. Harada, Y. Shindo, Rev. Sci. Instrum. 2001, 72 , 3802–3810. Y. Shindo, M. Nakagawa, Y. Ohmi, Appl. Spectrosc. 1985, 39 , 860–868. (a) F. H. Allen, Acta Crystallogr. 2002, B58 , 380–388. (b) I. J. Bruno, J. C. Cole, P. R. Edgington, M. Kessler, C. F. Macrae, P. McCabe, J. Pearson, R. Taylor, Acta Crystallogr. 2002, B58 , 389–397. (a) P. G. Jones, Acta Crystallogr. 1984, A40 , 663–668. (b) H. D. Flack, G. Bernardinelli, Chirality 2008, 20 , 681–690. Dictionary of Natural Products, v. 18.2, Taylor & Francis, 2010, http://dnp.chemnetbase. com/. (a) Y.-W. Guo, T. Kurt´an, K. Krohn, G. Pescitelli, W. Zhang, Chirality 2009, 21 , 561–568. (b) K. Krohn, I. Kock, B. Elisaesser, U. Fl¨orke, B. Schulz, S. Draeger, G. Pescitelli, S. Antus, T. Kurt´an, Eur. J. Org. Chem. 2007, 1123–1129. Z. Hassan, H. Hussain, V. U. Ahmad, S. Anjum, G. Pescitelli, T. Kurt´an, K. Krohn, Tetrahedron Asymm. 2008, 18 , 2905–2909. H. Hussain, K. Krohn, U. Fl¨orke, B. Schulz, S. Draeger, G. Pescitelli, S. Antus, T. Kurt´an, Eur. J. Org. Chem. 2007, 292–295. H. Hussain, K. Krohn, U. Fl¨orke, B. Schulz, S. Draeger, G. Pescitelli, P. Salvadori, S. Antus, T. Kurt´an, Tetrahedron Asymm. 2007, 18 , 925–930. K. Krohn, U. Farooq, U. Fl¨orke, B. Schulz, S. Draeger, G. Pescitelli, P. Salvadori, S. Antus, T. Kurt´an, Eur. J. Org. Chem. 2007, 3206–3211.

247

248

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

71. W. Zhang, K. Krohn, J. Ding, Z.-H. Miao, X.-H. Zhou, S.-H. Chen, G. Pescitelli, P. Salvadori, T. Kurt´an, Y.-W. Guo, J. Nat. Prod . 2008, 71 , 961–966. 72. W. Zhang, K. Krohn, Zia-Ullah, U. Fl¨orke, G. Pescitelli, L. Di Bari, S. Antus, T. Kurt´an, J. Rheinheimer, S. Draeger, B. Schulz, Chem. Eur. J . 2008, 14 , 4913–4923. 73. K. Krohn, S. F. Kouam, G. M. Kuigoua, H. Hussain, S. Claudius-Brand, U. Fl¨orke, T. Kurt´an, G. Pescitelli, L. Di Bari, S. Dr¨ager, B. Schulz, Chem. Eur. J . 2009, 15 , 12121–12132. 74. M. Saleem, H. Hussain, I. Ahmed, S. Draeger, B. Schulz, G. Pescitelli, T. Kurt´an, U. Fl¨orke, K. Karsten, Eur. J. Org. Chem. 2011, 808–812. 75. J. Dai, K. Krohn, U. Fl¨orke, G. Pescitelli, G. Kerti, T. Papp, K. E. K¨ov´er, A. C. B´enyei, S. Draeger, B. Schulz, T. Kurt´an, Eur. J. Org. Chem. 2010, 6928–6937. 76. J. Dai, H. Hussain, S. Draeger, B. Schulz, T. Kurt´an, G. Pescitelli, U. Fl¨orke, K. Krohn, Nat. Prod. Commun. 2010, 5 , 1175–1180. 77. E. M. K. Wijeratne, T. J. Turbyville, A. Fritz, L. Whitesell, A. A. L. Gunatilaka, Bioorg. Med. Chem. 2006, 14 , 7917–7923. 78. G. Snatzke, Angew. Chem. Int. Ed. Engl . 1979, 18 , 363–377. 79. (a) L. A. Naﬁe, Nat. Prod. Commun. 2008, 3 , 451–466. (b) T. B. Freedman, X. Cao, R. K. Dukor, L. A. Naﬁe, Chirality 2003, 15 , 743–758. (c) P. J. Stephens, F. J. Devlin, J.J. Pan, Chirality 2008, 20 , 643–663. (d) See also Chapter 11 in the present volume, and see Chapter 24 in Volume 1. 80. P. L. Polavarapu, Chirality 2008, 20 , 664–672. 81. (a) G. Hans, Helv. Chim. Acta 1977, 60 , 3039–3044. (b) A. Kobayashi, S. Yata, T. Hino, K. Kawazu, Agr. Biol. Chem. 1987, 51 , 2857–2860. 82. B. Franck, H. Flasch, Die ergochrome (Physiologie, isolierung, struktur und biosyntheses), in Progress in the Chemistry of Organic Natural Products, W. Herz, H. Grisebach, G. W. Kirb, eds., Springer-Verlag, New York, 1973. 83. W. Zhang, Y.-W. Guo, K. Krohn, Chem. Eur. J . 2006, 12 , 5122–5127. 84. G. Snatzke, H. Ripperger, C. Horstmann, K. Schreiber, Tetrahedron 1966, 22 , 3103–3116. 85. K. Krohn, C. Biele, K.-H. Drogies, K. Steingr¨over, H.-J. Aust, S. Draeger, B. Schulz, Eur. J. Org. Chem. 2002, 2331–2336. 86. R. Shan, H. Anke, M. Stadler, O. Sterner, Tetrahedron 1996, 52 , 10249–10254. 87. V. V. Borovkov, T. Harada, Y. Inoue, R. Kuroda, Angew. Chem. Int. Ed . 2002, 41 , 1378–1381. 88. T. Agatsuma, A. Takahashi, C. Kabuto, S. Nozoe, Chem. Pharm. Bull . 1993, 41 , 373–375. 89. (a) C.-P. Hsu, G. R. Fleming, M. Head-Gordon, T. Head-Gordon, J. Chem. Phys. 2001, 114 , 3065–3072. (b) B. Mennucci, J. Tomasi, R. Cammi, Phys. Rev. B 2004, 70 , 205212/1–11. (c) M. F. Iozzi, B. Mennucci, J. Tomasi, R. Cammi, J. Chem. Phys. 2004, 120 , 7029–7040. (d) F. Lipparini, B. Mennucci, J. Chem. Phys. 2007, 127 , 144706/1–6. 90. G. Kerti, T. Kurt´an, A. Borb´as, Z. B. Szab´o, A. Lipt´ak, L. Szil´agyi, Z. Illy´es-Tuende, A. B´enyei, S. Antus, M. Watanabe, E. Castiglioni, G. Pescitelli, P. Salvadori, Tetrahedron 2008, 64 , 1676–1688. 91. G. Pescitelli, D. Padula, T. Kurt´an, unpublished results. 92. A. Sch¨afer, H. Horn, R. Ahlrichs, J. Chem. Phys. 1992, 97 , 2571–2577. 93. See also Chapter 4 in Volume 1. 94. (a) J. D. Chisholm, J. Golik, B. Krishnan, J. A. Matson, D. L. Van Vranken, J. Am. Chem. Soc. 1999, 121 , 3801–3802; (b) L. Alcaraz, G. Macdonald, J. P. Ragot, N. Lewis, R. J. K. Taylor, J. Org. Chem. 1998, 63 , 3526–3527; (c) S. Matile, N. Berova, K. Nakanishi, J. Fleischhauer, R. W. Woody, J. Am. Chem. Soc. 1996, 118 , 5198–5206; (d) N. Berova, D. Gargiulo, F. Derguini, K. Nakanishi, N. Harada, J. Am. Chem. Soc. 1993, 115 , 4769–4775.

A B S O L U T E C O N F I G U R AT I O N O F N AT U R A L P R O D U C T S B Y S O L I D - S TAT E E C D

95. J.-B. Hou, G. Tang, J.-N. Guo, Y. Liu, H. Zhang, Y.-F. Zhao, Tetrahedron Asymm. 2009, 20 , 1301–1307. 96. T. Shimizu, H. Isono, M. Yasui, F. Iwasaki, N. Kamigata, Org. Lett. 2001, 3 , 3639–3641. 97. (a) J. Linderberg, J. Michl, J. Am. Chem. Soc. 1970, 92 , 2619–2625. (b) R. W. Woody, Tetrahedron 1973, 29 , 1273–1283. 98. T. Kurt´an, G. Pescitelli, P. Salvadori, A. Ken´ez, S. Antus, L. Szil´agyi, T.-Z. Illy´es, I. Szab´o, Chirality 2008, 20 , 379–385. 99. C. H¨attig, A. K¨ohn, J. Chem. Phys. 2002, 117 , 6939–6951. 100. A. Shimizu, T. Mori, Y. Inoue, S. Yamada, J. Phys. Chem. A 2009, 113 , 8754–8764. 101. S. Stecko, J. Frelek, M. Chmielewski, Tetrahedron Asymm. 2009, 20 , 1778–1790. 102. M. Legrand, R. Bocourt, Bull. Soc. Chim. Fr. 1967, 2241–2242. 103. A. Koslowski, N. Sreerama, R. W. Woody, Theoretical approach to electronic optical activity, in Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, pp. 55–95. 104. H. Hussain, I. Ahmed, B. Schulz, S. Draeger, U. Fl¨orke, G. Pescitelli, K. Krohn, Chirality. 2011, 23 , 617–623. 105. G. Pescitelli, Chirality 2012, accepted for publication.

249

7 DYNAMIC STEREOCHEMISTRY AND CHIROPTICAL SPECTROSCOPY OF METALLO-ORGANIC COMPOUNDS James W. Canary and Zhaohua Dai 7.1. INTRODUCTION This chapter aims to give a current overview of the dynamic stereochemistry of metalbased molecular switches together with the role that chiroptical spectroscopy has played in the development of the ﬁeld. Recent interest and the development of chiroptical spectroscopy tools have led to many exciting discoveries in the dynamic stereochemistry of metal-based chiroptical switches. These materials generally involve a scaffold built upon a metal-chelate complex that is capable of changing its interaction with polarized light upon exposure to an external stimulus. Such molecular switches could play a key role in the development of future optical displays, molecular electronics, and telecommunications materials. Many metal-based chiroptical switches have been synthesized within the past decade [1–6]. Most of these systems include a transition metal chelated to multiple atoms of an associated chiral ligand. The conformational state of the chelation complex is altered by a multitude of triggering agents such as photochemical irradiation, reducing–oxidizing agents, counterion exchange, and others. Regardless of which trigger is used, the interconversion between states must be efﬁcient, selective, and reversible. The mechanism may involve structural changes induced by the metal ion directly, such as in redox-triggered molecular switches, or the metal may play a structural role or participate in generation of the readout signal. The means of analysis of these metal-based chiroptical switches is also critical. A nondestructive read-out of the optically active system must be employed in order to gain accurate information about the different states of the switch. Thus, the tools used to accurately assess switching may be limited. Indeed, the presence of the metal may offer limitations on read-out, such as quenching ﬂuorescence of the organic ligand, or may offer new opportunities such as presenting additional chromophoric elements that can be monitored. Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

251

252

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Chiroptical spectroscopic measurements are powerful tools for the characterization of materials in this ﬁeld [6]. All chiroptical methods are based on the different interaction between an optically active compound and the left- and right-circularly polarized vector components of plane polarized, monochromatic light. These components interact with a chiral medium in two ways: (a) Difference in velocity through the medium results in a circular birefringence or anisotropic refraction (nL − nR = 0), which is observed as a rotation of the plane of polarization. (b) Difference in absorption by the medium results in a circular dichroic effect or anisotropic absorption (AL − AR = 0, or ε = 0). Anisotropic refraction has been utilized in polarimetry for the determination of optical rotation at a speciﬁc wavelength [7]. The rotatory power of a substance varies with the wavelength. The need to measure the optical rotation at different wavelengths led to instruments for recording the optical rotatory dispersion (ORD) over the UV–vis range, which gives information about the change in sign of the rotation on passage of the absorption band, the Cotton effect (CE). More recently, ORD has largely been replaced by electronic circular dichroism as an experimental technique. Circular dichroism (CD) is the most commonly used method due to the relatively strong information content of this technique as compared, for example, to single-wavelength optical rotation measurement and the widespread availability of sensitive CD instruments. Optical activity is strongly dependent on conformational equilibria, and, therefore, CD and ORD spectroscopy are two techniques that can give telling information on the spatial arrangement of molecules in solution. Various mechanisms may give rise to CD, including induced (ICD) and CD exciton chirality. A typical CD spectrum will appear similar to the corresponding UV–vis absorbance spectrum, with amplitude that depends on the strength of the electronic absorbance, the asymmetric distribution of electrons in the molecule, and the coupling between these two phenomena. This typical spectroscopy involving electronic transitions is generally called electronic circular dichroism (ECD), which includes electronic near infrared CD (NIR CD). Vibrational circular dichroism (VCD) and near-infrared vibrational circular dichroism (NIR VCD), which involve vibration transitions, have recently received considerable attention [8–15]. Unless the spectrum can be calculated explicitly, the features of an induced CD spectrum are difﬁcult to interpret on the basis of structural changes other than by comparing to similar systems. An exception is CD exciton chirality, which is particularly useful for solution studies [6, 16] and is covered extensively elsewhere in this volume. Few other spectroscopic signals report structural aspects of molecular conformation or intermolecular association so dramatically. By this feature, CD exciton chirality has been used as a sensitive probe of changes in molecular conformation or intermolecular association. If the metal-based chiroptical switch is ﬂuorescent, it can be investigated by ﬂuorescence-detected circular dichroism (FDCD) or differential circularly polarized ﬂuorescence excitation (CPE) [17], which is based on detection in emission, while conventional CD measures the intensity of transmitted light. Such metal-based switches can also be studied by circular polarized luminescence (CPL), which is an emission analogue of circular dichroism and probes chirality of electronically excited species [18]. The character of the chiroptical signal is only one aspect of chiroptical switches. In general, requirements for the design of an efﬁcient metal-based chiroptical molecular switch include: (i) stability of the optically active forms, (ii) chemical reversibility of the switching processes, (iii) high sensitivity of the chiroptical response, and (iv) potential application in multimode switching. There are many systems that achieve these aims. This chapter will be organized by the chiroptical responses of metal-base dynamic stereochemical systems: optical rotation/ORD, ECD, VCD, FDCD/CPE, and CPL. Since

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

the majority of the systems have been studied mainly by ECD, such systems will be suborganized by input mechanism: First, we will discuss selected examples of metal association and dissociation, which is a seemingly trivial means of triggering chiroptical response but for which several unique and interesting examples will be discussed. Changes in the environment of metallo-organic complexes including solvent, temperature, counterion, and pressure will be discussed. We will mention brieﬂy the Pfeiffer effect, a classic phenomenon that was not originally intended as a chiroptical switch yet involves chemistry that is highly relevant to metal-based chiroptical switches. Redox chemistry is particularly suited to metal-based systems, and therefore redox-triggered chiroptical molecular switches will be discussed at length. Fewer examples of optically induced metal-based chiroptical switches have been reported, and these will also be presented. Dynamic stereochemistry in the solid state will also be discussed. CD has more often been used to obtain stereochemical information from solutions, while its utility has so far remained limited in the solid state or aggregates. Methods for obtaining solid-state spectra are currently under development, but interpretation of CD data obtained from oriented media is very challenging and subject to error [19]. We will present examples of metal-based chiroptical switches in solid state, although their CD spectra might have been obtained in solution or in polycrystalline form.

7.2. METAL-BINDING INDUCED SWITCHES STUDIED BY OPTICAL ROTATION AND ORD As mentioned, optical rotation and ORD were available much earlier than other forms of chiroptical spectroscopy. In the absence of modern computational methods, limited structural conclusions were available from such data. However, very remarkable chemistry related to metal-based chiroptical switches was examined using these techniques. Several examples will be discussed here.

7.2.1. Foundational Studies: Octahedral Complexes of Transition Metals The ability to turn “off” a chiroptical signal by changing the environment was known from the earliest days of coordination chemistry and established by monitoring optical rotation. One of the two enantiomers of Co(en)3 3+ (where en = NH2 CH2 CH2 NH2 ) that Werner [20] resolved in 1912 rotates plane polarized light from the sodium D line (589.3 nm) toward the right while the other rotates the light by the same amount in the opposite direction. Plotting the values of [α]λ of (+)Co(en)3 3+ as a function of wavelength results in its ORD curve. Since the resolved enantiomer of Co(en)3 3+ may be racemized by boiling an aqueous solution of one of the enantiomers in the presence of activated charcoal, this dynamic stereochemistry was originally monitored by optical rotation or ORD. In this context, it should be mentioned that ORD and ECD are not truly independent methods [21] and theoretical ORD and ECD spectra can be obtained in a single quantum mechanical calculation [22]. Thus, in a sense, the ability to trigger “off” a metal-based switched was known from the beginning of coordination chemistry in part because optical rotation was available at this early time as a technique that could report the racemization reaction.

253

254

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

7.2.2. Metal-Based Chiroptical Switches Controlled by Polarized Light Optical rotation was used to show that partial optical resolution can be achieved by enriching one enantiomer when irradiating a racemic mixture with right- or left-circularly polarized light. This phenomenon is the consequence of differential absorptivity of circularly polarized light by enantiomers, resulting in the preferential isomerization of one enantiomer over the other. Yoneda et al. [23] used an argon ion laser to partially photoresolve tris(acetylacetonato)chromium (III) (Cr(acac)3 ) and related complexes. Linearly polarized light emitted from the laser was passed through a quarter-waveplate matched for the 514.5 nm wavelength of the laser to produce circularly polarized light. This process was monitored by following optical rotation of the sample as a function of time (Figure 7.1). At time t = 100 min, the sense of the circularly polarized light was inverted by rotating the quarter-waveplate 90◦ . The data clearly show that the photoresolution process is switchable. A 4.8% resolution was achieved and no measurable photodecomposition occurred.

20

α(10–3 °)

10

0 100

200 (min)

–10

–20

Figure 7.1. The time-dependence of the photoinversion of Cr(acac)3 [23]. (Reproduced by permission of the Chemical Society of Japan.)

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

255

200 pH = 6.7 150

L-His + ZnCl2 (1:1) L-His + ZnCl2 + EDTA (1:1:2)

[M] x 10–2

100

L-His Methyl Ester + ZnCl2 (1:1)

50

0

–50

–100

–150

Figure 7.2. Chelation of Zn(II) by histidine 200

220

240

260 λ (mm)

280

300

320

studied by ORD [24]. (Reproduced by permission of the American Chemical Society.)

7.2.3. Conformational Studies by Optical Rotatory Dispersion: L-Histidine Chelation In the context of studies probing the effect of metal ions in biological systems, ORD was used to show that metal ions induce a change in conformation of l-histidine upon chelation [24]. l-Histidine itself exhibits a positive Cotton effect in the absence of metal ions. Upon chelation with transition metals such as Zn(I1) (Figure 7.2) Co(II), Ni(II), and Cu(II), a negative Cotton effect of greater amplitude appeared. Addition of two equivalents of ethylenediaminetetraacetate (EDTA) restored the optical rotatory dispersion (ORD) curve of free l-histidine. Chelation with metal ions resulted in a change in conformation that was highly improbable for the free l-histidine in solution. These observations are consistent with the well-known sensitivity of ORD to conformation, which is highly responsive to metal chelation.

7.2.4. Metal-Binding Induced Dynamic Stereochemistry in a Biopolymer Monitored by Optical Rotation and Optical Rotatory Dispersion In another study of the effect of metals on biological molecules, ORD spectroscopy was used to study the conformational behavior of biopolymers. The tryptic peptide comprising residues 38–61 of oxidized bovine pancreatic ribonuclease A was isolated and shown to be convertible to a helical conformation in 2-chloroethanol- or 2,2,2triﬂuoroethanol–water mixtures [25]. The characteristic trough in the optical rotatory dispersion spectrum near 233 nm associated with the helical conformation was progressively reduced by addition of each of 3 eq of Cu2+ (Figure 7.3). The depth of the trough was regained quantitatively by addition of an equivalent amount of EDTA. The absorption

256

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

4

+1000 0

3 2

γ

–1000 –2000 –3000

Figure 7.3. Effect of stepwise addition of Cu2+

–4000

on ORD curves for peptide dissolved in 50% aqueous 2,2,2-triﬂuoroethanol. Curve 1, curve for the peptide at an apparent pH of 10.0; curves 2, 3, and 4, results after addition of 1, 2, and 3 equivalents of Cu2+ . •, values obtained after

1

–5000

220

240

260

280 300 λ (mm)

320

340

360

addition of 3 moles of EDTA [25]. (Reproduced by permission of the American Society for Chemistry and Molecular Biology.)

and optical rotatory spectra of the cupric peptide complexes were qualitatively similar to those of some smaller peptides in which chelation by amide nitrogen atoms of peptide bonds has been demonstrated. The effect of Cu2+ binding on the helix content was due to the conformational incompatibility of the chelate complexes with the α-helix.

7.3. METAL-BASED SWITCHES MONITORED BY ELECTRONIC CIRCULAR DICHROISM (ECD) ECD has been the dominant chiroptical technique used to probe metal-based switches in recent times. With the availability of many examples, this section will be organized by the triggering method, and the role of ECD in characterizing the dynamic stereochemical phenomena will then be discussed in context.

7.3.1. Metal-Binding Induced Dynamic Stereochemistry 7.3.1.1. Metal-Binding-Induced Dynamic Stereochemistry in Biological Polymers. In a more recent study of metal ions interacting with biological polymers, the Woolfson group designed and characterized a peptide that reversibly switches between a trimeric α-helical coiled coil and a zinc-bound folded monomer [26]. Koksch and coworkers [27] constructed a simple system to study of the impact of different metal ions on peptide secondary structures. The design was based on an α-helical coiled coil peptide. Histidine mutations were incorporated into the heptad repeat (Figure 7.4) to generate possible complexation sites in the target peptide CCM for Cu2+ and Zn2+ ions, which was shown to be a strong trigger for a secondary structure switch from an α-helix (structure 1) to a β-sheet (structure 2). The capture of Cu2+ and Zn2+ ions by EDTA reversed the β-sheet formation to an α-helical structure, as revealed by CD spectra. Therefore, switching between α-helix and β-sheet was easily controlled in both directions.

257

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

L L L L

S E H E

V

V K H K

L L L L

E H E S

EDTA

L L L L

(1)

V H V

E H E S

L

V L

L

V L V H H H L V L

V H H H

CCM CCM + 1 eq. Cu 2+ 2+ CCM + 1 eq. Cu + 1 eq. EDTA

40 30 20 10 0 –10 –20 –30 50

Cu2+/Zn2+

E E K L K L H E E K L K

Abz

Abz

50

K H K

[θ] (103 deg cm2 dmol–1)

L L L L

L

L V L V S S V S S

E E K L K L H E E K L K L

(2)

[θ] (103 deg cm2 dmol–1)

S E H E

V H V

CCM 2+ CCM + 1 eq. Zn 2+ CCM + 1 eq. Zn + 1 eq. EDTA

40 30 20 10 0 –10 –20 –30 190

(a)

200

210

230

220

240

λ (nm) (b)

Figure 7.4. CD spectra of peptide CCM in 40% TFE at pH 7.4 and at a peptide concentration of 0.1 mM with (a) 0.1 mM CuCl2 /EDTA and (b) 0.1 mM Zn(OAc)2 /EDTA [27]. (Reproduced by permission of the Royal Society of Chemistry.)

H O

t-BuO

O EtO

O

Ba2+ (2.5 e.q.) Ba2+ (1.3 e.q.) Ba2+ (0.6 e.q.) no guest

2 1

OEt

m n

3

300 350 400 450 Wavelength (nm)

0 500

[q]2nd x 10–3 (degree cm2 dmol–1)

y

x

H

10 8 6 4 2 0 –2 –4 –6 –8

e x 10–4 (cm–1 M–1)

[q] x 10–3 (degree cm2 dmol–1)

7.3.1.2. Metal-Binding-Induced Dynamic Stereochemistry in Synthetic Polymers.. Chiroptical response to metal ion binding has also been studied in synthetic polymer systems. Kakuchi and co-workers [28] reported that a poly(phenylacetylene) bearing a polycarbohydrate ionophore as a graft chain, copolymer 3, showed a split-type circular dichroism (CD) in the long absorption region of the conjugated polymer backbone (280–500 nm) which varied in response to different metal ions (Figure 7.5). The CD of 3 strongly depended on the radii of the guest metal cations. A metal cation whose ˚ or greater, such as Ba2+ , decreased the amplitude and eventuionic radius was 1.16 A ˚ such as Co2+ , only ally caused an inversion while those with radii smaller than 1.16 A, decreased the amplitude without inversion. This suggested that copolymer 3 underwent a helix–helix transition through the host–guest complexation with achiral inorganic metal cations.

6

Ba2+ Pb2+

4 Li+ 2 0 –2

Na+

Mg2+

Sr2+

Zn2+ Cu2+ Ca2+ Cd2+ Co2+ 0.8

Ni

2+

1.0

1.2

1.4

Ionic radius (Å)

Figure 7.5. Middle: CD (upper) and UV (lower) spectra of 3 with Ba(ClO4 )2 in 3/1 CHCl3 /acetonitrile at −30◦ C. Right: Plots of [θ ]2nd versus the metal-cation radii [28]. (Reproduced by permission of John Wiley & Sons.)

258

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

O

5

O 0

SiMe2 Me2S n 4

Absorbance

4

+

Li Na+ K+

2

0 250

θ (mdeg) Absorbance

O

O

2

(b)

5 CD

6

0 CD Cu2+ Zn2+

–5

θ (mdeg)

(a)

–2

ABS

ABS 300

350

Wavelength (nm)

–10 400

0 250

300

350

400

Wavelength (nm)

Figure 7.6. Absorption and CD spectra of 4 in dichloromethane in the presence of metal ions: (a) Li(OTf) (dotted line), Na(OTf) (dashed line), and K(OTf) (solid line). (b) Cu(OTf)2 (solid line) and Zn(OTf)2 (dotted line) [29]. (Reproduced by permission of the American Chemical Society.)

Sanji, Tanaka, and co-workers [29] reported that the σ(Si-–Si) –π conjugated poly(mphenylenedisilanylene) with a chiral tri(ethylene glycol) side chain, 4, displays optical activity arising from metal-coordination interactions (Figure 7.6). Interaction of the side chains with the metal ions is sensitive to the size and the coordination mode of the metal ions, selectively inducing twist sense bias of the helical conformation to display optical activity. Li+ induced a positive CE, while complexation with Na+ or K+ gave a negative CE. Very little CE was observed upon addition of Mg2+ or Ca2+ and a negative CE appeared when complexed with Cu2+ or Zn2+ . The metal-induced chirality in polymer 4 is reversible: No CD signals were observed when the solution was washed with water to remove the metal ions. 7.3.1.3. Electron-Switched Supramolecular Chiral Polythiophene Aggregates. In a system showing supramolecular complexity, Goto and Yashima [30] reported electron-switched supramolecular chiral polythiophene aggregates. Polythiophene oligomers with side chains containing remote chiral centers form chiral aggregates in certain solvents that exhibit induced circular dichroism (ICD) in the π –π * transition region. Addition of a Cu(II) salt resulted in oxidative doping of the polymer main chain and the disappearance of CD. Further addition of amines such as triethylenetetramine removed the doping by shifting the oxidation potential of the copper, extracting electrons from the polythiophene molecules and consequently regenerating the ICD signal. Doping also induced color and morphological changes as measured by electronic absorbance and atomic force microscopy. Besides using chirality to probe fundamental interstrand interaction phenomena [31], chiral oligothiophenes have been examined for applications involving circular polarized electroluminescence [32] as well as enantioselective sensors, electrodes, catalysts, and adsorbents [33, 34]. Electron-induced triggering of polythiophenes would bring new dimensions to such devices if adaptable in the solid state.

7.3.2. Environment-Induced Switches Chiroptical properties of metal complexes respond to a startling variety of environmental triggers including counterion, temperature, solvent, and pressure. In 1931, Pfeiffer and Quehl [35] reported that the optical rotation of a solution of an optically active compound (the “environment” compound) changes upon the addition of racemic mixtures of some

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

optically active coordination compounds, which is generally referred to as the “Pfeiffer effect.” This effect has been attributed to the observed shift of enantiomeric equilibrium due to an outer-sphere association of the racemic metal complex and the chiral additive. The perturbation of the equilibrium leads to an unequal diastereomer occupation [36]. We will not discuss this effect in detail here, but we will cite several recent environmentinduced stereodynamic systems that are conceptually related to the Pfeiffer effect. 7.3.2.1. Guest-Controlled Tripodal Ligand Chirality. Supramolecular conformational bias based on host–guest complexes includes nonempirical CD approaches for the determination of the absolute conﬁguration of primary amines [37]. Complexation of the conformational racemate [Cu(BQPA)](ClO4 )2 (BQPA = N , N -bis[2-quinolyl]methylN -[2-pyridyl]methylamine) [38] with various chiral amines gave ECCD spectra that were consistent with formation of a preferred conformational diastereomer. The zinc(II) complexes with solvent coordinated as the ﬁfth ligand exhibit conformational enantiomerism, with the chirality of the molecule originating from the helical structure of the ligand when associated with the metal ion. Introduction of a chiral guest molecule may displace the solvent and create diastereomers. It was observed that left-handed guests induced a lefthanded propeller conformation in the ligand, yielding a positive couplet in the CD spectra [39]. The details of the CD spectral assignment were elucidated with chiral ligands and are discussed later in this chapter. 7.3.2.2. Zinc Porphyrin Tweezers. Much more sensitive chirality “sensing” was achieved by forming supramolecular complexes of primary chiral amines with pentanediol-linked zinc porphyrins, referred to as a “zinc porphyrin tweezer” [16, 40, 41], as shown in Figure 7.7. Binding occurs ﬁrst between the chiral primary amine and carrier molecules, which, in turn, bind to the bidentate zinc porphyrin. The resulting 1:1 host–guest complex generates an ECCD spectrum. Based on the sign of the couplet in the split CD spectrum, the group order (small, medium, large) in the Newman projection is determined and the absolute conﬁguration at the chiral carbon of the monoamine can be resolved. This approach offers a sensitive, widely applicable nonempirical advancement for probing chirality of organic compounds. A variety of substrates including aromatic amines, cyclic and acyclic amines, amino esters, amides, and cyclic amino alcohols have been examined by this and similar approaches, and the method has been generalized to substrates with various potential points of ligation, including monodentate compounds [42]. A porphyrin dimer with a shorter ethane bridge linker was reported to display pronounced CD spectra in the presence of monofunctional amines [43, 44]. Supramolecular chirality induction occurs when amines bind to the “face-to-face” syn conformation of the achiral bis(zinc porphyrin) resulting in the extended anti conformer with an increased helical dislocation. These two forms are easily distinguished spectroscopically. Left- and right-handed screw twists of the bis(zinc porphyrin)-amine diastereomers are due to the absolute conﬁguration of the amines. Several amines with the (R) absolute conﬁguration induced ECCD spectra with negative couplets, and amines with (S ) absolute conﬁguration induced the opposite handedness. It should be noted, though, that this relationship does not necessarily follow since R, S conﬁgurational assignment follows the CIP-Rule where the atomic number of the substituent attached to the stereogenic center determines R or S assignment. However, it is steric factors that determine the interpophyrin helicity. The bulkier amines resulted in stronger CD signals, conﬁrming the role of steric factors in the mechanism of chiral induction.

259

260

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(R)

(R) (R)

(R) (R)

(R)

O O

O O

NH Zn

Zn

H

+

NH2

N

Zn

H

O

O

O

Zn

O

NH2 O

O

O

O

Conjugate/tweezer complex 15 10

N

N Zn N

+

20 15′

N

5

–

10′

N

N Zn N N

20′

>>>

5′ O

O

O

O

Favored conformation I

Unfavored conformation II

Predicted positive CD exciton couplet

Predicted negative CD exciton couplet

Figure 7.7. Macrocyclic 1:1 host–guest complex formed between guest and host. (Reproduced by permission of The Royal Society of Chemistry [16, 41].)

7.3.2.3. Practical Application of the Pfeiffer Effect for Analyzing Chiral Diamines. Recently, Anslyn took advantage of Pfeiffer-related phenomena to develop a rapid assay of enantiomeric excess. Chiral diamines were added to racemic Cu(I) or Pd(II) complexes, resulting in CD spectra corresponding to metal-to-ligand charge transfer (MLCT) bands of the metal complexes [45, 46]. An instrument interfaced to a robotic 96-well plate allowed rapid and convenient measurement of the CD spectra of the compound library. Linear discriminate analysis of the CD spectra then determined the identiﬁcation, concentration, and enantiomeric excess of the diamines. This study represents a practical application of a chiroptical sensor technique. 7.3.2.4. Anion-Controlled Switching of Amide Complexes. There are several interesting examples of chiroptical metalloswitches triggered by interaction with

261

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

OCH3 O

1.3 nm

0.3 nm

+2 H• (CF3SO3H)

2X

NO3–

2X

N * H OCH3

X X

•

–2 H N

NO3–

N

X: H2O, CH3CN or CF3SO3– “Extended-Λ-form”

H N

N*

H2L5

“Extended-Δ-form”

N* H

N

N*

OCH3

O OCH3

O H3C

“Contracted-Λ-form”

N

H N

CH3

O

H2L6

535 nm Contracted-Λ

Δε (dm3mol–1cm–2)

1

Extended-Λ

0.5

0

–0.5 Extended–Δ 400

600

800

1000

λ (nm)

Figure 7.8. Stretching and inverting dual motions of the CoII complex. Crystal structures CD spectra of [Co(L5)](left, • for CD) and [Co (H2 L5) (CF3 SO3 ) (H2 O)] (CF3 SO3 )–(CHCl3 ) (middle, ), as well as DFT-optimized structure of [Co(H2 L5)(NO3 )]+ (right, ), are illustrated [50]. (Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission.)

◦

anions. Interest in such systems is heightened because supramolecular recognition of anions developed later than that of cations, and the abundance of potential anionic analytes in biology and other areas. Yano et al. [47] reported chiral inversion induced around a seven-coordinated cobalt center by interaction between sugars and sulfate anions. New cage-type cobalt(II) complexes that consist of N -glycosides from mannose-type aldoses and tris(2-aminoethyl)amine (tren), [Co((aldose)3 tren)]X2 • nH2 O (X = Cl− , Br− ), and [Co((aldose)3 tren)]SO4 • nH2 O exhibited C3 helical conﬁguration inversion around the Co(II) center. The CD spectral characteristics of [Co((aldose)3 tren)]X2 • nH2 O changed dramatically with the addition of sulfate anions, and even inverted at high sulfate concentrations, suggesting ion pair formation which was conﬁrmed by a crystal structure. When sulfate ion is embedded into the cavity of the sugar hydroxyl groups, the complex adopts a conﬁguration, while the complex with the halogen anion exhibits a conﬁguration. When the sulfate anion approaches the sugar complex, the electrostatic attraction between the doubly negative and positive charges of the sulfate anion and complex cation causes the hydrogen bonds between the ligands to be interrupted and brings about a chiral inversion due to the sulfate embedding into the large complex cavity. Reversibility was exhibited when the sulfate ions were removed and replaced with halide ions. Miyake et al. [48] established that the helicity of a chiral tetradentate ligand (ligand L6 in Figure 7.8) chelated to Co(II) was readily inverted by the addition of nitrate anion. Preliminary studies suggest that two molecules of nitrate serve to invert the helicity

262

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

of the ligand. Chelation of the ﬁrst equivalent to the Co(II) center displaces the two tertiary nitrogens of the bound ligand, while the second equivalent of nitrate disrupts hydrogen bonding of the amide to solvent [49]. Circular dichroism studies indicated that the initial Co(II) complex exhibited a positive CD signal in the range of the d –d transition (around 530 nm). Upon gradual addition of Bu4 NNO3 the d –d transition the CD response changes from positive to negative. These early ﬁndings by Miyake led to the synthesis of a chemical device designed to exhibit dual mode motions [50]. This time, a modiﬁed version of chiral tetradentate ligand (ligand L5 in Figure 7.8) including 2,5-dimethoxybenzene moieties attached through amide linkages to the terminals of the ligand was employed. An acid–base reaction of the corresponding cobalt complex triggers an interconversion of coordinating atoms between amide nitrogen atoms and amide oxygen atoms, which causes a stretching (extension/contraction) molecular mode. Inversion of helicity (again from to ) after addition of ﬁve equivalents of Bu4 NNO3 accounts for the second device mode. CD studies of the [Co(L5)] complex in CH3 CN/CHCl3 = 1/9 exhibit positive signals at 433 nm and 918 nm and negative signals at 474 nm, 607 nm, and 1100 nm, which correspond to the contracted -form. Two equivalents of CF3 SO3 H (to form [Co(H2 L5)(CF3 SO3 )(H2 O)](CF3 SO3 ) • (CHCl3 ) then caused a rapid signal shift from 0 to positive in the 530-nm region (∼ 5 s) which indicated a switch to the extended -form. The helicity inversion caused by addition of ﬁve equivalents of Bu4 NNO3 gave rise to a negative CD signal around 530 nm. The similarity of this CD signal shift to the original H2 L6 ligand study supports the assertion that helicity is changed in the device. CD signals remained consistent after many deprotonation/protonation cycles, proving that robust reversibility was established. Such a kinetically labile Co(II) complex provides for a dynamic dual mode switch that could potentially be required for sophisticated supramolecular switching devices. Recently, pentapeptide chains were combined in such a chirality-switchable Co(II) complex. The peptide chains experienced helix inversion following the reconﬁguration of the octahedral metal center upon addition of the NO3 − anion stimulus (Figure 7.9) [51]. Similar peptide helix inversion by nitrate anion was shown to occur in analogous NiII and ZnII complexes. Selection between ZnII , CoII , or NiII allowed tuning of the rate of the inversion process to occur on a timescale from milliseconds (ZnII ) to hours (NiII ). The estimated half-lifetime (log τ1/2 ) of these metallo-peptide complexes showed a linear correlation with the water exchange lifetime of the aqueous metal cations.

P O O NO3– O O N M M O O Λ Helicity Δ Inversion

4

M Rapid

O N OM O Chirality Δ Transfer

log(t1/2)

P

Ni(II)

2 Co(II) 0 Zn(II)

–2 Λ, P-form

Δ, P-form

Δ, M-form

–8

–7

–6 log(τ)

–5

–4

Figure 7.9. (Left) Helicity inversion around a metal center and sequential chirality transfer to the pentapeptide helical tubes (-Aib-Phe-Aib-Phe-Aib-OCH3 ). (Right) Linear relationship between half-lifetime (t1/2 ) of helicity inversion and cation water exchange lifetime (τ /s) [51]. (Reproduced by permission of the American Chemical Society.)

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

Δε a b1

N S

R R

N R H

c1, d1 He1 g1, h1 R N f1 i1, j1 R i2, j2 R S N f2 g2, h2 N H e2 c2,d2

λ acetonitrile

N

H

Water Δε

b2 a

L7 (RRRR) (RSRS)

λ

Figure 7.10. (Left) Structures of one of the chiral macrocycle isomer L7(RRRR)(RSRS) . (Right) Solvent induced reversible helicity inversion with accompanying CD spectra [54]. The top spectra show the change in ellipticity over time for the compound on the left upon dissolution in acetonitrile; the bottom spectra are for the compound on the right in water. (Copyright American Chemical Society. Reproduced with permission.)

7.3.2.5. Solvent-Controlled Switching of Metal Complexes. Solvent can also induce signiﬁcant changes in chiroptical response in metal-based systems, either as a result of general differences in polarity or nonspeciﬁc solvation or as a result of coordination of solvent in the inner coordination sphere of a metal complex. In a novel study, the Lisowski group has examined chiroptical switches involving lanthanide complexes of chiral hexaazamacrocycles [52, 53]. The hexaazamacrocycle L7 [54] shown in Figure 7.10 was designed to complex large lanthanides such as Yb(III) and Eu(III). Upon addition of Yb(NO3 )3 • 5H2 O in acetonitrile, it was observed that the chiral ligand L7 wrapped around the Yb in a helical -form corresponding to the (R, R, R, R)-(S, R, S, R) L7 isomer. Crystal structure studies of [YbL7(NO3 )2 ]2 -[Yb(NO3 )5 ](NO3 )4 • 5CH3 CN show an improper torsion angle C2–C4–C15–C17 of −13.3◦ , which is unusually high for a lanthanide(III) hexaazamacrocycle complex. Solvation of the same complex in water, though, leads to ligand reorganization presenting a sharp shift in helicity as evidenced by an improper torsion angle C2–C4–C15–C17 of 87.2◦ for the (R, R, R, R)-(S, S, S, S ) isomer. CD studies conﬁrm helicity inversion by solvent effects, demonstrating quantitative conversion in 144 hours. The proposed mechanism of inversion involves an initial exchange of hydrate into a 10-coordinate metal inner sphere, which is followed by slow ligand reorganization into an 8-coordinate sphere. Lisowski argues that the “squeezed” (R, R, R, R)–(S, S, S, S ) isomer is more capable of accommodating smaller water axial ligands whereas the “open” (R, R, R, R)–(S, R, S, R) isomer preferentially binds the bulkier nitrate ligand in the axial position [54]. The study as a whole represents a rare case of reversible solvent induced helicity inversion for a metal-based complex. Recently, Muller, Lisowski, and co-workers [55] reported that a similar but larger chiral nonaazamacrocyclic amine wraps around the lanthanide(III) ions to form enantiopure helical complexes. The NMR and CD spectra show that kinetic complexation product of the (R, R, R, R, R, R) isomer prefers the (M )-helicity. However, the preferred helicity of the thermodynamic product is M for the early lanthanide(III) ions and P for the late lanthanide(III) ions. In solution, the late lanthanide(III) complexes slowly invert their helicity from the kinetically preferred M to the thermodynamically preferred P . The Mamula group reported the diastereoselective self-assembly of the enantiomerically pure pinene-bipyridine-based receptor, (−) or (+) L-, in the presence of Ln(III) ions

263

264

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Ln3+

Ln3+ N CH3OH

N

COO– (+)-L

CH3CN

CH3CN CH3CN + H2O

tris-Ln[(+)-L]2

tetra-Ln4[(+)-L]9

Figure 7.11. Divergent self assembly leading to the synthesis of interconverting trinuclear [Ln3 {(+)-L}6 (μ3-OH)]2+ and tetranuclear [Ln4 {(+)-L}9 (μ3-OH)]2+ complexes [56]. (Reproduced by permission of the American Chemical Society.)

(Figure 7.11) [56]. Upon exposure to La, Pr, Nd, Sm, Eu, Gd, and Tb ions in dry acetonitrile, it forms a C3 -symmetrical, pyramidal tetranuclear species with the general formula [Ln4 (L)9 (μ3 -OH)](ClO4 )2 ) (abbreviated as tetra-Ln4 L9 ). Three metal centers shape the base: an equilateral triangle surrounded by two sets of helically wrapping ligands with opposite conﬁgurations. The tetranuclear structure is completed by a capping, helical unit LnL3 whose chirality is also predetermined by the chirality of the ligand. The sign and the intensity of the CD bands in the region of the π –π * transitions of the bipyridine are highly inﬂuenced by the helicity of the capping fragment LnL3 . In methanol, it selfassembles to give the trinuclear species [Ln3 (L)6 (μ3 -OH)(H2 O)3 ](ClO4 )2 ) (abbreviated as tris-LnL2 ). The two related superstructures can be interconverted. As is shown by the CD evolution in Figure 7.12, in pure dry acetonitrile, pure tris-LnL2 disassembles and reorganizes gradually to form tetra-Ln4 L9 . If a certain amount of water is added, tris-LnL2 can be reassembled quantitatively. Water stabilizes the trinuclear species to the detriment of the tetranuclear ones. Reducing the amount of water by molecular sieves leads to the tetranuclear species. However, the number of these reversible cycles is limited due to partial decarboxylation of the ligand in the presence of water. Recent reports from Nitschke describe a Cu(I) based solvent-triggered molecular switch [57]. The initial synthesis of the chiral Cu(I) complex in methanol resulted in an equal mixture of both P - and M -diastereomers which was characterized by a weak circular dichroism spectrum bearing similarities to that of the free ligand. A featureless CD spectrum in the region of the MLCT band further established that there was no net diastereomeric excess formed. A stark contrast was then encountered when the Cu(I) complex was dissolved in dichloromethane-d2 and the CD spectrum revealed a positive CE in the MLCT region. Combined studies of CD and NMR suggest that the Cu(I) complex (Figure 7.13) fully converts to the P -diastereomer in nonpolar dichloromethane. Similar studies in DMSO then showed that the M -diastereomer of the complex (Figure 7.13) exists in 20% excess, setting the stage for a reversible metal-based chiroptical molecular switch. The solvent-induced conformational exchange was reasoned to be dominated by hydrogen bonding effects. A weakly polar solvent such as dichloromethane only weakly

265

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

300 pure tetra-Pr4L3

180 min 115 min

Δε (M–1 cm–1)

200

75 min

45 min

100 15 min 10 min

0 5 min

–100

pure tris-Pr[L]2 260

280

300 λ (nm)

320

340

360

Figure 7.12. Time-dependent evolution of the CD proﬁle of tris-Pr[(+)-L]2 in CH3 CN [56]. (Reproduced by permission of the American Chemical Society.)

Figure 7.13. Postulated structures in DMSO (left, M predominating) and CH2 Cl2 (right, P exclusively) [57]. (Reproduced by permission of the Royal Society of Chemistry.)

interacts with the hydroxyl groups of the ligand, allowing for intramolecular hydrogen bonding. Such hydrogen bonding serves to rigidify the structure and lock the complex into the P conformation. Polar solvents such as DMSO, though, interact strongly with the ligand hydroxyl group and push the hydroxyl groups apart, leading to a preference for the M conformation. The authors hypothesize that such a reversible solvent-triggered complex could serve as a means to control stereoselectivity in future metal-catalyzed reactions.

7.3.3. Redox-Triggered Systems The rich coordination chemistry literature offers many avenues for entry into the design of redox-sensitive metal complexes that display rich chiroptical spectra. Redox-active metal ions themselves often show useful electronic spectral changes. However, changes in CD spectra of the organic ligand are also very useful, particularly in complexes that display CD exciton chirality. 7.3.3.1. Iron Translocation in Triple-Stranded Helical Complexes. Shanzer and co-workers [58] reported the ﬁrst published example of a redox-mediated chiroptical redox switch. The system was based on chemical triggering of iron translocation in

266

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Fe2+

+e–

Fe3+

–e–

(a)

O N

N H

(b)

H N CH3

OH N O

O

O N H

N N 3

Figure 7.14. Tripodal ligand containing two binding sites and the redox-switched chirality of its iron complexes [58]. (Reprinted by permission from Macmillan Publishers Ltd. Copyright 1995.)

triple-stranded helical complexes (Figure 7.14). The design accommodated a single metal ion in one of two sites, either a “hard” binding N3 O3 cavity presenting three hydroxamate moieties or a “soft” N6 -cavity with three bipyridyl ligands. Chemical reduction of Fe(III) to Fe(II) induced the metal to translocate from the hydroxamate binding site to a bipyridyl site, because the “softer” Fe(II) favored the site with more nitrogen ligands. Redox switching of the complex was induced by reduction with ascorbate and oxidation with ammonium persulfate. Pronounced differences in UV–vis and CD spectra were observed corresponding to changes in absorbance associated with Fe(II) versus Fe(III) electronic spectra. A split CD spectrum in the UV region was observed that was three times more intense for the Fe(II) state, suggesting exciton interactions involving the bipyridyl moieties. Reduction was rapid, and oxidation gave the Fe(III) absorbance spectrum after a few minutes (several hours were required to achieve the original Fe(III) CD spectrum). The fact that metal exchange did not occur between control compounds with single metal binding sites suggested intramolecular translocation reaction. Variation of the structure resulted in signiﬁcantly different translocation rates. 7.3.3.2. Chiroptical Tripodal Ligands. Three-armed, or tripodal, ligand–metal complexes have been found to offer particularly rich stereodynamic behavior, especially when coupled with exciton chirality analysis. The development of redox-triggered chiroptical switches in the Canary laboratory began with the observation that tripodal, N4 ligands form stable coordination complexes with divalent metal cations [59, 60]. In the case of Zn(II) and Cu(II), the ligand, otherwise conformationally mobile with many conformations, wraps around the metal ion to form a propeller-like complex. In ligands with a single stereogenic center on one of the tripod arms, the helicity of the propeller formed by the

267

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

planes of the heterocycles is dictated by this stereocenter. A number of crystallographic structures [61], with few exceptions, established the relationship between the chiral carbon center and the propeller conﬁguration [62]. ECCD established the preponderance of a single propeller conformation in solution, and it tested whether the same conﬁguration was present relative to the carbon center as had been observed in solid-state studies [63]. It was conﬁrmed that Cu(II) and Zn(II) complexes showed ECCD spectra consistent with solid-state propeller-like structures. This method allowed the assignment of the absolute conﬁguration of secondary amines from the sign of the observed excitant couplet. Furthermore, the sign of the couplet discloses the sense of the propeller twist in solution. The dependence of CD exciton chirality upon the strength of the electronic transition moment, the proximity of the coupled transitions, and the angle between them led to the development of several interesting chiroptical molecular switches. An on/off system was studied involving a tripodal ligand containing three quinoline moieties [64]. The tris(quinoline) compound in Figure 7.15 forms a coordination complex with Cu(II) (8) involving the coordination of four nitrogen atoms and affording an exceptionally intense split CD spectrum that results from the additive effect of three ECCD couplets in one molecule. Reduction to the Cu(I) complex in the presence of strongly coordinating thiocyanate ion gave dissociation of one quinoline arm. This resulted in a much weaker ECCD spectrum due to two factors: (1) the dissociation of one quinoline reduces the number of ECCD couplets from three to one, and (2) the less sterically crowded environment around the copper ion allows unwinding of the ligand and reduces the magnitude

S C N N

S C N N N

+e–

CuII H

Cu I H

–e–

N

CH3

N CH3

N (a)

200

66000

100

33000

Δε 0

0 [Θ]

–100

–33000

–200

–66000

–300

–99000

–400

–132000

CuI –165000 CuII –600 –198000 200 210 220 230 240 250 260

0 [Cu(I)]

OFF

–100 –200

Δε

8

–500

N

N

–300 –400

[Cu(II)] –500

ON 0

1

2

3

4

5

Reversibility at 240 nm

λ (nm)

(b)

(c)

Figure 7.15. On/off chiroptical molecular switch. (a) One-electron reduction results in dissociation of an arm of the tripodal ligand. (b) CD spectra of Cu+ and Cu2+ complexes. (c) oxidation and reduction cycles with ascorbate and persulfate [64]. (Reproduced by permission of Wiley-VCH.)

268

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

of the dihedral angle and therefore diminishes the amplitude of the ECCD couplet. The overall effect is a very large difference in ECCD amplitude between Cu(I) and Cu(II) states. Dependence of the ECCD amplitude on the counterion supported the structural assignment [65]. The complex is highly reversible chemically upon oxidation of the Cu(I) complex with ammonium persulfate and reduction of the Cu(II) complex with sodium ascorbate. Temperature-dependent 1 H NMR studies of this system led to the conclusion that the two arms lacking the chiral carbon center are in rapid equilibrium between associated and dissociated states at room temperature, but slow on the 1 H NMR timescale at low temperature [66]. The arm containing the chiral carbon center, however, remains coordinated. Such tripodal ligands were found to act as chemosensor molecules by virtue of their ability to torque a nematic into a cholesteric liquid crystalline phase increased upon complexation with copper ion [67]. Changes in overall shape of the complexes induced by different metals and counterions were transferred sensitively to the supramolecular level, observed by proportionate changes in the degree of twisting. Redox changes (Cu(I)/Cu(II)) also gave large changes in twisting power. The handedness of the induced cholesteric phase was related to the stereochemistry of the ligand. Interestingly, a direct correlation was observed between helical twisting power and ECCD amplitude, consistent with each technique responding proportionately to the relative twist of the planes of the nitrogen heterocycles. Another related complex containing two chiral carbon centers within a piperidine ring (9) was reported (Figure 7.16) [68]. In this case, the rigidity of the ligand provided control as to which chair form of the piperidine was adopted. In the Cu(I) oxidation state, the ligand adopts a relatively stable cyclohexane chair conformation, with two equatorial and one axial substituent. This conformation places one pyridine moiety remote from the metal ion, but this is accommodated by the lower coordination number of the Cu(I) ion. In the Cu(II) state, strong binding to the higher-coordination number ion brings all three pyridines into association, which forces the piperidine to adopt a higher-energy chair with two axial and one equatorial substituents. The CD spectrum of the Cu(II) complex showed the largest amplitude of any complex in this series, but the Cu(I) spectrum did not give an exciton chirality spectrum. In this case, the Cu(I) structure was characterized by a series of 1 H NMR experiments. In these studies, CD exciton chirality served as a tool to gauge not only the conﬁguration of the propeller conformation but also the degree of twist of the molecule. Relatively few spectroscopic probes are available to report 3D molecular geometry, so it may be expected that this technique should be broadly applicable for the characterization of solution species [16]. Systematic exploration of amino acid derivatives by Canary and co-workers [69, 70] led to the discovery of a molecule that inverts helicity and CD couplet sign upon one-electron redox change [71]. A ligand derived from the amino acid methionine forms a tetradentate complex with Cu(II) involving three nitrogen atoms and a carboxylate (Figure 7.17). In this system, the propeller twist of the molecule is dictated by the asymmetric carbon center by virtue of a gearing mechanism between the methine and methylene carbon atoms, and it can be visualized when viewing down the bond between the tertiary amine nitrogen atom and the Cu(II) ion. Upon reduction to Cu(I), the ligand reorganizes and the sulﬁde moiety replaces the carboxylate, which is expected due to the preference of Cu(I) for this type of coordination. The reorganization requires a pivot about the bond between the tertiary nitrogen atom and the asymmetric carbon atom. This pivot destroys the gear previously mentioned; to retain the geared conformation, the

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

269

X N

CuI

N

N

X

H

H CuII HN

N H

N

N

N X = sovent 9

800 600 400

Free ligand CuI complex CuII complex CuI complex oxidized CuII complex reduced ZnII complex

Δε

200 0 –200 –400 20

12 8 4

200

210

220

230 240 λ (nm)

250

260

0 270

ε × 10–4

16

Figure 7.16. Redox-triggered inversion of one chair form of a piperidine ring into the other chair and corresponding CD (top) and UV–vis (bottom) spectra [68]. (Reproduced by permission of the American Chemical Society.)

two methylene carbon groups ﬂip, which, in turn, inverts the helical orientation of the two quinoline moieties and, therefore, the exciton chirality spectrum. The CD spectrum appears to give mirror images for the Cu(I) versus Cu(II) complexes. The switching was reversible with cyclical additions of ascorbate and ammonium persulfate. Crystallographic data supported the structural assignments [72]. The Cu(I)/Cu(II) complexes of other tripodal ligands also give inversion of the CD spectrum including derivatives of methioninol and S -methylcysteine [73]. 7.3.3.3. Redox-Controlled Molecular Flipper Based on a Chiral Cu Complex. Copper redox chemistry has also been explored in other ligand platforms for redox-dependent chiroptical effects. A molecular bipaddled ﬂipper based on a tetradentate chiral Cu complex was reported whose paddling motion could be controlled

270

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a)

(b) 40

N O CuII H N

H

Ar

H N ArO CO2-H CH3SCH2

H

N

S

+e– –e–

I

Cu N

N -O C 2 H

N

H

–60

[Cu(I)] [Cu(II)]

Cu

5

0

4

–100

3 2

II

Cu –200

(c)

7 6

I

100 Δε239

Ar

H –20 –40 –80

CO2–

S

Ar

CH3SCH2

N

20 0

0

1

2

3

ε × 10–4

X

Δε

+

–

1 210 220 230 240 250 260 λ

n

Figure 7.17. Redox-induced inversion of helicity. (a) As a result of the presence of gearing among the three arms of the tripod near the sterically crowded tertiary amine of the ligand, a pivot about a C–N bond results in the inversion of the propeller. (b) CD and UV–vis spectra of Cu+ and Cu2+ oxidation states. (c) chemical cycling with ascorbate and persulfate [71]. (Reproduced by permission of the American Association for the Advancement of Science.)

N (R) (R) N (R)

CH3

(R) N

CuI N

[Δ-10]+ +e–

[Λ-11]+ +e–

–e–

–e–

Figure 7.18. Crystal structure of [-10]+ (top left) and DFT structures of [-11]+ (top right), [-12]2+ (bottom

CH3

Δ-10PF6 [Δ-12]2+

[Λ-13]2+

left), and [-13]2+ (bottom right) [74]. (Reproduced by permission of the American Chemical Society.)

by reversible oxidation of the metal center (Figure 7.18) [74]. The isomeric pair of Cu(I) complexes -[CuI ((NR , NR , R, R)-L)]PF6 (-10PF6 ) and -[CuI ((NS , NS , R, R)-L)]PF6 (-11PF6 ) interconvert, a slight preference for -11PF6 , (Keq = 1.3), which can be monitored by time-dependent CD starting with a pure -10PF6 in CH2 Cl2 (Figure 7.19a). The amplitude of the Cotton effects decreased with time and eventually rested at smaller amplitudes with opposite signs within a couple of hours. The inversion of the chirality of the aliphatic N atoms resulted in the inversion of helicate chirality. The mechanical motion undergone by -10PF6 and -11PF6 could be switched on/off by reversible metal oxidation and reduction electrochemically or chemically. An oxidation experiment of species -10PF6 with AgBF4 was performed at different times of the isomerization process and the CD of the resulting Cu(II) species (-12PF6 and/or 13PF6 ) recorded (Figure 7.19b). The spectra varied, depending on when the oxidant was added, which corresponded to the degree of [-10]+ /[-11]+ isomerization. However, when AgBF4 was added to a pure -10PF6 sample, all of the -10PF6 was oxidized to -12PF6 and no isomerization to -13PF6 or -11PF6 was observed. The transition

271

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

7 3

40

CD (mdeg)

CD (mdeg)

80

0 –40

t = 30

t = 123

t=0

–7

× 10

–3

–80 –120 235

0

t = 123

t=0 t = 30

335

435 λ (nm) (a)

535

–1 235

436

635

835

λ (nm) (b)

Figure 7.19. (a) CD monitoring of the -10PF6 /-11PF6 isomerization reaction in CH2 Cl2 . (b) CD spectra of the oxidized -10PF6 /-11PF6 couple with AgBF4 at different isomerization times yielding [-12]2+ at t = 0 min and mixtures of [-12]2+ and [-13]2+ at t > 0 min [74]. (Reproduced by permission of the American Chemical Society.)

metal integrated into this device acts as a redox switch that permits one to start/stop the motion at will. 7.3.3.4. Redox-Switchable Pt-Bridged Cofacial Diporphyrins via Carbon– Metal σ Bonds. In one of several porphyrin-containing systems, Shinokubo, Osuka, and co-workers [75] reported the construction of a Pt(IV)-bridged cofacial diporphyrin architecture and its dynamic helical conformational change by reduction of the bridge to Pt(II) (Figure 7.20). Two stable Pt-C σ bonds supported by the pyridyl groups brought two porphyrin macrocycles to be in close proximity in each of these two complexes. The platinum bridge offers conformational ﬂexibility to the complexes due to the susceptibility of platinum toward redox reaction. These complexes also exhibit helical chirality. Reduction of M spiral of 14M mainly yielded the P spiral enantiomer of 15P . The Pt(IV) complex 14M exhibits exciton coupling in both UV–vis and CD, while there is no exciton coupling in the Pt(II) complex 15P . 7.3.3.5. Redox-Triggered Porphyrin Tweezers. Recently, a redox-triggered porphyrin tweezer was reported in an attempt to develop materials with optical properties in the visible region of the electromagnetic spectrum [76]. As shown in Figure 7.21, bis(porphyrin) methioninol derivative (16) gave a strong ECCD couplet upon metallation with Cu(II). The free ligand and Cu(I) complex did not give ECCD. The absence of an ECCD couplet in the Cu(I) complex was rationalized as resulting from relatively weak association of the metal under the conditions studied. The Cu(II) complex, however, showed very strong amplitude, affording an on/off chiroptical molecular switch. Other, nonchiral electrochemically responsive dimeric porphyrin systems have been reported where the redox changes occurred within the porphyrin moieties [77]. 7.3.3.6. Redox-Controlled Dinuclear Ruthenium-Based Switches Monitored by Electronic Near-IR CD. A system showing strong changes in near-infrared (NIR) CD spectra was reported recently [78]. NIR techniques are of interest for several reasons, including the beneﬁt of lower incident light energy on organic materials and greater transparency of NIR light in biological applications. Building on earlier studies of organic-based systems [79–81], the Wang laboratory studied dinuclear ruthenium complexes with 1,2-dicarbonylhydrazizo bridging ligands , -17 and , -17

272

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

N N

2e–

N Ni

N

CI

N N

N

N

N Ni

Ni

N

N

N N

CI N

N

Ni

N N

N

14M spiral

15P spiral

200

Δε (M–1cm–1)

100

0

–100 14M –200

15P

–300

–400 300

400

500

600

700

800

Wavelength (nm)

Figure 7.20. (Top) Reduction of the M spiral Pt(V) complexe 14M result in P spiral Pt(II) complex 15P. (Bottom) CD spectra. [75]. (Reproduced by permission of the American Chemical Society.)

(Figure 7.22) that are highly electrochromic with absorption bands near 500, 900, and 1200–1600 nm. Ligand-centered transitions in the UV region and redox-sensitive MLCT bands in the visible region dominate the CD spectra shown in Figure 7.22. A prominent band near 1115 nm observed in the Ru(II)/Ru(III) state, due to metal–metal charge transfer (MMCT), did not give a strong Cotton effect in the CD spectrum. The Ru(III)/Ru(III) state gave a strong MLCT band at 900 nm that gave a relatively strong Cotton effect in the CD spectrum. Reversible redox switch behavior was demonstrated by monitoring the CD signal at 890 nm and cycling up to seven times electrochemically between the Ru(II)/Ru(II) and Ru(III)/Ru(III) states. A variety of systems have thus been examined for redox-active metal ion triggered chiroptical molecular switches. The mechanisms reported involve translocation of a metal

273

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

N N C N N

Cu(CIO4)2

N HN

NH4SCN

NH N

N N

Cu

N

N

SCN

OH H

N HO HCu N N

16 N N 2

SCH3

SCH3

300 3eq of Cu(II) and NH4NCS free ligand

200 Δε

100 0 –100 –200

A

–300 0,16 0,14 0,12 0,10 0,08 0,06 0,04 0,02 0,00

360

390

360

390

420

450

480

420

450

480

λ (nm)

Figure 7.21. Redox-triggered reorientation of porphyrins [76]. (Reproduced by permission of the American Chemical Society.)

ion, changes of lability of ligand rearrangement, or inner sphere ligand rearrangement resulting from change in coordination number or hardness of the metal. The changes in amplitude of observed CD spectra can be dramatic, even leading to complete inversion of the sign of the ECCD couplet.

7.3.4. Photochemically Triggered Chiral Metal Switches Among many interesting studies, the Aida group used ECD to characterize a redoxtriggered system in which chemical or photoreduction of a chiral cerium bisporphyrinate double-decker complex resulted in racemization by acceleration of the porphyrin ligand

274

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

n+

N

N

HN

Ru

N O

N N

N

Pr

N

O N

Ru

NH

Pr

N N

3×10 2×10

5

RuII/RuII RuII/RuIII RuIII/RuIII

1×105 0 –1×105 –2×105 –3×105

300

400 500 600 Wavelength (nm)

Molar Ellipticity (deg x cm2/dmol)

Molar Ellipticity (deg x cm2/dmol)

Λ, Λ-17 5

700

3×104 RuII/RuII RuII/RuIII RuIII/RuIII

2×104 1×104

0 –1×104 –2×104 –3×104 600

800

1000

1200

1400

Wavelength (nm)

(a)

(b)

Figure 7.22. CD spectra of , -isomer of a diruthenium complex 17 at different oxidation states [78]. (Reproduced by permission of the Royal Society of Chemistry.)

rotation. They further showed that oxidation of a chiral zirconium complex resulted in deceleration of acid-induced racemization [82]. 7.3.4.1. Azobenzene-Based Molecular Scissors. The Aida group carried out the synthesis of other complex light-triggered chiroptical molecular switches. Life-sized scissors having a handle, pivot, and blades inspired the preliminary design of a pair of molecular “scissors” [83]. The chemical equivalents to these three units were found to be azobenzene as the handle, ferrocene as a pivot, and phenyl groups as the blades (Figures 7.23a and 7.23c). The operation of the molecular scissor is quite elegant. Under standard conditions, the azobenzene handle is predominantly in the trans state leading to “closed” blades. Under irradiation of UV light, the azobenzene undergoes isomerization to the cis isomer, which then causes a slight rotation of the cyclopentadienyl rings of the ferrocene pivot. This ﬁnally moves the attached phenyl rings away from one another, leading to an “open” scissor state. The scissors’ chirality (due to the planar 1,1 ,3,3 -tetrasubstituted ferrocene) allows both open and closed states to be seen using circular dichroism (Figure 7.23b). The authors explain that the trans-to-cis isomerization of [CD(−)280]-trans-18 upon UV-irradiation (λ = 350 nm) after 180 s gave rise to CD spectral changes at 240–300 nm due to the major adsorption of the tetraarylferrocene unit. Upon irradiation with visible light (λ > 400 nm), a reverse spectral change occurred, where the system quickly reached a photostationary state in 15 s. Effective reversibility was also exhibited by the system upon sequential irradiation with UV and visible light.

275

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

+2.0

Handle Pivot

N N

Fe 18 (a)

Δε

+1.0 0.0

–1.0

trans-18 10 sec 20 sec 60 sec 180 sec

–2.0 235 260 285 310 335 360 λ (nm) (b)

Blade

Azobenze Strap

(c)

Figure 7.23. (a) Structural representation of azobenzene controlled ‘‘molecular scissors.’’ (b) CD spectral changes of trans-18 upon irradiation with UV light. (c) Graphic conceptualization of the ‘‘molecular scissors’’ [83, 86]. [Reproduced by permission of the American Chemical Society (a, b) and the Royal Society of Chemistry (c).]

7.3.4.2. Host-Controlled Guest Chirality. Further research into Aida’s molecular scissors proved that they could be applied to the ﬁeld of host–guest chemistry [84]. When metallated porphyrins were attached to the 4-position of the phenyl blades, it was found that a diisoquinoline guest was able to chelate to the zinc porphyrin units. Upon irradiation of the host–guest complex with UV light (λ = 350 ± 10 nm), the trans-azobenzene unit again isomerizes to the cis-isomer, causing a long-distance conformational twist of the diisoquinoline guest (Figure 7.24). The guest molecule (19) in solution is initially achiral due to its conformational freedom; but when added to the host molecule (trans-20), it binds in a nonplanar CD-active chiral geometry. Overlap of CD bands (275–350 nm) from the host molecule required that differential CD spectra be used to examine the motion of the guest (Figure 7.24b). Irradiation with UV light caused the Cotton effects at 270–350 nm of the guest (19) to diminish and then vanish. It was reasoned that the disappearance of the CD band is caused by the guest molecule being forced into a nearly planar state when bound to the cis-isomer of the host compound. Sequential irradiation of host–guest complex (19 • trans-20) with UV and visible light proved that the complex was controllably reversible (Figure 7.24c). This represents the ﬁrst instance of a molecular machine causing chirality manipulation in a controllable and reversible manner. 7.3.4.3. Chirality Transfer via Ternary Complex. Recently, Aida and coworkers [85] have created many similar compounds incorporating the molecular scissor as a basis for more elaborate and complex systems. Such systems include a ternary compound, which includes a pyridine-appended dithienylethene derivative as a photochromic module that can again be used to transfer conformational information with UV and visible light as a trigger [85]. Extension of the pivotal ferrocene has also been adopted in reversible self-locking compounds shown in Figure 7.25. In the presence of trans1,2-bispyridine ethylene, the zinc–porphyrin moieties coordinate intramolecularly with the anilines to “lock” the molecule internally [86]. UV light is then used to isomerize to the cis-1,2-bispyridine ethylene that is then capable of coordinating to the zinc porphyrin units, “locking” the molecules externally. The process is again shown to be reversible by alternating UV and visible light irradiation. Such discoveries by Aida and co-workers could help to controllably transmit chiral and mechanical information through long molecular distances.

276

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

N Ar N Zn N N Ar

Ar

(a)

N Zn N N N

N

Ar

N

Ar N

19

N

Vis

N N Zn

N

N

N

UV

N

Fe

Ar

Fe

19

N N

N Zn N N Ar

Trans-20

Ar

Cis-20 19·Cis-20

19·Trans-20 100

(b)

[19]/Trans-[20]

Δε

50

0.0 0.3 0.6 1.0 2.0

0

–50

–100

250

300

350

400

450

500

Wavelength (nm) 6.0

(c) Δε

4.5 3.0 1.5

NN N Zn NN

0

Δε

150

N

NN Zn NN

100 50

[trans-20]

[trans-20] + [cis-20]

0 1.00 0.75 0.50 N N

0.25 0.00

0 10 20 30 40 50 60 70 80 90 Times (s)

Figure 7.24. (a) Photoisomerization of a 1:1 complex of molecular pedal 20 with rotary guest 19 (19 • 20). (b) CD spectral changes of trans-20 upon titration with 19. (c) CD visualization of the motions of guest-binding molecular pedals 19•(+)-20 and 19•(−)-20 triggered by light [84]. (Reprinted by permission from Macmillan Publishers Ltd: Copyright 2006.)

277

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

H2 N

NH2

N

N N Zn N

N

N N Zn N

N

UV N N N Zn

N N N Zn N

Vis

trans-22

N

N

N N H2 Internally Double-locked 21

NH2 Externally Locked 21 in 21⊃cis-22

Figure 7.25. Structures of internally double-locked 21 and externally locked 21 ⊃ cis−22, and the self-locking operation in response to photochemical isomerization of 22 [86]. (Reproduced by permission of the Royal Society of Chemistry.)

7.4. DYNAMIC STEREOCHEMISTRY MONITORED BY VCD VCD spectra have been applied to study transition-metal complexes since the pioneering work by Naﬁe and co-workers [8, 13, 87] VCD spectroscopy was applied to monitoring in situ the photoinduced rewind of supramolecular helices in a liquid crystal (Figure 7.26) [11]. A room-temperature liquid crystal, ZLI-1132, was doped with a chiral Cr(III) complex -[Cr(acac)2 (2C12)] (acac = acetylacetonate; 2C12 = 4, 4 -didodecyloxyated dibenzoylmethanate). The selective reﬂection wavelength λc = np (n is the average refractive index, p is the pitch length of an induced helix) of the nematic phase was determined to be 5.3 μm. At this wavelength, circularly polarized light reﬂects from or passes through the sample when it has the same or opposite sense of the induced helix, respectively. Under the illumination of UV light (365 nm), the photoracemization of the Cr(III) complex rewound helices in the chiral nematic phase. In response to this, the VCD spectrum of the system exhibited the transient change. Figure 7.26 shows the time course of the VCD spectrum recorded every 30 min. The peak at 1610 cm−1 increased intensity for the initial 1 h. The change reﬂected the elongation of the pitch maintaining the relation of λ/λc > 1. The peak underwent a drastic change at 1.5 h: the spectral shape transformed from a Gaussian to a biphasic one, which indicated the relation of λ/λc = 1 was fulﬁlled at this stage. After 2 h, the peak returned to a Gaussian shape with a negative sign, indicating the relation of λ/λc < 1. Reﬂecting the further elongation of the helical pitch, the position of a biphasic peak shifted toward the longer wavelength. When the VCD spectrum is regarded as a memory signal for the photoresponsive events in this liquid-crystalline system, it shows high signal-to-noise ratio (S/N) since it changes the sign as well as the intensity. The spectral change conveys the information of time memory because the position of the biphasic shaped peak shifts toward the longer wavelength on continuing irradiation. It

278

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

C12H25O O O

O O

O

Cr

Cr O

O O C12H25O

6h 5.5h 5h 4.5h 4h 3.5h 3h 2.5h 2h 1.5h 1h 0.5h 0h

OC12H25

O O

O O

Δ Cr(acac)2(2C12) Λ

OC12H25

1800

1600

1400 1200 Wavenumber (cm–1)

1000

Figure 7.26. VCD spectra of a chiral nematic sample of ZLI-1132 doped with 0.538 mol% -[Cr(acac)2 (2C12)] at various times after UV light (365 nm) irradiation [11]. (Reproduced by permission of Taylor and Francis.)

shows extremely stable memory since the racemization process is accompanied by an increase of entropy, thus irreversible.

7.5. DYNAMIC STEREOCHEMISTRY MONITORED BY FDCD AND CPE Fluorescence-detected circular dichroism (FDCD) is a method that can measure the CD response by detection in emission if the chromophore is also ﬂuorescent. This method was originally developed by Tinoco and co-workers [88, 89] and studied intensively more recently by Berova, Nakanishi, and co-workers [90–92]. While conventional CD measures the difference in a sample’s absorption of left- and right-circularly polarized light, FDCD measures the difference in ﬂuorescence intensity upon excitation by leftand right-circularly polarized light. Since it is usually true that the excitation spectrum of a ﬂuorophore parallels its absorption, the same circular dichroic information should be able to be extracted from both processes if the artifacts related to undesired ﬂuorescence anisotropy are eliminated by more advanced instrumentation [92]. FDCD has been shown to be more sensitive than absorption CD [90, 93], analogous to the fact that ﬂuorescence spectroscopy is more sensitive than the UV–vis absorption method because ﬂuorescence suffers no background interference from the incident light. Raw FDCD data measured by a JASCO circular dichroism system equipped with FDCD attachment and with the ﬂuorescence detector placed at 90◦ to the excitation beam (i.e., 90◦ to the CD detector) is recorded in two channels [90]. They represent excitation spectra and correspond to the difference in emission F (FL − FR ) and the total emission (FL + FR ) resulting from differential absorption of left- and right-circularly polarized light, respectively. Typically, the data are converted to CD spectra by established methods, which gives a normal CD spectrum if ﬂuorescence polarization is negligible: The FDCD and normal CD of zinc complex of chiral tripodal ligand 23 match perfectly. An adaptation of the FDCD technique can provide a unique and powerful new strategy for sensor applications by using the F , that is, (FL − FR ), component of FDCD data directly, without conversion to CD. To distinguish this new approach from traditional FDCD and to avoid confusion, this method was named differential circularly polarized ﬂuorescence excitation (CPE) [17], although no new instrument is required and all of the advantages and nature of FDCD still pertain. This is different from circularly polarized luminescence (CPL) because CPL is the differential spontaneous emission of left- and right-circularly polarized light and reﬂects the structural properties of the excited

279

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

state, while CPE is still an indirect reﬂection of the structural properties of the ground electronic state. The theoretical basis of CPE can be derived from long-established Eq. (7.1) [90]. F =

θ · (FL + FR ) · ln 10 . 33 · 2 · (10A − 1) · k

(7.1)

If A = AL − AR ≤ 0.1 and A/A ≤ 0.1, total emission (FL + FR ) should be proportional to ﬂuorescence induced by nonpolarized light, that is, FL + FR = k2 · F = k2 · F · I 0 · (1 − 10−A ),

(7.2)

Therefore Eq. (7.1) can be simpliﬁed as shown in Eq. (7.3), where K is a constant, which incorporates all other constants and F is the ﬂuorescence quantum yield. F =

θ · F · I 0 θ · F · I 0 · k2 · ln 10 = K · . 33 · 2 · 10A · k 10A

(7.3)

Materials with higher ellipticity θ and higher ﬂuorescence quantum yield F will lead to an even larger F . Substances lacking either ﬂuorescence or CD properties will not be observed. Shown in Figure 7.27 are the CD and CPE (F) responses of a chiral piperidine compound (23) to Zn2+ [17]. Apparently, FDCD and CPE may be used to monitor the chirality switching in such metal complexes.

N

N

N

N H

H 23

20 15

0.4 μM incremental 0 μM Zn2+

5

5 ΔF (V)

θ (mdeg)

10

0 –5

–10

5

Zn2+ 5.2 μM Zn2+

Zn2+

0 μM Zn2+ 0.4 μM incremental 5.2 μM Zn2+

–15 200 220 240 260 280 300 320 340 λ (nm) (a)

–5

0 μM Zn2+ 0.4 μM incremental

10 15

Zn2+ 5.2 μM Zn2+ 20 200 220 240 260 280 300 320 340 λ (nm) (b)

Figure 7.27. Spectral responses of 4.8 μM (R,R)-23 to 0–5.2 μM Zn(II) in acetonitrile. (a) CD. (b) CPE (700V, 81 deg, ﬁlter: 360 nm) [17]. (Reproduced by permission of the American Chemical Society.)

280

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

6 3+

Me

Ph

4

HN R

O

O N Me

N H

N

N

N

Ln

N

N

N

O NH Ph

(SSS)-Δ-[Ln·L25]3+

(IL − IR)

N

2 0 480

Ln = Eu, Gd, or Tb

−2

25a: R = H 25b: R = COOMe

−4

530

580

630

Me

Figure 7.28. (Left) CPL spectra for (SSS)--[Tb.L

Wavelength (nm)

25b 3+

] (solid curve) and in the presence of BSA (broken curve). (Right) Chiral lanthanide metal–ligand complex used to bind human or bovine serum albumin in ‘‘drug site II’’ [18]. (Reproduced by permission of the Royal Society of Chemistry.)

7.6. DYNAMIC STEREOCHEMISTRY MONITORED BY CPL Circularly polarized luminescence (CPL), the anisotropic emission of circularly polarized light originated from nonpolarized excitation, is the emission analogue to CD. The sign and magnitude of CPL are affected by the degree of helical twist of the complex, the nature of the ligand ﬁeld, and other factors. In this context, excited Ln(III) ions can be considered as “spherical” emitters and avoid the problems associated with anisotropy that can complicate some chiroptical analyses [94]. CPL reﬂects the time-averaged local helicity around the lanthanide(III) ion. The Parker group has recently utilized the chiral environment of drug site II of serum albumin to induce helicity inversion in complexes of terbium and europium (III) [18]. It was found that chiral complex (S , S , S )--[Tb.L25 ]3+ changed helicity to (S , S , S )--[Tb.L25 ]3+ upon addition of human or bovine serum albumin. Convincing data was supplied by circularly polarized emission (Figure 7.28). When the -isomer is exposed to BSA or HSA, there is an inversion of the sign of emission and 35% reduction of the signal intensity. The authors explain that the emission spectra “are consistent with the inversion of the helicity of the complex in the protein-bound form” [95]. Parallel experiments were run using the -isomer, but no change in emission spectra was found. These results give one of the few immediately biologically relevant examples of a metal-based chiroptical molecular switch. The system could potentially allow protein association to be tracked in vitro in real time.

7.7. SOLID-STATE METAL-BASED CHIROPTICAL SWITCHES 7.7.1. Pressure-based switches Pressure also has been reported to induce chiroptical responses in chiral metal complexes. In solution, high pressure can provide a powerful solvent effect since dispersive interactions depend strongly on density changes. In the solid state, crystal packing plays an additional role. The effect of pressure on circular CD spectra of the octahedral chiral -, -, and (, )tris{O, O -bis[(+)(S )-2-methylbutyl]dithiophosphate}Cr(III)

281

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

300

(Δ,Δ) − Cr[(S)(S)Mebdtp]3

200

S

Me O P O Me

O

Me

P S

S

Me O

O P

Δ − (S,S)(S,S)(S,S)

S M

S S

S

S

M S

1.7

Me

P O Me O Me

S

S P

S

Me O O Me

Me O P O Me

Δ − (S,S)(S,S)(S,S)

CD/mdeg

Me O

100 0

1.2

−100 −200

0.2 GPa

−300 400

500

600 λ (nm)

700

800

CD spectra at different pressures of solid -tris{O, O -bis[(+)(S)-2 -methylbutyl]dithiophosphate}Cr(III) and proposed , conversion mechanism [96]. (Reproduced by permission of Taylor and Francis.)

Figure

7.29.

complexes was studied in the pressure range 0–2.5 GPa (Figure 7.29) [96]. Results on polycrystalline samples dispersed in nujol show a pressure-induced -to- inversion of conﬁguration at the metal center above 1.2 GPa, which was suggested to arise from differential crystal packing in the solid-state structure of the diasteromeric complexes. The -form is conﬁrmed to be the most favored crystal packing among different ligand conformations of the chiral complex under high pressure. When the applied pressure exceeded 2.5 GPa, the CD band obtained from polycrystalline Nujol samples of chiral and -tris-[cyclic O,O , 1(R), 2(R)-dimethylethylene dithiophosphato]chromium (III) complexes inverted from negative to positive, which demonstrated inversion from the -form to the -form by means of pressure [97, 98]. To minimize artifacts, the spectra were obtained from average data from rotating the diamond anvil cell (DAC) around its optical axis, and the reference spectrum was normalized outside the absorption region of the sample. The cycle was reversible as demonstrated by applying repeating pressure cycles. However, the transition pressure varied and was dependent on the amount of -diastereomer present in the sample. Mechanistic explanation of the pressure induced chirality inversion could involve bond breaking or trigonal twisting around the metal center. It would be interesting to see if other solid-state data could be obtained to test the mechanistic hypothesis and exclude artifacts.

7.7.2. Temperature-Induced Dynamic Stereochemistry The compound α-Ni(H2 O)6 · O4 and its selenate derivative exhibit chirality only in the solid state. The Kuroda group observed a remarkable reversible sign inversion of CD in the 3 A2g → 3T1g (P) Ni(II) d –d transition at near liquid nitrogen temperatures, although the crystal structure hardly changes from 300 to 100 K (Figure 7.30) [99]. The change in Ni2+ electronic states at low temperatures might have altered the relative magnitude of the opposite sign ﬁrst- and second-order rotational strengths.

7.7.3. Photo-induced Switching Switching of molecular chirality under photoirradiation was studied in a cobaloxime complex crystal using CD (Figure 7.31) [100]. The (S -alkyl)(S -base) crystal was irradiated using two different wavelength bands, one with 439–499 nm covering the LMCT transition and the other with 640–900 nm covering the triplet d –d transition of Co(III). After irradiation with either wavelength band, the solid was dissolved in methanol and the changes in its CD spectrum were recorded. Excitation of the d –d transition of the Co(III) ion appeared to be much more effective in inducing the chirality change than

282

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

800 600

3

CD (mdeg)

296 k

A2g→3T1g(P)

3A →3T (F) 2g 1g

400 200 0 –200 300

83 k 400

500

600

700

Wavelength (nm) (a) 200

400 300

CD (mdeg)

CD (mdeg)

100

@384 cm

P41212 (0.062 cm)

200 100 0 –100

Cooling Heating

0 –100 –200

P41212 (0.06 cm)

–200

–300

–300 –400 0

–50

–100

–150

–200

–400

0

–50

–100

–150

Temperature(°C)

Temperature(°C)

(b)

(c)

–200

Figure 7.30. (a) Observed CD spectra of the same α-Ni(H2 O)6 · SO4 single crystal (P43 21 2, 0.062-mm thickness) at different temperatures. (b) CD signal at 384 nm plotted against temperature for the enantiomorphous α-Ni(H2 O)6 · SO4 single crystals. : P43 21 2 (0.62 mm thickness); : P41 21 2 (0.60 mm thickness). (c) Temperature dependence of the CD values of P41 21 2 crystal on cooling () and on heating (•) [99]. (Reproduced by permission of Elsevier.)

excitation of the ligand–metal charge transfer band, although the latter is more effective in breaking the Co–C bond that initiates the chirality switching. The chirality change versus irradiation time showed a step-like behavior suggesting that chirality switching of molecules occurred in correlation with their nearest neighbors. The same group made direct observation of a photoinduced chirality change or switching in the alkyl ligand of the cobaloxime complex in the hydrated and nonhydrated crystals of the cobaloxime complex by direct CD measurements in the solid phase in Nujol-mull and KBr pellets [101]. The CD spectra of the two crystal forms showed clear differences. Additional CD peaks in the spectra of the nonhydrated crystals seemed to arise from exciton splitting of the charge-transfer band. Photoirradiation induced chirality change or switching in the alkyl part of the molecule, but not in the crystal structure. The CD spectra well reﬂect such behavior.

283

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

(R-Alkyl)(S-Base)

(S-Alkyl)(S-Base)

O

O

H3C

CH3 hν * H O H O O N N Co N N O H O

H * CH3 O H O O N N Co N N O O H

H3C

H3C * HH2 H

H3C * NH2 H

(a) 100

0

–10 200

(S/R-Alkyl)(S-Base) 300

(S-Alkyl)(S-Base)

400 500 Wavelength (nm) (b)

600

75

25

50

50

25

75

(R-AlkyI) (%)

0

(R-Alkyl)(S-Base) (S-AlkyI) (%)

CD (mdeg.)

10

700 0 0

5 10 15 20 25 Irradiation time (hours)

100 30

(c)

Figure 7.31. (a) Molecular structures of a pair of diastereomers of cobaloxime complex that can be converted to each other by photoisomerization. (b) CD spectra of (S-alkyl)(S-base), (R-alkyl)(Sbase), (S/R-alkyl)(S-base) cobaloxime complexes in 1-mM methanol solution. (c) Time variation of photoinduced chirality change in the cobaloxime complex crystals of (S-alkyl)(S-base), (R-alkyl)(Sbase), and (S/R-alkyl)(S-base) under constant irradiation of light with wavelength of 640–900 nm [100]. (Reproduced by permission of the American Institute of Physics.)

7.8. CONCLUSIONS The future of metal-based chiroptical switches is bright, given the high degree of control available, a multitude of triggering mechanisms, and powerful chiroptical spectroscopy tools available for analysis. With the large number of recently discovered systems, these compounds and materials derived from them could potentially be used for applications including optical displays, complex molecular electronics, chiral resolution, and catalysis. The Pfeiffer effect and metal ion templated synthesis provided early chemistry relevant to more recently developed metal-based chiroptical switches. Environmentresponsive switches have been developed using a large variety of metals and ligands triggered by pressure, counterion alteration, light, and solvent changes. Redox triggered switches have been explored primarily using a tripodal ligand motif. Diazobenzeneferrocene systems were designed to reliably switch the conformations of a set of “molecular scissors,” which were then used in an array of interesting and complex supramolecular machines. Polymer systems have been explored illustrated by the use of metal dopants to cause chiroptical changes in oligothiophene polymers. The studies in this area have

284

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

provided much stimulating new chemistry and exemplify the power of modern molecular design and solution characterization techniques. There is no doubt that there are many more opportunities to develop even more imaginative systems. Many applications for these materials have been discussed, especially in the areas of electronics and sensors, and several of the available systems are poised to make a genuine contribution. Although early phenomena were studied by ORD, electronic circular dichroism experiments were used in nearly all experiments to analyze the conformational changes of the chiral compounds. Although other analytical techniques such as NMR are used to study these systems, it is readily apparent that CD experiments provide accurate and dependable read-out for chiral metal-based switches. The exciton chirality method has been particularly useful as a result of the fact that it gives a sizable and interpretable signal. Indeed, few other spectroscopic measurements give a direct report of three-dimensional shape as exciton chirality does for the orientation of chromophoric units. However, care should be taken in the assignment of CD data as arising from exciton coupling. Newer chiroptical spectroscopic methods are beginning to contribute to this ﬁeld. The ﬁrst reports of systems employing NIR-CD, which offers low-energy measurement, have appeared. FDCD can provide very sensitive detection if precautions are taken to avoid artifacts. FDCD or CPE may offer more speciﬁc information since it may arise from a subset of transitions compared to CD. VCD offers the possibility to probe vibrational phenomena as illustrated in studies of memory in liquid crystalline switches. All of these newer methods are ripe for further development in chiroptical switch detection strategies. It is also high time for computational methods coupled with chiroptical spectroscopy to play a greater role in not only characterization but also design of these systems. Most metal-based chiroptical switches reported to date were studied in solution, but many applications of chiroptical molecular switches involve the solid phase where chiroptical spectra are more difﬁcult to interpret. Fortunately, the development of solid phase characterization tools and accompanying theory is progressing. In this regard, computation has resulted in renewed interest in ORD and other classic methods due to the possibility of making structural conclusions by matching experiment with theory.

ACKNOWLEDGMENTS We are grateful to the National Science Foundation (CHE-0848234) for generous support of our work in this area. ZD thanks Research Corporation for Science Advancement and the donors of the American Chemical Society Petroleum Research Fund for support of this work.

REFERENCES 1. L. Fabbrizzi, A. Poggi, Transition Metals in Supramolecular Chemistry, Kluwer Academic Publishers, Dordrecht, 1994. 2. B. Champin, P. Mobian, J. P. Sauvage, Chem. Soc. Rev . 2007, 36 , 358–366. 3. H. Amouri, M. Gruselle, Chirality in Transition Metal Chemistry: Molecules, Supramolecular Assemblies and Materials, Wiley, Chichester, 2008. 4. H. Miyake, H. Tsukube, Supramol. Chem. 2005, 17 , 53–59. 5. M. Elhabiri, A. M. Albrecht-Gary, Coord. Chem. Rev . 2008, 252 , 1079–1092.

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

6. K. Nakanishi, N. Berova, R. Woody, Circular Dichroism: Principles and Applications, Wiley-VCH, New York, 2000 7. S. Allenmark, J. Gawronski, Chirality 2008, 20 , 606–608. 8. P. L. Polavarapu, L. A. Naﬁe, S. A. Benner, T. H. Morton, J. Am. Chem. Soc. 1981, 103 , 5349–5359. 9. L. A. Naﬁe, J. Mol. Spectrosc. 1995, 347 , 83–100. 10. T. B. Freedman, X. Cao, R. K. Dukor, L. A. Naﬁe, Chirality 2003, 15 , 743–758. 11. T. Taniguchi, K. Monde, S.-I. Nishimura, J. Yoshida, H. Sato, A. Yamagishi, Mol. Cryst. Liq. Cryst. 2006, 460 , 107–116. 12. H. Sato, D. Shirotani, K. Yamanari, S. Kaizaki, Inorg. Chem. 2010, 49 , 356–358. 13. D. A. Young, E. D. Lipp, L. A. Naﬁe, J. Am. Chem. Soc. 1985, 107 , 6205–6213. 14. S. Abbate, E. Castiglioni, F. Gangemi, R. Gangemi, G. Longhi, Chirality 2009, 21 , E242–E252. 15. C. Guo, R. D. Shah, R. K. Dukor, X. Cao, T. B. Freedman, L. A. Naﬁe, Charility 2005, 59 , 1114–1124. 16. N. Berova, L. Di Bari, G. Pescitelli, Chem. Soc. Rev . 2007, 36 , 914–931. 17. Z. Dai, G. Proni, D. Mancheno, S. Karimi, N. Berova, J. W. Canary, J. Am. Chem. Soc. 2004, 126 , 11760–11761. 18. C. P. Montgomery, E. J. New, D. Parker, R. D. Peacock, Chem Commun. 2008, 4261–4263. 19. T. Harada, H. Hayakawa, R. Kuroda, Rev. Sci. Instrum. 2008, 79 , 073103. 20. A. Werner, Ann. 1912, 386 , 1. 21. D. B. Shapiro, R. A. Goldbeck, D. Che, R. M. Esquerra, S. J. Paquette, D. S. Kliger, Biophys. J . 2005, 68 , 326–334. 22. M. Krykunov, M. D. Kundrat, J. Autschbach, J. Chem. Phys. 2006, 125 , 194110–194113. 23. H. Yoneda, Y. Nakashima, U. Sakaguchi, Chem. Lett. 1973, 1343–1346. 24. D. W. Urry, H. Eyrin, J. Am. Chem. Soc. 1964, 86 , 4574–4580. 25. W. T. Shearer, R. K. Brown, J. Biol. Chem. 1966, 241 , 2665–2671. 26. E. Cerasoli, B. K. Sharpe, D. N. Woolfson, J. Am. Chem. Soc. 2005, 127 , 15008–15009. 27. K. Pagel, T. Vagt, T. Kohajda, B. Koksch, Org. Biomol.. Chem. 2005, 3 , 2500–2502. 28. I. Otsuka, R. Sakai, T. Satoh, R. Kakuchi, H. Kaga, T. Kakuchi, J. Polymer Sci. A: Polymer Chem. 2005, 43 , 5855–5863. 29. T. Sanji, Y. Sato, N. Kato, M. Tanaka, Macromolecules 2007, 40 , 4747–4749. 30. H. Goto, E. Yashima, J. Am. Chem. Soc. 2002, 124 , 7943–7949. 31. O. Henze, W. J. Feast, F. Gardebien, P. Jonkheijm, R. Lazzaroni, P. Lecl`ere, E. W. Meijer, A. P. H. J. Schenning, J. Am. Chem. Soc. 2006, 128 , 5923–5929. 32. S. Haraguchi, M. Numata, C. Li, Y. Nakano, M. Fujiki, S. Shinkai, Chem. Lett. 2009, 38 , 254–255. 33. L. A. P. Kane-Maguire, G. G. Wallace, Chem. Soc. Rev . 2010, 39 , 2545–2576. 34. M. Funahashi, N. Tamaoki, Chem. Mater. 2007, 19 , 608–617. 35. P. Pfeiffer, K. Quehl, Ber. 1931, 64 , 2667. 36. H. G. Brittain, Polyhedron 1984, 3 , 1087. 37. G. A. Hembury, V. V. Borovkov, Y. Inoue, Chem. Rev . 2008, 108 , 1–73. 38. N. Wei, N. N. Murthy, K. D. Karlin, Inorg. Chem. 1994, 33 , 6093–6100. 39. J. W. Canary, A. E. Holmes, J. Liu, Enantiomer 2001, 6 , 181–188. 40. H. Ishii, Y. Chen, R. A. Miller, S. Karady, K. Nakanishi, N. Berova, Chirality 2005, 17 , 305–315. 41. N. Berova, G. Pescitelli, A. G. Petrovic, G. Proni, Chem. Commun. 2009, 5958–5980.

285

286

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76.

X. Li, B. Borhan, J. Am. Chem. Soc. 2008, 130 , 16126–16127. V. V. Borovkov, J. M. Lintuluoto, Y. Inoue, Org. Lett. 2000, 2 , 1565–1568. V. V. Borovkov, G. A. Hembury, Y. Inoue, Acc. Chem. Res. 2004, 37 , 449–459. S. Nieto, V. M. Lynch, E. V. Anslyn, H. Kim, J. Chin, Org. Lett. 2008, 10 , 5167–5170. S. Nieto, V. M. Lynch, E. V. Anslyn, H. Kim, J. Chin, J. Am. Chem. Soc. 2008, 130 , 9232–9233. S. Yano, M. Nakagoshi, A. Teratani, M. Kato, T. Onaka, M. Iida, T. Tanase, Y. Yamamoto, H. Uekusa, Y. Ohashi, Inorg. Chem. 1997, 36 , 4187–4194. H. Miyake, K. Yoshida, H. Sugimoto, H. Tsukube, J. Am. Chem. Soc. 2004, 126 , 6524–6525. H. Miyake, H. Sugimoto, H. Tamiaki, H. Tsukube, Chem. Commun. 2005, 34 , 4291–4293. H. Miyake, M. Hikita, M. Itazaki, H. Nakazawa, H. Sugimoto, H. Tsukube, Chem.-Eur. J . 2008, 14 , 5393–5396. H. Miyake, H. Kamon, I. Miyahara, H. Sugimoto, H. Tsukube, J. Am. Chem. Soc. 2008, 130 , 792–793. J. Gregolinski, J. Lisowski, Angew. Chem. Intl. Ed . 2006, 45 , 6122–6126. J. Gregolinski, T. Lis, M. Cyganik, J. Lisowski, Inorg. Chem. 2008, 47 , 11527–11534. J. Gregolinski, K. Slepokura, J. Lisowski, Inorg. Chem. 2007, 46 , 7923–7934. J. Gregolinski, P. Starynowicz, K. T. Hua, J. L. Lunkley, G. Muller, J. Lisowski, J. Am. Chem. Soc. 2008, 130 , 17761–17773. M. Lama, O. Mamula, G. S. Kottas, L. De Cola, H. Stoeckli-Evans, S. Shova, Inorg. Chem. 2008, 47 , 8000–8015. M. Hutin, J. Nitschke, Chem. Commun. 2006, 1724–1726. L. Zelikovich, J. Libman, A. Shanzer, Nature 1995, 374 , 790–792. J. W. Canary, C. S. Allen, J. M. Castagnetto, Y. H. Wang, J. Am. Chem. Soc. 1995, 117 , 8484–8485. J. W. Canary, Chem. Soc. Rev . 2009, 38 , 747–756. J. W. Canary, C. S. Allen, J. M. Castagnetto, Y. H. Chiu, P. J. Toscano, Y. H. Wang, Inorg. Chem. 1998, 37 , 6255–6262. X. D. Xu, K. J. Maresca, D. Das, S. Zahn, J. Zubieta, J. W. Canary, Chem. Eur. J . 2002, 8 , 5679–5683. J. M. Castagnetto, X. D. Xu, N. D. Berova, J. W. Canary, Chirality 1997, 9 , 616–622. S. Zahn, J. W. Canary, Angew. Chem. Intl. Ed . 1998, 37 , 305–307. S. Zahn, J. W. Canary, J. Am. Chem. Soc. 2002, 124 , 9204–9211. J. Zhang, K. Siu, C. H. Lin, J. W. Canary, New J. Chem. 2005, 29 , 1147–1151. S. Zahn, G. Proni, G. P. Spada, J. W. Canary, Chem. Eur. J . 2001, 7 , 88–93. J. Zhang, J. W. Canary, Org. Lett. 2006, 8 , 3907–3910. A. E. Holmes, S. Zahn, J. W. Canary, Chirality 2002, 14 , 471–477. S. Zahn, J. W. Canary, Org. Lett. 1999, 1 , 861–864. S. Zahn, J. W. Canary, Science 2000, 288 , 1404–1407. S. Zahn, D. Das, J. W. Canary, Inorg. Chem. 2006, 45 , 6056–6063. H. S. Barcena, A. E. Holmes, J. W. Canary, Org. Lett. 2003, 709–711. A. Company, M. Guell, D. Popa, J. Benet-Buchholz, T. Parella, X. Fontrodona, A. Llobet, M. Sola, X. Ribas, J. M. Luis, M. Costas, Inorg. Chem. 2006, 45 , 9643–9645. S. Yamaguchi, T. Katoh, H. Shinokubo, A. Osuka, J. Am. Chem. Soc. 2008, 130 , 14440–14441. A. E. Holmes, D. Das, J. W. Canary, J. Am.Chem. Soc. 2007, 129 , 1506–1507.

D Y N A M I C S T E R E O C H E M I S T RY A N D C H I R O P T I C A L S P E C T R O S C O P Y

77. G. Pognon, C. Boudon, K. J. Schenk, M. Bonin, B. Bach, J. Weiss, J. Am. Chem. Soc. 2006, 128 , 3488–3489. 78. D. Li, Z. Y. Wang, D. Ma, Chem. Commun. 2009, 1529–1531. 79. Z. Y. Wang, E. K. Todd, X. S. Meng, J. P. Gao, J. Am. Chem. Soc. 2005, 127 , 11552–11553. 80. J. Zheng, W. Qiao, X. Wan, J. P. Gao, Z. Y. Wang, Chem. Mater. 2008, 20 , 6163–6168. 81. J. Deng, N. H. Song, W. Liu, Q. F. Zhou, Z. Y. Wang, ChemPhysChem 2008, 9 , 1265–1269. 82. K. Tashiro, K. Konishi, T. Aida, J. Am. Chem. Soc. 2000, 122 , 7921–7926. 83. T. Muraoka, K. Kinbara, Y. Kobayashi, T. Aida, J. Am. Chem. Soc. 2003, 125 , 5612–5613. 84. T. Muraoka, K. Kinbara, T. Aida, Nature 2006, 440 , 512–515. 85. H. Kai, S. Nara, K. Kinbara, T. Aida, J. Am. Chem. Soc. 2008, 130 , 6725–6726. 86. K. Kinbara, T. Muraoka, T. Aida, Org. Biomol. Chem. 2008, 6 , 1871–1876. 87. Y. He, X. Cao, L. A. Naﬁe, T. B. Freedman, J. Am. Chem. Soc. 2001, 123 , 11320–11321. 88. D. H. Turner, I. J. Tinoco, M. Maestre, J. Am. Chem. Soc. 1974, 96 , 4340–4342. 89. I. J. Tinoco, D. H. Turner, J. Am. Chem. Soc. 1976, 98 , 6453–6456. 90. J.-G. Dong, A. Wada, T. Takakuwa, K. Nakanishi, N. Berova, J. Am. Chem. Soc. 1997, 119 , 12024–12025. 91. T. Nehira, C. A. Parish, S. Jockusch, N. J. Turro, K. Nakanishi, N. Berova, J. Am. Chem. Soc. 1999, 121 , 8681–8691. 92. T. Nehira, K. Tanaka, T. Takakuwa, C. Ohshima, H. Masago, G. Pescitelli, A. Wada, N. Berova, Appl. Spectr. 2005, 59 , 121–125. 93. L. Geng, L. B. McGown, Anal. Chem. 1992, 64 , 68–74. 94. J. L. Lunkley, D. Shirotani, Y. Yamanari, S. Kaizaki, G. Muller, J. Am. Chem. Soc 2008, 130 , 13814–13815. 95. R. A. Poole, C. P. Montgomery, E. J. New, A. Congreve, D. Parker, M. Botta, Org. Biomol. Chem. 2007, 5 , 2055–2062. 96. M. Benedetti, P. Biscarini, A. Brillante, High Pressure Res. 2000, 18 , 285–289. 97. M. Benedetti, P. Biscarini, A. Brillante, E. Castiglioni, Enantiomer 1999, 4 , 63–66. 98. M. Benedetti, P. Biscarini, A. Brillante, Physica B 1999, 265 , 203–207. 99. T. Harada, T. Sato, R. Kuroda, Chem. Phys. Lett. 2008, 456 , 268–271. 100. K. Ikeda, W. Liu, Y. R. Shen, H. Uekusa, Y. Ohashi, S. Koshihara, J. Chem. Phys. 2005, 122 , 141103. 101. K. Ikeda, W. Liu, Y. R. Shen, H. Uekusa, Y. Ohashi, S. Koshihara, Chem. Phys. Lett. 2006, 422 , 267–270.

287

8 CIRCULAR DICHROISM OF DYNAMIC SYSTEMS: SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY Angela Mammana, Gregory T. Carroll, and Ben L. Feringa

8.1. INTRODUCTION Chiral systems that undergo controlled dynamic processes including switching the conformation of molecules and the self-assembly of complex structures permeate molecular biology [1]. Studying these phenomena in model synthetic or semisynthetic systems holds great promise in gaining a better understanding of complex and highly organized biological systems. The creation, ampliﬁcation, and control of chirality [2] is a fundamental issue in chemical biology [3]. Hence, exploring some of the basic pillars regarding dynamic chiral systems at the molecular and supramolecular level, particularly those that undergo conformational switching upon given particular stimuli, provides a strong groundwork by which to develop paradigms to guide forays into the hinterlands of biomolecular and materials sciences. Reﬁning our knowledge regarding conformationally switching molecules and assemblies may provide guiding principles for gaining a deeper understanding of enzyme processes, molecular recognition and self-assembly, the mechanisms behind biological molecular machines [4, 5] and possibly clues regarding the speculative areas of the origins of homochirality and prebiotic structures [3]. Additionally, the study of switchable molecules and assemblies has the potential to provide knowledge innovation applicable to developing new molecular level technologies related to information storage, transport, and optics [6–8]. In this chapter we will explore molecular systems that undergo changes in chiral conformation and conﬁguration upon thermal, chemical, photochemical, or mechanical stimuli. A preview of the kinds of systems we will examine is presented in Figure 8.1. CD spectroscopy is a method “par excellence” to study dynamic chiral systems at different hierarchical levels ranging from molecules to the supramolecular and macromolecular scale. Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

289

290

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a)

(b)

(c)

Figure 8.1. Dynamic chirality at the molecular and supramolecular level detected by CD spectroscopy. (a) A chiral molecule can direct achiral molecules to self-assemble into chiral supramolecular structures. (b) A chiral molecular switch or motor undergoes conformational changes that include inversion of molecular helicity. (c) Chiral molecules can self-assemble into chiral supramolecular structures, the chirality of which is determined by the enantiomer in excess. (See insert for color representation of the ﬁgure.)

Our survey will emphasize supramolecular systems [9–12]. Chemists have made considerable strides in the ability to create and control covalent bond formation in molecules with a high degree of efﬁciency, selectivity, and control of chirality, enabling the creation of numerous unique structures. Although the number of reactions that chemists have developed greatly exceeds that utilized by nature, the natural molecular world, constructed over billions of years of evolution, is vastly more complex than synthetic systems. Nature’s qualitative hegemony can be attributed to its mastery of supramolecular chemistry. Supramolecular structures rely on the formation of intermolecular bonds through a variety of interactions including electrostatic, dispersive, hydrophobic, hydrogen bonding, π -stacking, adsorption, or simply entrapment. A variety of molecular spaces that can host a guest molecule and form a supramolecular complex include cyclodextrins, zeolites, buckminsterfullerenes, DNA, micelles, various aggregates, and even transient solvent cages [13]. Our understanding of the intermolecular (noncovalent) bond is much less advanced in comparison with covalent bond formation. Biology has utilized molecular recognition and self-assembly to develop the most complex molecular systems known. As chemistry moves forward in the twenty-ﬁrst century, a more precise understanding of intermolecular interactions will enable greater control over the properties of self-assembled materials. Supramolecular structures are comprised of two or more molecules forming weak bonds. The noncovalent nature of the interaction allows for greater ﬂexibility of the molecular constituents in comparison with covalent structures; however, polyvalent interactions allow for the stabilization of supramolecular species through the cooperation of many weak interactions [14]. Akin to the advancement of traditional organic and inorganic chemistry, much research has been performed in order to gain an understanding of the dynamics, conformations, and

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

stability of supramolecular structures. Among the various aspects of molecular chemistry that has a supramolecular analogue, chirality is of fundamental importance, particularly in regard to chemical biology. The same principles regarding chirality at the molecular level can be conceptually extended to the supramolecular level. A supramolecular system is chiral if the noncovalent units comprising the system are arranged in an asymmetric manner wherein its mirror image is nonsuperimposable, even if the constituents comprising the structure are achiral. A number of systems have been reported that demonstrate the formation and ampliﬁcation of supramolecular chirality from achiral molecules. Two current principles regarding chirality at the supramolecular level include the sergeant-andsoldiers [15, 16] and majority-rules [17] effects. In brief, a sergeant-and-soldiers effect involves a small amount of chiral material that enforces a chiral structure on an assembly composed predominantly of achiral molecules which is dictated by the chirality of the sergeant. The majority-rules effect states that in a chiral but nonracemic assembly of two enantiomers, the one in the greatest amount will dictate the chirality of the system. CD spectroscopy provides an invaluable tool in elucidating the underlying themes of molecular and supramolecular chiroptical switching [18]. We will focus on some speciﬁc examples in which CD spectroscopy is used to interrogate molecules and assemblies that undergo reversible changes in chirality.

8.2. THERMAL SYSTEMS Modiﬁcation of chirality through thermal processes is a well-known phenomenon, the most widely studied systems being biomacromolecules that lose their helical secondary structure upon melting [19, 20]. Similarly, synthetic systems show thermoresponsive supramolecular chirality. An organogel-based chiroptical system that reversibly assembles achiral porphyrin molecules into a chiral supramolecular assembly by a thermally controlled aggregation/deaggregation process was realized through the coassembly of glutamic diamide gelators, L1 or D1, and a tetra-alkyl-substituted porphyrin, TPPOC12 H25 (Figure 8.2) [21]. A very well-known property of many porphyrins is the ability to self-assemble under certain conditions, giving rise to two possible types of aggregates: edge-to-edge (J-aggregates) and face-to-face (H-aggregates). TPPOC12 H25 itself does not gelate; however, co-mixing with the gelator in DMSO followed by cooling the sample resulted in a supramolecular gel showing a positive exciton-type CD signal with a crossover at 443 nm and a positive Cotton effect with a maximum at 402 nm, in accordance with the Soret band absorption of J- and H-aggregates, respectively. Using a gelator of opposite chirality resulted in the reverse CD signal. The minimum ratio of gelator to porphyrin that displayed an ICD was 30:1 gelator:porphyrin. As more gelator was added, the signal was increased until a maximum was achieved at a ratio of 150:1. The induced chirality was attributed to assembling the porphyrin in a chiral environment. When porphyrins without long alkyl chains were used (i.e., TPPOH and TPPMe), an ICD could not be obtained, indicating that the cooperative assembly of the porphyrin is necessary for transmitting chirality. Furthermore, when TPPOC12 H25 and L1 were co-gelled in toluene, no ICD was obtained. The UV–vis spectra indicated that the porphyrin does not aggregate in toluene, whereas in DMSO both J- and H-aggregates were formed, indicating that aggregation of the porphyrin is essential for its chiral coassembly with the gelator. Upon melting the gel the CD signal was lost, but was regained upon cooling. The cycle could be repeated several times without a loss of signal.

291

292

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

OR

N

NH RO

OR 80

N

OR R=H R = CH3 R = C12H25 H37C18

NH

(TPPOH) (TPPOMe) (TPPOC12H25) HN

C18H37 O

O O

Mol. Ellip. (104)

HN

Gel

40 0

Solution

–40 –80 350

400

450 500 550 Wavelength (nm)

600

NH O L-1 or D-1

Figure 8.2. A supramolecular chiroptical switch comprising achiral porphyrins (top left) that form gels upon co-assembly with a chiral gelator (bottom left). The chiral assembly can be switched on and off by adjusting the temperature as conﬁrmed by CD spectroscopy. On the right, CD spectra of a mixture of achiral porphyrins (90 μM) and gelator (13 mM) in DMSO are shown. At high temperature the compounds do not gel and a CD signal is not obtained. Cooling the solution results in the formation of a chiral gel. The gel displays an exciton-type CD signal with a crossover at 443 nm and a positive Cotton effect at 402 nm, corresponding to the Soret band of J- and H-aggregates, respectively. (Reproduced by permission of The Royal Society of Chemistry [21].)

Temperature does not always show a simple relationship to the formation or disappearance of chiral structures. In order to understand the effect of temperature on the chiral ampliﬁcation of dynamic supramolecular polymers, a simple and well-studied building block, trialkylbenzene-1,3,5-tricarboxamide, which assembles by triple-hydrogen bonding, was employed (Figure 8.3) [22]. Both sergeant-and-soldiers and majority-rules experiments were performed. A fast dynamic equilibrium exists between monomers and hydrogen-bonded stacks, allowing for modiﬁed self-assembled structures to form within one minute of mixing external components into the solution. Sergeant-and-soldiers experiments conﬁrmed that (S )-TABTC could induce chirality in aggregates of achiral TABTC [23]. Similarly, the system was shown to follow the majority-rules principle. As the opposite (S )-TABTC enantiomer is mixed into the solution of the (R)-TABTC, the CD spectrum weakens. At 0% ee the CD disappears. As the ee is further increased to 40%, the intensity of the CD signal reaches a maximum. The authors studied the temperature dependence of the CD intensity at 223 nm as a function of fraction of sergeant (Figure 8.4a) and as a function of enantiomeric excess (Figure 8.4b). Similarly, the temperature dependence of the net helicity as a function of fraction of sergeant and enantiomeric excess was studied (Figures 8.4C and 8.4D). The phenomena were quantiﬁed by

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

Figure 8.3. On the top left is presented the structure of a discotic molecule based on benzene1,3,5-tricarboxamide (TABTC). The R substituents can be achiral (TABTC) or chiral [(R)-TABTC or (S)-TABTC]. When R is chiral, self-assembly results in chiral supramolecular polymers. On the right a proposed right-handed supramolecular helix is presented (note that the side chains were replaced by methyl groups for clarity). (Reproduced by permission of the American Chemical Society [22].)

calculating the free energy penalties associated with helix reversal (HRP) in a stack and the introduction of a chiral monomer into a stack of its unpreferred helicity (mismatch penalty, MMP). For both sergeant-and-soldiers and majority-rules experiments, the HRP is associated with the disruption of three hydrogen bonds. In contrast, the MMP has a different physical meaning for the two types of experiments. In the sergeant-and-soldiers experiment, the MMP corresponds to the incorporation of a chiral sergeant into a stack of achiral molecules of its unpreferred helicity. In the majority-rules experiment the MMP is associated with the incorporation of one chiral enantiomer into the helix formed by the opposite enantiomer. The strength of the noncovalent interactions decreases with temperature; however, even at elevated temperatures long stacks composed of 100 monomers were predicted for this cooperative self-assembling system. The HRP was found to change very little with temperature; however, the MMP was found to decrease due to a slight increase of the intermolecular distance which reduces unfavorable steric interactions. The effect of temperature on the MMP value explains why the degree of chiral ampliﬁcation is reduced for the sergeant-and-soldiers system while it is enhanced for the majority-rules system. A lower MMP reduces the authority of the sergeant in the former case. In the latter case a lower MMP makes it easier for the minor enantiomer to join the helix dictated by the major enantiomer. Consequently, for the sergeant-and-soldiers experiment, upon increasing temperature a higher fraction of sergeant is required in order to obtain a homochiral system. On the contrary, for the majority-rules experiment, a lower ee is required in order to obtain a homochiral system.

293

294

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

10°C 20°C 40°C 50°C

20

10

(b) CD (mdeg)

CD (mdeg)

(a) 30

0

10°C 20°C 40°C 50°C

30 20 10 0

0.00

0.04

0.08

0.12

0.16

0.20

0.0

1 0.8 0.6 0.4 0.2 0 0.00

0.04

0.08

0.12

0.16

(d)

1

Net helicity (–)

(c) Net helicity (–)

Fraction of sergeant (–)

0.8

0.20

Fraction of sergeant (–)

0.2 0.4 0.6 0.8 Enantiomeric excess (–)

1.0

0.6 0.4 0.2 0 0.0

10°C 20°C 40°C 50°C

0.2 0.4 0.6 0.8 Enantiomeric excess (–)

1.0

Figure 8.4. Effect of temperature on the CD signal on chiral supramolecular assemblies of benzene-1,3,5-tricarboxamide monomers containing either varying amounts of a chiral sergeant that induces chirality into the assembly of achiral soldiers or a mixture of two enantiomers at various ratios [22]. (a) CD intensity at 223 nm as a function of fraction of sergeant at 10◦ C, 20◦ C, 40◦ C, and 50◦ C. (b) CD intensity at 223 nm as a function of enantiomeric excess at 10◦ C, 20◦ C, 40◦ C, and 50◦ C. (c) Net helicity as a function of fraction of sergeant at 10, 20, 40 and 50◦ C. (d) Net helicity as a function of enantiomeric excess at 10◦ C, 20◦ C, 40◦ C, and 50◦ C. (Reproduced by permission of the American Chemical Society [22].)

8.3. PHOTOACTIVE SYSTEMS Systems that can be addressed by light offer many advantages over systems requiring thermal or chemical stimuli. Light provides a clean, traceless, and noninvasive reagent that leaves behind no byproducts. Second, photochemical reactions are localized to the chromophoric functionalities that absorb at the wavelength employed, allowing for highly speciﬁc transformations to occur within a molecule [24]. Finally, the use of light allows for spatially directed transformations of materials through the use of well-established photolithographic techniques [25]. Many photoactive overcrowded alkenes have been shown to contain helical structures that give rise to CD signals [26]. A particularly interesting class of overcrowded alkenes behave as molecular rotary motors [27]; that is, one-half of the molecule can undergo continuous 360◦ unidirectional rotation relative to the other half. Changes in the helicity of the motor during the stages of the rotary process make CD spectroscopy an invaluable tool in characterizing the rotary motion. The overcrowded alkene in Figure 8.5, (P , P )-trans1, contains two aromatic halves connected by a photoisomerizable double bond [28]. Each half also contains a stereogenic center in a pseudoaxial conﬁguration, which is crucial for obtaining unidirectional rotation. The helical nature of the molecule is reﬂected in the sign

295

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

+400 Meax

≥ 280 nm

Meeq

≥ 380 nm

+200

Meeq

(P, P)-trans-1

(M, M)-cis-2 20°C

60°C

Meeq ≥ 380 nm Meax Meeq

Δε (1mol–1 cm–1)

Meax

0

≥ 280 nm Meax –400

(M, M)-trans-1

(a) (b) (c) (d)

–200

(P, P)-cis-2

220

240 λ (nm)

260

280

Figure 8.5. Rotary cycle of a light-driven molecular motor. A combination of photochemical and thermal isomerizations results in a net 360◦ rotation. Each of the four isomers has a distinct P- or M-helicity and a unique CD signal [28]. The CD spectrum of the initial motor at the start of the cycle, (P, P)-trans-1, is shown on the right (a). Upon absorption of a photon, the motor undergoes trans–cis isomerization to form (M,M)-cis-2, which displays an inversion of the CD signal at 217 nm (b). Thermal isomerization generates (P, P)-cis-2, which shows an inversion of the CD signal at 217 nm (c) in comparison with (M, M)-cis-2. A second photochemical isomerization inverts the CD signal at 217 nm and forms (M, M)-trans-1 (d). The original CD spectrum (a) is restored upon thermal isomerization to generate the starting conformation, (P, P)-trans-1.

of the CD spectrum at 217 nm (Figure 8.5). Upon absorption of a photon, the molecular motor undergoes trans–cis isomerization to form (M , M )-cis-2. CD spectroscopy reveals that the resulting cis motor has inverted helicity relative to the trans. The photochemical isomerization is accompanied by a change of the stereogenic center to a pseudoequatorial orientation, which is more unstable than the original pseudoaxial orientation. The motor undergoes a thermodynamically favorable thermal isomerization and forms (P , P )-cis2, which restores the more stable pseudoaxial conformation and inverts the helicity as reﬂected in the CD spectrum. The large free energy change provided by the thermal isomerization results in an irreversible conformational change, preventing the motor from rotating in the opposite direction. A second photochemical isomerization followed by thermal isomerization regenerates the initial CD spectrum, indicating the initial stage of the rotary cycle. A key challenge in applying the motor to exert nanomechanical-like forces is to demonstrate the rotational motion of the motor while anchored to a macroscopic surface such as a solid substrate (Figure 8.6). Appending the motor to a solid macroscopic surface puts limitations on the characterization options compared to the traditional solution-phase measurements because both (a) a smaller quantity of molecules is present (typically approximately 1014 molecules per cm2 in monolayers [11]) and (b) the analytical methods are limited to techniques that are applicable to the solid state. Although the use of

296

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

S

Rotor Axle Stator O

Legs

O O

O

n

Au Surface

S

n S

2 CD (mdeg)

hν

hν

0 –2 200

240

280 λ (nm)

320

Figure 8.6. Assembly of thiol-terminated light-driven rotary molecular motors on a semitransparent gold ﬁlm provides a monolayer of chiroptical material that can be analyzed with CD spectroscopy. The CD signals invert between positive and negative bands, corresponding to changes in the helicity of the molecules comprising the monolayer upon the application of photons and thermal energy. The initial spectrum (solid black) inverts (dotted black) after irradiation with UV light (λmax = 365 nm) at room temperature. After heating the surface (70◦ C, 2 h) the spectrum inverts again to restore the original (solid gray). A second dosage of photons inverts the signal (dotted gray). Heating brings the rotors back to the original orientation relative to the substrate [30]. (See insert for color representation of the ﬁgure.)

CD spectroscopy to analyze monomolecular layers of organic molecules is rare [29], it provided an invaluable tool in characterizing the rotary motion of molecular motors attached to a semitransparent gold ﬁlm [30]. In order to use CD spectroscopy, only a very thin layer of gold could be used in order to minimize the optical absorbance of the system. Therefore, 5 nm of gold was deposited onto both sides of an aminosilane-coated quartz substrate. Despite the low amount of material present in monolayer systems, the motor provides a strong enough CD signal to demonstrate rotary motion, as evidenced by the inversion of the CD spectra upon given photochemical and thermal stimuli (Figure 8.6). In addition to providing evidence that the motor can access the four stages of the rotary cycle, CD spectroscopy was used to uncover the proper length of spacer required to minimize quenching of the photochemical isomerization by the gold ﬁlm. When spacers of eight atomic units were used, the CD signals did not change sign upon irradiation; however, when the motor contained spacers of 16 atoms, the gold-mounted chromophore was able to undergo photoinduced isomerization followed by thermal helix inversion. In addition to controlling the helicity of the chromophore, the overcrowded alkene can also exert intramolecular control over the helicity of a polymer. The stages of the rotary process have been shown to inﬂuence the twist sense of helical poly(hexyl isocyanate) (PHIC) when attached to the end-terminus of the macromolecule (Figure 8.7) [31]. Poly(isocyanates) (PIC) are stiff helical polymers that exist as a racemic mixture of P - and M -helices in the absence of an asymmetric inﬂuence [32, 33]. X-ray studies ˚ A similar on poly(n-butyl isocyanate) have revealed an 8/3 helix with a pitch of 5.14 A. helical structure is maintained in solution [32, 34, 35]. Chiral perturbations can favor the

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

Figure 8.7. A single light-driven molecular motor attached to the end-terminus of poly(hexyl isocyanate) (PHIC) is used to control the twist sense of the polymer. The trans isomer of the motor end-group exerts no chiral induction, allowing an equal probability for the P and M helices to form. Photochemical isomerization to the cis form induces a preferred handedness to the polymer backbone. Thermal isomerization inverts the handedness. Restoring the motor to its original trans form brings the polymer solution back to a racemic mixture with no preferred twist sense. (Reprinted by permission of John Wiley & Sons, Inc. [31].)

presence of one helical twist over the other. For example, when a PIC contains chiral repeat units, the two diastereomeric helices, which have the same stereogenic conﬁguration in the repeat units, can have a different energy, making one helix more favorable over the other. Many studies have shown that PICs containing chiral repeat units give rise to CD spectra that indicate that the backbone assumes a preferred helicity [36–38]. A rather striking example is the preference for one helicity from the subtle asymmetric presence of a deuterium in place of a hydrogen atom [33, 35]. The resulting PHIC is optically active and displays a CD spectrum with a band at approximately 250 nm in the region where the recurring amide groups of the backbone absorb. The preference for one helicity is attributed to cooperative effects among the repeat units which amplify a slight energetic preference for one helix. The use of a chiral solvent can also bias the P - or M -helicity [39]. The preparation of PHIC with the presence of a molecular motor at the end-terminus of the macromolecule provides an example in which the helical twist sense can be reversibly controlled [31]. A molecular motor containing a benzamide functionality in the lower half was used to initiate the polymerization of n-hexyl isocyanate from the benzamide’s sodium salt. The use of a molecular motor as the initiator results in a PHIC with a chiral environment at the α-chain end of the polymer that can inﬂuence the twist-sense of the macromolecule for certain stages of the rotary cycle in which the helical end-group can interact with the repeat units. In the trans conformation the upper naphthalene shows no detectable interaction with the polymer backbone, and so the helix has no bias (Figure 8.8). No detectable excess of P - or M -helicity is revealed by CD spectroscopy because the signal of the motor-PHIC polymerized from an enantiomerically pure motor matches the signal of the enantiomerically pure motor alone. Photoisomerization of the motor to the cis conﬁguration results in an increase in the intensity of the CD signal and is attributed to an induced preference for one helicity of the polymer. The induced helicity can be attributed to the encroaching naphthalene rotor, which imposes a chiral environment on the nearest repeat unit and favors one helicity over the other. Cooperative interactions along the chain would then amplify the chirality, resulting in a preferred helical twist of the polymer. After thermal

297

298

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

step 1

step 2

polymer: R N PIM = 50:50 Bz

365 nm

stable (2'S)-(M)-trans-2-PHIC (R = PHIC) stable (2'S)-(M)-trans-3 (R = COCH3)

R polymer: N excess M helicity Bz

unstable (2'S)-(P)-cis-2-PHIC (R = PHIC) unstable (2'S)-(P)-cis-3 (R = COCH3)

R polymer: N excess P helicity Bz stable (2'S)-(M)-cis-2-PHIC (R = PHIC) stable (2'S)-(M)-cis-3 (R = COCH3)

(b)

(a)

(c)

4

4

2

2

2

0 –2 –4

θ (mdeg)

4 θ (mdeg)

θ (mdeg)

Δ

hn

0 –2

–4

–4 220

280 λ (nm)

340

0 –2

220

280 λ (nm)

340

220

280 λ (nm)

340

Motor-polymer Motor (control)

Figure 8.8. The helicity of motor-terminated poly (hexyl isocyanate) (PHIC) can be followed with CD spectroscopy (Et2 O, −20◦ C). The CD spectra of the motor without PHIC (R = COCH3 ) and the motor attached to PHIC (R = PHIC) are shown. (a) In the trans form of the motor the CD spectra of the motor and motor polymer are the same, reﬂecting the lack of preferred helicity in the polymer backbone. (b) Photochemical isomerization (UV lamp, λmax = 365 nm) to the cis form induces a preferred handedness to the polymer helix, which is reﬂected in an increase in the intensity of the CD signal relative to the motor alone. (c) Thermal isomerization (20◦ C, 30 min) inverts the CD spectrum, indicative of a helix inversion. Reprinted by permission of John Wiley & Sons, Inc. [31].

isomerization the CD spectrum of the motor-PHIC is inverted and more intense than the motor alone. The change in the CD implies that the upper-half rotor remains close enough to the polymer backbone to maintain inﬂuence over the helical twist; however, the rotor is now on the opposite side of the polymer and induces the reverse twist. Photochemical and thermal isomerization bring the motor back to the starting state in which the polymer has no preferred handedness. The changes in the helicity of both the motor and polymer can be followed with CD spectroscopy (Figure 8.8). Recent follow-up experiments show that these kinds of systems can be used to control the pitch of cholesteric liquid crystals [40] and form interesting toroidal morphologies when dried on a solid substrate [41]. Similarly, incorporating photoisomerizable molecules into the backbone of a poly(isocyanate) can reversibly switch the helical sense of the polymer. The azobenzene chromophore, which can be reversibly photochemically switched between cis and trans isomers (and thermally isomerized from the cis to trans form), has a large body of literature related to its use as a photoswitch [42, 43]. A PIC containing chiral azobenzene pendant groups in the repeat units, AzoPIC, was shown to induce a preferred helicity when the chromophore was in the trans conformation (Figure 8.9) [44]. Photoisomerization to the cis isomer resulted in a reversal of the CD spectrum as shown in Figure 8.9. The dashed curve represents the polymer after irradiation and displays the mirror image of the spectra of the polymer before irradiation, indicating that the direction of the helicity was reversed. It was found that in order to induce a preferred helical twist the chiral group must be linked to the same phenyl ring of the azobenzene

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

Figure 8.9. A poly(isocyanate) (PIC) copolymer with repeat units that contain an azobenzene pendant group bearing a chiral chain displays a CD signal (THF, 0.5 mg/mL). Photoisomerization of the azobenzene unit reverses the sign of the CD signal. (Reproduced by permission of the American Chemical Society [44].)

as the isocyanate backbone. Variation of the location of the stereogenic center allowed the dominant direction of the helicity to be controlled. A variety of azobenzene PICs were synthesized, and their response to photoisomerization was found to depend on the stereogenic center [45]. The various systems synthesized showed different behavior in response to irradiation. Isomerization from the trans to the cis isomer in some polymers was shown to lead to an increase in the chiral interaction, detected by an increase in the CD spectrum, while in other polymers a decrease in the interaction was deduced based on an attenuation of the CD signal. Additionally, strong changes in the optical rotation were observed for some of the polymers during photoisomerization. Photoisomerization of azobenzene chromophores has been shown to affect the chirality of other helical systems as well. For example, a foldamer, a small oligomer that adopts a secondary structure, was designed by incorporating azobenzene into the core of an oligo(meta-phenylene ethynylene) derivative [46]. In the cis form of the azobenzene the foldamer can obtain a helical conformation showing a bisignate CD signal (Figure 8.10). Photoisomerization to the trans form denatures the helix because the coordinates of the foldamer components compromise the propensity to form a stable helical structure. An attenuation of the CD signal upon thermal isomerization from the trans to the cis form was attributed to depletion of the helical conformation through unfolding. Regeneration of the cis form by photo- or thermal isomerization restores the initial helical structure and is accompanied by the regrowth of the initial CD signal. Photoswitchable molecules have also been shown to modify the conformation of proteins in a reversible manner [47]. Such changes can be monitored by CD spectroscopy. For example, spiropyran modiﬁed poly(l-glutamate) (SPPGA) (Figure 8.11) has been shown to be effective in inducing such changes [48]. Spiropyran can exist as a neutral spiro form or as a zwitterionic merocyanine form [49]. The two states contain a considerably different geometry and polarity. SPPGA was shown to undergo conformational changes upon switching between the two forms. When the peptide-appended switch is in the merocyanine form (MEPGA), the polypeptide assumes a random coil conformation. Photoisomerization to SPPGA allows the polypeptide to undergo a transition to an α-helix. In hexaﬂuoro-2-propanol, the spiro form thermally converts to the merocyanine form with a half-life of 2.5 h at 25◦ C, regenerating the disordered conformation. The

299

300

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

CO2Tg

Tg

O

O

O

O

O

O

O

O

OMe

H 6

O N N

O

O

O

O

O

O

OMe

H 6

20

CO2Tg helix

θ (mdeg)

10 0 −10

hν, Δ

hν

−20 250

325 λ (nm)

400

Random coil

Figure 8.10. A photoswitchable foldamer composed of an azobenzene core bearing two oligo(meta phenylene ethynylene) pendant groups. The cis form of the switch can be accessed by irradiation with UV light (λmax ∼ 365 nm). The CD spectrum of the cis form is shown (acetonitrile, 5.6 × 10−6 M) and has been attributed to the foldamer obtaining a helical structure. Thermal isomerization to the trans form disrupts the chiral structure and shows a decrease in the CD signal. (Reprinted by permission of John Wiley & Sons, Inc. [46].)

changes in the conformation of the polypeptide manifest as changes in the CD spectrum. The characteristic negative CD signal of an α-helix with minima at 208 and 222 nm is generated and increases during irradiation. Subsequently, the negative CD signal becomes smaller in magnitude as SPPGA converts to MEPGA in the dark. UV–vis and ﬂuorescence measurements indicate that the merocyanine form of the switch dimerizes, which is the proposed driving force for the distortion of the structure. Dithienylethenes (DET) provide another example of an often exploited photoswitch (Figure 8.12) [50]. The dithienylethene unit can undergo reversible photochemical ringopening and closure reactions upon absorption of a photon of the appropriate wavelength. The two forms of the switch absorb at sufﬁciently different wavelengths to allow for a particular state to be selected through the use of conventional lamps emitting UV and visible light. The open form of the switch absorbs in the UV region of the electromagnetic spectrum, while the closed form of the switch absorbs in both the UV and visible region. DET-1o can be reversibly photochemically closed to form DET-1c, and reopened to form DET-1o without fatigue for at least ﬁve cycles, and possesses a thermal stability that prevents the ring-opening and closure reactions from occurring in the dark. In the open form the switch exists as a dynamic structure that rapidly interconverts between a P - and M -helicity. Photochemical ring closure locks the switch in equal amounts of the RR and SS enantiomers; however, the chirality can be controlled by photochemical switching in a chiral environment as will be discussed below.

301

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

NO2

dark N H (CH2)2 O CO (CH2)2 N H

CH

O

NO2

N+ (CH2)2 O CO

hν COOH

(CH2) 2

(CH2)2

N H

C H

N H

CO

SPPGA

CH

N H

–O

COOH (CH2)2 C H

CO

MEPGA

–20

–210

230 λ (nm)

light

–10 dark

[Θ]·10–3

0

250

Figure 8.11. A spiropyran-modiﬁed PGA (poly(L-glutamate)), SPPGA, undergoes a ring opening reaction in the dark to the merocyanine form of the dye, MEPGA. Formation of MEPGA is accompanied by a distortion of the helical structure of the peptide (lower right) which is reﬂected in a change in the CD spectrum (lower left). Top curve (positive signal), solid line: MEPGA in hexaﬂuoro-2-propanol before irradiation. Bottom curve (negative signal), solid line: SPPGA is generated after irradiation with sunlight. Intermediate spectra, dashed lines: thermal isomerization from SPPGA to MEPGA over a time period of 8 h. (Reproduced by permission of the American Chemical Society [47, 48].)

A dithienylethene photochromic unit functionalized with (R)-1-phenylethylaminederived amides (Figure 8.12) self-assembles into supramolecular structures through hydrogen-bond formation [51]. The stereogenic center causes the assembly to form a helical ﬁber. At room temperature, DET-1o forms a gel in organic solvents such as toluene, benzene, and hexane. Utilizing photochemical processes that control the molecular conformation of the switch and using thermal processes that control the macroscopic aggregation, it is possible to realize four supramolecular chiroptical states of the gel (Figure 8.13). The aggregated molecules form helical ﬁbers and show a CD band at approximately 320 nm (Figure 8.14). Only the negative half of the exciton band is shown because the positive half is obscured by the solvent, which in this case is toluene. The CD signal of the aggregate is attributed to locking of a selected molecular helicity, M or P , of the open form of the switch when it is conﬁned in the self-assembled structure. The enantiomeric (S )-1-phenylethylamine-derived DET-1o forms an aggregate that shows the opposite CD signal. The chirality of the stereocenters at the hydrogen-bonding components of the molecule determines the selection of the helical conformation of the

302

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

H N

S O

H H

S

O

DET-1o

Vis H N

S

UV

Vis

UV

H N

S

O

M – DET- 1o

P – DET- 1o

O (S,S) – DET- 1c

(R,R) – DET- 1c

DET-1c

Figure 8.12. A dynamically helical photoswitchable dithienylethene chromophore containing chiral amides can be photochemically switched between an open, DET-1o, and closed, DET-1c, form. The open form of the switch interconverts between two helical conformations. The helicity of the open form during photochemical ring-closure determines the stereochemistry of the two stereogenic centers on the photoswitchable unit of the molecule. The amides allow the molecule to self-assemble into chiroptical ﬁbers [51].

Sol 1

Gel (α) 1

Gel (β) 1 Vis

Vis UV Gel (α) 2 (PSS)

UV

Δ

Sol 2 (PSS)

Gel (β) 2 (PSS)

Figure 8.13. Scheme of the aggregation and switching processes of a DET-1 gel (1–4 mM, toluene), which can access four chiroptical states. Upon cooling an isotropic solution of the open form of DET-1 (Sol 1), a stable Gel (α) 1, with a P-helicity, is obtained. It is possible to reversibly close and open the central ring of DET-1, cycling between the stable Gel (α) 1 and the metastable Gel (α) 2 (PSS) with high diastereoselectivity (96% DE) and P-helicity, by irradiating with UV light (313 nm) and visible light (> 420 nm), respectively. Note that during the photochemical step the helicity of the gel is preserved; however, the photochemical ring-opening or ring-closure changes the rigidity and chirality (ﬁxed or dynamic) of the central unit and as a consequence the stability of the chiral aggregate. Heating of the metastable Gel (α) 2 (PSS) leads to an isotropic solution of 2 [sol 2 (PSS)], which, upon cooling, results in stable Gel (β) 2 (PSS) with M-helicity. The thermal processes (50◦ C required to fully melt gel) irreversibly convert the gel from a metastable to a stable aggregate with an inversion of helicity. Again, the photochemical step is reversible and occurs with retention of the supramolecular chirality on going from the stable Gel (β) 2 (PSS) to the metastable Gel (β) 1, and vice versa. Finally, heating of the metastable Gel (β) 1 gives the isotropic solution of DET-1 (sol 1) closing the cycle shown [51].

core, M or P , during aggregation. The CD signal disappears when the aggregate is melted by heating to 50◦ C (Figure 8.15) and reaches a maximum at temperatures below 0◦ C. The attenuation of the CD bands correlates with changes in the molar fraction of DET-1o existing as a free monomer compared to the amount aggregated as measured by NMR. When completely dissolved, DET-1o photochemically cyclizes with no diastereomeric

303

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

–100

0

(mdeg)

(mdeg)

200

–200

0

–400

–100

–600

–200

300

350 400 Wavelength (nm) (a)

450

300

400 500 600 Wavelength (nm)

700

(b)

Figure 8.14. (a) CD spectra of stable gel DET-1o obtained by cooling a hot solution of DET-1o (1.8 mM) in toluene (solid line) and an unstable gel of DET-1o obtained by photochemically ringopening (λ > 420 nm) a stable gel of DET-1c (3.6 mM) (dotted line). The dashed line corresponds to DET-1o (0.35 mM) in solution and shows that when DET-1o is not aggregated, it does not produce a CD spectrum. (b) CD spectra of an unstable gel of DET-1c with 96% DE obtained by irradiating a stable gel of DET-1o (3.6 mM) (solid line), a stable gel of DET-1c with 96% DE obtained by heating (100◦ C) and cooling (0◦ C) the aforementioned unstable gel of DET-1c (dashed line) and a gel of DET-1c with no DE obtained by photochemical ring-closure (λ = 313 nm) of DET-1o and cooling (dash–dotted line) [51].

Figure 8.15. Temperature dependency (−15◦ C to 70◦ C) on the intensity of the CD band of DET1o at 320 nm. The extent of aggregation as probed by NMR and the %DE during photochemical cyclization of DET-1o to DET-1c decrease with increasing temperature in correlation with the attenuation of the CD signal [51].

excess (DE); however, in the gel form, DET-1o undergoes photochemical cyclization to DET-1c with 96% DE. The DE correlates with the intensity of the CD and the molar fraction of aggregated DET-1o. In the closed form the chirality of the methyl substituents at the photochromic core are locked. The aggregate of DET-1c can be photochemically ring-opened to regenerate the aggregate of DET-1o.

304

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Aggregation of a solution of DET-1c containing 0% DE forms a gel that gives a similar CD spectrum obtained upon photochemical ring closure of the aggregate of DET-1o. NMR reveals that the aggregate contains only one diastereomer of the closed switch. Interestingly, the diastereomer of DET-1c that gels is opposite to that formed upon stereoselective photocyclization of DET-1o in the aggregated state. Melting an aggregate of DET-1c formed by photochemical ring closure of an aggregate of DET1o, followed by cooling to regenerate the gel, results in an aggregate with the opposite helicity of that present after photochemical ring closure. The results suggest that in the open form, the chirality of the peripheral amide groups dictates the chirality of the selfassembled aggregate. However, in the closed form the chirality of the central photoactive part of the switch dictates the thermodynamically most stable helical conformation of the aggregate. Photochemical ring closure in the gel results in a metastable aggregate of DET-1c. Melting of the gel and re-cooling is necessary in order to access the more thermodynamically stable form of the supramolecular ﬁber. Similarly, when a stable aggregate of DET-1c is formed, photochemical ring opening forms a metastable aggregate of DET-1o. The CD signal of the metastable aggregate inverts upon successive heating and cooling. In addition to studies of switch DET-1, switch DET-2 (Figure 8.16), which differs from switch DET-1 by the presence of two methylene units between the outer phenyl rings and the stereogenic center, can also form chiral gels [52]. Interestingly, the gels formed by DET-2 have the opposite chirality of those formed by DET-1. The combination of ﬁxed stereogenic centers in the hydrogen bonding units and a dynamic helicity that can be switched “on” and “off” allows for a four-state switching system in which molecular chirality and supramolecular chirality communicate. The self-assembling system described shows how the molecular and supramolecular chirality in a chemical system can inﬂuence each other. The self-assembling DET system described above was further shown to amplify chirality in aggregates of the isostructural, achiral switch DET-3o, shown in Figure 8.16 [52]. While both DET-1o and DET-2o can form chiral gels, DET-3o, which lacks a stereocenter, forms an achiral aggregate. Photochemical ring closure of the gel results in equal amounts of the RR and SS isomers. However, the chirality of DET-1o or DET-2o can induce DET-3o to aggregate in a chiral

H N O

H N n

H N

S

S

n

n

H N

S

O

O

H N

S

H N O

O

S

O

DET-1o n = 0 DET-2o n = 2

DET-3o (M)

DET-3o (P)

UV Vis

UV Vis

UV Vis

S

S

O O DET-1c n = 0 DET-2c n = 2

H N n

H N

S O

H N

S O

DET-3c (SS)

H N

S O

H N

S

H N

S O

DET-3c (RR)

Figure 8.16. Structure of photoresponsive organogelators for dynamic chiral selection and ampliﬁcation. The compounds aggregate via hydrogen bond formation. Irradiation with UV (λmax = 313 nm) and visible (λ > 420 nm) light [52] produces photochemical ring-closure and ring-opening reactions. DET-3o lacks a stereocenter and forms achiral aggregates; however, in the presence of DET-1o or DET-2o, DET-3o aggregates in a chiral manner as shown in Figure 8.17 [52].

305

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

manner. When DET-3o is coassembled with DET-1o or DET-2o, a supramolecular chiral aggregate results, with the chirality transferring and propagating to achiral DET-3o. When DET-3o is coassembled with DET-1o or DET-2o, the resulting CD spectrum shows a more intense signal compared to DET-1o or DET-2o assembled alone (Figure 8.17), indicating that DET-3o coassembles with its chiral analogue and that the selected helicity of the chiral switch is imparted to the achiral switch. The chirality induced in DET-3o was locked by photochemical ring closure and shown to proceed with 94% ee of DET-3c. As the ratio of DET-3o increases, the intensity of the CD signal decreases and correlates with a decrease in ee. The transfer to and ampliﬁcation of molecular chirality in DET3o by DET-1o or DET-2o is in line with previous sergeant-and-soldiers systems. The contrasting effects of DET-1o and DET-2o, which differ only in the presence or absence of two pairs of methylene units, show how subtle differences in molecular structure can inﬂuence both the resulting molecular and supramolecular properties of the system. The dithienyl systems described above utilize switches containing an inherent chirality. The resulting chiral supramolecular structures in turn inﬂuence the chirality of the molecules constituting the assembly. In order to initiate the chirality transfer, a stereogenic center needs to be present within the constituents. It is also possible to induce chirality in an achiral DET by using an external helical template. dsDNA provides a well-studied, chiral nanoscale building block that affords a high level of control over the structure [53, 54] and possesses charge-transport capabilities [55]. Its use as a scaffold for building new classes of organic materials merits attention [56]. Among the number of ways to functionalize dsDNA, electrostatic binding provides perhaps the simplest approach [57]. The construction of a DET switch bearing two primary amines provided a suitable candidate for binding studies (Figure 8.18) [58]. Molecular models show that the amines can access coordinates that closely match the locations of negative charge density on the DNA base pairs due to the phosphate groups. Similar to the DET switches described above, both DET-4o and DET-4c do not display CD spectra when dissolved in solution. However, addition of a poly(dGdC)2 2+3 150

CD (mdeg)

100

2

50 0 –50 –100

1 1+3

–150 300

320

340

360

380

400

Wavelength (nm)

Figure 8.17. CD spectra of chiral gels of DET-1o (1.3 mM, toluene) (solid line 1, negative band), DET-2o (1.2 mM, toluene) (solid line 2, positive band) and coassemblies of the achiral DET-3o (1 equiv.) with the respective chiral switch DET-1o (dotted line 1 + 3) or DET-2o (dotted line 2 + 3). In both cases the addition of DET-3o increases the CD intensity of gels containing DET-1o or DET-2o. (Reproduced by permission of the American Chemical Society [52].)

306

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

UV O

S

Vis

HN

NH H2N

O

O

S

DET-4o

NH2

S

S

H2N

+

O HN

NH

NH2

DET-4c

DET-4o

–

+ –

dG

dC

Figure 8.18. A photoswitchable chiroptical DNA complex. At the top is shown the photoequilibrium between the open (DET-4o) and closed (DET- 4c) forms of a dithienylethene molecular switch that contains pendant ammonium groups to confer water solubility and allow the switches to bind electrostatically to the polyanionic backbone of DNA when the amine is protonated. Photochemical ring closure was accomplished through the use of a UV lamp (λmax = 365 nm) using a 340-nm cutoff ﬁlter. Photochemical ring opening was performed with visible light using a 520-nm cutoff ﬁlter. Molecular models (created using Hyperchem®) show that the distance between the terminal ammonium functionalities closely resembles the distance between the anionic phosphate groups of a guanosine (G)–cytosine (c) base pair [58]. (See insert for color representation of the ﬁgure.)

oligonucleotide to a solution containing DET-4o or DET-4c results in an ICD corresponding to the switch (Figure 8.19). The chirality of the DNA double helix is transmitted to the orientation of the switches comprising the supramolecular complex. Additionally, the CD signal corresponding to the DNA attenuates, indicating that both components modify the structure of the other. The intensity of the CD grows until the amount of switch is approximately 79% the amount of base-pairs. The results are similar to studies regarding the binding of simple mono- and divalent cations to DNA, of which the charge compensation was found to be no larger than 85% [59]. Similar results were found for both the open and closed form. The titration experiments were used to calculate the binding constants of DET-4o and DET-4c, both of which are 2 × 105 . The CD measurements were complemented by UV–vis absorption measurements. The spectra of DET-4o and DET-4c show a hypochromic effect and a red shift. The strength of the binding could be controlled by adjusting the pH of the solution. As the pH was increased, the CD signal decayed, consistent with the hypothesis that the switch binds via electrostatic interactions between the ammonium groups of the switch and the negatively charged phosphate groups exposed at the outer surface of the DNA double helix. When the pH increases above 9.12, the CD signal corresponding to the switch is completely removed. The switches were also shown to bind to poly(dAdT)2 . Interestingly, the interaction showed enhanced chiroptical activity. As found for the poly(dGdC)2 , when poly(dAdT)2 is added to a solution of DET-4o or DET-4c the UV–vis absorption spectra undergo hypochromic effects and the CD spectra show ICDs corresponding to the isomer of the

307

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

– – – (a)

(c)

– – – – (b)

(d)

Figure 8.19. Changes in both the CD (a) and absorption (c) spectra of the dithienylethene (DET-4)–poly (dGdC)2 complex due to cyclically performing photochemical ring-opening and ring-closure reactions. All spectra were taken at room temperature in aqueous buffer at a pH of 6.5. Photochemical ring closure was accomplished through the use of a UV lamp (λmax = 365 nm) and a 340-nm cutoff ﬁlter. Photochemical ring opening was performed with visible light, using a visible light-emitting lamp equipped with a ﬁber optic and a 520-nm cutoff ﬁlter. The CD spectrum of the open DET- 4o- poly (dGdC)2 complex shows a clear ICD corresponding to the open form of the switch. Irradiation with visible light results in the attenuation of the band corresponding to the open DET- 4o and the growth of a signal in the visible region corresponding to the closed DET- 4c. Several cycles of photochemical switching can be performed as indicated by the reversible changes in the CD signal (b) at 350 nm and the UV–vis signals (d) at 331 nm (DET4o) and 560 nm (DET- 4c) [58].

switch. The optimum ICD was obtained at a ratio of 1:1.7 switch:base pairs compared to a ratio of 1:1.3 for the poly(dGdC)2 complex. The ICD spectra for the poly(dAdT)2 appear more intense, especially when the bisignate signals for the closed form of the switch are compared. Also, because the signal for the DNA is slightly blue-shifted, a clear bisignate signal for the open form is visible. A more complicated relationship between the ratio of switch to DNA was observed. The intensity of the ICDs did not simply increase and plateau as DNA was added. Furthermore, for some ratios, unique ICD signals could be obtained, indicating that intermolecular interactions between the bound switches seem to play a role. The higher charge density associated with poly(dAdT)2 may allow for multiple orientations of the switch, the most stable of which depends on interactions between the bound switches. Regardless of the sequence used, both DET-4o and DET4c retain their photochemical switching abilities when complexed to the DNA. The chiroptical response can be modulated with photons as shown by the reversible changes in the CD spectra, allowing for a multitude of unique chiral states to be generated (Figure 8.19).

308

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

8.4. CHEMICAL AND MECHANICAL SYSTEMS A porphyrin system displaying a dynamic memory in which supramolecular chirality can be reversibly stored and released provides an elegant example of a switchable molecular assembly that responds to chemical stimuli (Figure 8.20) [60]. Achiral porphyrins have been shown to form chiral aggregates in the presence of chiral noncovalent amino acid polymers. The use of cationic and anionic porphyrins permits hetero-aggregation to occur via electrostatic interactions of the oppositely charged porphyrins. The strength of the interactions allows the assembly of achiral molecules to maintain supramolecular chirality after removal of the asymmetric template. By utilizing porphyrins containing

–

–

SO3

O3S

–

N

N

NHHN

N–H H–N

N

N

O3S

–

H2TPPS anionic porphyrin

SO3

N

N

NH

+4H+

N

–4H+

N HH N

+H N

N H+ TpyP4+

H2TpyP neutral porphyrin

route a

+

HN

N

+

–H+

+

N

N

H6 cationic porphyrin

+H+

monomers nonchiral aggregate

route b

chiral aggregate

+H+

cationic porphyrins anionic porphyrins neutral porphyrins

monomers

chiral seeds chiral aggregate

Figure 8.20. On the top are shown the structures of, respectively: the anionic meso-tetrakis (4-sulfonatophenyl) porphyrin (H2 TPPS), the neutral meso-tetrakis (4-pyridyl) porphyrin (H2 TpyP), and its protonated form, the cationic (H6 TpyP)4+ . On the bottom the system is described schematically. Cationic and anionic porphyrins form a chiral aggregate in the presence of one enantiomer of the chiral template phenyl alanine (either L or D). After removal of the template, the aggregate maintains chirality. The aggregate can be disassembled by deprotonation. Two routes of disassembly can be hypothesized. In route a, the aggregate fully disassembles, leaving only monomers in solution. Reprotonation would result in an achiral aggregate. In route b, the aggregate disassembles; however, the disassembly is not complete due to the slow rate of deprotonation and leaves small chiral aggregates in solution. Upon reprotonation, the small chiral aggregates can be thought of as chiral ‘‘seeds’’ that direct the regrowth of the template in a chiral manner. (Reproduced by permission of the American Chemical Society [60].)

309

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

ionizable groups in the meso positions, a system for which it is possible to imprint, store, release, and restore chirality was developed. The porphyrins employed contain four protonated pyridine groups in the meso positions. Deprotonation results in the loss of the cationic charge and consequent disassembly of the chiral aggregate. The CD signal of the aggregate disappears upon increasing the pH (Figure 8.21). The deprotonation step is kinetically slow, allowing for chiral seeds to remain in solution, provided that 24 hours have not elapsed. The chiral seeds are very efﬁcient templates for their own selfpropagation. Reprotonation therefore leads to the regrowth of the supramolecular chiral assembly, which displays a CD signal matching the signal of the initial chiral aggregate. The two states can be cyclically addressed even in the absence of the chiral template. When a template containing the opposite chirality is employed, the CD spectrum shows the mirror image. The above concept was extended to anionic porphyrin J-aggregates templated from or -[Ru(phen)3 ]2+ [61]. It is known that H2 TPPS4 is zwitterionic at a pH lower than 3 and in the presence of a millimolar concentration of salt. Under these conditions H2 TPPS4 self-assembles to give both H- and J-aggregates, displaying absorption maxima at 422 and 490 nm, respectively. The authors showed that J-aggregates formed in the presence of one enantiomer of the title ruthenium complex can maintain memory of the chirality imprinted at pH 6 and can be switched on and off by simply modulating the pH. The chiroptical

15 a1,2,3

10 L

5

b1,2

0

0

–40

–5 D –10 –15 350

450 λ (nm)

a 0.0 370 390 410 430 450 λ (nm) (b)

–80

a1,2,3

400

b

0.8

40 Δε

CD (mdeg)

80

Abs (a.u.)

1.6

500

550

(a)

Figure 8.21. (a) CD spectroscopy is used to detect the disappearance and reappearance of the chiral porphyrin aggregate. The sign of the CD signal is dependent on the chirality of the template that is used to initially imprint the chirality. CD spectra of the erasure and restoration of the chiral aggregate upon changing the pH are shown. Aggregates were made using either L- or D-phenyl alanine templates at a pH of 2.3 and show CD spectra that are mirror images. Increasing the pH to 9 results in the disappearance of the CD signal (curve a1 becomes b1 for either the L- or Reducing the pH back to 2.3 restores the signal (a2 ). The cycle can be continued to

D-aggregate).

give b2 (no signal) and a3 (restored signal). At least 10 cycles can be performed. (b) The absorption spectra of the aggregate at pH 2.3 (a) and at pH 9 in which the aggregate disassembles to predominantly monomer (b). (Reproduced by permission of the American Chemical Society [60].)

310

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

response could be controlled by cycling the pH between 2.5 and 6 of a solution containing chiral J-aggregates of H4 TPPS4 templated from -[Ru(phen)3 ]2+ in the presence of an excess of the -enantiomer of the complex. Upon raising the pH to approximately 6, the ICD of the J-aggregate disappears. Decreasing the pH back to 2.5 restores the CD signal with the same sign of the exciton couplet as that of the starting J-aggregate. As in the previous example, the propensity to reproduce the chiral conformation of the original aggregate during regrowth despite the introduction of a template that possesses the opposite chirality of the original is due to the remarkable inertness of these chiral aggregates and consequently to the presence of chiral seeds that have a stronger driving force to reconstruct the memorized supramolecular architecture compared to the ability of the template of opposite chirality to direct the converse structure. A thin ﬁlm based on the layer-by-layer (LBL) assembly of DNA and poly(allylamine hydrochloride) (PAH) was shown to induce chirality in tetrakis(N -methylpyridinium-4yl)porphine upon its addition to the preexisting ﬁlm (Figure 8.22) [62]. The induced chirality was evident by the appearance of a bisignate CD signal in the Soret band region with positive and negative Cotton effects at 423 and 445 nm, respectively, and a crossover at 432 nm. The authors attribute the Cotton effects to intercalation into the DNA and electrostatic binding [63]. A negative Cotton effect at 268 nm was assigned to DNA. Differences between the DNA CD band in the ﬁlm compared to solution were attributed to polymer-salt-induced aggregation during assembly with PAH. The induced

Θ (mdeg)

12

(a)

0

–12

NH3 and H2O

HCI

–24 HCI Absorbance

(b) 0.6 As-prepared film (ICD)

A (no ICD)

B (ICD)

0.3 TMPyP 0.0 200

300

400 500 Wavelength (nm)

600

Protonated TMPyP

700

Figure 8.22. A chiral switch based on a layer-by-layer assembled DNA/poly (allylamine hydrochloride) ﬁlm containing tetrakis (N-methylpyridinium-4-yl) porphine (TMPyP) additives is shown. The interaction between the dye and DNA is characterized by the appearance of an induced CD (ICD) signal corresponding to the main absorption band (Soret band) of the porphine. On the right is schematically presented the initial ﬁlm (As-prepared ﬁlm) that shows an ICD signal; the ﬁlm after exposure to HCl gas (a), which breaks the interaction between the DNA and dye (TMPyP) and consequently cancels the ICD signal; the ﬁlm after exposure to ammonium followed by H2 O (B), which restores the interaction and the ICD signal. On the left are shown the CD (a) and UV–vis (b) spectra of the ﬁlms exposed to the same conditions described in the scheme on the right: signal of the ﬁlm as prepared (dashed line); after exposure to HCl gas (dotted line); after subsequent exposure to NH3 gas (dashed–dotted line); and ﬁnally after exposure to water vapor (solid line). (Reproduced by permission of the American Chemical Society [62].)

311

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

chirality could be switched off by exposure to HCl, which can protonate both the DNA and dye [62]. The loss of the CD signal was attributed to deintercalation of the dye due to possible electrostatic repulsions and changes in the structure of the DNA. Loss of the ICD was accompanied by a change in the color of the ﬁlm from yellow to green. Exposure to ammonium gas restores the yellow color of the ﬁlm; however, the ICD was not recovered until the ﬁlm was subsequently exposed to water. A pH-switchable DNA complex was designed that utilizes an achiral naphthalene derivative (P) bearing a diaminopurine hydrogen bonding unit that can bind to an oligothymine template (Figure 8.23) [64]. Upon mixing a ssDNA template with the diaminopurine, the UV–vis spectrum underwent a blue shift and hypochromic effect. The CD spectrum of the complex showed a positive Cotton effect in the region where naphthalene absorbs, with a zero-point crossing at 338 nm. The authors deduced that the naphthalene guests are arranged in a right-handed helix [65]. When the pH is decreased from 9 to 2, the Cotton effect reverses, suggesting that a left-handed helix is formed. The Cotton effect begins to reverse at pH 5. A pH titration revealed that the pKa of P is 4.8, which is in the range of the value at which the helix inversion occurs, suggesting that protonation of the purine induces a rearrangement in the complex that results in an inversion of the helicity. P

Tn

H H N

O N

H N H

HO

O

O

O

N

OH

O OPO O

N N O

O

O

O

O n-1H

OH

O N N H O O

N O

O N

H

75 pH = 9

CD (mdeg)

50 25 0 –25 pH = 2

–50 –75 200

250

300

350

400

450

500

λ (nm)

Figure 8.23. A naphthalene–diaminopurine derivative, P (top left), forms a dynamic, helical hydrogen bonding assembly with an oligothymine template, Tn (top right), where n is the number of residues. At higher pH values a right-handed helix forms. At lower pH values a left-handed helix forms. CD spectra (bottom) of the complex ([P] = 2[T]T40 = 0.5 mM) at 268 K at various pH values show that the handedness of the complex can be controlled by pH. As the pH is changed from 9 to 2, the sign of the CD bands invert. (Reprinted by permission of John Wiley & Sons, Inc. [64].)

312

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Amphiphilic thin ﬁlms composed of achiral molecules have shown supramolecular chirality, possibly resulting from mechanical or other effects during ﬁlm preparation (Figure 8.24) [66]. Regardless of the mechanism, the induced chirality can be controlled by chemical stimuli. 5-Octadecyloxy-2-(2-thiazolylazo)phenol (TARC18) amphiphile forms a chiral ﬁlm at the air–water interface that gives rise to a CD signal (Figure 8.25). The origin of the chirality is thought to arise from a spontaneous overcrowded packing of the functional groups into a helical sense during compression of the ﬁlm. Two different types of chiral ﬁlms could be formed; however, no control over the selection of the type of chiral ﬁlm could be achieved. FT-IR analysis indicated that the ﬁlms differed in their orientation to the plane of the ﬁlm as well as their trans/gauche conformation. Interestingly, the chirality of the ﬁlm could be erased and restored upon exposure to HCl gas and air. Several cycles could be repeated before the ﬁlm began to peel from the substrate. An explanation for the observed behavior is that exposure to HCl gas protonates the nitrogen of the thiazolyl group, changing the conjugation of the molecule that is accompanied by a change in the packing of the ﬁlm to a state where the chirality is lost. Exposure to air restores the conjugation and allows the ﬁlm to reform the original chiral arrangement. Although the initial direction of the chirality is selected at random, switching only occurs between the initial direction of chirality and an achiral structure. The opposite chirality is not obtained upon exposing the ﬁlm to air after loss of chirality through exposure to HCl. Films composed of tetrakis(4-sulfonatophenyl)-porphine or an amphiphilic benzthiazolyl derivative also showed supramolecular chirality; however, these ﬁlms could not restore the chiral structure after its removal with HCl [67, 68]. Interest in the relationship between macroscopic spinning motion and chirality can be traced back to Louis Pasteur’s attempts at controlling optical activity by performing chemical reactions in a centrifuge or growing plants while rotating in a given direction [2]. Although no induction or inversion of chirality could be detected, more recently the relationship between vortices or macroscopic rotation and the chirality of supramolecular assemblies has shown interesting results that may lead to new insights regarding the origins of homochirality in nature [69–72]. Several reports have dealt with this

Interface

Interface

Compression

M-Chiral

Compression

Achiral

P-Chiral

Figure 8.24. A supramolecular chiroptical switch composed of achiral amphiphiles. Space-ﬁlling structures of achiral amphiphile (TARC18), which forms a Langmuir–Schaefer ﬁlm at the air–water interface, and chiral supramolecular structures formed upon interface compression. (Reprinted by permission of John Wiley & Sons, Inc. [66].) (See insert for color representation of the ﬁgure.)

313

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

0.6 b

OC18H37

12 a

a N

c

N

6 CD (mdeg)

Absorbance

OH

0.4

N S

0.2

c

0 –6

b

–12 0.0

300

400 500 Wavelength (nm)

600

(a)

300

400

500

600

Wavelength (nm) (b)

Figure 8.25. (a) Absorption and (b) CD spectra of 70 layer ﬁlms on quartz or CaF2 substrates. The same spectra are obtained regardless of substrate used; however, the resulting ﬁlms show one of two possible structures, both of which are formed by chance: (a) ﬁlm I and (b) ﬁlm II. The chirality can be removed by exposure to HCl gas: (c) ﬁlm II after exposure to HCl gas. (Reprinted by permission of John Wiley & Sons, Inc. [66].)

challenging topic and include both static and dynamic systems. In the static systems the spinning sense of a solution results in the formation of one enantiomeric form of an assembly. For example, rotoevaporation of a solution of aggregating porphyrins showed a supramolecular chirality that was dependent on the direction of rotation of the spinning ﬂask [73]. Similarly, spin-coated ﬁlms of porphyrin–dendrimer wedges displayed a chirality dependent on the spinning direction of the substrate, while ﬁlms prepared without spinning showed no chiral selection [69]. In dynamic systems the chirality of the system changes upon removal or reversal of the spinning perturbation. Stirring of noncovalent Jaggregates of protonated meso-tetrakis(4-sulfonatophenyl) porphyrin (H2 TPPS4) shows both static and dynamic induced asymmetry (Figure 8.26) [71]. The authors hypothesized that J-aggregates are inherently chiral and exist as racemic mixtures and their distribution can be inﬂuenced by chiral vortices. CD spectra taken of solutions stirred in a clockwise (CW) direction showed that aggregates were preferentially formed. Reversing the direction of spinning to counterclockwise (CCW) showed a CD signal that indicates that the aggregates are transformed to aggregates. When the solutions are stirred for 24 h, a static induced chirality is realized. It was shown that in a mixture of J-aggregate enantiomers, the predominant supramolecular enantiomer deposits on the wall of the cuvette after stirring for 24 h. The monomers and minor enantiomer remain in solution. Applying a macroscopic vortex to a solution of J-aggregates results in the adhesion of one enantiomeric J-aggregate. Spinning in the opposite direction results in the deposition of the opposite enantiomer. The results indicate that stirring affects the distribution of enantiomers present in an initially racemic solution. Clockwise (CW) stirring prefers aggregates, whereas counterclockwise (CCW) stirring prefers aggregates. The effect of stirring in the presence of a chiral stimulus that favors the enantiomer that is opposite to that favored by the vortex was tested by the addition of a chiral ruthenium complex, [Ru(phen)3 ]2+ (vide supra), with a small enantiomeric excess of the -enantiomer. Stirring in a CW direction resulted in the deposition of -aggregates, despite the opposing driving force to form -aggregates. However, upon increasing the concentration of

314

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

1000

600

JΛ

E

Λ

Δ

(A) No Stirring

CD

400 CD (mdeg)

JΔ

CCW CW

200

λ (nm)

λ (nm) JΛ JΔ

0

JΔ

E

–200

CD

800

(B)

E

JΛ (C)

–400 CW

–600

CCW

–800 –1000 460

470

480 490 λ (nm)

500

510

Figure 8.26. The meso-tetrakis(4-sulfonatophenyl)porphyrin (H2 TPPS4) (10 μM) in aqueous solution at pH = 3 and [NaCl] = 0.3 M forms J-aggregates. On the left are shown CD spectra of the J-aggregates recorded at different stirring conditions: without stirring (continuous line), clockwise (CW) stirring (dashed–dotted line), and counterclockwise (CCW) stirring (dotted line). On the right, energy diagrams for the three mentioned conditions are shown: (a) In the absence of stirring, neither enantiomer is favored; (b) CW stirring favors the -enantiomer as shown by the negative CD couplet shown in the small inset; (c) CCW stirring favors the -enantiomer as shown by the positive CD couplet shown in the small inset. (Reprinted by permission of John Wiley & Sons, Inc. [71].)

-[Ru(phen)3 ]2+ , the -aggregate was deposited when CW spinning was performed. The results show that the conformational fate of a supramolecular aggregate is dependent on a competition between macroscopic mechanical forces and molecular level chiral interactions.

REFERENCES 1. B. Alberts, A. Johnson, J. Lewis, M. Raff, K. Roberts, P. Walter, Molecular Biology of the Cell , 4th ed ., Garland Science, New York, 2003. 2. B. L. Feringa, R. A. van Delden, Angew. Chem. Int. Ed . 1999, 38, 3419–3438. 3. W. A. Bonner, Origins of Life and Evolution of Biospheres 1991, 21, 59–111. 4. H. C. Berg, R. A. Anderson, Nature 1973, 245, 380–382. 5. T. Kubori, N. Shimamoto, S. Yamaguchi, K. Namba, S. Aizawa, J. Mol. Biol . 1992, 226, 433–446. 6. B. L. Feringa, J. Org. Chem. 2007, 72, 6635–6652. 7. V. Balzani, A. Credi, F. M. Raymo, J. F. Stoddart, Angew. Chem. Int. Ed . 2000, 39, 3348–3391. 8. T. Verbiest, S. Van Elshocht, M. Kauranen, L. Hellemans, J. Snauwaert, C. Nuckolls, T. J. Katz, A. Persoons, Science 1998, 282, 913–915. 9. J. Elemans, A. E. Rowan, R. J. M. Nolte, J. Mater. Chem. 2003, 13, 2661–2670. 10. A. Ulman, Chem. Rev . 1996, 96, 1533–1554.

SWITCHING MOLECULAR AND SUPRAMOLECULAR CHIRALITY

11. A. Ulman, An Introduction to Ultra-thin Organic Films: From Langmuir–Blodgett to SelfAssembly. Academic Press, Boston, 1991. 12. J. N. Israelachvili, Intermolecular and Surface Forces: With Applications to Colloidal and Biological Systems. Academic Press, New York, 1992. 13. N. J. Turro, Proc. Natl. Acad. Sci. USA 2005, 102, 10766–10770. 14. M. Mammen, S.-K. Chio, G. M. Whitesides, Angew. Chem. Int. Ed . 1998, 37, 2755–2794. 15. M. M. Green, M. P. Reidy, R. J. Johnson, G. Darling, D. J. Oleary, G. Willson, J. Am. Chem. Soc. 1989, 111, 6452–6454. 16. H. Gu, Y. Nakamura, T. Sato, A. Teramoto, M. M. Green, S. K. Jha, C. Andreola, M. P. Reidy, Macromolecules 1998, 31, 6362–6368. 17. M. M. Green, B. A. Garetz, B. Munoz, H. P. Chang, S. Hoke, R. G. Cooks, J. Am. Chem. Soc. 1995, 117, 4181–4182. 18. N. Berova, R. Koji Nakanishi, W. Woody, Circular Dichroism: Principles and Applications. 2nd ed., Wiley-VCH, Weinheim, 2000. 19. W. Saenger, Principles of Nucleic Acid Structure. Springer-Verlag: New York, 1984. 20. A. Mammana, G. Pescitelli, T. Asakawa, S. Jockusch, A.G. Petrovic, R.R Monaco, R. Purrello, N.J. Turro, K. Nakanishi, G.A. Ellestad, M. Balaz, N. Berova, Chem. Eur. J . 2009, 15, 11853–11866. 21. Y. G. Li, T. Y. Wang, M. H. Liu, Soft Matter 2007, 3, 1312–1317. 22. M. M. J. Smulders, I. A. W. Filot, J. M. A. Leenders, P. van der Schoot, A. R. A. Palmans, A. Schenning, E. W. Meijer, J. Am. Chem. Soc. 2010, 132, 611–619. 23. L. Brunsveld, A. Schenning, M. A. C. Broeren, H. M. Janssen, J. Vekemans, E. W. Meijer, Chem. Lett. 2000, 292–293. 24. N. J. Turro, Modern Molecular Photochemistry. University Science Books: Sausalito, CA, 1991. 25. E. J. Park, G. T. Carroll, N. J. Turro, J. T. Koberstein, Soft Matter 2009, 5, 36–50. 26. B. L. Feringa, R. A. van Delden, N. Koumura, E. M. Geertsema, Chem. Rev . 2000, 100, 1789–1816. 27. B. L. Feringa, R. A. van Delden, M. K. J. ter Wiel, Pure Appl. Chem. 2003, 75, 563–575. 28. N. Koumura, R. W. Zijlstra, R. A. van Delden, N. Harada, B. L. Feringa, Nature 1999, 401, 152–155. 29. C. Nuckolls, T. J. Katz, T. Verbiest, S. Van Elshocht, H. G. Kuball, S. Kiesewalter, A. J. Lovinger, A. Persoons, J. Am. Chem. Soc. 1998, 120, 8656–8660. 30. G. T. Carroll, M. M. Pollard, R. v. Delden, B. L. Feringa, Chem. Sci . 2010, 1, 97–101. 31. D. Pijper, B. L. Feringa, Angew. Chem. Int. Ed . 2007, 46, 3693–3696. 32. A. J. Bur, L. J. Fetters, Chem. Rev . 1976, 76, 727–746. 33. M. M. Green, N. C. Peterson, T. Sato, A. Teramoto, R. Cook, S. Lifson, Science 1995, 268, 1860–1866. 34. U. Shmueli, W. Traub, K. Rosenhec, J. Polym. Sci. A 1969, 7, 515–524. 35. M. M. Green, C. Andreola, B. Munoz, M. P. Reidy, K. Zero, J. Am. Chem. Soc. 1988, 110, 4063–4065. 36. M. Goodman, S. C. Chen, Macromolecules 1970, 3, 398–402. 37. M. Goodman, S. C. Chen, Macromolecules 1971, 4, 625–629. 38. M. M. Green, J.-W. Park, T. Sato, A. Teramoto, S. Lifson, R. L. B. Selinger, J. V. Selinger, Angew. Chem., Int. Ed . 1999, 38, 3139–3154. 39. M. M. Green, C. Khatri, N. C. Peterson, J. Am. Chem. Soc. 1993, 115, 4941–4942. 40. D. Pijper, M. G. M. Jongejan, A. Meetsma, B. L. Feringa, J. Am. Chem. Soc. 2008, 130, 4541–4552.

315

316

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65.

66. 67. 68. 69. 70. 71. 72. 73.

G. T. Carroll, M. Jongejan, D. Pijper, B. L. Feringa, Chem. Sci . 2010, 1, 469–472. A. Natansohn, P. Rochon, Chem. Rev . 2002, 102, 4139–4175. G. S. Kumar, D. C. Neckers, Chem. Rev . 1989, 89, 1915–1925. G. Maxein, R. Zentel, Macromolecules 1995, 28, 8438–8440. M. Muller, R. Zentel, Macromolecules 1996, 29, 1609–1617. A. Khan, C. Kaiser, S. Hecht, Angew. Chem. Int. Ed . 2006, 45, 1878–1881. O. Pieroni, A. Fissi, N. Angelini, F. Lenci, Acc. Chem. Res. 2001, 34, 9–17. F. Ciardelli, D. Fabbri, O. Pieroni, A. Fissi, J. Am. Chem. Soc. 1989, 111, 3470–3472. H. Duerr, H. Bouas-Laurent, Photochromism: Molecules and Systems, revised edition. Elsevier, Amsterdam, 2003. M. Irie, Chem. Rev . 2000, 100, 1685–1716. J. J. D. de Jong, L. N. Lucas, R. M. Kellogg, J. H. van Esch, B. L. Feringa, Science 2004, 304, 278–281. J. J. D. de Jong, T. D. Tiemersma-Wegman, J. H. van Esch, B. L. Feringa, J. Am. Chem. Soc. 2005, 127, 13804–13805. N. C. Seeman, Mol. Biotechnol . 2007, 37, 246–257. N. C. Seeman, Biochemistry 2003, 42, 7259–7269. C. J. Murphy, M. R. Arkin, Y. Jenkins, N. D. Ghatlia, S. H. Bossmann, N. J. Turro, J. K. Barton, Science 1993, 262, 1025–1029. N. C. Seeman, Chem. Biol . 2003, 10, 1151–1159. A. D’Urso, A. Mammana, M. Balaz, A. E. Holmes, N. Berova, R. Lauceri, R. Purrello, J. Am. Chem. Soc. 2009, 131, 2046–2047. A. Mammana, G. T. Carroll, J. Areephong, B. L. Feringa, J. Phys. Chem. B , 2011, 115 , 11581–11587. G. S. Manning, Q. Rev. Biophys. 1978, 11, 179–246. A. Mammana, A. D’Urso, R. Lauceri, R. Purrello, J. Am. Chem. Soc. 2007, 129, 8062–8063. R. Randazzo, A. Mammana, A. D’Urso, R. Lauceri, R. Purrello, Angew. Chem. Int. Ed . 2008, 47, 9879–9882. S. Jiang, M. Liu, Chem. Mater. 2004, 16, 3985–3987. J. Lang, M. Liu, J. Phys. Chem. B 1999, 103, 11393–11397. P. G. A. Janssen, A. Ruiz-Carretero, D. Gonzalez-Rodriguez, E. W. Meijer, A. P. H. J. Schenning, Angew. Chem. Int. Ed . 2009, 48, 8103–8106. P. G. A. Janssen, S. Jabbari-Farouji, M. Surin, X. Vila, J. C. Gielen, T. F. A. De Greef, M. R. J. Vos, P. H. H. Bomans, N. A. J. M. Sommerdijk, P. C. M. Christianen, P. Lecl`ere, R. Lazzaroni, P. D. Van Schoot, E. W. Meijer, A. P. H. J. Schenning, J. Am. Chem. Soc. 2009, 131, 1222–1231. P. Guo, L. Zhang, M. Liu, Adv. Mater. 2006, 18, 177–180. L. Zhang, Q. Lu, M. Liu, J. Phys. Chem. B 2003, 107, 2565–2569. P. Guo, M. Liu, Langmuir 2005, 21, 3410–3412. T. Yamaguchi, T. Kimura, H. Matsuda, T. Aida, Angew. Chem. Int. Ed . 2004, 43, 6350–6355. A. Tsuda, M. A. Alam, T. Harada, T. Yamaguchi, N. Ishii, T. Aida, Angew. Chem. Int. Ed . 2007, 46, 8198–8202. A. D’Urso, R. Randazzo, L. Lo Faro, R. Purrello, Angew. Chem. Int. Ed . 2010, 49, 108–112. M. Wolffs, S. J. George, Z. Tomovic, S. C. J. Meskers, A. P. H. J. Schenning, E. W. Meijer, Angew. Chem. Int. Ed . 2007, 46, 8203–8205. J. M. Rib`o, J. Crusats, F. Sagu`es, J. Claret, R. Rubires, Science 2001, 292, 2063–2066.

9 ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS Cheng Yang and Yoshihisa Inoue

9.1. INTRODUCTION The last decade has witnessed rapid progress in supramolecular chirality research [1]. Chiral phenomena in supramolecular systems, such as chirality induction, chiral association, recognition, memory and ampliﬁcation, which primarily originate or develop from molecular chirality, are often much more complicated and therefore more challenging than those at the molecular level. Besides the signiﬁcantly larger number of incorporated building blocks (atoms and molecules), supramolecular chiral structures integrated by noncovalent interactions, such as electrostatic, hydrogen bonding, and van der Waals, π –π stacking interactions, are free from the strict geometrical restrictions of covalent bonds dominating the molecule chirality. Therefore, chiral supramolecular structures are usually more ﬂexible, diverse, and adjustable in geometry, leading to novel chiral phenomena and architectures beyond the limitations of individual molecules. Furthermore, supramolecular chirality is not necessarily based on molecular chirality, but can also be generated upon aggregation of achiral molecules. Elucidating such unconventional higher-order chiral phenomena is crucial for understanding a wide variety of chiral architectures in the nature. Biomolecules by themselves are chiral supramolecular systems, and most of the physical and chemical events occurring in biosystems, including enzyme catalysis, molecular recognition, pharmacological action, and helix formation in DNA and proteins, are intrinsically chiral and supramolecular. Studies on supramolecular chirality provide insights into the biological superstructures and functions, and they may shed light on the origin of biomolecular homochirality. Supramolecular chirality is also important in chemical and materials science and technology, and it ﬁnds practical

Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

317

318

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

applications in asymmetric catalysis, chiral separation and sensing, data storage, optical devices, and liquid crystal displays. Electronic circular dichroism (ECD) is the most powerful and versatile tool among various spectroscopic techniques employed in the study of molecular and supramolecular chiral phenomena. ECD spectroscopy, relying on the differential absorption of rightand left-handed circularly polarized light, is indispensable in particular for the study of enantiomeric supramolecular systems and events that produce signiﬁcant changes in chiroptical properties. Other spectroscopic methods, such as NMR, UV–vis, ﬂuorescence, and IR spectroscopy, are achiral in nature and hence applied to the study of diastereomeric supramolecular systems. X-ray crystallography can directly provide the absolute structures of chiral molecules and molecular assemblies, but is applicable only to crystalline samples and not suitable for the observation of dynamic chiral behavior. In contrast, ECD is more widely applied not only to liquid samples but also to amorphous and crystalline samples and even to gaseous samples, and it can monitor the kinetic and dynamic chiral processes under a variety of conditions. Other chiroptical techniques that use circularly polarized light include optical rotatory dispersion (ORD) [2, 3], vibrational circular dichroism (VCD) [4], ﬂuorescence-detected circular dichroism (FDCD) [5, 6], and circularly polarized luminescence (CPL) [7]. ECD is undoubtedly more widely applicable than these methods: ECD has obvious advantageous over ORD, especially for relatively weak transitions; VCD in the infrared region provides chiral information about the relevant covalent bonds but is less sensitive than ECD; FDCD and CPL are more sensitive than ECD but applicable only to ﬂuorescent species. ECD plays an irreplaceable role in supramolecular chirality research as a highly sensitive tool for determining the chiral sense of the relative orientation of chromophores in a supramolecular system. Chiral chromophores are inherently CD-active and often exhibit signiﬁcant changes in CD intensity upon aggregation or complexation with other molecules, enabling us to quantitatively investigate the kinetics, thermodynamics, and structural changes associated with such processes. Achiral chromophores incorporated in a chiral supramolecular environment may also show appreciable induced CD through extrachromophoric chiral perturbation upon noncovalent interactions with chiral entities in supramolecular system. ECD spectral studies can provide insights into the relative orientation of the relevant chromophore and chiral center and is therefore a highly efﬁcient tool for analyzing complicated chiral supramolecular phenomena. The signiﬁcance of ECD in supramolecular chirality research is underscored by the vast number of publications devoted to the application of ECD to supramolecular systems. In this chapter, we will review the ECD studies on chiral supramolecular systems mainly in solution phase, which are sorted by the chiral event or method in supramolecular system. Several important rules and interesting phenomena relevant to the ECD spectra of chiral supramolecular systems will also be introduced with examples.

9.2. CHIRALITY SENSING WITH ACHIRAL CHROMOPHORE Detecting the chiral sense of asymmetric molecule or molecular assembly with achiral chromophore through noncovalent interactions represents an important application of ECD in supramolecular chemistry [8, 9]. Achiral chromophore is CD-silent by itself but can function as a chirality sensor or reporter when situated in chiral environment. This strategy can be used for the determination of the absolute conﬁguration of a chiral molecule or the enantiomeric excess (ee) of a chiral compound. The CD intensity induced

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

to achiral chromophore is not very strong in general, but can be signiﬁcantly ampliﬁed by introducing two chromophores to the chiral system to cause an exciton coupling interaction. Such ampliﬁcation enables sensing of chiral molecules that are barely CDactive or limited in quantity. Possible overlap with the inherent CD signals at shorter wavelengths may also be avoided by choosing an appropriate chromophore that absorbs at longer wavelengths. An ideal chirality-sensing host should bear a strong binding site(s) highly selective to the target chiral molecule and intensely absorbing chromophore(s) near the binding site(s). Hydrogen-bond and coordination are the most frequently employed interactions in the design of a sensing host, primarily due to their strong binding and directing properties. Other weak interactions, such as π –π , hydrophobic, and electrostatic interactions, may play major or supplementary roles under certain conditions. The exciton chirality method, originally proposed by Harada and Nakanishi, is the most frequently used method for analyzing supramolecular chirality by ECD [10, 11]. Exciton-coupling interaction of two transition moments leads to a splitting of absorption band (Davydov splitting) in UV–vis and ECD spectra. The coupling of two transitions arranged in P (plus, right-handed)-helicity gives a “positive” couplet, displaying a positive Cotton effect peak at a longer wavelength and a negative one at a shorter wavelength, and vice versa for the M -helicity. The couplet amplitude produced, being a function of the distance and angle of the two coupling transitions, is usually very strong and enables chirality sensing of a very small amount of chiral sample.

9.2.1. Chirality Sensing Through Hydrogen-Bonding Interaction Molecular recognition of free-rotating achiral biphenyl-2,2 -diols 1a–c with chiral diamines 2a–c (Figure 9.1) through hydrogen-bonding interaction was investigated by using ECD [12, 13]. Complexation of stereolabile 1 with enantiopure 2 hinders the free rotation about the interaromatic bond in 1 to achieve the point-to-axial chirality transfer from 2 to 1. A negative CD couplet centered at 324 nm was induced to the major transition band of 1a, which is assignable to the 1 La band of phenol units, upon complexation with (1R, 2R)-2a in toluene at a 1a/2a ratio of 2, while antipodal (1S , 2S )-2a gave a mirror-imaged spectrum (Figure 9.2). The CD amplitude was a critical function of the binding strength and the steric interactions between 1 and 2, and therefore it was maximized for the most bulky diamine 2a in toluene. However, the use of polar solvents, such as acetone, acetonitrile, and ethanol, signiﬁcantly reduced the CD signal due to the weakening of the hydrogen bonds. Interestingly, the nature of the interaction between 1 and 2 is altered from hydrogenbonding to electrostatic by lowering the temperature. The proton transfer from 1a to 2a was conﬁrmed in toluene at −80◦ C by the appearance of a phenolate absorption band at

Figure 9.1. Chirality sensing of (1R, 2R)-diaminocyclohexane 2a–c with biphenyl derivatives 1a–c.

319

320

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

1.5

(1S,2S)-2a

Δε (M–1cm–1)

1.0 0.5 0.0 –0.5 –1.0 –1.5

(1R,2R)-2a 280

320

400 360 Wavelength (nm)

440

Figure 9.2. Circular dichroism spectra of the complexes of host 1a with guest (1R,2R)- and (1S,2S)2a in toluene at 25◦ C. [1a] = 6.83 × 10−5 M, [2a] = 1.37 × 10−4 M. (Reprinted with permission from reference 14. Copyright Royal Society of Chemistry.)

412 nm [14]. The phenolate band and the corresponding Cotton effect were observed only upon addition of an excess amount of 2a (1.2 equivalents or more). The stoichiometric study suggested that 1a and 2a form a 1:2 complex at low temperatures. The anisotropy factor for the 1:2 proton-transfer complex is one order of magnitude larger than that for the 1:1 hydrogen bonding complex, suggesting that the axial chirality is more effectively ﬁxed in the 1:2 complex. Although a solution containing 1a and 2a in 1:1 ratio did not show any CD signal, addition of an equimolar amount of achiral diisopropylamine to this solution induced CD signals, intensity of which was half of that for a 1:2 complex of 1a with 2a. 1,8-Naphthyridine derivatives 3a,b (Figure 9.3), in which the naphthyridine moiety acts as a proton acceptor while the pyrrole or indole moiety as a proton donor, were prepared as sensors for detecting monosaccharides [15, 16]. Complexation of 3a,b with 4c–h in dichloromethane led to signiﬁcant changes in UV–vis, CD and ﬂuorescence

OH

HO O

O

HO HO

OC8H17 C8H17O

OH

OH OH

HO

4b

4a N

N 3a

NH

OH HO

OH

HN

O

HO HO

OH

4c

HO HO

O OC8H17

O-C8H17

4d

OH OH

NH

N 3b

HO

Me

O

O

N

OH

OC8H17

C8H17O

OH HO

4e

HN

HO HO

4f O

O OC8H17

OH

4g

OH

CH2OH

HO HO

OCH3 OH

4h

Figure 9.3. Chirality sensing of monosaccharides 4a–h with achiral naphthyridines 3a,b.

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

spectra. The two pyrrole rings in 3 takes a dual inward conformation when bound to chair-form pyranoside. Complexation of 3b with octyl-β-d-glucopyranoside 4a induced a positive CD signal at 442 nm, a strong negative one at 344 nm, and a much weaker one at 317 nm, while antipodal octyl-β-l-glucopyranoside 4b gave the mirror-imaged CD. However, the conjugated chromophore in 3b, involving a naphthyridine and two indoles connected with ethynylene linkers, appears to hinder the unambiguous analysis of the observed CD signals due to the overlapped transitions. Nevertheless, these sensing hosts may be used as a tool for distinguishing the monosaccharide enantiomers.

9.2.2. Chirality Sensing through Coordination A large number of achiral porphyrin-based sensors have been developed, exploiting the strong coordination ability and the large extinction coefﬁcients of the Soret band. Nakanishi and Berova employed achiral bis(zinc porphyrin)s linked with a ﬂexible tether as chirality probes for enantiomeric diamines and aminoalcohols [17–24]. Figure 9.4 shows typical tweezer porphyrin 5, which binds a variety of chiral diamines and aminoalcohols, including 6a–n, to produce bisignate CD signals at the Soret band as a result of the exciton coupling interaction of the twisted porphyrin transitions. The sign of CD couplet is nicely correlated with the absolute conﬁguration of guest 6a–h, except for l-lysinol 6f, which, however, gives a consistent result by acylating the hydroxyl to give 6g [17]. As shown in Figure 9.5, an extremely large amplitude of > 1000 cm−1 M−1 was observed upon complexation of 6l with 5, for which the 1:1 stoichiometry was conﬁrmed by a Job plot [18]. The couplet intensity was related to the difference in size between the smallest and largest substituents at the stereogenic center. Recently, this methodology was theoretically investigated by computer simulation with the potential optimized for liquid simulation [22, 23]. These studies not only provide

Figure 9.4. Chiral diamines and aminoalcohols 6a–n examined by chirality sensor ZnZn-5.

321

322

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

16

435 nm + 578

400 200 Δε

ACD = + 1078

0 –200 –400 423 nm - 500

400

MCH

420 440 Wavelength (nm)

CD amplitude (mdeg)

(a)

(b)

12 8 4 0

0

0.2

0.4

0.6

0.8

1.0

Molar fraction of host 5

Figure 9.5. CD spectrum of tweezer complex of 6l (20 equiv.) with ZnZn-5 (1 μM) in methylcyclohexane; ACD denotes the amplitude of the CD exciton couplet. (b) Job plot of ACD as a function of the molar fraction of host 5; [5] + [6l] = 1 μM. (Reprinted with permission from reference 18. Copyright American Chemical Society.)

an important support for the tweezer approach but also offer crucial insights into the structural factors governing the complexation and the couplet amplitude. The computational method was found useful for predicting the absolute conﬁguration of more complicated chiral compounds with multiple stereogenic centers, such as cis- and trans-3-hydroxyl4-aryl/alkyl-β-lactams 6m and 6n, indicating that the remote stereogenic center has only a secondary effect on the inter-porphyrin twist. Inoue, Borovkov, and co-workers [25–38] proposed the chirality sensing by achiral bis(metalloporphyrin) 7 (Figure 9.6), relying on the exciton chirality method. As illustrated in Figure 9.6, coordination of an enantiopure monodentate ligand (Figure 9.7) leads

Figure 9.6. Chirality sensing mechanism for bis(metalloporphyrin)s. (Reprinted with permission from reference 36. Copyright American Chemical Society.)

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

Figure 9.7. Mono- and diamines 8 and 9 examined with chirality sensor ZnZn-7.

to the conformational switching of bis(zinc octaethylporphyrin) ZnZn-7 from stacked syn to twisted anti [39]. Due to the steric repulsion between the ligand’s substituents and the ethyl substituents on the adjacent porphyrin, the anti -form thus produced is not linear but twisted to afford strong bisignate CD signals (Figure 9.8) at the Soret band [25–27]. The sign of the couplet is uniquely correlated with the absolute conﬁguration of chiral ligand—that is, positive couplets for monodentate (S )-ligands and negative ones for (R)-ligands. Solvent does not signiﬁcantly affect the CD spectra of ZnZn-7 complexes with simple amines 8a–d. However, the CD spectral behavior of the complexes with amino acid esters 8e–g strongly depended on the solvent polarity, affording a negative couplet in nonpolar solvents, in good agreement with the results for 8a–d, but a positive couplet in polar media [34], for which the increased effective size of the ester moiety by solvation would be responsible (Figure 9.9). The ECD spectrum is signiﬁcantly temperature-dependent in this chirality sensing system, showing a gradual increase of the couplet amplitude at lower temperatures, due to the increased afﬁnity of chiral alcohol and amine ligands. Other factors, such as the central metal and the binding stoichiometry, are also crucial in determining the sign and intensity of the induced ECD [35, 40]. MgMg-7, rather than ZnZn-7, exhibits a higher afﬁnity to chiral alcohols to display a strong CD couplet even at room temperature [41]. Bis-porphyrin ZnZn-7 forms a 1:1 tweezer complex with suitable bidentate ligands, such as trans-1,2-(1R, 2R)-diaminocyclohexane 9a, exhibiting a bisignate CD signal. The tweezer complex is transformed to a 1:2 host–ligand complex in the presence of an excess amount of the ligand. The CD amplitude (ε) of the tweezer complex amounts to 500 M−1 cm−1 as a result of the rigid and optimized geometry, while the corresponding

323

324

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

20

0

Δε (cm–1 M–1)

10

ε (105cm–1 M–1)

–10

–20

2

0 300

400

500 Wavelength (nm)

600

Figure 9.8. UV–vis and CD (inset) spectra of 7 in the absence (dotted lines) and presence of 8a (solid lines) and 8e (dashed lines) in CH2 Cl2 (black lines) and in cyclohexane. (Reprinted with permission from reference 34. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.).

1:2 complex still gives a huge, but signiﬁcantly smaller, ε of 200 M−1 cm−1 , due to the more ﬂexible and dynamic structure. Similar stoichiometry-dependent chiroptical behavior is seen upon chirality sensing by tweezer bis(Zn porphyrin)s 10a–c with longer linkers (Figure 9.10) [42, 43]. All of these tweezers show high afﬁnities (>105 M−1 ) to trans-1,2-diaminocyclohexane (9a) upon 1:1 complexation to give strong bisignate CD signals. However, further addition of 9a led to a red shift of the Soret band and a dramatic decrease of CD intensity with sign inversion, indicating switching of the bis-porphyrin conformation from tweezer to open form upon 1:2 host–ligand complexation (Figure 9.11).

9.2.3. Chirality Sensing through Other Noncovalent and Covalent Interactions Supramolecular complexation through electrostatic, hydrophobic, and van der Waals interactions also induces ECD signals to achiral chromophores. Water-soluble achiral calix[n]arenes 11n (Figure 9.12) form host–guest complexes with chiral ammonium ions through electrostatic and cation–π interactions. Addition of (R)-12 to 114 produced a strong negative CD couplet (Figure 9.13) [44], which was interpreted by a twisted array of the benzene rings in 114 induced primarily by the largest naphthylethyl substituent in the chiral guest.

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

Figure 9.9. Effects of strongly and weakly interacting solvents on the mechanism of supramolecular chirality induction in 7. (Reprinted with permission from reference 34. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.).

Figure 9.10. Chirality sensing with tweezer bis(Zn porphyrin)s 10a–c.

Boronic acids bind to organic 1,2- and 1,3-diols with high afﬁnities through the reversible boronate formation. As shown in Figure 9.14, the diboronic acid-bearing 13 gives 1:1 adducts with several monosaccharides, including glucose, mannose, galactose, and talose, through the synergic formation of two boronate esters [45, 46]. Most of the

325

326

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a)

(b)

Figure 9.11. UV–vis (top) and CD (bottom) spectral changes of (a) 10a and (b) 10b in the Soret region upon titration with (1R, 2R)-9a (0, 1, 2, 10, 50, 100 equivalent). (Reprinted with permission from reference 42. Copyright Royal Society of Chemistry.).

Figure 9.12. Achiral calixarene 11 for chirality sensing of 12.

d-saccharides, except for d-galactose, give a negative CD couplet upon esteriﬁcation with 13, while the l-form afforded a positive couplet. Rosini and co-workers [47] used 4-biphenylboronic acid 14 for sensing chiral 1-arylethane-1,2-diols via boronate ester formation. The conformationally rigid ﬁvemembered boronate ring can ﬁx the two aromatic groups in a well-deﬁned orientation to produce a strong CD couplet. Upon esteriﬁcation with 13, all of the examined (R)and (S )-1-arylethane-1,2-diols consistently exhibited negative and positive couplet, respectively.

9.3. CHIRAL CONFORMATION OF MACROMOLECULE Macromolecules are usually conformationally diverse in solution, and ECD is a crucial tool for elucidating the static properties and dynamic processes of macromolecules, in particular those with helical conformation. Intense ECD signals are often observed for macromolecules composed of chiral subunits. Macromolecules made from achiral monomers may also take chiral conformations upon interaction with external chiral effectors. Even if the target macromolecule has no UV–vis absorption, CD spectral techniques may be applied by covalently or noncovalently attaching a chromophore. ECD spectroscopy is an important tool for analyzing the structures of biopolymers, such as DNA and proteins. Readers should consult (a) the sections in Chapter 4 of the present volume

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

327

10000

[θ] (deg cm2dmol–1)

5000 114 – (S) –12 0

–5000 114 – (R) –12

Figure 9.13. Induced CD spectra of 114 in the presence of (R)- and (S)-12 in pH 7 phosphate

–10000

200

300 λ (nm)

400

buffer at 25◦ C. (Reprinted with permission from reference 44. Copyright Royal Society of Chemistry.)

Figure 9.14. Chirality sensing through boronate ester formation with chiral 1,2- and 1,3-diols.

2 and (b) relevant literatures for comprehensive information regarding the application of ECD to various biomolecules [48]. In this section, we will focus on the application of ECD to synthetic macromolecules, such as polymers and dendrimers.

9.3.1. Inherently Chiral Macromolecule Chiral polymer is obtained by polymerizing monomers with chiral substituent(s) or by inducing chirality in the main chain upon polymerization of achiral monomers under the inﬂuence of chiral initiator, additive, and so on. Copolymers of chiral and achiral monomers also show optical activities; and two intriguing phenomena, leading to the

328

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

15

[Θ] x 10–4 deg cm2 dmole–1

10

5

0

–5

–10

–15 180

200

220

240

260

280

300

λ (nm)

O

O N Δ

H H 3C

N

C H 3C H

(R)

(S)

O

O C

N

N

C Y

X H H3C

C

H 3C (R)

H

(S)

15 X

Y

O

49

51

×

56

44

Figure 9.15. CD spectra of polyisocyanate with chiral side chain of different enantiomeric ratio: R:S = 100 : 0 (triangle), 0:100 (square), 49:51 (circle), and 56:44 (cross). (Reprinted with permission from reference 49. Copyright American Chemical Society.)

“majority rule” and the “soldiers and sergeants principle,” have been discovered upon ECD measurement of mixed chiral or chiral–achiral copolymers. The majority rule was ﬁrst experimentally demonstrated by Green et al. [49] in their study on the polymerization of isocyanates with chiral substituent. The polyisocyanates thus obtained formed helices in solution, handedness of which was controlled by the ee of the chiral side chain. They found that the CD intensity of chiral copolymer 15 (Figure 9.15) obtained from a mixture of (R)- and (S )-monomers in varying ratios is not proportional to the ee of the used monomer, but behaves nonlinearly against the ee.

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

Figure 9.16. Chiral copolymers 16 that obey the 16

majority rule.

Thus, the CD spectrum of 2% (S )-rich copolymer (i.e., R/S ratio = 49 : 51) was identical in sign and shape to that of homopolymer of (S )-monomer, but the CD intensity was unexpectedly strong, amounting to one-third of the homopolymer (Figure 9.15). It was interpreted that the energy barrier for helix reversal is much higher than the chiral bias caused by the pendant group [50], and therefore the helical sense will be controlled by the conﬁguration of the majority pendant conﬁguration. Thus, the CD spectrum of 12% (R)-rich copolymer (i.e., R/S ratio = 56 : 44) became exactly the same as that of (R)-homopolymer, indicating that the helix reversal was completely suppressed at that ee. The “soldiers and sergeants” principle is another interesting phenomenon related to the chirality induction in copolymer of chiral and achiral monomers. According to the principle, a minute amount of chiral monomer incorporated in copolymer with achiral monomer can dominate the overall helical sense of the copolymer. This phenomenon was ﬁrst observed by Green and Reidy in their study of the speciﬁc rotation and CD intensity of copolymer 16 (Figure 9.16) obtained from the copolymerization of chiral and achiral monomers in various ratios [51, 52] Copolymer 16 containing 4% chiral component (x = 0, y = 4, and z = 96) showed a speciﬁc rotation comparable to that of 100% chiral homopolymer (x = 0, y = 100, z = 0), while by incorporating 1% chiral unit (x = 1, y = 0, z = 99) 16 achieved half the speciﬁc rotation of homopolymer (x = 100, y = 0, z = 0). Strong CD signal was observed even at a chiral unit content as low as 0.12% (x = 0.12, y = 0, z = 99.88), demonstrating that a minute amount of the chiral unit embedded in a long sequence of achiral N -hexylisocyanate units can appreciably bias the equilibrium between the right- and left-handed helices. The chiroptical properties of the copolymer did not change even if the chiral unit content was reduced to 15%. These observations indicate that a very small number of chiral inductor units, acting as the sergeants, can command a much large number of achiral soldier units in the copolymer to align in one direction. Dendrimers are repeatedly branched, monodisperse, and usually highly symmetric spherical macromolecules. Its highly branched 3D structure features high surface functionality and versatility. Dendrimers often show good conformational cooperativity and can convey local structural or chiral information to the next hierarchical level of structural organization. Dendrons 17–19 (Figure 9.17) are moderately soluble in water and highly soluble in organic solvents [53]. An intense negative couplet was observed for dendron 17 in both THF and water (Figure 9.17), indicating an M -helical conformation of anthranilate components in the dendron. The second-generation dendron 18 showed a negative couplet in THF but a positive one in water, indicating an M -to-P helix transition

329

330

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Me O O

O

O NH CO2R

O

OMe CI

O

(a)

N

O O

NH O

OR

Me

CI

O

OR

17 O

P

N HN

O O

O

HN

NH

O

O

RO

O

Me O

O

OR

Me

Me 18

OR

RO RO

OR

Me

O

OR

O HN O

O

O

1 0 –1 –2

240 260 280 300 320 340 360 380 400 λ (nm)

O

O O

O

2.4

RO

O

O

N H

H N

O

O NH N

O HN O RO

N

N HN

O

O O

Me

NH O Me RO

O

O

N HN

CI

θ/(105deg cm2 mol–1)

HN

O

17 THF 17 H2O 18 THF 18 H2O

O

NH N

19

18 UV (THF) π π*

2

HN

N

H N

HN N

O

Me

O

NH

O

Me

O O

Me

Me

O

M

(b)

O

O

RO

N

RO

OR O

NH N Me

H2O

O O O HN N O OH

HN

NH

O

O

O

N

RO THF

N O O N H HO O N

θ/(105deg cm2 mol–1)

R:/

OR

O

NH

UV (THF) 1.6

π

π*

19 THF 19H2O

0.8 0 –0.8 –1.6

O O

Me

230 250 270 290 310 330 350 370 390

OR

λ (nm)

Figure 9.17. (a) Direction of the electronic transition moment of the π –π ∗ transition at 316 nm and the sign of the corresponding CD couplet. (b) CD spectra of dendrons 17-19 in H2 O and THF, normalized for the concentration and the number of chiral terminal groups. (Reprinted with permission from reference 53. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.).

in water. Similar M -to-P transition was observed with the third-generation dendron 19 upon going from THF to water. In good agreement, the experimental IR spectral studies and the computational simulations indicate an increase of the gauche C–C and anti C–O bonds of the poly(oxyethylene) chains on going from organic to aqueous solution. The solvent-induced ECD change reﬂects the conformational ﬂuctuations of the terminal chains that are coupled with the dendron’s helical secondary structure through correlated dendron-chain motions.

9.3.2. Chirality Induction in Achiral Polymer Polymerization of achiral monomer may give polymers with helical segments, handedness of which is, however, randomly populated in the polymer chain to give no ECD signal on

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

Figure 9.18. Condensation polymerization of 21 with 22 in chiral liquid crystal 20.

Figure 9.19. Chirality induction in achiral polymer 24 by chiral alcohols 25 and 26.

the whole. Polymerization with chiral catalysts or in chiral environment can give chiral polymer efﬁciently [54–56]. Akagi and co-workers [57] reported copolymerization in chiral liquid crystal 20 of achiral monomers 21 and 22 (Figure 9.18). The nematic phase of 20 is stable over a wide temperature range and remains stable upon addition of 21 and 22. Copolymers 23 obtained by condensation in (+)-20 and (−)-20 showed CD spectra with positive and negative exciton couplet, respectively. However, the couplet of copolymer 23 seems to originate from the chiral aggregation rather than the helical structure of the main chain (axial chirality) [57]. Achiral polymer can adopt a chiral conformation upon noncovalent interaction with chiral molecules. Fujiki and co-workers [58] reported the induced CD for achiral polysilylene 24 upon complexation with chiral alcohols 25 and 26 (Figure 9.19). Notable negative or positive CD couplet was induced when 24 was mixed with (S )- or (R)-25, for which the hydrogen-bonding interaction between the hydroxyl group in 25 or 26 and the ether oxygen in 24 is responsible. Interestingly, the majority rule does not work in this system, because a good linear correlation was observed between the ee of 24 and the induced CD intensity, which, however, offers a possibility to quantitatively determine the optical purity of chiral alcohols. Sada and co-workers [59] investigated the aggregation of chiral bis(dioxazolylpyridinyl)porphyrin 27 with achiral poly(trimethylene iminium) 28 (Figure 9.20). The formation of double-stranded structure, in which bidentate ligand 27 bridges two polymer chains of 28, was revealed by AFM and UV titration, while the ECD spectra provided the information about the 3D structure and the handedness of the supramolecular chirality. In the absence of polymer 28, 27 exhibited an apparent positive CD couplet in the Soret region, which was attributed to the dipole coupling between the porphyrin and terminal dioxazolylpyridine transitions. Upon addition of 28 to 27, the positive couplet was inverted in sign and decreased in intensity. Such an inversion was not caused by adding monomeric 29. This conﬁrms the formation of helical double-stranded structure, in which the porphyrin rings are twisted in order to avoid the steric hindrance between the isopropyl groups in 27.

331

332

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Helical aggregate of porphyrins

O N N

Hydrogen-bonding

H N NH

O

Achiral synthetic polymer

Hydrogen-bond pair

Double-stranded helix

Figure 9.20. Double-helical chiral aggregation of achiral polyamine 28 with bidentate chiral inductor 27. (Reprinted with permission from reference 59. Copyright 1998 American Chemical Society.)

30

31

Figure 9.21. Helical tubular structure formation of polythiophene 31 with schizophyllan 30.

ECD was used for detecting the assembling of water-soluble achiral polythiophene 31 with chiral polysaccharide, schizophyllan 30 (Figure 9.21) [60]. Upon addition of 30 to an aqueous solution of 31, both UV–vis and ﬂuorescence peaks of polythiophene showed signiﬁcant bathochromic shifts, reﬂecting the increased effective conjugation length of the polythiophene backbone. Formation of a tubular structure, in which polythiophene is included in the helical schizophyllan tube, was proposed based on the UV–vis titration and AFM measurement. An intense positive CD couplet observed in the polythiothene’s π –π * transition region indicates a right-handed twist of the backbone of 31 in the tubular structure.

333

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

9.3.3. Chiral Memory in Polymer Supramolecular chirality of conjugated polymers often exhibits unique dynamic features. Yashima and co-workers [61–78] investigated the chiral properties of a series of conjugated polymers (Figure 9.22). Achiral conjugated polymer 32, originally a mixture of at least four conformers of cis-transoid, cis-cisoid, trans-transoid , and trans-cisoid , forms a single dynamic helical structure upon complexation with chiral amines, showing a characteristic CD couplet (Figure 9.23) [62]. The amplitude of the couplet is augmented with increasing bulkiness of the chiral amine and also with decreasing distance between the chiral center and the amino group, indicating the importance of steric effect in the chirality transfer. One of the most intriguing phenomena associated with this conjugated polymer–chiral amine system is the memory of macromolecular chirality, which is preserved even after replacing the chiral amine attached to 32 with an achiral amine. As shown in Figure 9.23, addition of (R)-naphthaleneethylamine 35 to a solution of polymer 32 induces a negative CD couplet [64, 70]. Further addition of chiral amino alcohol (S )-36 to this solution leads to an inversion of the couplet sign from negative to positive. Surprisingly, the original negative couplet of [32·(R)-35] complex is not affected by the addition of achiral amine 37 and the subsequent removal of (R)-35 from the solution by gel permeation chromatography, and even upon further addition of (S )-36 [64]. The CD

Figure 9.22. Achiral conjugated polymers 32–34, which give helical conformation upon complexation with chiral amines.

334

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(b) (32-(R)-35 complex) + (S)-36

3

[θ] (104 degree cm2 dmol–1)

2 1 0 –1

Figure 9.23. CD spectra in DMSO of (a) [32·(R)-35] complex, (b) a mixture of [32·(R)-35] complex with (S)-36, (c) a mixture of [32·(R)-35] complex with 37,

(d) Fractionated 32

–2

and (d) 32 isolated by GPC with a DMSO eluent containing 37 (0.8 M); [32] = 3.0 mg/mL (20.4 mmol in monomer unit/mL) for traces a–c and 0.13 mg/mL

(c) (32-(R)-35 complex) + 37 –3

(a) (32-(R)-35 complex) 310

350

400

450

Wavelength (nm)

500

550

for trace d. (Reprinted with permission from Macmillan Publishers Ltd. [64], copyright 1999.)

signal lasts for a long period of time to show only a 5% decrease in intensity after 3 months. The memory efﬁciency of the macromolecular helicity of 32 preserved by amino alcohols is a critical function of the number of methylene groups in amino alcohol but almost independent of its afﬁnity to the carboxyl group. Similarly, intense CD was induced to achiral polymer 33 by keeping its solution containing chiral aminoalcohol at 50◦ C for 29 days. Once the helical polymer structure was induced, the helicity was memorized even after removal of the chiral amine and only about 10% decrease in CD intensity was observed after 29 days without any assistance of achiral amine [70, 78, 79]. Polymer 34 with bulky aza-18-crown-6 pendants forms a single helical structure upon complexation with chiral amines 39–41 (Figure 9.22), displaying a characteristic split-type CD at the absorption band of the backbone of 34.[76] Achiral cyanine dye 38 is entrapped in polymer 34 to form J aggregates, which afford CD signals when L-39 is coincluded. Interestingly, the supramolecular chirality of the J-aggregates was maintained even after the chirality of the polymer backbone was inverted by adding D-39.

9.3.4. Chiral Molecular Recognition with Polymer Inouye and co-workers [80] studied the CD spectral detection of saccharides with achiral water-soluble polymer 42 (Figure 9.24) in aqueous protic media. Meta-ethynylpyridine polymers 42 form helical structures in protic media through intramolecular solvophobic interactions. Hydroxyl groups of saccharide form a hydrogen bond to the pyridine nitrogens of 42 even in MeOH–H2 O mixture. In contrast to the weak CD signal induced upon complexation with d-glucose, 42 exhibits much stronger induced CD for octylβ-d-glucopyranoside. As shown in Figure 9.25a, the CD signal is inverted in sign by changing the solvent composition from MeOH/H2 O = 5 : 1 to 10:1, due to the varied ratio of d- and l-glucose in solutions of different composition. Furthermore, the CD spectra of 42 obtained immediately after the addition of d- and l-glucose are opposite in sign but gradually reduced in intensity to eventually converge to a common spectrum

335

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

OR

complexation N

I

N

N

N

N

N

N

N

NOH

N OH

N mutarotation of glucose

N

N

42a: R = (C2H4O)8CH3 42b: R = n-C4H9

complexation

N

N

n

42

N

N

N

N

N OH

N N N

OH

left-handed helical complex

right-handed helical complex

Figure 9.24. Chiral self-aggregation of achiral polymer induced by a saccharide. (Reprinted with permission from reference 80. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.) (See insert for color representation of the ﬁgure.)

3 MeOH / water

CD/(mdeg)

CD/(mdeg)

2

1

0

–1 300

8

5:1 6:1 7:1

8:1 9:1 10:1 310

320

330

340

350

360

6 β-glucose +7 +2.4 mdeg 4 (337 nm) 2

time 0h 1h 3h 8h 15 h 24, 48 h 15, 24, 48 h 8h

0

5h 3h α-glucose –2 –3 1h +2.4 mdeg 0h (337 nm) –4 300 310 320 330 340 350 360

λ (nm)

λ (nm)

(a)

(b)

Figure 9.25. (a) Induced CD spectra of a mixture of 42a (1 mM in monomer unit) and D-glucose (0.3 M) in 5:1–10:1 MeOH/H2 O at 25◦ C. (b) Time-dependent CD spectra of a mixture of 42a (1 mM in monomer unit) and α- or β-D-glucose (0.3 M) in 5:1 MeOH/H2 O at 25◦ C. (Reprinted with permission from reference 80. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.) (See insert for color representation of the ﬁgure.)

after standing the solution at 25◦ C for 24 h (Figure 9.25b). These phenomena were rationalized in terms of the anomerization of d-glucose between the α- and β-forms in the protic solvent.

9.4. SUPRAMOLECULAR COMPLEXATION WITH CHIRAL MOLECULAR HOST ECD is a powerful tool for studying the supramolecular complexation behavior of inherently chiral host molecules, because achiral chromophoric guests often become CD-active when bound to a chiral host. The CD signals thus induced provide crucial information about the spatial arrangement of the chromophoric guest(s) included. On the other hand, the guest inclusion may also cause a change in host conformation, which can be reﬂected in the CD spectrum. Hence, ECD spectral study enables us not only to qualitatively detect the supramolecular complexation and guest orientation but also to quantitatively determine the binding stoichiometry and afﬁnity.

336

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

9.4.1. Binding with Cyclodextrin Cyclodextrins, a family of cyclic oligosaccharides typically composed of 6–8 glucose units linked by α-(1 → 4) glycosidic bonds, are water-soluble truncated cone-shaped macrocyclic hosts with a hydrophobic cavity that can include a wide range of organic guests through hydrophobic interactions. Cyclodextrin is inherently chiral, and the complexation of chromophoric guest in its cavity induces appreciable CD signals in the guest’s absorption region. By using the ECD response as a tool for detecting supramolecular interaction, Tokura and co-workers [81] investigated the complexation behavior of benzoylbenzoates with cyclodextrins in mid-1970. Mason [82] predicted that the anisotropy (g) factor induced to an achiral chromophore that is oriented randomly to a chiral molecule will be much smaller (10−5 –10−6 ) than the one in ﬁxed orientation (10−2 –10−3 ). Harata and Uedaira [83, 84] attempted to calculate the sign of ECD induced to a chromophoric guest complexed with cyclodextrin by using Kirkwood’s [85] oscillator theory. Later, Kajtar et al. [86] proposed empirical “sector rule” to correlate the sign of induced CD signal with the orientation of transition moment of a chromophoric guest complexed with cyclodextrin. According to the sector rule (Figure 9.26), a positive CD signal is induced when the transition moment of a chromophore is located in the conical sector along the cavity axis, while a negative one to a more slanted transition located outside the cone. Kodaka and others further discussed the correlation of induced CD with the location and orientation of a chromophore complexed by various cyclodextrins [87–91]. Thus, the “Kodaka rule” says that when a chromophoric guest is accommodated inside the cavity, a transition parallel to the cavity axis induces positive CD and a perpendicular transition causes negative CD, but exactly the opposite is true for a chromophore located outside the cavity. These rules have been well-conﬁrmed by the experimental results and therefore used as a standard tool for analyzing or interpreting the orientation of chromophoric guest in and around the cyclodextrin cavity [92–98]. For instance, methyl orange 43 with an azobenzene chromophore (Figure 9.27) exhibits a positive CD upon inclusion by β-cyclodextrin with its π –π * transition being aligned along the cavity axis, whereas the azobenzene moiety in compound 44, being laterally positioned above the cyclodextrin portal, affords a positive CD for the π –π * transition [99]. Brinker and co-worker [94]

Figure 9.26. Sector rule for predicting the sign of induced CD upon complexation in cyclodextrin cavity.

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

Figure 9.27. Methyl orange 43 azobenzene-bridged 43

cyclodextrin 44.

Δε (M–1cm–1)

44

λ (nm)

Figure 9.28. Induced CD and orientation of azi-adamantane 45 upon complexation with β- (top) and γ -cyclodextrin (bottom). (Reprinted with permission from reference 94. Copyright 1998 American Chemical Society.)

investigated the orientation of azi-adamantane 45 (Figure 9.28) included in the cavity of α-, β-, and γ -cyclodextrin by ECD. Complexation of 45 with β-cyclodextrin induced positive CD, while γ -cyclodextrin complex only afforded a much weaker negative CD (Figure 9.28). By taking into account the size and shape of 45, the azo chromophore is deduced to be located near the portal of β-cyclodextrin or inside the γ -cyclodextrin cavity, as illustrated in Figure 9.28, which induces the positive or negative CD, respectively, according to the sector rule applied to the π –π * transition which is perpendicular to the N=N bond. The weaker CD intensity observed with γ -cyclodextrin is ascribed to the higher guest mobility in the γ -cyclodextrin cavity. CD spectral titrations with 45 gave the association constants of 6150 and 2740 M−1 for β- and γ -cyclodextrin, respectively.

337

338

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

4

HO

θ/(mdeg)

2

O

OH O OH HO

O HO O OH

0

O OH O HO

HO

–4

O

–6

HO

O

46

OH O OH

–2

OH OOH

O O

OH O OH MeO OH OH O O

OH

O

200

220 240 260 280 Wavelength (nm)

HO

Figure 9.29. CD spectral changes of m-methoxybenzyl-β-cyclodextrin (0.1 mM) upon addition of cyclooctene.

θ/(mdeg)

By appending a chromophore, cyclodextrin becomes CD-active. In aqueous solution, the chromophore appended to cyclodextrin is often included in its own cavity, but can be driven out of the cavity by adding an appropriate guest as a competitor. The CD spectral changes thus induced can be used as a measure of guest inclusion. As illustrated in Figure 9.29, 6-O-m-methylbenzoyl-β-cyclodextrin shows a negative CD at the 1 Lb band and a weak positive CD at the 1 La band, indicating shallow penetration of the benzoate moiety into the cavity. Addition of cyclooctene leads to a gradual increase of CD intensity at both the 1 Lb and 1 La bands, suggesting the change of orientation of the methylbenzoyl moiety caused by the guest inclusion. A quantitative CD spectral titration gives the association constant of 11,120 M−1 [100, 101]. γ -Cyclodextrin, possessing a cavity larger than its lower homologues, can accommodate two planar aromatic guests, which are usually stacked in a chiral fashion to produce a split ECD. Inoue and co-workers [97, 102–110] investigated the 1:2 host–guest complexation of anthracenecarboxylic acid (AC) with γ -cyclodextrin derivatives, before examining the asymmetric photocyclodimerization of AC. The intimate stacking of ACs in the cavity caused signiﬁcant changes in UV–vis, NMR, ﬂuorescence, and CD spectra. As shown in Figure 9.30, the complexation of AC with monoaltro-γ -cyclodextrin 47 led

Wavelength (nm)

Figure 9.30. CD spectral change upon addition of AC (0–0.3 mM) to the solution of monoaltroγ -cyclodextrin 47 (2 mM) in pH 9.0 aqueous buffer.

339

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

to a strong positive couplet at the 1 Bb band, indicating the P -helical arrangement of two ACs in the cavity.

9.4.2. Binding with Biomolecule DNA forms three types of duplexes with different helical structures (i.e., A-, B-, and Z-DNA) through multiple hydrogen-bonding and stacking interactions of A–T and G–C ˚ width × 8-A ˚ depth nucleobase pairs. B-DNA has the major and minor grooves of ∼13-A ˚ ˚ and 4.5-A width × 6-A depth, respectively. The major groove comprises more nucleobase substituents and phosphodiesters, while the minor groove is walled by the hydrophobic part of sugar. A guest molecule may be bound to the major or minor groove or intercalate in between two base pairs. The CD intensity induced to a groove-bound chromophore is thought to be one or two orders of magnitude greater than that of an intercalated one [111]. Because of the existence of multiple binding sites, the complexation behavior with DNA is sometimes complicated and a critical function of the guest concentration and the ionic strength [112, 113]. For example, acridine orange intercalates to DNA to give negative CD at low concentrations, but much stronger positive CD at higher concentrations due to groove binding. Furthermore, the intercalating and groove-binding ligands couple to each other to give a split CD. The CD signal induced to a chromophore upon interaction with DNA crucially depends on the DNA structure. As shown in Figure 9.31, addition of sulfonated Ni porphyrin 48 to right-handed B-DNA of poly(dG-dC)2 induces no appreciable CD signal at the Soret band (∼400 nm) of 48 [114]. However, once the B-DNA is converted to left-handed Z-form by adding spermine, an intense negative couplet emerges at the Soret band, enabling selective sensing of Z-DNA. The electrostatic interaction of negatively charged 48 with the protonated spermine bound to Z-DNA is the main driving force for the ternary complexation of 48 with a Z-DNA–spermine complex, in which the exciton

20 15

– O3S

48 + Z-DNA 48 + B-DNA

SO3

–

N N Ni

10

N

CD/(mdeg)

N

5

– O3S

– SO3

48

0 –5 –10 –15

250

300

350

400

450

500

550

Wavelength (nm)

Figure 9.31. CD spectrum of NiTPPS 48 (4 μM) in the presence of poly(dGdC)2 (50 μM) in righthanded B-form (gray) and in left-handed Z-form induced by adding spermine (black). (Reprinted with permission from reference 114. Copyright 2009 American Chemical Society.)

340

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(M)-49

(P)-49

Figure 9.32. (M)- and (P)-helicene 49.

coupling arises from the through-space interaction of bound porphyrins. The induced CD of a ternary DNA–spermine-48 complex disappears by raising the pH to 8.2, but is recovered by lowering the pH back to 6.9, suggesting that the induced CD can be modulated reversibly by pH. Enantioselective complexation with DNA is expected to occur for chiral species, and ECD is a powerful probe for investigating the chiral recognition upon complexation with DNA [115–117]. Sugiyama and co-workers studied the complexation behavior of (P )and (M )-helicene 49 with the B- and Z-form DNA of d(CGCm8 GCG)2 (Figure 9.32). The CD intensity of (P )-49 was reduced by 70% upon complexation with Z-DNA, but no appreciable change was induced by B-DNA. In contrast, antipodal (M )-49 did not show any chiral discrimination upon interaction with Z-DNA, for which the ﬁve-fold smaller afﬁnity for (M )-49 would be responsible. Complexation with protein is a crucial issue in various biological phenomena, such as enzyme catalysis, antibody–antigen interaction, and drug delivery. Folding of polypeptide chain often produce crevices or cavities that function as binding sites for organic guest molecules. These binding sites are usually hydrophobic in nature and surrounded by a set of amino acid side chains that are arranged to optimize the noncovalent interactions with speciﬁc ligands. Commonly, more than one binding site will be created near the surface of a protein, and therefore the complexation of a guest with protein is elaborate in general. The binding afﬁnity and stoichiometry of a guest in different binding sites rely on the size, shape, and functional group of the guest and the noncovalent interaction operating. In view of the intrinsically chiral nature of protein, ECD is one of the most crucial and widely employed tools for studying the interaction of organic guests with proteins. Inoue and co-workers [118–120] have studied the binding of AC to bovine (BSA) and human serum albumin (HSA) by means of ECD. As illustrated in Figure 9.33, complexation of AC with BSA induced well-structured CD at 330–400 nm. The intensity of positive CD induced was almost proportional to the AC concentration to reach a maximum at AC/BSA = 1 and then decreased gradually to eventually give negative CD upon further addition of AC. Detailed Job plot and titration experiments using CD, UV–vis, and ﬂuorescence spectroscopy revealed the presence of four independent binding sites for AC in BSA, which respectively accommodate 1, 3, 2, and 3 AC molecules in the following order of afﬁnity: K = 5.3 × 107 , 1.3 × 105 , 1.4 × 104 , and 3.0 × 103 M−1 . Similarly, a CD spectral titration with HSA showed four inﬂection points at AC/HSA = 1, 2, 5 and 10, indicating the presence of ﬁve binding sites that accommodate 1, 1, 3, 5, and >10 AC molecules in the order of decreasing afﬁnity [120].

9.4.3. Binding with Synthetic Chiral Host Shinkai and co-workers [121–123] investigated the binding of achiral guests 50 and 51 by per-(S )-2-methylbutylated calixarenes of different ring sizes, 52[n] (Figure 9.34). In

341

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

AC/BSA 1

(a)

10

5

q (mdeg)

10

(c)

0

0

5 (b) 1

5

q391 (mdeg)

10

0

0 –5 –5 10 –10

–10 300

350

400

0

2

Wavelength (nm)

4

6

8

10

AC/BSA

Figure 9.33. CD spectral change upon addition of AC to a phosphate buffer solution (pH 7) of BSA (0.08 mM) at 25◦ C; (a) [AC] = 0–0.08 mM (from bottom to top); (b) [AC] = 0.08–0.8 mM (from top to bottom); (c) CD intensity at 391 nm as a function of AC/BSA ratio. (Reprinted with permission from [118]. Copyright 2003 American Chemical Society.)

50

51

52[n]

Figure 9.34. Chiral calixarenes 52[n] and achiral azobenzene guests 50 and 51.

the absence of guest, positive CD is observed for 52[4], but split CD for 52[6] and 52[8]. Addition of aliphatic alcohols did not appreciably change the CD spectrum of 52[4], but considerably reduced the CD intensities of 52[6] and 52[8]. 4-Cyano-4 -(diethylamino) azobenzene 51 showed a negative CD couplet upon complexation with 52[6]. However, a positive couplet was observed with 52[8], suggesting that the CD-active species are not monomeric but are instead aggregates of 51, which are arranged counterclockwise with 52[6] and clockwise with 52[8]. Synthetic chiral host 53 shows a dramatic CD change upon complexation with sulfate anion [124]. In the absence of sulfate, chiral guanidium host 53 exhibits a simple positive Cotton effect. Addition of sulfate leads to the formation of 2:1 complex (Figure 9.35),

342

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

N N

N O N H O

N O

N H

OTBDPS

SO42– O

OTBDPS

N H O

N H O S

53 TBDPSO

O H N

O H N

O O

N

N

Figure 9.35. 1:2 Complexation of sulfonate anion with chiral guanidine host 53.

inducing a bisignate Cotton effect. Further addition of sulfate switches the complex stoichiometry from 2:1 to 1:1 with accompanying CD spectral change from bisignate to simple negative one.

9.5. CHIRAL MOLECULAR ASSEMBLY Chiral chromophoric compounds often form aggregates with accompanying CD spectral changes. Thus, CD spectral study provides not only the evidence for aggregation but also the structural information of the supramolecular assembly [125]. Achiral chromophore included in well-deﬁned chiral aggregates, such as liquid crystal, may also display induced CD signals, from which the chiral supramolecular structure can be deduced [126–129].

9.5.1. Chiral Homo-aggregate Cyclodextrins modiﬁed with a long rigid chromophore of appropriate size are prone to thread together to give linear self-aggregates or supramolecular polymers, in which the chromophore group penetrates into the cavity of other cyclodextrin. Harada and coworkers [130] revealed that β-cyclodextrin 4-aminocinnamate 54 (Figure 9.36) forms a tail-to-tail dimeric aggregate in aqueous solution, while 4-(trinitroanilino)cinnamate 55, possessing a bulkier terminal group, forms head-to-tail aggregates to give a gel. The induced CD signals at 220–350 nm (Figure 9.37) are attributed to the inclusion of the chromophore in the β-cyclodextrin cavity. Interestingly, addition of urea (2 M), which is known to break the hydrogen bond, does not alter the CD spectrum, but the supramolecular polymer is disassembled by adding a better guest, adamantanecarboxylic acid, to give much weaker CD.

9.5.2. Chiral Hetero-aggregate The sergeants and soldiers principle and the majority rule originally found for conventional copolymers are also applied to supramolecular polymers [131–149]. Meijer and co-workers [145] studied the chiral supramolecular aggregation driven by hydrogenbonding and π –π stacking interactions. C3 -symmetric 56 (Figure 9.38), possessing nine identical chiral lipophilic chains at the periphery, forms a columnar structure through intermolecular hydrogen-bonding and stacking interactions. The anisotropy (g) factor of a solution of 56 is not proportional to the ee of 56 but obeys the majority rule (Figure 9.39). The free energy penalty for helix inversion is eight-fold larger than that

343

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

OR

R:

54

55

Figure 9.36. Chromophore-appended β-cyclodextrins 54 and 55.

(a)

30

AdCA 25 20 (b) θ (mdeg)

15 Urea 10 5 0 –5 –10 250

300 λ (nm)

350

400

Figure 9.37. Circular dichroism spectra of 1 mM 55 (black line), in the presence of 2 M urea (gray line) and in the presence of an excess of adamantanecarboxylic acid (AdCA) (dashed line). The inset shows the gel-to-sol transition upon addition of (a) 40 mM AdCA and (b) 2 M urea to a 20 mM solution of 55. (Reprinted with permission from reference 130. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.)

for incorporating an antipodal monomer with keeping the original helicity, which is the origin of the majority-ruled behavior of this system. The self-assembling mechanism of structurally resembling chiral (R)-57 and achiral 58 (Figure 9.40) was also investigated by ECD spectroscopy [136, 147]. A positive CD couplet was observed for (R)-57 aggregates, indicating the formation of right-handed

344

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

N

Figure 9.38. C3 -symmetrical disk-shaped dendrimer (R)- and (S)-56.

helical columnar structure. Temperature-dependence behavior of the CD intensity of (R)57 revealed two distinct regimes of nucleation and elongation [150]. Above the critical elongation temperature, (R)-57 exists in the nonhelical nucleation state and gives almost no CD, and the chiral aggregates start to elongate only below that temperature. A mixture of (R)-57 and 58 obeys the sergeants and soldiers principle, because all supramolecular columns are in the same handedness upon addition of only 4% (R)-57 to 58,

9.5.3. Chirality Memory in Supramolecular Assembly Chiral supramolecular assembly originally constructed from chiral subunits may preserve its chiral character even after the chiral component is removed or replaced by an achiral substitute, provided that the kinetic and dynamic requirements are met. Reinhoudt and co-workers [151, 152] discovered an interesting chirality memory phenomenon in the hydrogen-bonded aggregation of calix[4]arene dimelamines 59 (Figure 9.41) with chiral cyanurates. Calix[4]arene dimelamine 59 and (R)-barbiturate ((R)-BAR) forms M -helical complex (M )-[593 ·(R)-BAR6 ] in benzene through multiple hydrogen-bonding interactions, while the assembly of 59 with (S )-BAR affords antipodal (P )-[593 ·(S )BAR6 ]. The chiral barbiturate components of (M )-[593 ·(R)-BAR6 ] can be substituted for achiral cyanurates. Thus, addition of achiral butylcyanurate (BuCYA) to a solution of (M )-[593 ·(R)-BAR6 ] leads to the displacement of (R)-BAR by BuCYA with only accompanying a slight CD change, indicating preservation of the helical structure and sense. The rate-determining step of the racemization involves the dissociation of 59 from an intact assembly, followed by a quick disk-rotation and reassembling to the antipodal assembly.

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

345

100 80

Δε (L/mol.cm)

60 40 20 0 –20 –40 –60 –80 –100 200

250

300

350

400

450

500

λ (nm) (a)

0.002

g

0.001

0.000

–0.001

Figure 9.39. (a) CD spectra of octane solutions of dendrimer (S)-56 (open circles) and (R)-56 (closed circles); c = 2.49 × 10−5 M. (b) Anisotropy (g) factor as a function of the enantiomeric excess of

–0.002 –100

–50

0

50

100

Enantiomeric excess [%] (b)

56 at 20◦ C (closed circles) and 50◦ C (open circles). (Reprinted with permission from reference 145. Copyright 2005 American Chemical Society.).

9.6. SPONTANEOUS SYMMETRY BREAKING IN SUPRAMOLECULAR SYSTEM Controlling microscopic chiral event through macroscopic operation provides an important and intriguing tool for readily manipulating molecular and supramolecular chirality. Several recent studies based on ECD spectral analysis have drawn much attention to this possibility. Sodium chlorate crystallizes in l- and d-chiral forms in statistically equal numbers when crystallized from an aqueous solution without stirring. Interestingly, the crystals formed from a stirred solution are exclusively l- or d-chiral, although the handedness is not controlled [153]. It is believed that the autocatalytic secondary nucleation—that is, the formation of new crystal nuclei in the vicinity of an existing parent crystal—is responsible for this chiral symmetry-breaking processes. Chiral

346

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

R HN

O

O

H N R

NH O

R

57: R =

Figure 9.40. Self-assembling chiral 58: R =

benzenetricarbamides (R)-57 and achiral 58.

N HN R

2

NH2

H2N

N

N

N NH

1

R

N NH

N

HN

R

R

2

1

(R)-BAR O

BuCYA

O O O

59: R1 = NO2; R2 = (CH2)3CH3 O HN O

X: NH

X

N

C CH3

O BuCYA (R)-BAR (S)-BAR

Figure 9.41. Chirality memory in a hydrogen-bonded assembly.

autocatalysis was also observed with the random generation of large ee in the crystallization of 1,1 -binaphthyl melt [154]. Ribo and co-workers [155, 156] investigated the spontaneous chiral symmetry breaking in the vortex motion-induced chiral J-aggregation of 5,10,15-tris(4-sulfonatophenyl)20-phenylporphyrin 60 during rotary evaporation, with accompanying strong exciton coupling ECD signals at the Soret band. Thus, the clockwise/counterclockwise rotation affords negative/positive CD couplet, respectively, at high 85% probabilities, while unstirred aggregation leads to no chirality dominance. It was concluded that the motion of oligomeric blocks formed during the aggregation of 60 is diastereotopic, when a particular vortex direction is externally prescribed, and such a preferential asymmetric accretion is imprinted into the aggregated material as the newly arriving blocks weld at deﬁnite arrangements.

347

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

Me O

Me O

Me

Me O

Me O

Me O

Me O

O Me

O

O

O

Me

O

O

Me O

O

H O

O

O O

N

O

Zn N

Me O

O

O Me

O N

O Me

O

O

O

O O

O H

O O Me

O

O Me

O Me O

O

O Me

O

O

O

O

O Me

O Me

O

O

N

O

Me O

O Me

O

Me O

O

Me O O

O Me O O

Me O

O Me

O Me

O Me

O O

Me

O Me O Me

O Me

61

200 CCW

CD (mdeg)

100 0 OFF –100 CW –200 350

400

450

500 λ (nm)

550

600

650

Figure 9.42. CD spectra emerged upon rotary stirring in clockwise (CW) and counterclockwise (CCW) directions and without stirring (OFF). (Reprinted with permission from reference 157. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.)

Aida and co-workers [157] reported the induction of negative/positive CD couplet at the Soret band upon clockwise/counterclockwise rotary stirring of a solution containing dendric zinc porphyrin 61 (Figure 9.42). This stirring-induced exciton coupling was attributed to the macroscopic chiral alignment of the nanoﬁbers formed by J-aggregation of 60 upon rotary stirring.

9.7. SUMMARY Supramolecular chirality is a highly intriguing, rapidly growing area of chemistry and biology. Investigations of supramolecular chirality not only provide valuable insights into the chiral phenomena occurring in natural and artiﬁcial supramolecular systems, but also

348

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

offer guiding principles for designing advanced supramolecular materials and for better understanding and mimicking the biological and pharmacological processes. This is also a challenging interdisciplinary subject that requires theoretical and experimental knowledge and techniques in quantum chemistry, synthetic chemistry, stereochemistry, supramolecular chemistry, and analytical chemistry. ECD spectroscopy is a powerful indispensable analytical tool for investigating chiral supramolecular phenomena, owing to the development of theoretical and instrumental CD spectral tools applicable to supramolecular systems. Nakanishi and Harada’s exciton chirality theory is the most widely applied approach that allows the determination of chiral spatial arrangement of chromophores as well as the design and construction of chiral supramolecular architectures. Other principles, such as the sector rule for cyclodextrin complexation, are also crucial for elucidating the detailed supramolecular orientation and conformation in certain systems. ECD is applicable to most of the chiral supramolecular phenomena, including the complexation of chiral/achiral hosts with achiral/chiral guests and the aggregation of molecules with chiral elements. Chromophores with high extinction coefﬁcients are normally favored in ECD measurement, and chiral/achiral chromophores are commonly introduced to such supramolecular systems that lack absorption at appropriate wavelengths. Quantitative interpretation and prediction of supramolecular ECD are still a signiﬁcant challenge at the moment. There are reasons to believe that broader implementation in the near future of the fast advancing quantum mechanical methodologies for predicting the chiroptical properties will bring considerable success to the ﬁeld.

ACKNOWLEDGMENTS The authors are grateful to the supports of this work by PRESTO, JST (CY) and JSPS (YI).

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

G. Seeber, B. Tiedemann, K. Raymond, Top. Curr. Chem. 2006, 265 , 147–183. H. Eyring, H. Liu, D. Caldwell, Chem. Rev . 1968, 68 , 525–540. K. Wong, J. Chem. Educ. 1974, 51 , 573. T. Keiderling, Appl. Spectros. Rev . 1981, 17 , 189–226. D. Turner, I. Tinoco, Jr., M. Maestre, J. Am. Chem. Soc. 1974, 96 , 4340–4342. J. Dong, A. Wada, T. Takakuwa, K. Nakanishi, N. Berova, J. Am. Chem. Soc 1997, 119 , 12024–12025. J. Riehl, F. Richardson, Chem. Rev . 1986, 86 , 1–16. T. Bell, N. Hext, Chem. Soc. Rev . 2004, 33 , 589–598. A. Hembury Guy, V. Borovkov Victor, Y. Inoue, Chem. Rev . 2008, 108 , 1–73. N. Harada, K. Nakanishi, Circular Dichroic Spectroscopy: Exciton Coupling in Organic Stereochemistry, University Science Books, Mill Valley, CA, 1983. K. Nakanishi, N. Berova, Exciton chirality method: Principles and applications, in Circular dichroism—Principles and Applications. Wiley-VCH, New York, 2000, pp. 337–382. T. Mizutani, H. Takagi, O. Hara, T. Horiguchi, H. Ogoshi, Tetrahedron Lett. 1997, 38 , 1991–1994. J. Etxebarria, H. Degenbeck, A. S. Felten, S. Serres, N. Nieto, A. Vidal-Ferran, J. Org. Chem. 2009, 74 , 8794–8797.

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

14. H. Takagi, T. Mizutani, T. Horiguchi, S. Kitagawa, H. Ogoshi, Org. Biomol. Chem. 2005, 3 , 2091–2094. 15. J.-M. Fang, S. Selvi, J.-H. Liao, Z. Slanina, C.-T. Chen, P.-T. Chou, J. Am. Chem. Soc. 2004, 126 , 3559–3566. 16. W. Lu, L.-H. Zhang, X.-S. Ye, J. Su, Z. Yu, Tetrahedron 2006, 62 , 1806–1816. 17. X. Huang, B. Rickman, B. Borhan, N. Berova, K. Nakanishi, J. Am. Chem. Soc 1998, 120 , 6185–6186. 18. X. Huang, N. Fujioka, G. Pescitelli, F. Koehn, R. Williamson, K. Nakanishi, N. Berova, J. Am. Chem. Soc 2002, 124 , 10320–10335. 19. X. Huang, B. Borhan, B. Rickman, K. Nakanishi, N. Berova, Chemistry—Eur. J . 2000, 6 , 216–224. 20. G. Proni, G. Pescitelli, X. Huang, K. Nakanishi, N. Berova, J. Am. Chem. Soc 2003, 125 , 12914–12927. 21. G. Pescitelli, S. Gabriel, Y. Wang, J. Fleischhauer, R. Woody, N. Berova, J. Am. Chem. Soc 2003, 125 , 7613–7628. 22. Y. Chen, A. Petrovic, M. Roje, G. Pescitelli, M. Kayser, Y. Yang, N. Berova, G. Proni, Chirality 2010, 22 , 140–152. 23. A. Petrovic, Y. Chen, G. Pescitelli, N. Berova, G. Proni, Chirality 2010, 22 , 129–139. 24. X. Huang, K. Nakanishi, N. Berova, Chirality 2000, 12 , 237–255. 25. V.V. Borovkov, J. M. Lintuluoto, Y. Inoue, J. Phys. Chem. A 2000, 104 , 9213–9219. 26. V. V. Borovkov, J. M. Lintuluoto, Y. Inoue, Org. Lett. 2000, 2 , 1565–1568. 27. V. V. Borovkov, J. M. Lintuluoto, Y. Inoue, J. Am. Chem. Soc. 2001, 123 , 2979–2989. 28. D. Hesek, G. A. Hembury, M. G. B. Drew, V. V. Borovkov, Y. Inoue, J. Am. Chem. Soc. 2001, 123 , 12232–12237. 29. V. Borovkov Victor, T. Harada, Y. Inoue, R. Kuroda, Angew. Chem. Int. Ed . 2002, 41 , 1378–1381. 30. V. Borovkov Victor, M. Lintuluoto Juha, H. Sugeta, M. Fujiki, R. Arakawa, Y. Inoue, J. Am. Chem. Soc. 2002, 124 , 2993–3006. 31. V. Borovkov Victor, M. Lintuluoto Juha, M. Sugiura, Y. Inoue, R. Kuroda, J. Am. Chem. Soc. 2002, 124 , 11282–11283. 32. V. V. Borovkov, T. Harada, Y. Inoue, R. Kuroda, Angew. Chem., Int. Ed . 2002, 41 , 1378–1381. 33. V. Borovkov Victor, T. Harada, A. Hembury Guy, Y. Inoue, R. Kuroda, Angew. Chem. Int. Ed . 2003, 42 , 1746–1749. 34. V. Borovkov Victor, A. Hembury Guy, Y. Inoue, Angew. Chem. Int. Ed . 2003, 42 , 5310–5314. 35. V. Borovkov Victor, I. Fujii, A. Muranaka, A. Hembury Guy, T. Tanaka, A. Ceulemans, N. Kobayashi, Y. Inoue, Angew. Chem. Int. Ed . 2004, 43 , 5481–5485. 36. V. V. Borovkov, G. A. Hembury, Y. Inoue, Acc. Chem. Res. 2004, 37 , 449–459. 37. V. V. Borovkov, Y. Inoue, Top. Curr. Chem. 2006, 265 , 89–146. 38. P. Bhyrappa, V. Borovkov Victor, Y. Inoue, Org. Lett. 2007, 9 , 433–435. 39. V. V. Borovkov, J. M. Lintuluoto, Y. Inoue, J. Phys. Chem. B 1999, 103 , 5151–5156. 40. V. V. Borovkov, J. M. Lintuluoto, G. A. Hembury, M. Sugiura, R. Arakawa, Y. Inoue, J. Org. Chem. 2003, 68 , 7176–7192. 41. J. Lintuluoto, V. Borovkov, Y. Inoue, J. Am. Chem. Soc 2002, 124 , 13676–13677. 42. J. Etxebarria, A. Vidal-Ferran, P. Ballester, Chem. Commun. (Cambridge, U.K.) 2008, 5939–5941. 43. J. Etxebarria, H. Degenbeck, A. S. Felten, S. Serres, N. Nieto, A. Vidal-Ferran, J. Org. Chem. 2009, 74 , 8794–8797.

349

350

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

44. T. Morozumi, S. Shinkai, J. Chem. Soc., Chem. Commun. 1994, 1994 , 1219–1220. 45. K. Kondo, Y. Shiomi, M. Saisho, T. Harada, S. Shinkai, Tetrahedron 1992, 48 , 8239–8252. 46. Y. Shiomi, M. Saisho, K. Tsukagoshi, S. Shinkai, J. Chem. Soc., Perkin Trans. 1 1993, 1993 , 2111–2117. 47. S. Superchi, M. Donnoli, C. Rosini, Org. Lett 1999, 1 , 2093–2096. 48. G. Fasman, Circular Dichroism and the Conformational Analysis of Biomolecules, Plenum Press, New York, 1996. 49. M. Green, B. Garetz, B. Munoz, H. Chang, S. Hoke, R. Cooks, J. Am. Chem. Soc. 1995, 117 , 4181–4182. 50. K. Ute, Y. Fukunishi, S. Jha, K. Cheon, B. Munoz, K. Hatada, M. Green, Macromolecules 1999, 32 , 1304–1307. 51. M. Green, N. Peterson, T. Sato, A. Teramoto, R. Cook, S. Lifson, Science 1995, 268 , 1860. 52. M. Green, M. Reidy, R. Johnson, G. Darling, D. O’Leary, G. Willson, J. Am. Chem. Soc. 1989, 111 , 6452–6454. 53. A. Hofacker, J. Parquette, Angew. Chem. 2005, 117 , 1077–1081. 54. K. Akagi, G. Piao, S. Kaneko, K. Sakamaki, H. Shirakawa, M. Kyotani, Science 1998, 282 , 1683–1686. 55. H. Goto, K. Akagi, Macromolecules 2005, 38 , 1091–1098. 56. R. Ohta, F. Togashi, H. Goto, Macromolecules 2007, 40 , 5228–5230. 57. H. Goto, K. Akagi, Angew. Chem. 2005, 117 , 4396–4402. 58. H. Nakashima, J. Koe, K. Torimitsu, M. Fujiki, J. Am. Chem. Soc 2001, 123 , 4847–4848. 59. T. Sugimoto, T. Suzuki, S. Shinkai, K. Sada, J. Am. Chem. Soc. 2007, 129 , 270–271. 60. C. Li, M. Numata, A.-H. Bae, K. Sakurai, S. Shinkai, J. Am. Chem. Soc. 2005, 127 , 4548–4549. 61. E. Yashima, K. Maeda, T. Nishimura, Chemistry—Eur. J . 2004, 10 , 42–51. 62. E. Yashima, T. Matsushima, Y. Okamoto, J. Am. Chem. Soc 1997, 119 , 6345–6359. 63. E. Yashima, Y. Maeda, Y. Okamoto, J. Am. Chem. Soc. 1998, 120 , 8895–8896. 64. E. Yashima, K. Maeda, Y. Okamoto, Nature 1999, 399 , 449–451. 65. H. Goto, Y. Okamoto, E. Yashima, Chemistry—Eur. J . 2002, 8 , 4027–4036. 66. H. Goto, Y. Okamoto, E. Yashima, Macromolecules 2002, 35 , 4590–4601. 67. H. Goto, E. Yashima, J. Am. Chem. Soc. 2002, 124 , 7943–7949. 68. M. Ishikawa, K. Maeda, E. Yashima, J. Am. Chem. Soc. 2002, 124 , 7448–7458. 69. H. Goto, Q. Zhang Hao, E. Yashima, J. Am. Chem. Soc. 2003, 125 , 2516–2523. 70. K. Maeda, K. Morino, Y. Okamoto, T. Sato, E. Yashima, J. Am. Chem. Soc. 2004, 126 , 4329–4342. 71. T. Nishimura, K. Tsuchiya, S. Ohsawa, K. Maeda, E. Yashima, Y. Nakamura, J. Nishimura, J. Am. Chem. Soc. 2004, 126 , 11711–11717. 72. E. Yashima, K. Maeda, Y. Takeyama, H. Mochizuki, K. Sakajiri, PMSE Prepr. 2004, 91 , 125. 73. M. Ikeda, Y. Tanaka, T. Hasegawa, Y. Furusho, E. Yashima, J. Am. Chem. Soc. 2006, 128 , 6806–6807. 74. T. Kajitani, K. Okoshi, S.-i. Sakurai, E. Yashima, Polym. Prepr. (Am. Chem. Soc., Div. Polym. Chem.) 2006, 47 , 153. 75. S.-I. Sakurai, K. Okoshi, J. Kumaki, E. Yashima, J. Am. Chem. Soc. 2006, 128 , 5650–5651. 76. T. Miyagawa, M. Yamamoto, R. Muraki, H. Onouchi, E. Yashima, J. Am. Chem. Soc. 2007, 129 , 3676–3682. 77. T. Maeda, Y. Furusho, S.-I. Sakurai, J. Kumaki, K. Okoshi, E. Yashima, J. Am. Chem. Soc. 2008, 130 , 7938–7945.

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

78. Y. Hase, K. Nagai, H. Iida, K. Maeda, N. Ochi, K. Sawabe, K. Sakajiri, K. Okoshi, E. Yashima, J. Am. Chem. Soc. 2009, 131 , 10719–10732. 79. M. Ishikawa, K. Maeda, Y. Mitsutsuji, E. Yashima, J. Am. Chem. Soc 2004, 126 , 732–733. 80. M. Waki, H. Abe, M. Inouye, Angew. Chem. Int. Ed . 2007, 46 , 3059–3061. 81. S. Takenaka, N. Matsuura, N. Tokura, Tetrahedron Lett. 1974, 15 , 2325–2328. 82. S. F. Mason, Chem. Phys. Lett. 1975, 32 , 201–203. 83. K. Harata, H. Uedaira, Bull. Chem. Soc. Jpn. 1975, 48 , 375–378. 84. K. Harata, Bioorganic Chemistry 1981, 10 , 255–265. 85. J. Kirkwood, J Chem Phys 1937, 5 , 479. 86. M. Kajtar, C. Horvath-Toro, E. Kuthi, J. Szejtli, Acta Chim. Acad. Sci. Hung. 1982, 110 . 87. M. Kodaka, T. Shimizu, M. Hatano, Bull. Chem. Soc. Jpn. 1983, 56 , 523–527. 88. M. Kodaka, T. Fukaya, Bull. Chem. Soc. Jpn. 1986, 59 , 2032–2034. 89. M. Kodaka, T. Fukaya, Bull. Chem. Soc. Jpn. 1989, 62 , 1154–1157. 90. M. Kodaka, J. Phys. Chem. 1991, 95 , 2110–2112. 91. M. Kodaka, J. Am. Chem. Soc. 1993, 115 , 3702–3705. 92. M. Barra, C. Bohne, J. C. Scaiano, J. Am. Chem. Soc. 1990, 112 , 8075–8079. 93. S. Hamai, T. Ikeda, A. Nakamura, H. Ikeda, A. Ueno, F. Toda, J. Am. Chem. Soc. 1992, 114 , 6012–6016. 94. D. Krois, U. H. Brinker, J. Am. Chem. Soc. 1998, 120 , 11627–11632. 95. B. Mayer, X. Zhang, W. M. Nau, G. Marconi, J. Am. Chem. Soc. 2001, 123 , 5240–5248. 96. M. Miyauchi, Y. Takashima, H. Yamaguchi, A. Harada, J. Am. Chem. Soc. 2005, 127 , 2984–2989. 97. C. Yang, T. Mori, Y. Origane, Y. H. Ko, N. Selvapalam, K. Kim, Y. Inoue, J. Am. Chem. Soc. 2008, 130 , 8574–8575. 98. Y.-L. Zhao, W. R. Dichtel, A. Trabolsi, S. Saha, I. Aprahamian, J. F. Stoddart, J. Am. Chem. Soc. 2008, 130 , 11294–11296. 99. A. Ueno, H. Yoshimura, R. Saka, T. Osa, J. Am. Chem. Soc. 1979, 101 , 2779–2780. 100. R. Lu, C. Yang, Y. Cao, Z. Wang, T. Wada, W. Jiao, T. Mori, Y. Inoue, Chem. Commun. (Cambridge, U.K.) 2008, 374–376. 101. R. Lu, C. Yang, Y. Cao, L. Tong, W. Jiao, T. Wada, Z. Wang, T. Mori, Y. Inoue, J. Org. Chem. 2008, 73 , 7695–7701. 102. C. Yang, M. Nishijima, A. Nakamura, T. Mori, T. Wada, Y. Inoue, Tetrahedron Lett. 2007, 48 , 4357. 103. C. Yang, A. Nakamura, T. Wada, Y. Inoue, Org. Lett. 2006, 8 , 3005. 104. C. Yang, A. Nakamura, G. Fukuhara, Y. Origane, T. Mori, T. Wada, Y. Inoue, J. Org. Chem. 2006, 71 , 3126. 105. C. Yang, T. Mori, Y. Inoue, J. Org. Chem. 2008, 73 , 5786. 106. C. Yang, C. Ke, K. Fujita, D.-Q. Yuan, T. Mori, Y. Inoue, Aust. J. Chem. 2008, 1–4. 107. C. Yang, G. Fukuhara, A. Nakamura, Y. Origane, K. Fujita, D.-Q. Yuan, T. Mori, T. Wada, Y. Inoue, J. Photochem. Photobiol. A: chem. 2005, 173 , 375. 108. A. Nakamura, Y. Inoue, J. Am. Chem. Soc. 2005, 127 , 5338. 109. A. Nakamura, Y. Inoue, J. Am. Chem. Soc. 2003, 125 , 966. 110. C. Ke, C. Yang, T. Mori, T. Wada, Y. Liu, Y. Inoue, Angew. Chem. Int. Ed . 2009, 48 . 111. R. Pasternack, Chirality 2003, 15 , 329–332. 112. B. Nord n, T. Kurucsev, J. Mol. Recognit. 1994, 7 , 141–155. 113. D. Fornasiero, T. Kurucsev, J. Phys. Chem. 1981, 85 , 613–618.

351

352

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

114. A. D’Urso, A. Mammana, M. Balaz, A. Holmes, N. Berova, R. Lauceri, R. Purrello, J. Am. Chem. Soc 2009, 131 , 2046–2047. 115. J. Barton, J. Biomol. Struct. Dyn. 1983, 1 , 621. 116. B. Nord´en, F. Tjerneld, FEBS Lett. 1976, 67 , 368. 117. Y. Xu, Y. X. Zhang, H. Sugiyama, T. Umano, H. Osuga, K. Tanaka, J. Am. Chem. Soc. 2004, 126 , 6566–6567. 118. T. Wada, M. Nishijima, T. Fujisawa, N. Sugahara, T. Mori, A. Nakamura, Y. Inoue, J. Am. Chem. Soc 2003, 125 , 7492–7493. 119. M. Nishijima, T. Pace, A. Nakamura, T. Mori, T. Wada, C. Bohne, Y. Inoue, J. Org. Chem 2007, 72 , 2707–2715. 120. M. Nishijima, T. Wada, T. Mori, T. Pace, C. Bohne, Y. Inoue, J. Am. Chem. Soc 2007, 129 , 3478–3479. 121. T. Arimura, H. Kawabata, T. Matsuda, T. Muramatsu, H. Satoh, K. Fujio, O. Manabe, S. Shinkai, J. Org. Chem. 1991, 56 , 301–306. 122. S. Shinkai, T. Arimura, H. Kawabata, H. Murakami, K. Iwamoto, J. Chem. Soc., Perkin Trans. 1 1991, 1991 , 2429–2434. 123. T. Arimura, S. Shinkai, Bull. Chem. Soc. Jpn. 1991, 64 , 1896–1900. 124. K. Kobiro, Y. Inoue, J. Am. Chem. Soc 2003, 125 , 421–427. 125. G. Gottarelli, S. Lena, S. Masiero, S. Pieraccini, G. Spada, Chirality 2008, 20 , 471–485. 126. F. Saeva, J. Wysocki, J. Am. Chem. Soc. 1971, 93 , 5928–5929. 127. F. Saeva, G. Olin, J. Am. Chem. Soc. 1973, 95 , 7882–7884. 128. K. Tang, M. Green, K. Cheon, J. Selinger, B. Garetz, J. Am. Chem. Soc 2003, 125 , 7313–7323. 129. T. Kato, T. Matsuoka, M. Nishii, Y. Kamikawa, K. Kanie, T. Nishimura, E. Yashima, S. Ujiie, Angew. Chem. Int. Ed . 2004, 43 , 1969–1972. 130. W. Deng, H. Yamaguchi, Y. Takashima, A. Harada, Angew. Chem. Int. Ed . 2007, 46 , 5144–5147. 131. L. J. Prins, P. Timmerman, D. N. Reinhoudt, J. Am. Chem. Soc. 2001, 123 , 10153–10163. 132. M. Mateos-Timoneda, M. Crego-Calama, D. Reinhoudt, Chem. Soc. Rev . 2004, 33 , 363–372. 133. M. Crego-Calama, D. N. Reinhoudt, M. G. J. ten Cate, Top. Curr. Chem. 2005, 249 , 285–316. 134. M. A. Mateos-Timoneda, M. Crego-Calama, D. N. Reinhoudt, Supramol. Chem. 2005, 17 , 67–79. 135. A. Mateos-Timoneda Miguel, M. Crego-Calama, N. Reinhoudt David, Chemistry 2006, 12 , 2630–2638. 136. L. Brunsveld, A. Schenning, M. Broeren, H. Janssen, J. Vekemans, E. Meijer, Chem. Lett. 2000, 29 , 292–293. 137. A. R. A. Palmans, J. A. J. M. Vekemans, E. E. Havinga, E. W. Meijer, Angew. Chem. Int. Ed . 1997, 36 , 2648–2651. 138. L. Brunsveld, J. H. K. K. Hirschberg, A. R. A. Palmans, J. A. J. M. Vekemans, E. W. Meijer, Book of Abstracts, 217th ACS National Meeting, Anaheim, California, March 21–25 1999, POLY-300. 139. B. Langeveld-Voss, R. Waterval, R. Janssen, E. Meijer, Macromolecules 1999, 32 , 227–230. 140. L. Brunsfeld, B. G. G. Lohmeijer, J. A. J. M. Vekemans, E. W. Meijer, J. Inclusion Phenom. Macrocyclic Chem. 2001, 41 , 61–64. 141. L. Brunsveld, E. W. Meijer, R. B. Prince, J. S. Moore, J. Am. Chem. Soc. 2001, 123 , 7978–7984. 142. L. Brunsveld, J. A. J. M. Vekemans, J. H. K. K. Hirschberg, R. P. Sijbesma, E. W. Meijer, Proc. Natl. Acad. Sci. U.S.A. 2002, 99 , 4977–4982.

ELECTRONIC CIRCULAR DICHROISM OF SUPRAMOLECULAR SYSTEMS

143. J. J. van Gorp, J. A. J. M. Vekemans, E. W. Meijer, J. Am. Chem. Soc. 2002, 124 , 14759–14769. 144. M. Masuda, P. Jonkheijm, P. Sijbesma Rint, E. W. Meijer, J. Am. Chem. Soc. 2003, 125 , 15935–15940. 145. J. van Gestel, A. R. A. Palmans, B. Titulaer, J. A. J. M. Vekemans, E. W. Meijer, J. Am. Chem. Soc. 2005, 127 , 5490–5494. 146. A. J. Wilson, M. Masuda, R. P. Sijbesma, E. W. Meijer, Angew. Chem. Int. Ed . 2005, 44 , 2275–2279. 147. M. M. J. Smulders, A. P. H. J. Schenning, E. W. Meijer, J. Am. Chem. Soc. 2008, 130 , 606–611. 148. M. M. J. Smulders, I. A. W. Filot, J. M. A. Leenders, P. van der Schoot, A. R. A. Palmans, A. P. H. J. Schenning, E. W. Meijer, J. Am. Chem. Soc. 2010, 132 , 611–619. 149. M. M. J. Smulders, P. J. M. Stals, T. Mes, T. F. E. Paffen, A. P. H. J. Schenning, A. R. A. Palmans, E. W. Meijer, J. Am. Chem. Soc. 2010, 132 , 620–626. 150. P. Jonkheijm, P. van der Schoot, A. Schenning, E. Meijer, Science 2006, 313 , 80. 151. L. J. Prins, J. Huskens, F. De Jong, P. Timmerman, D. N. Reinhoudt, Nature (London) 1999, 398 , 498–502. 152. L. J. Prins, F. De Jong, P. Timmerman, D. N. Reinhoudt, Nature (London) 2000, 408 , 181–184. 153. D. Kondepudi, R. Kaufman, N. Singh, Science 1990, 250 , 975. 154. D. Kondepudi, J. Laudadio, K. Asakura, J. Am. Chem. Soc 1999, 121 , 1448–1451. 155. J. Ribo, J. Crusats, F. Sagues, J. Claret, R. Rubires, Science 2001, 292 , 2063. 156. C. Escudero, J. Crusats, I. Dez-Prez, Z. El-Hachemi, J. Rib Angew. Chem. 2006, 118 , 8200–8203. 157. A. Tsuda, A. Alam Md, T. Harada, T. Yamaguchi, N. Ishii, T. Aida, Angew. Chem. Int. Ed . 2007, 46 , 8198–8202.

353

10 THE ONLINE STEREOCHEMICAL ANALYSIS OF CHIRAL COMPOUNDS BY HPLC-ECD COUPLING IN COMBINATION WITH QUANTUM-CHEMICAL CALCULATIONS ¨ Gerhard Bringmann, Daniel Gotz, and Torsten Bruhn

10.1. INTRODUCTION The combination of HPLC with analytical methods, like NMR, MSn , and CD is one of the most powerful analytical tools for the structural elucidation of chiral compounds, especially in natural product chemistry, where one often has to deal with complex mixtures, small product quantities, and/or chemically or stereochemically unstable substances [1]. In 1980 Mason and co-workers [2] reported the ﬁrst hyphenation of HPLC with a CD dichrograph. By this combination it became possible to measure absorptions and optical activities simultaneously, in the on-ﬂow mode. Salvadori et al. [3] were the ﬁrst to describe the determination of absolute conﬁgurations of simple and known chiral compounds by recording their CD signals at a suited single wavelength and interpreting the obtained CD effect by empirical or nonempirical rules like, for example, the octant rule or the exciton chirality method. A huge step forward was provided by Mannschreck and co-workers [4, 5] in 1992: They succeeded for the ﬁrst time in the online measurement of full CD spectra by HPLC in the stopped-ﬂow mode. The main advantage of the technique is that stereoisomers no longer need to be separated in a time-consuming semipreparative way for ofﬂine CD measurements. In an analogous way, the HPLC-NMR technique was reﬁned in the 1980s and 1990s and in 1998 Bringmann et al. were the ﬁrst to report on HPLC-ROESY-NMR measurements in the online structural analysis of natural products [6]. Together with the HPLC-MSn technique the concept of the “analytical triad” LC-MSn -NMR-CD was born [7], permitting the elucidation of full absolute stereostructures directly from the peak in crude mixtures, thus saving the often laborious, time- and money-consuming isolation procedures. In the following years, several convincing examples that evidenced Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

355

356

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

the efﬁciency of this method were reported, such as the online structural elucidation of the conﬁgurationally semi-stable biaryl alkaloid dioncophylline E [8] and the dimeric metabolite ancistrogrifﬁthine A, which possesses two stereocenters and two chiral axes [9]. Another interesting application of the analytical triad is the direct monitoring of biotransformations of metabolites in plant cell cultures and their structural analysis as described by Iwasa et al. [10–12]. HPLC-CD has also become a valuable analytical method in the pharmaceutical industry for drug screening and quality management [13, 14]. Thus, the scope of this chapter is, to demonstrate the huge potential of HPLC-CD coupling in modern analytical chemistry.

10.2. HIGH-PERFORMANCE LIQUID CHROMATOGRAPHY Between 1960 and 1970, high-performance liquid chromatography (HPLC) was developed as a new and efﬁcient analytical tool. Today, chiral resolution by HPLC is the most widely practiced analytical method for determination of optical purity, being applicable even to samples that include many impurities [15]. In addition, liquid chromatography is the only technique that permits the separation and identiﬁcation down to femtomolar components in complex matrices, but also allows for the isolation and puriﬁcation of synthetic industrial products in ton quantities [13, 16]. With the high safety standards, optical-purity analysis of pharmaceutical agents and agrochemicals is nowadays strictly required, since the presence of “undesired” stereoisomers, even in small quantities, may sometimes lead to harmful side effects [17, 18]. In this context, HPLC-CD coupling—that is, the hyphenation of a CD dichrograph to an HPLC device—offers a unique potential for the stereochemical investigations on chiral analytes, even if occurring in trace concentrations and accompanied by further byproducts. Important information that can be obtained from an HPLC-CD experiment are the determination of the enantiomeric (or diastereomeric) excesses [19], the study of isomerization processes of stereochemically unstable analytes [5, 20], the determination of the elution order of stereoisomers, and the measurement of full CD spectra of even minor compounds from crude extracts or reaction mixtures. The current chapter focuses on the latter three issues.

10.3. THE HPLC-CD DEVICE In general the setup of common HPLC-CD interfaces as schematically depicted in Figure 10.1 has essentially remained unchanged since its ﬁrst introduction in 1980 by Mason et al.: The outlet of a standard HPLC system is connected to a ﬂow cell installed within a “normal” CD detector. In the early days of HPLC-CD coupling, only individually constructed instruments were used by a small number of experts in the ﬁeld. But meanwhile the technique has become broadly available to nonspecialized end-users and some companies offer benchtop solutions for HPLC-CD applications at affordable costs. Usually an optical detector and a chiroptical one are connected in series to record UV and CD spectra simultaneously. Similar to usual UV detection in chromatography, a phase-sensitive CD trace can be obtained by monitoring the differential absorption of left- and right-circularly polarized light (A = AL − AR ) at a ﬁxed wavelength λ, and the resulting output is thus a plot of A against time. The hyphenation of the chromatographic system to the spectrometer can be achieved by using a motor valve, permitting to stop the solvent ﬂow through the measurement

357

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

Standard HPLC device Motor valve

Column UV detector

CD detector

Recorder output Waste UV signal A ‘On-ﬂow ’ chromatograms at a fixed wavelength λ

A

ΔA

CD signal at λ1 t

UV

A

UV

Recorder ΔA

CD

λ

λ1

Full online spectra recorded in the λ “stopped-ﬂow” mode

CD

λ1

λ λ

Figure 10.1. General schematic representation of a standard HPLC-CD device.

cell—for example, by redirecting the eluent from the HPLC pump directly into a waste ﬂask. Consequently, full CD and UV spectra can be recorded in the stopped-ﬂow mode with commercially available ECD detectors usually covering the spectral range from 200 to 850 nm. In addition, some HPLC pumps offer the possibility to keep the current eluent composition constant during the stopped-ﬂow measurement thus permitting to subsequently proceed with the analysis of further substances within the same run. The huge potential and wide application range of HPLC-CD hyphenation is based on the simple experimental setup and, in particular, on the fact that CD itself is a quite sensitive method that can detect trace amounts of chiral compounds as long as they contain sufﬁciently UV-absorbing chromophores [21, 22]. In order to further improve the sensitivity of CD measurements also more specialized CD detection devices applying laser-beam sources [23, 24] or using phase-sensitive FDCD (ﬂuorescence detection circular dichroism) [25–27] have been developed. Of these, however, only the laser-based CD detectors have so far been used in HPLC-CD hyphenation [23]. Beyond the outstanding potential of HPLC-CD alone, the additional hyphenation of high-performance liquid chromatography with further spectroscopic methods—that is, with tandem mass spectrometry (HPLC-MS/MS) and NMR spectroscopy (HPLCNMR)—has led to the “analytical triad” LC-MS/MS-NMR-CD [7]. This combined methodology has, during the past years, been applied to the full structural elucidation of—even complex—natural products right from crude extracts, as part of the strategy of a spectroscopy-guided search for structurally novel metabolites (Figure 10.2).

10.4. CHOICE OF THE CHROMATOGRAPHIC SYSTEM Mixtures of diastereomers can, in principle, be separated by HPLC on achiral phases. The chromatographic resolution of racemic mixtures, by contrast, requires a chiral auxiliary. It can be achieved by various approaches [13, 28], for example by the conversion of

358

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

HPLC

HPLC

NMR

MS Fragmentation patterns by MS/MS or exact mass from HRMS - molecular formula - constitution HPLC Polarizer

Resolution of crude extracts or reaction mixtures by HPLC

CD

1D and 2D experiments (1H, 13C, TOCSY, COSY, NOESY, HMQC, ROESY)

NMR - constitution - relative configuration

'On-flow' monitoring of chiral analytes and recording of full CD spectra in the 'stopped-flow' mode - enantiomeric excesses (ee) Detector - absolute configuration - ...

CD Quantum-chemical calculations NMR and CD calculations absolute configuration

Full absolute stereostructures right from the peak of the chromatogram of a crude mixture!

Figure 10.2. The fruitful interplay of HPLC-MS/MS, HPLC-NMR, and HPLC-CD within the ‘‘analytical triad’’ combined with quantum-chemical calculations.

the enantiomers into diastereomeric derivatives using chiral reagents and subsequent separation on an achiral column, by using a chiral mobile phase, or by complex formation with chiral additives (like, for example, cyclodextrins [29]). The derivatization, however, requires an additional synthetic step prior to chromatography and can be hampered by different reaction rates of the enantiomers (leading to a kinetic racemic resolution). The use of chiral solvents or additives usually causes substantial disadvantages, especially in hyphenated HPLC-CD applications, since (depending on the detection wavelength) the chiral auxiliary itself may give its own CD response and may, thus, falsify the overall signal. This can imply even more serious drawbacks if a solvent gradient is applied—that is, with a varying composition of the mobile phase. Consequently, HPLC on a chiral stationary phase (CSP) is the most common and broadly applicable method for online HPLC-CD analysis. To date, a plethora of both normal- and reversed-phase CSPs have been developed, of which more than 100 are commercially available [30]. The chiral resolution is based on diastereomeric interactions of enantiomers with the CSP, namely their differential adsorption, resulting in different retention times for the two enantiomers. For a given resolution problem the separating capacity and recognition ability of a CSP can be evaluated qualitatively and quantitatively by two main characteristic values: The separation (or selectivity) factor α and the resolution factor R (t1 and t2 are the retention times of the faster and the more slowly eluting enantiomers, respectively; t0 is the retention time of a nonretained compound, that is, the dead time; k1 and k2 are the retention factors; w1 and w2 are the peak widths at their bases) [31,

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

32]: α=

k2 (t2 − t0 ) = , (t1 − t0 ) k1

R=

2(t2 − t1 ) , w1 + w2

k1 =

(t1 − t0 ) , t0

(t2 − t0 ) . t0

k2 =

The separation factor α reﬂects the selectivity of the CSP, namely, the afﬁnity of the selected column for the individual enantiomers. The column performance is expressed by the plate number N ; thus the more efﬁcient the column, the smaller will be the peak width w at a given retention time t for a component:

t N = 16 w

2 .

The fundamental equation for optimizing HPLC separation conditions relates the resolution R to the number of theoretical plates N , the selectivity factor α, and the retention factor k2 : √ k2 N α−1 R= . α 1 + k2 4 Thus, R is a basic measure of the efﬁcacy of the chromatographic system in separating two components in a mixture and in order to provide a good resolution, the three terms have to be maximized [32, 33]. Optimization of the experiment usually involves manipulation of column and mobile-phase parameters to alter the relative migration rates of the components in the mixture and to reduce peak broadening. While baseline separation between two peaks usually requires an R value >1.5, a resolution around R = 1.0 may sometimes be sufﬁcient for HPLC-CD measurements. The reason for this seemingly higher resolution of LC-CD compared to LC-UV lies in the fact that CD spectra contain one (half) dimension more than UV spectra, in having positive and negative signals. Since enantiomers exhibit opposite Cotton effects, the CD signals of overlapping enantiomeric peaks partially compensate each other. Thus, in contrast to UV detection, two residual—opposite—signals may still remain at the edges of the seemingly unresolved peak in the UV chromatogram. Already during optimization of the chromatographic system the CD trace may provide a ﬁrst hint at partial success, long before the resolution of the two peaks becomes visible by UV detection [34, 35]. Increasing N by lengthening the column leads to a longer retention time and augmented peak broadening, which may not be desirable. Alternatively, the number of theoretical plates can effectively be increased by reducing the size of the stationary-phase particles. In addition, separations may be improved by controlling the retention factor k . The retention factors should normally lie between 2 and 5, but for complex mixtures a larger range may be required to resolve all components. The value of the retention factor for a given compound depends on its chemical properties and the following experimental variables: 1. Flow rate 2. Composition of the mobile phase (including pH value)

359

360

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

3. Column temperature 4. Stationary phase In many cases, the quality of the resolution is most efﬁciently improved by manipulating the selectivity factor α. If α is close to 1.0, optimization of k and increase of N may not be sufﬁcient to achieve good separations with reasonable retention times. In these cases, k is optimized ﬁrst and then α is increased, again by changing the mobilephase composition or the column temperature or by switching to a different stationary phase. Detailed procedures for optimizing the chromatographic system for an individual application have intensely been reviewed in various papers and book contributions [30, 36–38].

10.5. CHOICE OF APPROPRIATE DETECTION WAVELENGTHS The choice of a suitable detection wavelength is of utmost importance for HPLC-CD measurements, since it determines the response factor of the detector, which, in turn, affects both its sensitivity and selectivity. Concerning the selectivity, one may choose one of the following two “adjustment modes,” depending on the desired information: A nonselective detector will monitor the majority of different components of a mixture, while the optimization of the detection wavelength to a selected compound might preferentially record a response arising from one single species in a crude extract. Detection wavelengths below 250 nm are generally more suitable to reﬂect a plethora of different chiral substances in complex mixtures, since a large number of chromophores exhibit signiﬁcant absorptions in this wavelength region. On the other hand, this also means that UV and CD detection are often hampered by interfering contributions of the solvents used as the mobile phase, especially if they are not completely devoid of absorbing contaminants. While the mobile phase has of course to be chosen primarily according to an optimum solubility of the analytes and the efﬁcacy of the separation, a ﬁne-tuning of the solvent composition may be advantageous or can even become necessary, especially for the measurement of full online CD spectra by HPLC-CD coupling: Then it may be advisable to substitute the eluent by a solvent with a similar polarity but different UV properties. Sometimes it also may be important to substitute a hydrogen-bond donating solvent like MeOH by a nonprotic one, to achieve spectra more similar to the gas-phase spectra and to the results of quantum-chemical predictions. For example, acetonitrile can often be used instead of methanol to minimize undesired absorptions by the solvent in the region around 200 nm. Table 10.1 lists approximate cutoff wavelengths below which the eluent absorbance may become unacceptable. Below the wavelength λ0 , the absorption of the solvent exceeds 0.05 absorbance units (relative to water) with a pathlength of 10 mm (i.e., A1cm > 0.05), while the absorption of the solvent is even 20 times higher at λ1 (A1cm > 1.0). Online CD measurements can be performed without problems down to λ0 , while CD curves should be interpreted with caution in the region between λ0 and λ1 , especially if the UV curve of the eluent rises steeply and/or if the observed/expected Cotton effects are small. Below λ1 CD measurements may yield ambiguous and sometimes badly reproducible results due to the strongly interfering UV absorption of the eluent. If at all possible, the detection of CD curves below λ1 should thus be avoided. In conclusion, the detection wavelength has to be adjusted carefully and sometimes a change of the mobile phase (if applicable) can offer an alternative to obtain CD spectra of highest possible quality.

361

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

TAB L E 10.1. Solvent Cutoff Wavelengths of Some Commonly Used HPLC Eluents [39] Eluent Dichloromethane Methanol or 2-propanol n-Hexane Acetonitrile Water

λ0 (nm)

λ1 (nm)

245 240 225 200 190

230 205 195 190 185

On the other hand, the sensitivity of the HPLC device is associated with the detection limit (LOD), which is strongly dependent on the spectroscopic properties of the analyte, but also on the stationary phase and the applied eluent. Since CD spectroscopy monitors the difference between the absorption of left- and right-circularly polarized light, the best signal-to-noise ratio (S/N) is usually obtained if ε is large and, at the same time, ε is comparatively small. Consequently, the most intense absorption may sometimes not provide the wavelength of choice for the detector setting. In general, CD detection affords the optimum S/N if the anisotropy factor g (see Section 10.6) is maximized. This is, however, possible only if the UV and CD properties of the compounds to be analyzed are known or can be estimated reasonably well. Frequently, the detection wavelength is more easily adjusted if the CD signal is known to arise from a speciﬁc transition like, for example, in the case of n → π * transitions in saturated ketones (around 300 nm). Furthermore, if the CD signal is expected to result from an exciton coupling—that is, the dipole–dipole interaction of locally excited states in adjacent, (ideally) identical chromophores—the CD detector is usually set to a wavelength that is red-shifted by about 10 nm as compared to the maximum absorption of the racemate. Since, by deﬁnition, the CD effect tends to zero at the UV maximum, this arbitrary shift of the CD detection wavelength usually ﬁts the lowenergy extreme value of the respective couplet quite well. It is noteworthy that the CD response will switch sign if the detection wavelength is blue-shifted as compared to the corresponding UV absorption (for details see the chapter about the exciton chirality method). If two separate instruments are used for UV and CD detection, these can be adjusted to different wavelengths (as long as the g factor is not necessarily required). This might be advantageous since both sensitivity and selectivity can to some degree be tuned independently by the choice of the respective detection wavelengths of the UV (sensitivity) and CD (selectivity) detectors.

10.6. QUANTITATIVE ANALYSIS AND EE DETERMINATION USING HPLC-CD DETECTION: THE ANISOTROPY FACTOR G A common drawback in chromatography arises from the fact that the absolute quantity of a compound is hard to determine, especially at a single point of the chromatogram. However, with simultaneous UV and CD detection the so-called anisotropy or dissymmetry factor g can be derived from the ratio of the dichroic signal and absorbance [2, 34, 40, 41]: A . g= A

362

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The g factor does not depend on the sample concentration and is linearly related to the enantiomeric excess (ee), provided that the pathlength is the same for the UV and the CD measurement: ee . g = g max × 100 Thus, if the anisotropy factor g of the pure enantiomer (g max ) is known, the enantiomeric composition of the eluates can easily be determined at every single point of the chromatogram directly during the HPLC analysis: ee =

g g max

× 100

The direct measurement of the enantiomeric excess in a scalemic mixture is fundamental, for example, for the optimization of fraction collection in preparative liquid chromatography if two enantiomers can be resolved only partially on a given chiral stationary phase: The g factor of a pure enantiomer has a well-deﬁned, constant value (g max ), which decreases when the ﬁrst eluted peak in a partially resolved racemate becomes contaminated by the more slowly eluting peak—that is, by the oppositely conﬁgured analyte. Thus, monitoring of the anisotropy factor g permits to collect the largest-possible fractions of optically pure material from a partially resolved racemic or scalemic mixture [40].

10.7. GENERAL INTERPRETATION OF CD SPECTRA There are several strategies for the stereochemical analysis of experimental CD spectra, from merely empirical methods to quantum-chemical calculations. The most common (and merely experimental) approach to determine the absolute conﬁguration of a novel chiral substance is the comparison of its CD spectrum with that of a structurally closely related, conﬁgurationally known compound. As simple and straightforward as this seems at ﬁrst glance, the method implies hidden—and thus dangerous!—traps, since it is often difﬁcult to judge whether the chosen reference structure is really ‘comparable’ or not. Of course the compared substances have to possess identical chromophores, with a similar stereo-orientation to each other. The pivotal effect of the conformation of the chromophores (e.g., of a phenyl substituent) on the overall CD can be seen in rocaglamide AE (1) versus its close, but cyclic, analogue cyclorocaglamide (rocaglamide AN, 2) [7, 42, 43], which shows a nearly opposite CD spectrum, despite the identity of the absolute conﬁguration at all ﬁve stereocenters! The bridging in 2 stabilizes one particular conformational array, which is also present—but less populated—in 1, where chiroptically opposite conformers prevail. Furthermore, it is indispensable to know the inﬂuence of the different substituents of the chromophores on the CD spectrum. In most cases, simple substituents such as OH, OMe, or Me groups usually have no signiﬁcant impact on a CD curve (Figure 10.3)—as long as their effects are overlayed by more dominant chromophores such as the naphthyl ring in dioncopeltine A (3) and habropetaline A (4) and if they do not have an effect to the conformation. By contrast, more strongly electron-withdrawing or -donating groups can effect a signiﬁcant inﬂuence on the spectrum by altering the electronic structure and polarity of the subunits, thus changing the electron distribution in the chromophores, and the assignment of absolute conﬁgurations by a mere comparison of the CD spectra may become doubtful in such cases [44].

363

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

40 Cyclorocaglamide (2)

OMe OH OH

O

O HO

S R R R S

C

NMe2 O

Ph

CD [mdeg]

O

OMe

O OO

0 –20

MeO

O

20

–40 200

C

NMe2 O

Ph MeO

Rocaglamide AE (1)

OMe

OMe Rocaglamide AE (1)

OH

SRR RS

Cyclorocaglamide (2) 300

250

350

wavelength λ [nm] (a)

Me HO M MeO HO

R R

NH

OH Me 5'

CD [mdeg]

50 Dioncopeltine A (3)

25

Me HO M

0 Habropetaline A (4) –25

MeO MeO

Dioncopeltine A (3)

R

NH OH Me R

5'

Habropetaline A (4) –50 200

250

300

350

wavelength λ [nm] (b)

Figure 10.3. (a) Comparison of the CD spectra of rocaglamide AE (1) and cyclorocaglamide (2). Although they are constitutionally very closely related and possess the same absolute conﬁguration, the CD spectra are nearly mirror-image like! (b) Dioncopeltine A (3) and habropetaline A (4). Changing the substituent at C5 from OH to MeO does not have any signiﬁcant effect on the spectrum.

The use of semiempirical approaches (like the octant rule for saturated ketones) or of the nonempirical exciton chirality method may be a good alternative to derive absolute conﬁgurations from experimental CD curves. The octant rule and the exciton chirality approach are described in detail in other chapters of this volume. However, these methods are again limited to speciﬁc structures or to a detailed knowledge of transitions in the chromophores. The octant rule can only be applied to cyclic saturated ketones (or aldehydes) with known, rigid conformation and is not valid in the presence of an additional stronger chromophore, while the exciton chirality method has a broader range of application. In any case, however, the knowledge about the mutual orientation of the dipole moments of the chromophores is essential: Without knowing the exact and energetically relevant conformation(s) of the investigated structure, it is not possible to unambiguously assign absolute conﬁgurations. Thus, conformational analyses using, for example, DFT methods, are usually applied to give information about the different possible conformations of a new structure and then the exciton chirality method may become applicable [45].

364

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Chiral product of natural or synthetic origin

Experimental UV spectrum

+

Experimental CD spectrum 8

Comparison

exp.

0 8

–8

0

–16

Single UV spectra Starting geometry

Conformational analysis

Overall UV spectrum

Boltzmann weighting Single CD spectra

–24 200

calcd.

–8 250 16 24200 250 300 350

Absolute stereostructure

Comparison UV correction

Overall CD spectrum

Corrected overall CD spectrum

Figure 10.4. Flowchart of a general approach to calculate UV and CD spectra and of the determination of the absolute conﬁguration by the comparison of experimental and computational results.

When completely new compounds with unknown chromophores are analyzed, or when there is a doubt in the applicability of certain helicity rules, the use of quantumchemical calculations is often the only way to determine the absolute stereostructure. This is achieved by comparing the experimental spectra with the ones quantum-chemically predicted for the respective stereoisomers (usually enantiomers). Which of the semiempirical, ab initio, or DFT methods will be the most appropriate has to be carefully decided and will be discussed in other chapters so that only some key facts of the basic, general approach (Figure 10.4) will be mentioned here: The ﬁrst—and mandatory—step in any calculation of CD spectra has to be a solid conformational analysis with suited methods that yield reliable energies for the investigated class of compounds. CD spectra are very sensitive to even slightest conformational changes of the chiral molecule, and it can happen that two conformations with the same absolute conﬁguration give rise to nearly mirror-image like CD curves. For example, the dihedral angle at the biaryl axis of a (P )-1,1 -binaphthyl has a drastic inﬂuence on the CD spectrum of the compound. With an angle between 50◦ and 100◦ a positive exciton couplet in the CD spectrum is produced, whereas an angle above 120◦ will give a negative one, although the absolute conﬁguration of the chiral axis remains the same [45]. The observed experimental CD spectrum is the macroscopic result of the CD spectra originating from all individual molecules in a population—that is, the energy-weighted summation of the single CD spectra of all possible conformations of the measured structure according to their percental occurrence in the equilibrium mixture. The contribution of a single conformer to the overall CD (or UV) spectrum can be calculated based on a Boltzmann statistical weighting of the energies of the conformers found during the conformational analysis. In general, every conformation within an energy range of ∼12 kJ/mol above the global minimum may contribute to a signiﬁcant degree, and so all conformations with an energy above this value can usually be disregarded for the subsequent calculation of the CD and UV spectra [46]. These CD computations will not give rise to full curves, but only to bar spectra (one single value for the energy of each of the excited states). To achieve a result that is optically comparable with the experiment, the single values have to be overlaid with

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

Gaussian-shape functions (sometimes Lorentz curves will be used instead) [47]. All of the single spectra thus obtained will then be summed up, taking into account the Boltzmann statistics of the conformational analysis, ﬁnally providing the full predicted CD and UV curves. In principle, it would be preferable to use the highest-level method available for the computation of the excited states of the chiral molecule. In practice, however, this is not feasible and one has to ﬁnd a compromise between accuracy and computational costs.

10.8. APPLICATION OF THE ‘‘UV SHIFT’’ A simple empirical approximation, the so-called UV shift [46, 48], can sometimes help to keep the calculation time low. It is based on the following considerations: The most common software packages for quantum-chemical calculations will always compute rotational and oscillator strength values at certain excitations in one run. For both values the same method is used to obtain the energy of the excited state and thus the wavelength of the excitation. This means that any systematic error with respect to the wavelength determination will be the same for both spectra. Identiﬁcation of the extent of this error is much easier for the UV than for the CD curve. By comparing the experimental UV spectrum with the calculated one, it is possible to determine the difference between the lambda values of the experimental and the calculated maxima—that is, the “UV shift” This empirical factor can then be applied to the CD spectrum, thus providing a better match between the experimental and the calculated data. If this match is not evident, the chosen calculation method is either unsuited or the absolute stereostructure (maybe even the constitution) for which the computation was done does not correspond to the analyzed compound. The use of the UV shift is, however, only recommended if the UV spectrum is already of sufﬁcient quality, which, in simple words, means that the number of maxima, their relative intensities, and the energy differences between them should be the same in the experimental spectrum as in the computed one. Application of the UV shift often permits to tolerate larger systematic errors—for example, by using smaller basis sets, which can allow for signiﬁcantly reduced computational time.

10.9. HPLC-CD IN PRAXIS In the following, the still underestimated potential of HPLC-CD analysis and its possible application areas will be demonstrated by a detailed description of two representative conﬁgurational assignments of chiral natural products, in particular the determination of the absolute conﬁguration by a combination of HPLC-NMR, HPLC-CD, and quantumchemical calculations. Through these examples the reader will get an idea of how to perform a full conﬁgurational elucidation of chiral substrates by HPLC hyphenation techniques and will gain a feeling about possible difﬁculties encountered during the structural analysis and receive useful advice to overcome them. This provides a general guideline of how to proceed for the full stereochemical assignment of novel-type chiral compounds even from crude mixtures.

10.9.1. Ancistrocladium B—A Conﬁgurationally Semi-Stable, Axially Chiral Biaryl Ancistrocladinium B (5) is a novel-type metabolite isolated from an as yet not fully identiﬁed, possibly new Ancistrocladus species from the rainforest in the Democratic

365

366

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Republic of Congo [49]. It is one of the ﬁrst of the recently discovered N , C -coupled representatives of the naphthylisoquinoline alkaloids. Its typical occurrence as a double peak in the HPLC chromatogram in combination with the observation that all isolation attempts always led to the same 45:55 mixture according to NMR and analytic HPLC indicated the presence of two diastereomers slowly interconverting at room temperature (Figure 10.5). That these were the respective atropo-diastereomers—that is, with different conﬁgurations at the (apparently conﬁgurationally semi-stable) novel N , C axis—became evident from the fact that both had the same conﬁguration at the stereocenter at C3, S , as determined by oxidative degradation [50]. Due to the semi-stable conﬁguration at the N , C axis, all ofﬂine spectra (NMR, CD) always recorded the diastereomeric mixtures, hampering an unambiguous elucidation of the absolute stereostructures with common methods, making ancistrocladinium B an excellent example of how to apply the so-called “analytical triad”, that is, the combination of HPLC-NMR, HPLC-MS, and HPLC-CD. 10.9.1.1. Optimizing the HPLC Separation Conditions. As for all HPLC hyphenation techniques, the ﬁrst and most important step to the unequivocal elucidation of the constitution and the stereostructure of ancistrocladinium B (5) was the elaboration of a reliable method for the full resolution of this compound by HPLC. To get baseline-separated peaks, several columns (normal phase, RP-C18 and -C8 ) and solvents (acetonitrile, methanol, water) were tested. In addition, the pH value of the mobile phase and the temperature of the column had to be optimized to provide a good separation of the two presumed atropo-diastereomers, preferably with short retention times. After intense efforts, the best separation was achieved by using a Symmetry-C18 column (Waters, 4.6 × 250 mm; 5 μm) at 10◦ C with an isocratic solvent system of methanol and water (60:40, acidiﬁed with 0.05% TFA, ﬂow rate: 0.8 mL/min). These conditions yielded a clear baseline separation with short retention times (Figure 10.5), giving one peak at about 17 minutes (Peak A) and a second one at about 19 minutes (Peak B). Still, a preparative separation of the two peaks to give fully pure diastereomers was not possible due to their slow interconversion at room temperature.

LC-UV 231 nm Peak A Peak B

MeO

*

Me OH

OMe

t [min]

N * OMe Me

17

19 LC-CD 235 nm

Me

Ancistrocladinium B (5) * stereogenic elements of initially unknown absolute configuration

Figure 10.5. HPLC-UV and HPLC-CD t [min] 17

19

(on-ﬂow) chromatograms of ancistrocladinium B (5).

367

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

MeO

Me H S OH OMe N 6′

MeO

3

P

OMe Me H

Me H H N M 7′

3 S

Figure 10.6. Online-ROESY NMR

6′ 7′

MeO (P,3S)-5 (Peak A)

correlations diagnostically indicative for the relative conﬁguration of the

OMe Me HO Me (M,3S)-5 (Peak B)

Me

atropo-diastereomers of ancistrocladinium B.

10.9.1.2. Information Obtained by HPLC-MS and HPLC-NMR Measurements. This interconversion was, however, slow enough to permit a complete stereochemical analysis online, right from the peak in the chromatogram, by applying the analytical triad for further structural investigations, using the optimized separation conditions described above. By HPLC-MS, each of the two peaks gave a mass of m/z = 406. HPLC-HRMS(ESI) experiments showed that both compounds were cationic, having the same molecular formula, C25 H28 NO4 + . This again conﬁrmed that ancistrocladinium B was indeed a mixture of two isomers, probably atropo-diastereomers. For both peaks, 1 H, 13 C, COSY, ROESY, HMBC, and HMQC spectra were recorded by online HPLC-NMR measurements, using a Bruker Cryoprobe for higher 13 C sensitivity and a ﬂow insert (CryoFit, Bruker) for HPLC-NMR hyphenation. The two compounds showed nearly identical NMR spectra, except for the fact that some of the signals displayed slightly different chemical shifts. The results corroborated the anticipated constitution of ancistrocladinium B (Figure 10.6); in addition, based on the online-ROESY correlations, even the relative conﬁgurations of the compounds were unambiguously assigned: The faster eluting Peak A displayed a diagnostically signiﬁcant interaction between H7 and the proton at the stereocenter C3, showing these two protons to be both on the same side of the molecule—that is, both up (as drawn in Figure 10.6, left) or both down. The more slowly eluting Peak B, by contrast, had a strong correlation between H7 and the protons of Me3, hinting at the opposite relative conﬁguration axis versus center (Figure 10.6, right). Therefore, the faster eluting atropo-diastereomer of ancistrocladinium B had to be (P , S )- or (M , R)-conﬁgured, while the more slowly eluting one had the (P , R)- or the (M , S )-conﬁguration. Together with the (S )-conﬁguration at C3 as already known from the degradation experiment, the two (3R)-conﬁgured stereoisomers were excluded. 10.9.1.3. HPLC-CD and Quantum-Chemical CD Calculations. To get an independent and unambiguous proof for the above assignments HPLC-CD measurements in combination with quantum-chemical calculations were performed. The full CD spectra of the two peaks were recorded online in the stopped-ﬂow mode (Figure 10.7), and these spectra were nearly mirror-image like. It is known that the chiral axis in biaryls often dominates the CD spectrum and that additional stereocenters usually do have a negligible effect only (for exceptions due to the strong chromophores close to the stereogenic center, see reference 51). Thus, the CD spectra alone would already have provided a clear hint that the two peaks in ancistrocladinium B correspond to atropo-diastereomeric compounds. Assuming that the exciton chirality method (see Chapter 4 in this volume) was valid for ancistrocladinium B, the positive couplet around 350 nm would have predicted the (P )-conﬁguration for Peak A and, vice versa, the (M )-conﬁguration for Peak B with its negative couplet at 350 nm. However, without knowing the exact conformations of

368

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

6

6 exp. Peak A

3 CD [mdeg]

CD [mdeg]

3

0 calcd. for (P,S)

–3

exp. Peak B

0

–3 calcd. for (M,S)

–6 200

–6 250 300 350 wavelength λ [nm]

MeO S

400

250 300 350 wavelength λ [nm]

200

Me + OH OMe N 6′

Me

MeO S

P

MeO

MeO Me

M

6′

MeO

Me Me

(P,3R)-5

Me

6 exp. Peak A

calcd. for (P,R)

exp. Peak B

3 CD [mdeg]

CD [mdeg]

+ OH OMe N 6′ P

MeO

6

3

Me

Me R

Me HO

(M,3R)-5

MeO

MeO

Me R

N+

6′

Me HO

(M,3S)-5

MeO

MeO

N+ M

Me

(P,3 S)-5

400

0

–3

0

–3 calcd. for (M,R) –6

–6 200

250 300 350 wavelength λ [nm]

400

200

250 300 350 wavelength λ [nm]

400

Figure 10.7. Assignment of the absolute conﬁguration to the two—conﬁgurationally semistable—atropo-diastereomers of ancistrocladinium B (5) by comparison of the experimental LC-CD spectra (stopped-ﬂow) of Peak A (left) and Peak B (right) with the spectra calculated for (P, 3S)-5, (M, 3R)-5, (P, 3R)-5, and (M, 3S)-5 by using TDDFT with subsequent UV-shift correction.

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

369

the diastereomers and the orientation of the dipole moments of the two chromophores, this was not a reliable assignment and thus computational calculations were done. To elucidate the absolute conﬁguration of the diastereomers of 5 independently, the CD spectra of all possible conﬁgurations, (P , 3S )-5, (M , 3R)-5, (P , 3R)-5, and (M , 3S )5, were calculated and compared with the experimental results [46]. A conformational analysis was performed using B3LYP/6-31G* [52–55] yielding four relevant conformers for each diastereomer. For the subsequent TDAB3LYP/SVP [56, 57] calculations, solvent effects were taken into consideration by using the COSMO [58] approach, using an epsilon value of 56.52 and a refraction index of 1.33. By comparison of the calculated UV spectra with the experimental curves, a UV shift of 24 nm was determined and applied to the calculated CD curves. The CD spectrum computed for (P , 3S )-5 did ﬁt quite well to the experimental one of Peak A, while the spectrum calculated for the (M , 3R)-enantiomer did not match the measured one, proving that Peak A was (P , 3S )conﬁgured. In the case of Peak B the comparison of the calculated spectrum of (M , 3S )-5 with the experiment showed a good agreement, while the curve calculated for (P , 3R)-5 again did not ﬁt, corroborating the (M , 3S )-conﬁguration of Peak B. 10.9.1.4. Further Investigations: Estimation of the Rotational Barrier. As already mentioned above, the iminium-aryl axis of ancistrocladinium B is conﬁgurationally semi-stable at room temperature. For an estimation of the barrier of rotation around the chiral axis, again HPLC experiments were carried out. For each of the two atropo-diastereomers the isomerization process was monitored by HPLC-UV measurements. The decrease of the diastereomeric excess of freshly puriﬁed fractions enriched in the respective (P )- or (M )-atropisomer was measured at three different temperatures

Peak B Peak A

t [min] 17

19

Isolation of Peak A

Isolation of Peak B

t = 330 min t

Figure 10.8. Determination of the axial isomerization

t = 180 min

t

rates of (P, S)-5 and (M, S)-5 by their chromatographic

t [min]

resolution and subsequent thermal equilibration over time (here exemplarily shown at 65◦ C), monitored by HPLC-UV on an achiral C18 phase.

t = 0 min 17

19 t [min]

17

19

370

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(55◦ C, 65◦ C, and 85◦ C, Figure 10.8), permitting determination of the respective rate constants of the isomerization process. By applying the Eyring equation, the Gibbs free energies of activation at room temperature (25◦ C) were calculated. The experimental values thus obtained were Grot = 105.8 kJ mol−1 for the isomerization from P to M and Grot = 105.7 kJ mol−1 for the conversion of M to P .1 In summary, the structure elucidation of ancistrocladinium B (5) impressively demonstrates the value of HPLC hyphenation techniques in modern natural products chemistry. Besides the establishment of the full constitution and relative conﬁguration by HPLCNMR and HPLC-MS, also the absolute conﬁgurations of the two naturally occurring atropo-diastereomers of 5 were determined. Online HPLC-CD measurements were performed with trace quantities of crude material. Due to the unprecedented structure of the N , C -coupled chiral aryliminium cation, the CD spectra thus obtained could not be interpreted by empirical comparison with other compounds, but inevitably required quantum-chemical CD calculations. This ﬁnally permitted the unambiguous assignment of the full absolute stereostructures of the two diastereomers. Now knowing their elution order, HPLC techniques also served to determine the atropisomerization barriers by monitoring the decrease of the diastereomeric excess of freshly prepared samples enriched in one of the two rotational isomers. The absolute stereostructures of the two atropo-diastereomers of ancistrocladinium B (5) have meanwhile been conﬁrmed by total synthesis [59].

10.9.2. The Absolute Axial Conﬁguration of Knipholone and Knipholone Anthrone Knipholone (6) and knipholone anthrone (7) are well-known representatives of the class of naturally occurring phenylanthraquinones [60]. They are interesting biosynthetically (origin from eight plus four acetate units), pharmacologically (anti-infective and anti-tumoral properties), and, in particular, stereochemically, due to their rotationally hindered and thus conﬁgurationally stable biaryl axis (Figure 10.9) [60]. They have been initially discovered by Steglich, Dagne, and Yenesew in 1984 (knipholone) [61] and 1993 (knipholone anthrone) [62], but the elucidation of their correct absolute conﬁgurations succeeded in 2007 only [63]. Previous assignments using semiempirical calculations had deduced absolute conﬁgurations at the biaryl axis, which were in contrast to those expected from the results of a later stereoselective total synthesis using the lactone method [64, 65]. In nature, knipholone mostly occurs as a scalemic mixture (i.e., enantiomerically enriched, not enantiopure). In this chapter only, the stereostructure of the main enantiomer, (+)knipholone, will be discussed. In 2007 the correct absolute conﬁguration was determined unambiguously and independently by renewed experimental work in combination with quantum-chemical calculations using higher-level methods [63]. The individual challenges arising in the course of these investigations will be described in the following paragraph, again demonstrating the value of the fruitful interplay of HPLC-CD measurements with computational work. 10.9.2.1. Remeasuring the CD Spectra of Knipholone and Knipholone Anthrone. As mentioned above, the initial assignment of the absolute conﬁguration of knipholone (6) and knipholone anthrone (7) was in contrast to the results of the total Note that the two Grot values are not identical, although concerning the same transition state, due to the different energies of the two interconverting atropisomers, which in this case are diastereomers. This is also reﬂected by their experimentally determined isomeric ratio of 45:55 of M : P . 1

371

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

HO

O

HO

OH

O

OH

Me

Me X HO

M

X HO

OH

P OH

Me

Me MeO O (M)-6 X = O (M)-7 X = H2

O OMe (P)-6 X = O (P )-7 X = H2

Figure 10.9. Structures of knipholone (6) and knipholone anthrone (7).

(+)–6

4 Δε [cm2/mol]

Δε [cm2/mol]

5

0

–5

8

from B. capitata

(+)–7 from synthesis from K. foliosa

0 –4

from B. frutescens

–8 200

300 400 wavelength λ [nm]

500

200

300 400 wavelength λ [nm]

500

Figure 10.10. Experimental CD spectra of (+)-knipholone (6, left) and (+)-knipholone anthrone (7, right) from different origins (measured in methanol).

synthesis by applying the lactone concept [64].2 Several independent approaches were followed to clarify the origin of this discrepancy. First of all, the CD spectra of 6 and 7 isolated from different natural sources were remeasured to provide new and reliable experimental data (Figure 10.10). In addition, the inﬂuence of different solvents and solvent mixtures (n-hexane, methanol, acetonitrile:water 60:40) was investigated, since these may have a signiﬁcant effect on a CD spectrum, due to the polarity and the hydrogen-donor or -acceptor properties of the solvent but also due to possible aggregation of the analyte in solution. In the case of knipholone, however, the spectra did not show any signiﬁcant differences, regardless of whether the compounds were isolated from varying natural sources or measured in different solvents (Figure 10.11).

2

Note that in this experimental work the use of an S -conﬁgured catalyst in the lactone-cleavage reaction unexpectedly seemed to give the P -conﬁguration (as erroneously deduced from the initially wrongly attributed absolute (P )-conﬁguration of (+)-knipholone). This seeming contradiction was solved by the new—and correct—assignment of the absolute conﬁguration of (+)-knipholone as being M -conﬁgured in 2007 [64].

372

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

8

(+)–6

(+)–7

in n-hexane in n-hexane Δε [cm2/mol]

Δε [cm2/mol]

5

0

–5

in acetonitrile:water (60:40)

0

–8

300 400 wavelength λ [nm]

in methanol

–4

in methanol 200

4

200

500

in acetonitrile:water (60:40) 300 400 wavelength λ [nm]

500

Figure 10.11. Experimental CD spectra of (+)-knipholone (6, left) and (+)-knipholone anthrone (7, right) in various solvent systems.

8

8

(+) –7

(+) –7 Δε [cm2/mol]

Δε [cm2/mol]

fresh sample 4 0

4 0 –4

–4 partially decomposed

(+)–6 –8

–8 200

300

400

wavelength λ [nm]

500

200

300

400

500

wavelength λ [nm]

Figure 10.12. Experimental CD spectra of freshly prepared (+)-knipholone anthrone (7) and of a partially decomposed sample (left) and comparison of the CD spectra of (+)-knipholone and (+)-knipholone anthrone (right).

The newly recorded CD spectrum of knipholone anthrone (7), by contrast, obviously differed signiﬁcantly from the spectra measured earlier (Figure 10.12): The freshly prepared sample of 7 showed a quite intense CD couplet at about 300 nm with a positive ﬁrst Cotton effect, while this couplet was almost completely absent in the spectrum measured previously [65]. The question why this intense couplet had not been observed during the initial investigations was answered by the following experiment: Keeping the freshly puriﬁed 7 for 2 h in methanol at room temperature under ambient light led to a drastic change in the measured CD spectrum due to an intensity loss of the couplet around 300 nm. This clearly hinted at a partial decomposition of the compound to give a new chiral—and likewise chiroptically active—product, because the whole CD spectrum would just have decreased in intensity if simply racemization had occurred. The loss of only one couplet might be explained by the oxidation of knipholone anthrone to knipholone: While both compounds show a similar negative Cotton effect (CE) around 210 nm, their CEs around 300 nm are opposite. Consequently, these couplets will largely

373

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

1000 O

HO

OH

O

OH

HH HO

P

HO

800 Me O HO

600

Me

P OH

Me

Me 6

400

OH

O OMe

O OMe

7 *

unknown decomposition products

* 200 *

*

* *

0 3

4

5

6

32

7

8

9

10 t [min]

0h

CD [mdeg]

16 6h

24 h

Figure 10.13. HPLC-UV chromatogram of

0

a partially decomposed sample of knipholone anthrone (7) after 2 h at rt and ambient light (top) and full online HPLC-CD spectra of knipholone anthrone

−16 −32 200

300

400

wavelength λ [nm]

500

(7) after 0, 6, and 24 h in the ﬂow cell (Chromolith column, solvent gradient of acetonitrile/water, both with 0.05% TFA).

compensate each other in a mixture of knipholone and knipholone anthrone resulting in a selective decrease of the intensity of the couplet at 300 nm. The above investigations exemplify that a detailed knowledge of the chemical stability of a given compound can be highly important for the measurements and interpretation of CD spectra: One may not only observe a decrease of the intensity of the entire CD curve, as in the case of a (partial) racemization, but also new signals may appear or genuine ones may disappear if new chiral compounds are formed in situ, thus changing the overall measured CD spectrum, which may easily lead to ambiguous or even wrong interpretations. Such falsiﬁcations of CD spectra by new (or existing) chiroptically active impurities can often be overcome by HPLC-CD: As an example Figure 10.13 shows the HPLC-UV chromatogram of partially decomposed knipholone anthrone (7) and full online CD spectra of 7. By HPLC-CD in the stopped-ﬂow mode, the peak corresponding to knipholone anthrone was kept in the ﬂow cell for several hours and the full CD curve was recorded after 0, 6, and 24 h. Since the sample was not exposed to light in the ﬂow cell and air oxygen was largely excluded, no signiﬁcant change of the CD spectrum was observed (Figure 10.13) even after 24 h, in contrast to the earlier presented ofﬂine measurement (cf. Figure 10.12), where decomposition may have signiﬁcantly decreased the couplet at 300 nm already within a time span of 2 h.

374

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

8 exp. for (+)–6

DFT/MRCI calcd. for (P)–7

4 Δε [cm2/mol]

Δε [cm2/mol]

5

0

0 –4

–5

exp. for (+)–7

DFT/MRCI calcd. for (P)–6 –8 200

300 400 wavelength λ [nm]

500

200

300 400 wavelength λ [nm]

500

Figure 10.14. Determination of the absolute conﬁgurations of (+)-knipholone (6) and (+)knipholone anthrone (7) by comparison of the experimental CD spectra with the CD curves calculated by DFT/MRCI.

10.9.2.2. Elucidation of the Absolute Conﬁguration by QuantumMechanical Calculations. According to the exciton chirality method, the “newly found” couplet now also hinted at a (P )-conﬁgured biaryl axis of compound 7, again in contrast to the initial assignment of the absolute conﬁguration as M [65]. Thus, in parallel to the renewed experiments, the computational investigations were rechecked leading to the conclusion that the previous semiempirical approaches had by far not been sufﬁcient to unequivocally determine the absolute conﬁguration 6 and 7. In the case of knipholone anthrone even TDDFT calculations led to ambiguous results and thus the DFT/MRCI approach developed by Grimme [66] was chosen for these phenylanthraquinones. Finally, the absolute conﬁgurations of the two compounds had to be revised: (+)-Knipholone (6) and (+)-knipholone anthrone (7) clearly do have the (P )-conﬁguration (Figure 10.14), which is now in full accordance with the results of the total synthesis by using the lactone concept, and—in the case of 7—also with the attribution of the absolute conﬁguration based on the exciton chirality method applied to the newly measured CD spectrum. 10.9.2.3. Stereochemical Correlation of 2 and 3 by Coelution Experiments. Additional support for the new assignment of the absolute conﬁguration of the two compounds was expected from their stereochemically unambiguous interconversion by reduction of (P )-6 to give—if correctly assigned—(P )-7, and, vice versa, the oxidation of (P )-7 to deliver (P )-6, to be analyzed by chromatography on a chiral phase with online CD coupling (LC-CD). These investigations were, however, hampered by the partial racemization of the compounds under the applied conditions. Still, the enantiomers of the two compounds were easily resolved on a chiral OD-H column. Interestingly, they showed inverted elution orders: Thus, the (P )-enantiomer of knipholone (6) corresponded to the faster eluting peak of the two enantiomers, while for knipholone anthrone (7) the M -conﬁgured enantiomer had the shorter retention time (Figure 10.15). The described results highlight that elution orders of similar, but not identical, compounds do not necessarily provide a hint at their absolute conﬁgurations, because even two so closely related compounds—like 6 and 7—can show opposite elution orders of their identically conﬁgured enantiomers.

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

HO

OH

O

O

HO

Me

Me P

O HO

OH

OH

SnCl2, HOAc Me

Me 7

OH

KOH, air

P

HH HO

375

6

O OMe

O OMe

(b)

(a)

P

P

M LC-UV

LC-UV KOH, air

LC-CD

LC-CD

t [min]

t [min] 25 35 0:100

30 40 69:31 (c)

(d) M

P

P M

LC-UV

LC-UV

Figure 10.15. Proof of the stereochemical

SnCl2, HOAc

identity of (+)-knipholone as P (b), since it was obtained by oxidation of enantiomerically pure

LC-CD

LC-CD

t [min]

t [min] 25 35 48:52

30 40 75:25

knipholone anthrone (100:0 P to M) (a), and, vice versa, enantiomeric resolution of—mainly (P)-conﬁgured—knipholone anthrone (7) (d) obtained by reduction of authentic knipholone (75:25P to M) (c).

An authentic sample of enantiomerically highly pure synthetic (+)-knipholone anthrone (Figure 10.15a) was oxidized to knipholone showing only a ratio of 69:31P (rapid) to M (slow). And, vice versa, an authentic sample of knipholone (6) isolated from B. capitata, which had a ratio of 75:25P to M , was reduced to knipholone anthrone (7), whose now more slowly eluting main peak coeluted with the peak of the pure (P )-enantiomer of 7. The interconversion showed that the more rapidly eluting peak of knipholone anthrone (7) was (M )-conﬁgured, while in the case of knipholone (6) the faster eluting one was P . These results are in agreement with the fact that at the chosen wavelength of 290 nm (P )-conﬁgured 6 shows a positive Cotton effect, while the (P )-enantiomer of 7 displays a negative one! The example reveals that even seemingly marginal structural changes can be accompanied by an inversion of the CD response, at least at a given wavelength λ. Due to the near-racemic character of the knipholone anthrone sample obtained by reduction of (P )-knipholone (Figure 10.15d), additional evidence of the identity of the

376

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

P

P

M

P

t [min]

t [min]

t [min] 30 40

30 40

Near-racemate 45:55

Pure + ‘racemate’

30 40 Pure Knipholone anthrone

LC-UV

M

Figure 10.16. Spiking LC-CD t [min]

t [min] 30 40

t [min] 30 40

30 40

experiments to prove the identity of peaks, exemplarily for knipholone anthrone (7) of different enantiomeric ratios from different sources.

respective peaks, and thus of the reversed HPLC behavior of 7 in comparison to 6, was acquired by a spiking experiment by adding enantiomerically pure knipholone anthrone to that near-racemic sample of 7 with subsequent chromatographical analysis by using HPLC-UV and HPLC-CD (Figure 10.16). This experiment showed that after addition of pure (P )-conﬁgured knipholone anthrone to the racemate, the second-eluting peak increased in intensity. This clearly conﬁrmed, once again, that the elution order for the diastereomers of 7 is inversed as compared to the chromatographic behavior of compound 6. 10.9.2.4. Knipholone and Knipholone Anthrone: Instructive Examples of the Manifold Aspects of HPLC-CD. The investigations on knipholone (6) and knipholone anthrone (7) exemplify several advanced aspects of HPLC-CD hyphenation: A general disadvantage might be that the solvent used for the HPLC-CD measurement may inﬂuence the experimental CD curves signiﬁcantly, although such effects were shown to be negligible in the case of 6 and 7. In ofﬂine CD measurements, such aggregation phenomena can be evaluated by simply using different solvents. In online CD investigations, by contrast, hints at possible spectral changes originating from aggregated species can be obtained from a dilution series. Such a “dilution experiment” can sometimes be achieved by simply measuring the CD spectrum of the investigated compound at different positions of its peak in the UV-monitored chromatogram (e.g., left versus right slope). Another useful lesson one can learn from the racemate resolution and the online CD analysis of knipholone (6) and knipholone anthrone (7) is that even slight structural changes may lead to an inversion of the elution order of the respective enantiomers. Thus, the absolute conﬁguration of structurally related analogues, and, in particular, of unknown compounds can usually not be determined for sure by simple comparison of relative retention times. Beyond the elucidation of the full absolute conﬁguration by LC-CD in combination with empirical and nonempirical methods or quantum-chemical calculations, spiking (= coelution) experiments can help to establish the elution order of stereoisomers of known substances if stereochemically pure or enriched authentic

377

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

material (e.g., from a total synthesis) is available. The chemical stability of the analyte has to be considered when performing CD investigations since chiral decomposition products can largely contribute to the overall CD response, thus possibly leading to an ambiguous or even wrong assignment of absolute conﬁgurations. Again online HPLC-CD measurements in the stopped-ﬂow mode can be helpful to overcome such problems by selectively measuring only the correct, still structurally intact analyte (even in a mixture of a plethora of other CD-active products) and, in addition, by minimizing reactions of the analyte by exclusion of light and air oxygen within the HPLC-CD ﬂow cell.

10.10. FURTHER EXAMPLES The value of the method is emphasized by a broad variety of further stereochemically intriguing examples from most different classes of compounds, with stereogenic centers or elements of axial or planar chirality, whose absolute conﬁgurations were established by HPLC-CD in combination with quantum-chemical CD calculations and/or by applying empirical or nonempirical rules. A few selected examples are shown in Figure 10.17, among them the 3,8 -linked biﬂavonoid 14 from Gnidia involucrate [67], whose absolute conﬁguration was determined using Gafﬁeld’s isochroman helicity rule [68, 69], and ﬂavanthrin (8), a 9,10-dihydrophenanthrene dimer isolated from Pholidota chinensis [70], which was structurally elucidated by HPLC-CD in combination with quantum-mechanical

Me

SMe CN

R

NH OH Me Phylline (9) R

OMe HO

OH M

OH Flavanthrin (8)

HO

A synthetic quateraryl, 10

MeO calcd. S NH N H CCl3 (S)-TaClo (11)

OH

calcd.

O R S

t

Ph

Ph

HO Ph

N

N

N

Ph

Ph

β,β '-Coupled bisporphyrins, 12

N

Zn

N N

Ph

R

P

P

M

R

H

R

R

Ph

P

H

H OH O O

N Zn

OMe

exp.

exp.

N

Me OMe

M

HO R

Ph R = CH2OH

OH R

O

OH

OH A biflavonoid, 14

A configurationally semi-stable bi[10]paracyclophane, 13

Figure 10.17. Selection of structurally most diverse compounds with one or more stereogenic centers, axes, and/or elements of planar chirality from different research groups, stereochemically investigated by HPLC-CD hyphenation techniques.

378

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

CD calculations. Further conﬁgurational assignments succeeded with synthetic quateraryls like 10 [71], the neurotoxin TaClo (11) [72], and the anti-malarial drug artemisinin [73], and all were investigated through the hyphenation of HPLC and CD. Some additional selected applications of HPLC-CD hyphenation will be highlighted in the following pages, involving online CD investigations on axially chiral bisporphyrins like 12, likewise synthetic paracyclophanes of type 13, and the natural alkaloid phylline (9) as illustrative examples.

10.10.1. The Stereostructure of Intrinsically Axially Chiral β,β -Bisporphyrins The utility of the online CD analysis in combination with quantum-chemical CD calculations for the fast conﬁgurational assignment of chiral compounds has also been succesfully applied to unnatural, merely synthetic compounds possessing unprecedented stereostructures—for example, to the stereochemical characterization of the ﬁrst intrinsically axially chiral bisporphyrins with a rotationally hindered direct β,β -linkage [74, 75]. Among a whole series of β,β -linked porphyrin dimers of type 12, several racemic free-base representatives could not be resolved, although a variety of different chiral HPLC phases were tried under various conditions [75]. By contrast, a couple of fully metalated dimeric porphyrins like, for example, rac-12a, rac-12b, and rac-12c (Figure 10.14) gave a clear baseline separation of the respective atropo-enantiomers at room temperature after extensive optimization of the separation conditions (Chirex-3010 column, n-hexane/CH2 Cl2 60/40). Note that for the online LC-CD measurements the CD detection wavelength was set to 435 nm, that is, red-shifted by about 10 nm as compared to the UV maximum of the racemate (as outlined above), since the CD spectrum was expected to arise from an exciton coupling of the two adjacent identical chromophores. The full online CD spectra were perfectly mirror-imaged and showed a positive ﬁrst Cotton effect around 450 nm for the faster enantiomer (Peak A) and a negative one for the more slowly (Peak B) eluting atropisomer (Figure 10.18). The absolute conﬁgurations at the biaryl axis of the enantiomers (P )-12a and (M )-12a were established by HPLCCD experiments in the stopped-ﬂow mode in combination with quantum-chemical CD calculations (ZINDO/S-CI//BLYP-D/SVP), revealing the (P )-atropo-enantiomer to be the faster eluting one. As expected, the CD spectra of 12b and 12c (differing from 12a in the metalation pattern and/or the meso-substituents) were strongly related to those of 12a, permitting, in this case, conﬁgurational assignment by comparison of these curves with the ones calculated for the enantiomers of the parent dimer 12a, (P )-12a, and (M )-12a. This clearly evidenced that the chromatographically faster atropo-enantiomers of both, 12b and 12c, were (P )-conﬁgured and that, consequently, the slower ones had the M conﬁguration at the porphyrin–porphyrin axis, here proving that the substitution pattern and, in this case, also the type of the central metals may have only minor effects on the overall CD spectrum. On the other hand, the fact that the palladium(II) and copper(II) derivatives of rac12a (structures not shown) revealed an inverse elution order (with the M -enantiomer being the faster eluting one), as compared to that of rac-12a, rac-12b, and rac-12c, again shows that even small structural differences can, within the same class of compounds, lead to substantial changes of the chromatographical behavior, highlighting that a conﬁguration assignment cannot be based on elution orders alone, but must be assisted by, for example, online CD investigations.

379

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

LC-UV 425 nm

Resolution on a chiral HPLC phase (Chirex-3010®, rac-12a, CH Cl /n-hexane) 2 2 rac-12b, or rac-12c

Peak A Peak B

6

LC-CD t [min] 435 nm A

300

B

exp. for 12a 12b 12c

calcd. for (P)-12a

400

500

Online-CD (stopped flow)

6

CD [mdeg]

CD [mdeg]

Online-CD (stopped flow) 40 20 0 –20 –40 –60 –80

8 t [min]

8 t [min]

80 60 40 20 0 –20 –40

exp. for 12a 12b 12c calcd. for M-12a

600

300

wavelength λ [nm] A

Ar1 N

Ar1 N

Ar1

N

M1

Ar1 N

P

Ar2 Ar1 (P )-12a-c

Ar1

Ar2 N 2 N M N N

Ar2

400

500

600

wavelength λ [nm]

Ar2

N

N M1 N

B

N

Ar1 Ar2

Ar2

N 2N M N N

M Ar1

Ar2

Ar2

(M )-12a-c

Figure 10.18. Stereochemical characterization of the axially chiral β,β -bisporphyrins 12a–c by online HPLC-CD measurements on a chiral phase and comparison of the experimental spectra with those obtained by quantum-chemical CD calculations [rac-12a (M1 = M2 = Zn, Ar1 = Ar2 = phenyl), rac-12b (M1 = Zn, M2 = Ni, Ar1 = Ar2 = Phenyl), rac-12c (M1 = M2 = Zn, Ar1 = Ar2 = 4methoxyphenyl)].

10.10.2. BI [10] Paracyclophane: An Axially Chiral, Yet Conﬁgurationally Semi-Stable Biphenyl Meso-compounds are constitutionally symmetric molecules that do possess pairs of stereogenic elements, but of opposite conﬁgurations each. They are, thus, achiral—at least on the time average: Even if possessing chiral conformations (maybe even exclusively), these may rapidly interconvert in ﬂexible systems, so that the molecule will, macroscopically, appear as achiral above a certain temperature [76–78]. The bi[10]paracyclophane 13 (Figure 10.17) was synthesized by Tochtermann and co-workers [79, 80] since it was expected to constitute an unprecedented borderline case between an achiral meso-compound and an axially and planar chiral compound. It has two elements of planar chirality, whose absolute conﬁgurations were known to be opposite to each other from its synthesis so that one-half of the molecule was (pP )- and the other one (pM )-conﬁgured. In this remarkable molecule, rotation at the central axis will, despite the presence of the two planar-chiral elements, lead to enantiomers, (pM , aM , pP ) versus (pM , aP , pP ), not diastereomers. Different from the situation with similar, more hindered

380

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

HPLC-UV at 210 nm

HPLC-CD at 210 nm

t [min] 8 10 12 14

HPLC-CD of Peak A

HPLC-CD of Peak B 10

12

calcd. for (pP,aM,pM)-13

6

Δε [cm2/mol]

8

exp. Peak B

4

2 0 –2

0 –4 –8

exp. Peak A –6

–12 200

250

300

wavelength λ [nm]

350

–10 200

calcd. for (pP,aP,pM)-13 250

300

350

wavelength λ [nm]

Figure 10.19. Resolution of an enantiomeric mixture of 13 by HPLC and determination of the absolute conﬁguration by comparison of online CD spectra with the calculated CD curves.

(and thus C1 symmetric) analogues investigated earlier [80], the axis connecting these two moieties was conﬁgurationally semi-stable and, as a consequence, 13 should be a genuine meso-compound at higher and a “simple” C1 -symmetric compound at lower temperature. Due to the observed atropisomeric interconversion at room temperature, LC-CD was the method of choice for the investigation of the stereochemical properties of 13. Apparently due to the low rotational barrier at the central axis, the resolution was most unsatisfactory at ambient temperature, making it necessary to perform the separation at lower temperature: The best results were obtained using a Chiralcel ODRH column (Daicel) at 5◦ C and with acetonitrile/water (68:32) as the mobile phase. That the two peaks thus observed (Figure 10.19) indeed corresponded to the expected atropo-enantiomers of 13 was evidenced by their online CD analysis, which resulted in almost opposite CD spectra. It is noteworthy that although the two peaks of the enantiomeric mixture were not baseline separated, it was possible to obtain CD spectra of good quality. One reason is that, different from CD spectra of diastereomers, those of enantiomers are fully opposite and retain their qualitative appearance even if the sample is not enantiomerically pure (see also Section 10.4). Therefore, even enantiomers with nearly identical retention times, which thus give only one—seemingly unresolved—LCUV peak, may sometimes give full CD spectra by using the HPLC-CD method [34, 35].

381

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

The chromatographic separation of the two enantiomers of 13 and their online CD measurement permitted assignment of their absolute conﬁgurations by quantum-chemical CD calculations (CNDO-S-CI//AM1 method) and comparison of the computed CD spectra with the experimental curves. As can be seen in Figure 10.19, the experimental CD curve of the more rapidly eluting Peak A showed a good agreement with the spectrum calculated for the (pP , aM , pM )-enantiomer of 13, while that of the slower Peak B matched well with the CD spectrum predicted for (pP ,aP ,pM ), thus permitting an unambiguous assignment of the two peaks, A and B, to the corresponding enantiomers. As another powerful application of the LC-CD hyphenation, the decrease of the CD curve of Peak A was monitored directly after resolution for the determination of the half-life (t1/2 ) of the racemization process at the biaryl axis of 13. Because of the relatively unstable axial conﬁguration and, therefore, fast vanishing of the CD effect, the time for scanning the CD spectrum had to be reduced dramatically. This was achieved by minimizing the spectral width from 200–230 nm down to 30 nm. On the basis of these experiments, a t1/2 of about 70 s at room temperature was roughly estimated, which ﬁtted quite well with the value obtained from previous NMR experiments [79].

10.10.3. Phylline, Structure Elucidation Directly from the Crude Extract Phylline (9) is a nice example of the application of the analytical triad LC-MS/MS-NMRCD. The HPLC-UV chromatogram of an extract of the rare tropical liana Habropetalum dawei from Sierra Leone showed several peaks (Figure 10.20, left), some of them corresponding to known compounds such as dioncophylline A, but also hinting at the presence

Extract of H. dawei LC-UV at 266 nm

mV 30

new compound

20

mV positive CD effect

30

LC-CD at 266 nm

20 10

x2

10

0

0 0

10

20

30

t [min]

0

10

20

30

t [min]

LC-MS/MS-NMR Me trans NH OH

Me

Constitution, relative configuration

Me R 3 R NH 1

OH Me Natural Phylline (9)

Me S S NH

OH Me Synthetic ent-Phylline (1S,3S-9)

Absolute 1R,3R-configuration

Figure 10.20. Structure elucidation of natural phylline (9) directly from the crude extract by using the analytical triad LC-MS/MS-NMR-CD.

382

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

of new natural products. According to HPLC-MS/MS and HPLC-NMR measurements, one of them was the naphthalene-devoid tetrahydroisoquinoline part of dioncophylline A, thus later simply named phylline [7]. Online ROESY experiments evidenced that the two methyl groups were trans to each other, so that natural phylline had to possess either the 1R, 3R- or the 1S , 3S -conﬁguration. From a synthetic S , S -conﬁgured reference (ent-9) [81], it was known that the R, R-enantiomer should have a positive Cotton effect at 266 nm. Thus, already from the HPLC-CD chromatogram as detected at a single wavelength (266 nm), the 1R, 3R-conﬁguration was deduced for the natural phylline since the corresponding peak showed a clear positive signal in the LC-CD trace (Figure 10.20, right).

10.11. CONCLUSIONS Due to the nowadays commercial availability and the reduced costs of HPLC-CD systems, the hyphenation of HPLC with CD becomes more and more interesting for industrial and academic research. While LC-CD is still mainly used for quality management in the pharmaceutical industry, it has several further substantial advantages in other research areas. In combination with HPLC-MSn and HPLC-NMR, it is a time- and work-saving method for the structure elucidation of novel natural products. Crude extracts of plants, bacteria, or fungi can be directly analyzed without the need of previous isolation work. Thus, already known products will easily be disregarded and no further time and money are wasted for unnecessary isolation work and one can focus on truly new substances or even new classes of compounds. In addition, chemically or conﬁgurationally unstable or semi-stable substances can more easily—that is, directly—be analyzed by these methods. Yet, CD spectroscopy is, in general, still largely neglected in chemical and biochemical studies and we, therefore, plead for the urgently needed change of mind here. Especially by the help of quantum-chemical calculations, it is nowadays quite easy to interpret the CD spectra of novel compounds, even with unknown, unprecedented chromophores. This has further enhanced the value of CD as an efﬁcient tool for the elucidation of absolute conﬁgurations, which is true not only for ECD but also for vibrational circular dichroism (VCD). Unfortunately, due to detection limits and problems with eluent absorption, it is still not possible to hyphenate HPLC with VCD measurements, which would be a very desirable additional tool for the determination of absolute conﬁgurations of new structures, especially for substances that lack suitable UV-active chromophores. Another promising combination of techniques could be HPLC-ORD [82–84], but here are again some problems that have to be solved before. The quality of these measurements in many cases is not sufﬁcient (i.e., because of the still too high detection limit). An additional problem is the unknown concentrations of substances during HPLC runs. Of course, there is the possibility to measure HPLC-CD instead of HPLC-ORD, and convert the spectra by the Kramers–Kronig transformation, but without knowing the concentrations the results are not very helpful. Furthermore, the rotatory power α is known to be strongly inﬂuenced by the eluent composition, which hampers applications of HPLC-ORD hyphenation [85].

10.12. SOFTWARE RECOMMENDATIONS There are several possibilities to analyze and visualize the results of experiments and calculations in this ﬁeld. Normally, every software delivered from the vendor for HPLC

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

systems or spectrometers is capable of exporting chromatograms or spectra as pure ASCII ﬁles (often with the *.txt ﬁle ending). These ﬁles can be imported either into commercially available software or can be processed with free software. Among the numerous commercially available software for the Windows® OS, the authors only have experience with Excel from Microsoft® and Origin from OriginLab® and thus cannot give a full overview over all available program packages. Comparing Excel and Origin, we would recommend Origin since its main focus lies on the data analysis of work of scientists and engineers. Alternatively, in the following a few examples of freely available software will be given, showing that the preparation of publishable graphics does not necessary require expensive software tools. For the processing of statistical or scientiﬁc results Gnuplot (http://www.gnuplot.info) is one of the best-known free tools and is available for several platforms, including Linux and Windows. Unfortunately, it is not an easy-to-use software, because it is command-line driven, but there are third-party tools that make use of the high-quality plotting capabilities of Gnuplot. Knowing that experimental data can be plotted with Gnuplot, there is still the need to get spectra from the output of quantum-chemical calculations in a format that is usable for plotting software. Thus, one needs a program that can do a Gauss or Lorentz curve generation and afterwards gives readable ﬁles. The quantum-chemical package ORCA (http://www.thch.uni-bonn.de/tc/orca) delivers a tool (orca_mapspc) to generate these curves and to produce ﬁles that can be used for further plotting. Using Gaussview, curves of UV and CD spectra can be obtained from Gaussian calculations (http://www.gaussian.com). Gaussum (http://gausssum.sourceforge.net) or Gabedit (http://gabedit.sourceforge.net) are two examples that can do this work as well and will then export the curves as text ﬁles that can be used for plotting with other software. Both are available for different platforms (e.g., Windows or Linux) and can handle the results of Gaussian, ADF (http://www.scm.com), and GAMESS (http://www.msg.chem.iastate.edu/gamess). In addition, Gabedit can in parts read the results of ORCA and QChem (http://www.q-chem.com) calculations. However, the results here have to be prepared “manually” for Gnuplot, and thus it is quite laborious to get publishable graphics of comparisons of experimental and calculated UV or CD curves. An alternative is the freely available software SpecDis as developed in our lab (http://www-organik.chemie.uni-wuerzburg.de/lehrstuehlearbeitskreise/bringmann/ specdis). It can directly compare results from UV/CD calculations performed with Gaussian, TURBOMOLE (http://www.turbomole.com), ORCA, or DFT/MRCI with experimental results in the ASCII format. The only drawback of SpecDis is that it is currently only available for Windows. However, it can be used to determine a UV shift or to do a Boltzmann weighting, and it generates ﬁles that can be processed with Gnuplot to get publishable graphics in the eps format. SpecDis can also handle chromatogram data ﬁles (in ASCII format). All graphics in this chapter that show spectra or chromatograms were initially prepared with SpecDis and Gnuplot.

REFERENCES 1. A. Gergely, J. Pharm. Biomed. Anal . 1989, 7 , 523–541. 2. A. F. Drake, J. M. Grould, S. F. Mason, J. Chromatogr. A 1980, 202 , 239–245. 3. P. Salvadori, C. Rosini, C. Bertucci, J. Org. Chem. 1984, 49 , 5050–5054.

383

384

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

4. G. Brandl, F. Kastner, R. Fritsch, H. Zinner, A. Mannschreck, Monatsh. Chem. 1992, 123 , 1059–1069. 5. A. Mannschreck, Chirality 1992, 4 , 163–169. 6. G. Bringmann, C. G¨unther, J. Schlauer, M. R¨uckert, Anal. Chem. 1998, 70 , 2805–2811. 7. G. Bringmann, K. Messer, M. Wohlfarth, J. Kraus, K. Dumbuya, M. R¨uckert, Anal. Chem. 1999, 71 , 2678–2686. 8. G. Bringmann, K. Messer, K. Wolf, J. M¨uhlbacher, M. Gr¨une, R. Brun, A. M. Louis, Phytochemistry 2002, 60 , 389–397. 9. G. Bringmann, M. Wohlfarth, H. Rischer, J. Schlauer, R. Brun, Phytochemistry 2002, 61 , 195–204. 10. K. Iwasa, W. Cui, M. Sugiura, A. Takeuchi, M. Moriyasu, K. Takeda, J. Nat. Prod . 2005, 68 , 992–1000. 11. W. Cui, K. Iwasa, M. Sugiura, A. Takeuchi, C. Tode, Y. Nishiyama, M. Moriyasu, H. Tokuda, K. Takeda, J. Nat. Prod . 2007, 70 , 1771–1778. 12. K. Iwasa, W. Cui, T. Takahashi, Y. Nishiyama, M. Kamigauchi, J. Koyama, A. Takeuchi, M. Moriyasu, K. Takeda, J. Nat. Prod . 2010, 73 , 115–122. 13. Y. Zhang, D.-R. Wu, D. B. Wang-Iverson, A. A. Tymiak, Drug Discov. Today 2005, 10 , 571–577. 14. O. McConnell, A. B. II, C. Balibar, N. Byrne, Y. Cai, G. Carter, M. Chlenov, L. Di, K. Fan, I. Goljer, Y. He, D. Herold, M. Kagan, E. Kerns, F. Koehn, C. Kraml, V. Marathias, B. Marquez, L. McDonald, L. Nogle, C. Petucci, G. Schlingmann, G. Tawa, M. Tischler, R. T. Williamson, A. Sutherland, W. Watts, M. Young, M.-Y. Zhang, Y. Zhang, D. Zhou, D. Ho, Chirality 2007, 19 , 658–682. 15. A. M. Krstulovic, Chiral Separations by HPLC: Applications to Pharmaceutical Compounds, Ellis Horwood, Chichester, 1989. 16. K. K. Unger, R. Ditz, E. Machtejevas, R. Skudas, Angew. Chem., Int. Ed . 2010, 49 , 2300–2312. 17. D. E. Drayer, Clin. Pharmacol. Ther. 1986, 40 , 125–133. 18. T. Andersson, Clin. Pharmacol . 2004, 43 , 279–285. 19. O. Zerbinati, R. Aigotti, P. G. Daniele, J. Chromatogr. A 1994, 671 , 281–285. 20. P. Salvadori, C. Bertucci, G. Ascoli, G. Uccello-Barretta, E. Rossi, Chirality 1997, 9 , 495–505. 21. H. Jiang, X. Huang, K. Nakanishi, N. Berova, Tetrahedron Lett. 1999, 40 , 7645–7649. 22. A. Kawamura, N. Berova, V. Dirsch, A. Mangoni, K. Nakanishi, G. Schwartz, A. Bielawska, Y. Hannun, I. Kitagawa, Bioorg. Med. Chem. 1996, 4 , 1035–1043. 23. Z. Rosenzweig, E. S. Yeung, Appl. Spectrosc. 1993, 47 , 2017–2021. 24. R. E. Synovec, E. S. Yeung, Anal. Chem. 1985, 57 , 2606–2610. 25. T. Nehira, C. A. Parish, S. Jockusch, N. J. Turro, K. Nakanishi, N. Berova, J. Am. Chem. Soc. 1999, 121 , 8681–8691. 26. D. H. Turner, I. Tinoco, Jr., M. Maestre, J. Am. Chem. Soc. 1974, 96 , 4340–4342. 27. I. Tinoco, Jr., D. H. Turner, J. Am. Chem. Soc. 1976, 98 , 6453–6456. 28. T. J. Ward, B. A. Baker, Anal. Chem. 2008, 80 , 4363–4372. 29. A. Le´on, A. Olives, M. Mart´ın, B. del Castillo, J. Inclusion Phenom. Macrocyclic Chem. 2007, 57 , 577–583. 30. G. Subramanian, Chiral Separation Techniques, 2nd ed., Wiley-VCH, Weinheim, 2001. 31. J. J. Kirkland, Modern Practice of Liquid Chromatography, Wiley, New York, 1971. 32. C. Horv´ath, W. R. Melander, J. Chromatogr. Libr. 1983, 22A, 27–135. 33. D. Song, J. Wang, J. Pharm. Biomed. Anal . 2003, 32 , 1105–1112.

T H E O N L I N E S T E R E O C H E M I C A L A N A LY S I S O F C H I R A L C O M P O U N D S

34. P. Salvadori, C. Bertucci, C. Rosini, Chirality 1991, 3 , 376–385. 35. G. Bringmann, S. Tasler, H. Endress, J. Kraus, K. Messer, M. Wohlfarth, W. Lobin, J. Am. Chem. Soc. 2001, 123 , 2703–2711. 36. T. Tokunaga, M. Okamoto, K. Tanaka, C. Tode, M. Sugiura, Anal. Chem. 2010, 82 , 4293–4297. 37. G. Liu, D. M. Goodall, A. T. Hunter, P. R. Massey, Chirality 1994, 6 , 290–294. 38. A. Berthod, S. C. Chang, D. W. Armstrong, Anal. Chem. 1992, 64 , 395–404. 39. Kaye & Laby Online, Tables of Physical & Chemical Constants. 2.1.4 Hygrometry. Version 1.0 (2005), www.kayelaby.npl.co.uk. 40. C. Bertucci, E. Domenici, P. Salvadori, J. Pharm. Biomed. Anal . 1990, 8 , 843–846. 41. P. Horv´ath, A. Gergely, B. Nosz´al, Talanta 1997, 44 , 1479–1485. 42. P. Salvadori, L. Di Bari, G. Pescitelli, in Circ. Dichroism, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, pp. 797–817. 43. G. Bringmann, J. M¨uhlbacher, K. Messer, M. Dreyer, R. Ebel, B. W. Nugroho, V. Wray, P. Proksch, J. Nat. Prod . 2003, 66 , 80–85. 44. U. Kiehne, T. Bruhn, G. Schnakenburg, R. Fr¨ohlich, G. Bringmann, A. L¨utzen, Chem. Eur. J . 2008, 14 , 4246–4255. 45. N. Berova, L. Di Bari, G. Pescitelli, Chem. Soc. Rev . 2007, 36 , 914–931. 46. G. Bringmann, T. Bruhn, K. Maksimenka, Y. Hemberger, Eur. J. Org. Chem. 2009, 2717–2727. 47. P. J. Stephens, N. Harada, Chirality 2010, 22 , 229–233. 48. G. Bringmann, S. Busemann, in Natural Product Analysis: Chromatography, Spectroscopy, Biological Testing, P. Schreier, M. Herderich, H.-U. Humpf, and W. Schwab, eds., Vieweg, Wiesbaden, 1998, pp. 195–211. 49. G. Bringmann, I. Kajahn, M. Reichert, S. E. H. Pedersen, J. H. Faber, T. Gulder, R. Brun, S. B. Christensen, A. Ponte-Sucre, H. Moll, G. Heubl, V. Mudogo, J. Org. Chem. 2006, 71 , 9348–9356. 50. G. Bringmann, R. God, M. Schaeffer, Phytochemistry 1996, 43 , 1393–1403. 51. F. Bracher, W. J. Eisenreich, J. M¨uhlbacher, M. Dreyer, G. Bringmann, J. Org. Chem. 2004, 69 , 8602–8608. 52. A. D. Becke, J. Chem. Phys. 1993, 98 , 1372–1377. 53. C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37 , 785–789. 54. P. C. Hariharan, J. A. Pople, Theor. Chim. Acta 1973, 28 , 213–222. 55. M. M. Francl, W. J. Pietro, W. J. Hehre, J. S. Binkley, M. S. Gordon, D. J. DeFrees, J. A. Pople, J. Chem. Phys. 1982, 77 , 3654–3665. 56. A. Dreuw, M. Head-Gordon, Chem. Rev . 2005, 105 , 4009–4037. 57. A. Sch¨afer, H. Horn, R. Ahlrichs, J. Chem. Phys. 1992, 97 , 2571–2577. 58. A. Klamt, COSMO-RS: From Quantum Chemistry to Fluid Phase Thermodynamics and Drug Design, Elsevier, Amsterdam, 2005. 59. G. Bringmann, T. Gulder, B. Hertlein, Y. Hemberger, F. Meyer, J. Am. Chem. Soc. 2010, 132 , 1151–1158. 60. G. Bringmann, J. Mutanyatta-Comar, M. Knauer, B. M. Abegaz, Nat. Prod. Rep. 2008, 25 , 696–718. 61. E. Dagne, W. Steglich, Phytochemistry 1984, 23 , 1729–1731. 62. E. Dagne, A. Yenesew, Phytochemistry 1993, 34 , 1440–1441. 63. G. Bringmann, K. Maksimenka, J. Mutanyatta-Comar, M. Knauer, T. Bruhn, Tetrahedron 2007, 63 , 9810–9824. 64. G. Bringmann, D. Menche, Angew. Chem., Int. Ed . 2001, 40 , 1687–1690.

385

386

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

65. G. Bringmann, J. Kraus, D. Menche, K. Messer, Tetrahedron 1999, 55 , 7563–7572. 66. S. Grimme, M. Waletzke, J. Chem. Phys. 1999, 111 , 5645–5655. 67. J. Ferrari, C. Terreaux, T. Kurt´an, A. Szikszai-Kiss, S. Antus, Jerome D. Msonthi, K. Hostettmann, Helv. Chim. Acta 2003, 86 , 2768–2778. 68. W. Gafﬁeld, Tetrahedron 1970, 26 , 4093–4108. 69. G. Kerti, T. Kurt´an, T.-Z. Illy´es, K. E. K¨ov´er, S. S´olyom, G. Pescitelli, N. Fujioka, N. Berova, S. Antus, Eur. J. Org. Chem. 2007, 2007 , 296–305. 70. S. Yao, C.-P. Tang, Y. Ye, T. Kurt´an, A. Kiss-Szikszai, S. Antus, G. Pescitelli, P. Salvadori, K. Krohn, Tetrahedron: Asymmetry 2008, 19 , 2007–2014. 71. A. Goel, F. V. Singh, V. Kumar, M. Reichert, T. A. M. Gulder, G. Bringmann, J. Org. Chem. 2007, 72 , 7765–7768. 72. G. Bringmann, D. Feineis, R. God, K. Maksimenka, J. M¨uhlbacher, K. Messer, M. M¨unchbach, K.-P. Gulden, E.-M. Peters, K. Peters, Tetrahedron 2004, 60 , 8143–8151. 73. J. Chen, J. Gu, R. Zhao, R. Dai, J. Wang, Chromatographia 2009, 69 , 361–363. 74. G. Bringmann, S. R¨udenauer, D. C. G. G¨otz, T. A. M. Gulder, M. Reichert, Org. Lett. 2006, 8 , 4743–4746. 75. G. Bringmann, D. C. G. G¨otz, T. A. M. Gulder, T. H. Gehrke, T. Bruhn, T. Kupfer, K. Radacki, H. Braunschweig, A. Heckmann, C. Lambert, J. Am. Chem. Soc. 2008, 130 , 17812–17825. 76. K. Mislow, Introduction to Stereochemistry, W.A. Benjamin, New York, 1966. 77. E. L. Eliel, S. H. Wilen, L. N. Mander, Stereochemistry of Organic Compounds, Wiley, New York, 1994. 78. R. W. Hoffmann, Angew. Chem., Int. Ed . 2003, 42 , 1096–1109. 79. G. Bringmann, T. A. M. Gulder, K. Maksimenka, D. Kuckling, W. Tochtermann, Tetrahedron 2005, 61 , 7241–7246. 80. W. Tochtermann, D. Kuckling, C. Meints, J. Kraus, G. Bringmann, Tetrahedron 2003, 59 , 7791–7801. 81. G. Bringmann, R. Weirich, H. Reuscher, J. R. Jansen, L. Kinzinger, T. Ortmann, Liebigs Ann. Chem. 1993, 877–888. 82. D. M. Goodall, Trends Anal. Chem. 1993, 12 , 177–184. 83. Z. Wu, D. M. Goodall, D. K. Lloyd, J. Pharm. Biomed. Anal . 1990, 8 , 357–364. 84. D. K. Lloyd, D. M. Goodall, Chirality 1989, 1 , 251–264. 85. J. L. Sims, J. A. Carreira, R. A. Bragg, Chirality 1997, 9 , 122–125.

11 DETERMINATION OF THE STRUCTURES OF CHIRAL NATURAL PRODUCTS USING VIBRATIONAL CIRCULAR DICHROISM Prasad L. Polavarapu

11.1. INTRODUCTION Circular dichroism in electronic transitions, referred to as electronic circular dichroism (abbreviated as ECD, and most often simply as CD), has dominated the chemical sciences in much of the twentieth century for chiral molecular structural determination [1]. The measurement of circular dichroism associated with molecular vibrational transitions is referred to as vibrational circular dichroism (VCD).1 VCD is supported by all chiral molecules, just as ECD. However, the number of molecular vibrations that can be studied in the accessible infrared region is far greater than the number of electronic transitions that can be studied in the accessible ultraviolet–visible region. Furthermore, vibrational transitions are expected to be more sensitive than electronic transitions to the conformational details of molecules. Therefore the information content that is derived through VCD spectra is anticipated to be much more detailed. Advances in quantum chemical methods (see Chapter 24 by Kenneth Ruud in Volume 1) provided reliable approaches for predicting the VCD spectra of chiral molecules [2]. Subsequent development of quantum chemical software packages [3], for calculating the VCD spectra, provided a convenient means to analyze the experimental VCD spectra. For chiral compounds of known absolute conﬁgurations, the quantum chemically predicted VCD spectra provided impressive agreements with corresponding experimentally measured VCD spectra. These ﬁndings revolutionized the outlook and applications for 1

Although circular dichroism in the infrared region was reported in 1972, the ﬁrst reference to circular dichroism in vibrational transitions appeared in 1974. See: G. Holzwarth, E. C. Hsu, H. S. Mosher, T. R. Faulkner, A. Moscowitz, J. Am. Chem. Soc. 1974, 96 , 252–253.

Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

387

388

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

O

O

O

H

COOH COOH

HO

OR1

COOH

O

OR2

H

OH

HOOC

(2S,3S)-1

H

O

O

H

O

O

(2S,3R)-2

(R)-3

(R)-4

(1R,5R)-5 H

O O O

O

O

OR H

(1R)-6

(1S)-7

(S)-8

(4S)-9

(4S)-10

(1R,2S,4R)-11 H

H 7 C N B 2 E N D H 21 20

O O

O

F O

H

A

O

O

N N

H

O

O

12

13

(R)-14

H

H N

(2R,7S,20S,21S)-18

H H N N

H3CO

O

O

(2R,7S,20S,21S)-17

7

HO

6

N

5

4

2 3

A

O 2’ R1 = bzl =

H

OH

R2O

7 6

2 H CN 21 E N D 20

O

N

20

5

CH3 1 4

O 2

R1 = R2 = Acl =

3 OR1

O

21

7

B

H3CO OH

(2R,19R,20S,21S)-19

CH3

1

H

N H

O

N H

H3CO

(2S,7R,20R,21R)-16 N

N

O

H3CO

O

H3CO

N H

H3CO

(2R,7S,20S,21S)-15

22

23

Scheme 11.1. Structures of compounds 1–23 discussed in the text.

VCD spectroscopy in the twenty-ﬁrst century [4]. As a result, the number of laboratories adapting VCD spectroscopy for structural applications of chiral molecules has been increasing in recent years. This chapter will ﬁrst provide a tutorial for chiral molecular structure determination using VCD and then provide a comprehensive review of the applications of VCD to chiral natural products. An earlier review on VCD spectroscopy of natural products was published in 2008 [5]. Some of the natural products discussed here are shown in Schemes 11.1–11.4.

11.2. DETERMINATION OF CHIRAL MOLECULAR STRUCTURES USING VCD: A TUTORIAL The procedure for determining the chiral molecular structures using VCD spectra is dependent on one basic criterion: If the experimentally observed VCD spectrum for an enantiomer of a chiral compound is satisfactorily reproduced by quantum chemically

389

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

O R2O

N

Sen= O

OH

Tgl=

O

OMe

H

OR1

N

O

O

(1R,3R,5S,6R)-24 [R1 = H; R1 = H]

Cl

OMe O

Ang=

O

H

N

Haemanthamine 29

(2S,3S,4aS,11S)-30

(1R,3R,5S,6R)-25 [R1 = Sen; R1 = H] (1R,3R,5S,6R)-26 [R1 = H; R1 = Tgl] (1R,3R,5S,6R)-27 [R1 = H; R1 = Sen] (1R,3R,5S,6R)-28 [R1 = H; R1 = Ang] O H

OCH3

O H

OCH 3 O H 5 C HD 89 O O B 1 A 10 O H H O 14

O

O

O O H

H

O

OCH3 H

H

H O

O H

OCH3

O

O H

O

H

O O H O

H

O

OH (1R,5S,8S,9S,10S)-32

(1R,5S,8S,9S,10S)-31

CH3

O

(1R,5S,8S,9S,10S)-33

(1R,5S,8S,9S,10S)-34 O

O O H AcO

3

OAc

RO

H 11 H

HO 1

O O (1S,11S,12S)-39

OH

HO

H

HO

O

H

H O

O

OAc

H

(1R,2R,8R,11R)-41 O

H 3 1 OH 5 10 H 6 8 O OH 11

O

O

H

(1R,2R,5S,8R,11R)-40

H H

H H

H H

2 5

OR

RO

(3S,5R,8R,9R,10S,13S,14S)-37, R = H (3S,5R,8R,9R,10S,13S,14S)-38, R = Ac

36

8

H

H

(R)-35

H

H

OAc

H

2 3 4

14

(1R,2R,5S,8R,11R)-42

14

11

1

2

10 9 5 6 78

3 4

15 O

13

12

15

11 1 10

5 6

7

13

9 8 12

O (1R,2R,4R,8R,11R)-43

44

(4R,9R,10R)-45

(4R,9S,10R)-46

Scheme 11.2. Structures of compounds 24–46 discussed in the text.

predicted VCD spectrum for only one of all possible absolute conﬁgurations, then the absolute conﬁguration used to calculate that satisfactorily predicted VCD spectrum can be assigned to the enantiomer used for experimental measurements. It is imperative that the calculated VCD spectra be obtained using reliable quantum chemical methods.

11.2.1. Experimental VCD Spectra Historically, VCD measurements began in the near-infrared region (10,000–4000 cm−1 , where overtone vibrations appear) and then extended to the O–H, N–H and C–H stretching vibrational region (4000–2500 cm−1 ). The experimental and theoretical aspects of near-infrared VCD have been presented by Sergio Abbate and his co-workers in

390

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Br

Br 9

10

3 O H 11

CI Br

OH

Br

Br

O

OH

48

Br

14

O O

R 1O O

O OR

(S)-53 [R1 = H, R2 = CH3]

14

(S)-54 [R1 = Ang, R2 = CH3]

52

OAng OH

57 [R = H] 58 [R = Ac]

(S)-55 [R1 = Sen, R2 = CH3] (S)-56 [R1 = H, R2 = H]

HO

1

O O

5 H

AcO 59

Me

Me

OH

OH

15 Me

HO

H

10

H Me

Me 13

11 H 15 O 12 O 60

Me

Me Me

Me

1

11 15

HO

3

Me Me

5

13

HO

SCH3

R O

O

7

H H 12

12

NH

OR2

9 5

H

8

14 63

62 S

OAc

11

1

Me

61

O H

9 H

HO

14

O

6

OAc

OO 12 H 5 11 10 7 1 6 9 8

13

13

OR

15

4

2

15

(7S,10S)-50 (R = H) (7S,10S)-51 (R = Ac)

(2R,5R,5aR,8R,9aS)-49 OR2

3

4

5

1

7 8

10 9

13 (2R,5R,5aR,7S,8S,9aS)-47

2

H

12

Br

8

7

5a OH

5

Br

O H

O H

2

OR1

OAc

(4R,5S,7R,9R,10R,11R)-64

N H

H N

(4R,5S,7S,8S,10R,11R)-65 R1 = Ang; R2 = i-val (4R,5S,7S,8S,10R,11R)-66

67

68 [R = CH2CN] 69 [R = CH3]

R1 = R2 = Ac H3CS

SCH3

N

N

S

N H

OCH3 (S)-70

H

(R)-71

CH3

O OCH3

O N

O

S

S O

SCH3

N

N OCH3 (2R,3R)-72

S

O

N H (S)-73

Scheme 11.3. Structures of compounds 47–73 discussed in the text.

Chapter 10 of Volume 1. The measurements subsequently were extended to cover the heavy atom stretching (–C=O, –C–C, –C–N, etc.) and hydrogen (–X–H, X=C, N, O, etc.) bending vibrations that absorb infrared radiation in the 2000- to 900-cm−1 region. The details of VCD measurements and instrumentation have been provided by Laurence Naﬁe in Chapter 5 of Volume 1. With the realization that the vibrational bands in the 2000- to 900-cm−1 region are better resolved, and can be predicted reliably using quantum chemical methods, most of the structural determination studies using VCD continue to be conducted in the 2000- to 900-cm−1 region. The experimental VCD measurements are usually conducted for liquid solutions in a suitable infrared transmitting solvent. The solvents usable in the 2000- to 900-cm−1 region include CCl4 , CH2 Cl2 , CHCl3 , CH3 CN,

391

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

OCH3

O R1 O HO R3

O

O

O

O

OCH3

OCH3

R2

H3CO

81-(a) 74 [R1= H, R2 = R3 = CH3] 75 [R1 = R2 = R3 = CH3] 76 [R1 = Ac, R2 = CH3, R3 = CH3] 77 [R1 = H,R2 = C2H5,R3 = CH3] 78 [R1 = H,R2 = CH3, R3 = C2H5]

OCH3 O

HO O

O

79 [R1 = CH3,R2 = C2H5, R3 = CH3] 80 [R1 = CH3, R2 = CH3, R3 = C2H5]

HO

OCH3 81-(b)

O

81-(c)

CH2

O

O

8

OH

O

O

9 O

O

O

O

HO

7 O

5 O 18

O

84

85

(1S,2S,3R)-83 82 CH3

CH3

CH3

OH H 3C

H3C

CH3

CH3 HO

OH

HO

H3C

H2C

CH3 H 3C

H3C

(P)-87 CH3

O HO CH3

HO

CH3

O

O

H H OH HO

OH

OH

OH OH

HO

(M)-88 CH3

OH H3C

CH3

HO

H3C

86

OH

CH3

O

OH HO

O

H3C OH (P)-89

(M)-90

OH

O

(aS)-91

Scheme 11.4. Structures of compounds 74–91 discussed in the text.

(CH3 )2 SO, CH3 OH, and/or their deuterated analogues. Measurements can also be done for water solutions [6], but in such cases high concentrations and lower pathlengths are required to minimize the interference from strong H2 O absorption at 1650 cm−1 . The use of D2 O clears up the region around 1650 cm−1 , but strong absorption of D2 O at ∼1250 cm−1 will preclude the measurements around that region. It is preferable to use low sample concentrations to avoid solute aggregation effects (such as formation of dimers) and to avoid solvents that form hydrogen bonds with solute molecules. However, practical considerations such as lower solubility and smaller vibrational molar extinction

392

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

coefﬁcients of the sample being studied may not leave much choice, in some cases, for the solvents and concentrations to be used. The enantiomers are usually labeled by the signs of their experimental optical rotations at 589 nm as (+)589 or (−)589 . The wavelength subscript is not often included, and by default these labels are written simply as (+) or (−). If both enantiomers of a chiral compound are available, then it is advisable to measure the VCD spectra for both of them. Using one-half of the difference between the VCD spectra of enantiomers, as [(+)-(−)]/2, provides better signal to noise and reduces the spectral and baseline artifacts for the resulting spectrum. As an alternative, if only one enantiomer is available along with its racemic mixture, then subtraction of the VCD spectrum of racemic mixture from that of enantiomer will reduce the spectral and baseline artefacts. If the racemic mixture is also not available, then the VCD spectrum of the solvent may be subtracted from that of the enantiomer solution, but one should be wary of the residual artifacts in such cases. Measurements on ﬁlms [7], mulls, and KBR pellets [8] are also possible, but one should take extreme care in minimizing the spectral artifacts. Unless the investigators can recognize and minimize the spectral artifacts, VCD measurements in solid state are not recommended. To undertake the VCD measurements, the needed sample concentration and cell pathlength should be determined by measuring the infrared (IR) vibrational absorption (VA) spectrum of the sample ﬁrst. A spectral band, for which VCD measurement is needed, should not have absorbance greater than ∼1.0. The optimal range for the absorbance of a band of interest is ∼0.3–0.6. The presence of a VCD band where there is no corresponding absorption band may indicate a possible spectral artifact, so the experimental VCD spectrum should be presented along with the corresponding VA spectrum, preferably one above the other, both on the same wavenumber scale. Representative experimental VCD and absorption spectra are shown in Figure 11.1.

11.2.2. Quantum Chemical VCD Predictions Density functional theoretical (DFT) method is the preferred choice for VCD predictions [9], and the B3LYP functional [10] is the most often used functional. The smallest basis set to be preferred is 6-31G(d) (also labeled as 6-31G*) [11]. The use of any other smaller basis set must be calibrated against molecules with known absolute conﬁgurations. Quantum chemical calculations of VCD yield vibrational band positions and integrated VCD intensities. These numbers have to be morphed into spectra, for a comparison with the experimentally observed spectra, by associating a band shape at each of the calculated vibrational band positions. Lorentzian (or Gaussian) band shapes are normally assumed with a constant width at all band positions. Since experiments are generally performed for liquid solutions, gas-phase calculations are not true representatives of the solution-phase experimental measurements. If the solute does not form hydrogen bonds, or complexes, with solvent, then solution-phase calculations using a reliable solvation model are to be investigated. Polarizable Continuum Model (PCM) [12] has been the most widely used solvation model. The older implementations of PCM suffered [13] from discontinuities and approximations in the deﬁnition and the implementation of the free energy functional in solution and of the corresponding analytical derivatives. The recently developed continuous surface charge PCM [14] overcomes these limitations [15]. If the solute forms hydrogen bonds, or complexes, with solvent, then solute–solvent clusters are to be investigated (vide infra).

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

600 noise

1138

1254

400

Δε x 104

1119 1203

200

1065 1801 VCD

0

1034

1755

1300

–200

1227

1092

(+)-Garcinia acid dimethyl ester in CD2Cl2

–400 1900

1700

1500

1300

1100

900

cm–1 1000 (+)-Garcinia acid dimethyl ester in CD2Cl2 800

1805

1751

600 ε 400

1254 12111153 1439

200 Absorption 0 1900

1700

1308 1350

1500

1300

1068 1119 1092 1100

900

cm–1

Figure 11.1. Experimental VA (bottom panel) and VCD (top panel) spectra for (+)-garcinia acid dimethyl ester in CD2 Cl2 solvent. The noise trace at the top represents the reproducibility level in the VCD spectrum collected for 1 h.

11.2.2.1. Conformational Analysis. A reliable prediction of VCD requires a thorough conformational analysis. The signs of predicted VCD bands can vary among different conformers (with the same absolute conﬁguration), and therefore it is important that all predominant conformations are identiﬁed. In the earlier days, conformational analysis used to be conducted manually by constructing the structures of different possible conformers and optimizing their geometries. This laborious task can now be automated, although without certain risk, using commercial conformational analysis programs, which include SPARTAN [16], CONFLEX [17], HYPERCHEM [18], and MACROMODEL [19]. These programs generate conformers by rotating atoms around different bonds or by wagging chosen atoms, and they use molecular mechanics force constants to optimize their structures. These rotations can be selected systematically (using rules predeﬁned within the software) or randomly (Monte Carlo method). When these programs are used, it is important to be aware of the restrictions built into these programs, which vary among the programs. In addition, molecular mechanics force constants are associated with classically deﬁned bonds (single, double, etc.), and rotations can occur only around single bonds. Therefore the starting structural deﬁnitions that are entered as input into

393

394

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

the molecular mechanics-based conformational analysis programs are to be identiﬁed with classical bonds.2 The energies calculated with molecular mechanics methods are not accurate, but as long as a large energy window is used to identify the possible conformations, the conformers generated with these commercial programs will serve as a reasonable set of starting conformers for further geometry optimizations using DFT theory. However, it should be noted that some conformational analysis programs do not recognize, by default, the variation of ﬁve-member-ring puckering angles as a viable option for the conformational search. The default use of such programs must therefore be supplemented with manual construction of additional conformers and/or relaxed potential energy scans (PESs) at least at the B3LYP/6-31G∗ level, to examine if any conformations are missed by the conformational analysis programs. During a relaxed PES, a desired dihedral (or torsion) angle can be changed in certain increments and the rest of the structural parameters optimized to minimize the energy. A trough in the plot of energy versus dihedral angle will indicate a stable conformer. It is recommended that the options employed in the use of a chosen conformational analysis program be stated clearly, so that the reader can assess the completeness of the reported conformational search. 11.2.2.2. Molecules with a Single Source of Chirality. If the molecule under investigation has a single source of chirality (only one chiral center or only one helical segment), then one of the two possible absolute conﬁgurations is chosen for the theoretical studies (because calculated VCD spectra for the two possible conﬁgurations are mirror images of each other). For that chosen conﬁguration, a set of low-energy conformations are identiﬁed as described above. Then the geometries of these conformers need to be re-optimized at least at the B3LYP/6-31G(d) level of theory. Based on the internal energies at these optimized energies, a better idea of low-energy conformers in the gas phase can be gained. Then a VCD calculation for the gas-phase predicted conformers, at these optimized energies, is performed. The absence of imaginary vibrational frequencies indicates that the conformers used are stable. These stable conformers may then be used for simulating the VCD spectrum. On the other hand, the presence of one or more imaginary vibrational frequencies indicates that the conformation used is an intermediate or a transition state. Such conformations are not used for VCD spectral simulation. The relative populations of stable conformers are calculated using Gibbs free energies obtained in the VCD calculations. Then a population-weighted VCD spectrum is obtained from the individual conformer VCD spectra and populations of stable conformers. The predicted VCD spectrum is referred to as the gas-phase predicted VCD spectrum. If the solvent used for experimental measurements does not form hydrogen bonds, or react, with solute molecules, then quantum chemical geometry optimizations mentioned in the previous paragraph are to be repeated with an appropriate solvation model to account for solvation inﬂuences and are referred to as solution-phase predicted conformers. VCD predictions for the solution-phase predicted conformers, using a solvation model, are used to generate population-weighted solution-phase predicted VCD spectrum. Note that the relative energy ordering of conformers predicted for the gas phase and the solution phase can be different, so population of a given conformer can change from the gasphase prediction to the solution-phase prediction. Most of the literature studies are based 2

The geometry ﬁles created by quantum mechanical programs do not necessarily identify the chemical bonds as single, double etc; or may identify the bonds as single, double etc using predeﬁned standards. Such ﬁles have to be edited before importing into conformational search programs.

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

on the contention that the gas-phase predictions can be compared to the solution-phase experimental measurements. This choice is clearly inadequate, and this practice should be avoided . If the solvent used for experimental measurements forms hydrogen bonds, or reacts, with solute molecules, then quantum chemical geometry optimizations mentioned in the previous paragraphs are to be undertaken for solute–solvent clusters [20a] or for resulting reaction complexes. For such cases the use of solvation models is inappropriate. 11.2.2.3. Molecules with Multiple Chiral Centers or Multiple Sources of Chirality. For a molecule with n chiral centers (or a mixture of chiral centers and helical segments), there will be 2n diastereomers, one-half of which are the enantiomers of the other half. Therefore one must undertake theoretical investigations for 2n−1 diastereomers. The conformational analysis, population-weighted gas-phase, and/or solution-phase VCD spectral predictions (as mentioned for molecules with single source of chirality) must now be undertaken for all 2n−1 diastereomers.

11.2.3. Absolute Conﬁguration Determination The VCD spectra calculated for different sets of conﬁgurations (that describe all chiral centers and helical segments) are compared with the experimental VCD spectrum of an enantiomer. Calculated VCD spectrum, which satisfactorily reproduces the experimental VCD spectrum, identiﬁes the absolute conﬁguration of the enantiomer as that employed in that calculated spectrum. To emphatically state that a chiroptical spectroscopic method unequivocally determined the correct absolute conﬁguration, it is necessary to predict the appropriate spectra for all possible diastereomers (each with appropriate low-energy conformations) and demonstrate that all but one conﬁguration do not yield spectra that can agree with the experimental spectra. Such studies, although computationally demanding, are not emphasized in the literature but should be required. The abovementioned VCD spectral comparisons should be accompanied by the corresponding VA spectral comparisons for many reasons. If the predicted absorption spectrum does not match the observed absorption spectrum, then the corresponding VCD comparison carries less merit. Representative predicted spectra are shown in Figure 11.2.

11.3. APPLICATIONS TO CHIRAL NATURAL PRODUCTS 11.3.1. Carboxylic Acids Carboxylic acids, and molecules with O–H functional groups, often exist as hydrogenbonded complexes, due to the possibility of intermolecular hydrogen bonding between solute molecules or with solvent molecules. Theoretical prediction of chiroptical spectra for such compounds becomes very labor-intensive due to the possibility of formation of intermolecular aggregates [20b]. To avoid these aggregates and associated conformational issues, corresponding esters (or appropriate derivatives) are used for experimental studies. A few chiral carboxylic acids with biological function are found in nature. 11.3.1.1. Garcinia Acid (1) and Hibiscus Acid (2). Garcinia Cambogia (tamarind fruit) is used as a medicine for weight loss and for controlling the appetite in South Asian countries. The skin of the tamarind fruit is rich in 2-hydroxycitric acid, which was isolated as (2S ,3S )-tetrahydro-3-hydroxy-5-oxo-2,3-furandicarboxylic acid

395

396

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

600

400

(2S,3S)-Garcinia acid dimethyl ester B3LYP/aug-cc-pVDZ-PCM

Δε x 104

1264

1141

200

1196

1813 0

–200

1029 1094

1311

1758

1235 –400 1900

1700

1500

1300

1100

900

cm–1 1000

1816 1757 (2S,3S)-Garcinia acid dimethyl ester

800

B3LYP/aug-cc-pVDZ-PCM

600 ε

1257 400

1096 1066

1300 200

0 1900

1453

1700

1215

1359

1500

1159

1124 1300

1100

900

cm–1

Figure 11.2. The conformer population weighted calculated VA (bottom panel) and VCD (top panel) spectra for (2S,3S)-garcinia acid dimethyl ester at the B3LYP/aug-cc-pVDZ level. The incorporation of CD2 Cl2 solvent inﬂuence, using the PCM model as implemented in Gaussian 09 program, makes the two predicted carbonyl stretching band positions at 1816 and 1757 cm−1 match well with the corresponding experimental band positions at 1805 and 1751 cm−1 .

or garcinia acid, 1 (see Scheme 11.1). Its diastereomer, (2S,3R)-tetrahydro-3-hydroxy-5oxo-2,3-furandicarboxylic acid, or hibiscus acid, 2, is extracted from the Roselle plant, which is a common ingredient in many herbal tea blends. The known absolute conﬁgurations of these acids can be used to test the predictive abilities of VCD spectroscopy. For this purpose, the combined experimental and theoretical VCD spectral investigations were undertaken on the corresponding dimethyl esters. Each of these dimethyl esters can exist in eight different conformations. The experimental VCD spectrum of garcinia acid dimethyl ester (GADE) (Figure 11.1), with positive optical rotation at 589 nm, was reproduced by the population-weighted predicted VCD spectrum of garcinia acid dimethyl ester with known (2S ,3S ) conﬁguration [15] at the B3LYP/aug-cc-pVDZ-PCM level (Figure 11.2). Similarly, the experimental VCD spectrum of the dimethyl ester of hibiscus acid (HADE), with positive optical rotation at 589 nm, was reproduced by the population weighted predicted VCD spectrum of hibiscus acid dimethyl ester with known (2S ,3R) conﬁguration [20c] at B3LYP/aug-cc-pVDZ-PCM level. However,

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

as mentioned earlier, the experimental VCD spectra have to be compared to the calculated VCD spectra for all stereoisomers. When this was done, it was noted [20c] that the experimental VCD spectra for (+)-GADE can be correlated to the calculated VCD spectra of both (2S,3S ) and (2R,3S ) stereoisomers. Based on the comparison of experimental and calculated absorption spectra, however, only (2S,3S ) could be correlated with (+)-GADE. Similarly, the experimental VCD spectra for (+)-HADE can be correlated to the calculated VCD spectra of both (2S,3R) and (2R,3R) stereoisomers. Again, based on the comparison of experimental and calculated absorption spectra, only (2S,3R) could be correlated with (+)-HADE. Since chemical conversion of acids into esters does not inﬂuence the absolute conﬁguration, the absolute conﬁgurations of the parent acids are inferred to be same as those of corresponding dimethyl esters. 11.3.1.2. Hexylitaconic Acid and Its Methyl Esters [21]. Although most natural products exist as single enantiomers, both enantiomers of hexylitaconic acid (3 with R1 = R2 = H) are found in nature: (+)-Hexylitaconic acid was isolated from Aspergillus niger, and (−)-hexylitaconic acid was derived from marine endophytic fungus Apiospora montagnei . A comparison of the experimental VCD spectra of both enantiomers of hexylitaconic acid methyl ester (3 with R1 = R2 = CH3 ) (in CDCl3 ) with the population-weighted predicted VCD spectrum at the B3PW91/6-311+G(d , p) level indicated the absolute conﬁguration to be (R)-(−). In these predictions it was assumed that the C–C–C–C dihedral angles in alkyl carbon chain are 180◦ . The populationweighted spectrum was generated from 32 low-lying conformers that were identiﬁed using a conformational analysis program [17]. To reduce the conformational uncertainties, the diester was converted to a cyclic lactone, hexyl-3-methylenedihydrofuran-2(3H )-one (4), with a more rigid structure. A similar analysis of the experimental and predicted VCD spectra of this lactone suggested the absolute conﬁguration of lactone to be (+)-(R). Stereochemical transformation from (−)-hexylitaconic acid mono methyl ester to (−)hexylitaconic acid dimethyl ester and from (−)-hexylitaconic acid mono methyl ester to (+)-hexyl-3-methylenedihydrofuran-2(3H )-one was used to conﬁrm the assignment of (R)-(−) hexylitaconic acid dimethyl ester and, as a consequence, of the parent acid as (R)-(−) hexylitaconic acid.

11.3.2. Monoterpenes and Derivatives VCD spectra of some monoterpenes have been used to test the instrumental performance and to investigate normal mode versus local mode behavior [22, 23]. VCD spectra of some other monoterpenes have been used as test cases to assess the predictive abilities of, or for determining their molecular structures with, VCD spectroscopy. 11.3.2.1. α-Pinene (5), Camphor (6), and Fenchone (7) [24–28]. α-Pinene and camphor have been the widely used chemicals for VCD studies, starting from very early stages. Because of the large VCD signals observed for these molecules, they have been used for checking the correct function of the VCD instruments and for calibration purposes. Since the absolute conﬁgurations of these molecules are known for a long time, these molecules have also been used as standards for evaluating the reliability of quantum theoretical levels for VCD predictions. Based on the comparisons between experimental and quantum chemical predictions of VCD spectra, Hartree–Fock theoretical level was considered to be inadequate for VCD predictions. Similarly, the use of DFT (with functional such as B3PW91 or B3LYP) with 6-31G∗ basis set was considered to be

397

398

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

minimal level for placing any reliability on VCD predictions [24, 25]. Excellent agreement between B3PW91/6-31G∗ -predicted VCD spectra and experimental VCD spectra for (1R)-α-pinene and (+)-α-pinene, (1R)-camphor and (+)-camphor, and (1S )-fenchone and (+)-fenchone indicated, in as early as 1996, that the absolute conﬁgurations of chiral natural products can be predicted reliably using combined experimental and quantum chemical VCD studies. 11.3.2.2. Carvone (8) [29]. Conformational analysis of 8, at the B3LYP/cc-pVDZ level, indicated that there are six conformers for this molecule with three conformations being predominant. These dominant conformations have isopropenyl group in equatorial position and differ in the relative orientation of the isopropenyl group with respect to the cyclohexenone ring. The population-weighted predicted VCD spectrum for (S )-carvone, at the B3LYP/cc-pVDZ level, was considered to match well with the experimental VCD spectrum of (+)-carvone liquid, thereby suggesting that VCD can be used for determining the absolute conﬁgurations. It was also pointed out that, while ordinary IR and Raman spectra may not be able to distinguish the conformers, VCD spectral analysis helps in identifying the conformations present in liquid solutions. 11.3.2.3. Limonene Oxide (9) [30]. 9 is an atmospheric pollutant resulting from the oxidation of other terpenes (limonene, α-pinene, etc.). Conformational analysis at the B3LYP/cc-pVDZ level indicated that there are a total of 12 conformers, with six of them containing isopropenyl group in equatorial position that account for 97% population. Three of them have the oxirane ring in trans, and the other three in cis, orientation with respect to the hydrogen atom at the chiral carbon (to which isopropenyl group is attached). The lowest-energy conformer has the oxirane ring in trans orientation. By comparing the experimental IR, Raman, and VCD spectra of the liquid sample with the corresponding population-weighted gas-phase predicted spectra, at the B3LYP/cc-pVDZ level, it was concluded that ﬁve of the abovementioned conformers are present in the liquid state, with two trans conformers accounting for ∼50% population. 11.3.2.4. Perilladehyde 10 [31]. This monoterpene with aldehyde functional group has 12 different conformations. Isopropenyl group can be in equatorial or axial position, with three orientations of isopropenyl group in each position. Furthermore, aldehyde group can take two different orientations with respect to the cyclohexene ring. Three equatorial conformers are the most stable ones among the conformers investigated. The population-weighted gas-phase predicted VCD spectrum at the B3LYP/cc-pVDZ level for (S )-conﬁguration was compared with the experimental VCD spectrum of (−)perillaldehyde liquid sample. Based on this comparison, the perillaldehyde molecules in liquid sample are suggested to exist in three conformers with (S )-(−) absolute conﬁguration. 11.3.2.5. endo-Borneol (11 with R = H) and Its Derivatives [32]. While the absolute conﬁguration of 11 (with R = H) has been known for a long time, the issue of how many predominant conformations this molecule has could only be answered through a reliable conformational analysis. The O–H group can rotate around its connecting bond to C atom, which yields at least three possible conformers. This conformational freedom can be restricted by chemically converting the O–H to methoxy (O–CH3 ), acetyloxy (O–COCH3 ), t-butoxy (O-C(CH3 )3 ), or trimethyl silane (O–Si(CH3 )3 ) derivative. The inﬂuence of this derivatization on structural predictions using VCD has been investigated

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

[32]. The experimental IR and VCD spectra of the O–H, acetyloxy, t-butoxy, or trimethyl silane derivatives were considered to compare well with the corresponding predicted spectra at the B3LYP/TZ2P and B3PW91/TZ2P levels of theory, for the conformations identiﬁed. Through this comparison, the (1R, 2S , 4R)-(+)-conﬁguration was conﬁrmed, and it was further concluded that chemical derivatization of hydroxyl groups increases the conformational rigidity and eliminates intermolecular hydrogen bonding. 11.3.2.6. Myrtenal 12 [33]. This molecule has a rigid cyclic structure with conformational mobility coming from the aldehyde group. Even though two orientations of aldehyde group are possible, the trans orientation has nearly 99% population. The small molecular size and conformational rigidity prompted VCD spectroscopic investigations with a view to evaluate the predictions with different size basis sets. The experimental VCD spectra were measured for (−)-12 in CCl4 (6.1 mg/150 μL). Theoretical VCD predictions were obtained for (1R)-12 using a B3LYP functional and 6-31G*, 6-31+G ∗∗ , 6-311+G∗∗ , and DGDZVP basis sets. Another calculation using B3PW91 functional and DGTZVP basis set was also undertaken. After a comparison of all these predictions with experimental VCD spectra, it was concluded that the absolute conﬁguration of (−)-12 is (1R) and that there were no signiﬁcant differences of concern among these different levels of predictions and that computationally less intensive calculation at the B3LYP/DGDZVP level provides computationally economical choice without compromising the predictive abilities. 11.3.2.7. 3-Oxo-1,8-cineole or 1,3,3-Trimethyl-2-oxabicyclo[2.2.2]octan5-one (13) [34]. The pure enantiomers, (−)-13 and (+)-13, were prepared by oxidation of (+)- and (−)-3-hydroxy-1,8-cineole. The absolute conﬁguration of (−)-3-hydroxy-1,8cineole was determined by preparing Mosher esters. The oxidation of (−)-3-hydroxy-1,8cineole and (+)-3-hydroxy-1,8-cineole with pyridinium chlorochromate yielded, respectively, (1R, 4R)-(+)-13 and (1S , 4S )-(−)-13. VCD spectroscopy was used to obtain independent evidence for the absolute conﬁguration of 3-oxo-cineole. Conformational search for (1S , 4S )-13 indicated a single conformer. The VCD spectrum predicted for (1S , 4S )-13 at the B3LYP/DGDZVP level was considered to compare well with the experimental VCD spectrum of (−)-13, leading to the assignment of (1S , 4S )-(−)-13, which is in agreement with that obtained from Mosher ester analysis. 11.3.2.8. β-Pinene, Nopinone, Menthene, Limonene, and Menthenol, and Menthol. VCD spectra of these compounds [35, 36], except limonene, were investigated before the development of DFT methods for VCD and are not discussed here. A recent investigation [37] using experimental IR, Raman and VCD spectra of (R)-(+)-limonene, along with B3LYP/aug-cc-pVDZ calculations, suggested that three different equatorial conformers exist for this molecule. 11.3.2.9. Camphor–CDCl3 Complex [38]. For (R)-(+)-6 dissolved in achiral CDCl3 solvent, a negative VCD band was observed at 2254 cm−1 , which is associated with the C–D stretching vibration of CDCl3 . Opposite-signed VCD band was seen at the same position for (S )-(−)-6 in CDCl3 . VCD observed for the C–D stretching vibration of CDCl3 was referred to as induced VCD or chirality transfer (from chiral solute to achiral solvent). The VCD induced in CDCl3 was rationalized as due to D. . .O bond formation between D atom of CDCl3 and O atom of 6. Induced VCD however was not visible in CD2 Cl2 solvent. The optimized geometry of the 1:1 camphor–CDCl3 complex, using

399

400

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

B3LYP functional and cc-pVDZ basis set, indicated a stable structure with D. . .O distance of 2.21 A. The predicted VCD spectrum for (R)-(+)-6-CDCl3 complex reproduced the experimental observation of negative VCD for the C–D stretching mode. 11.3.2.10. Pulegone (14)–CDCl3 Complex [39, 40]. For (R)-(+)-14 dissolved in achiral CDCl3 solvent, a positive VCD band was observed at 2250 cm−1 , which is associated with the C–D stretching vibration of CDCl3 . Opposite-signed VCD band was seen at the same position for (S )-(−)-14 in CDCl3 . The IR absorption for C–D stretching vibration itself exhibited enhanced absorption intensity in the presence of 14, which indicates the presence of interaction, possibly D. . ..O bond formation between D atom of CDCl3 and O atom of 14. Induced VCD, however, was not visible in CD2 Cl2 solvent, which was associated with less acidic character of D atom in CD2 Cl2 that does not favor D. . .O bond. To explain the induced VCD in the C–D stretching vibration of CDCl3 , 1:1 complex of pulegone–CDCl3 and solvation inﬂuence (incorporated via PCM) were considered. Incorporating the 1:1 complex in solvated model was considered to yield most satisfactory predictions. When induced VCD can be measured in an otherwise achiral solvent, one would think that the relative orientations of chiral solute and achiral solvent molecules can be inferred from the experimental observations. A detailed theoretical analysis by Nicu et al. [40] suggested that such information cannot be deduced with conﬁdence because the interaction between solute and solvent molecules is very weak. They found that the angle between electric and magnetic dipole transition moments, for the C–D stretching vibration of CDCl3 , as well as for the C=O stretching vibration of pulegone, is close to 90◦ . As a result, the VCD signs predicted for these vibrational modes depended on computational variations (such as tight geometry versus assumed or default geometry, basis set, functional, etc.), and therefore one cannot place much reliance on the predicted VCD signs of these bands.

11.3.3. Alkaloids 11.3.3.1. Cinchonidine [41]. Three different conformers were suggested for cinchonidine based on B3LYP/6-31G∗ calculations and solvent-dependent NMR studies. The population-weighted calculated VCD spectrum at the B3LYP/6-31G∗ level was considered to reasonably match the experimental VCD spectrum of cinchonidine in CDCl3 , leading to the suggestion that the ﬁndings from NMR and VCD are consistent. 11.3.3.2. Schizozygane Alkaloids [42]. The group of alkaloids isolated from the East African plant, Schizozygia caffaeoides, are referred to as schizozygane alkaloids. The structure of schizogygine 15, but not its absolute conﬁguration, was deduced from chemical reactions and spectroscopic properties. The absolute conﬁguration of a closely related alkaloid, strempeliopine (16), isolated from Cuban plant Strempeiopsis strempelioides, was determined to be (−)-(2S , 7R, 20R, 21R). 15 and 16 differ only in the absence of ring F and the hydrogenation of C14 –C15 bond. Since the [α]D of 15 in CHCl3 is +15.5, which is of opposite sign to that of 16, one could speculate the absolute conﬁguration of 15 to be (2R, 7S , 20S , 21S ), but such correlation can be dangerous. Therefore the absolute conﬁguration of 15, either (2R, 7S , 20S , 21S ) or (2S , 7R, 20R, 21R), remained undetermined until VCD spectroscopic analyses were undertaken. The experimental VCD spectra (at concentrations 0.127–0.425 M) and optical rotatory dispersion (ORD) spectra (0.0116 M) were measured in CDCl3 solvent, while ECD (12.5 mM) spectra were measured in CHCl3 . The conformational analysis indicated only two conformations, one with

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

ring C in chair conformation and another with ring C in boat conformation. Both conformations had ring F in near planar conformation (this was probably an artifact because SPARTAN program [16] used here does not pucker the ﬁve-member rings unless forced by the user). Re-optimization of these structures at the B3LYP/6-31G∗ level indicated that ring F is puckered. A potential energy scan varying the ring F puckering angle indicated the existence of two puckering angles for ring F. Thus a total of four conformations were found and their VCD spectra predicted at B3LYP/TZ2p and B3PW91/TZ2p levels. The population-weighted predicted VCD spectrum for (2R, 7S , 20S , 21S ) conﬁguration was considered to match well with the experimental VCD spectrum while that for (2S , 7R, 20R, 21R) was not. Therefore the absolute conﬁguration was suggested to be (2R, 7S , 20S , 21S )-(+)-15. This assignment was conﬁrmed by comparing the predicted ECD and ORD spectra with the corresponding experimental spectra. Two more alkaloids, schizogaline (17) and schizogamine (18), also isolated from Schizozygia caffaeoides, have structures similar to that of 15. Assuming that the biosynthetic pathway for these compounds is identical to that of 15, the absolute conﬁgurations were suggested to be (2R, 7S , 20S , 21S )-(+)-17, and (2R, 7S , 20S , 21S )-(−)-18. In the same manner, the absolute conﬁguration of naturally occurring 6,7-dehydro-19β-hydroxyschizozygine (19) was also suggested to be (2R, 19R, 20S , 21S )-19. 11.3.3.3. Iso-schizozygane Alkaloids [43]. Iso-schizogaline (20) and Isoschizogamine (21) had [α]D values of -260 and 262, respectively. To determine their absolute conﬁgurations, VCD spectral predictions were made for (2R, 7R, 20S , 21S )-20 and (2R,7R,20S ,21S )-21. It is not clear how this particular conﬁguration, among many other possible diastereomers, was chosen. The conformational search for hexa-cyclic core indicated three conformations originating from changes in the conformations of rings C and E. The conformational analysis of methoxy benzene indicated two stable conformers with C1 C2 OC3 dihedral angles of 0◦ (cis) and 180◦ (trans). As a result, a total of six conformers were investigated for (2R, 7R, 20S , 21S )-20 and found to be stable at B3LYP/6-31G∗ level. Only three of these conformers were considered to be signiﬁcantly populated in the gas phase. The conformational search done for orthodimethoxy benzene indicated seven stable conformations, which, in combination with three conformations of rings C and E, yielded 21 conformations for the case of (2R,7R, 20S , 21S )-21. The 14 lowest-energy conformations were further optimized at B3LYP/6-31G∗ , B3LYP/TZ2P, and B3PW91/TZ2P levels. Three of these conformers account for > 90% population. The population-weighted predicted VCD spectrum of (2R, 7R, 20S , 21S )-20 was considered to compare well with the experimental VCD spectrum (0.255 M in CDCl3 ) of naturally occurring (−)-20. In the same manner, the population-weighted predicted VCD spectrum of (2R, 7R, 20S , 21S )-21 was considered to compare well with the experimental VCD spectrum (0.225 M in CDCl3 ) of naturally occurring (−)-21. Based on these observations, the absolute conﬁgurations were assigned as (2R, 7R, 20S , 21S )-(−)-20 and (2R, 7R, 20S , 21S )-(−)-21. These assignments were conﬁrmed with a comparison of predicted ECD and ORD spectra with corresponding experimental spectra.

11.3.4. Tropane Alkaloids 11.3.4.1. 6β-Hydroxyhyoscyamine (22). Two diastereomers of 22 with positive and negative optical rotations have been isolated from the plant materials. The literature assignments of absolute conﬁgurations of these compounds were tentative, and

401

402

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

this uncertainty led to VCD studies on these two diastereomers for determining their absolute conﬁgurations [44]. Also, it was of interest to ﬁnd the spectral signature that will differentiate the two diastereomers. Using a B3LYP functional and a 6-31G∗ basis set, it was suggested that eight different conformers of each of the diastereomers contribute to the conformationally averaged VCD and VA spectra. The VCD bands in the 1300–1200 cm−1 exhibited signiﬁcant dependence on the conformations, while those in the 950–1100 cm−1 showed little variation with conformation. But the predicted VCD bands in the 950- to 1100-cm−1 region gave opposite signs for the two diastereomers. These bands were suggested to originate from the tropane ring and provide differentiation between the diastereomers. The comparison between experimental and predicted VCD spectra was used to suggest the absolute conﬁguration assignments as (−)-(3S , 6S , 2 S )-22 and (+)-(3R,6R, 2 S )-22. 11.3.4.2. 3α,6β-Diacetoxytropane (23). (−)-23 is a semisynthetic product obtained from (−)-(3S ,6S ,2 S )-22. The absolute conﬁguration of (−)-23 was not known until VCD spectroscopy was used [45] to determine its absolute conﬁguration as (3S , 6S )-(−)-23. Theoretical analysis was conducted with B3LYP functional and DGDZVP basis set. Because a tropane ring is conformationally rigid, the conformational freedom in this molecule originated from the disposition of N -methyl group and two acetyl groups. The equatorial orientation of the N -methyl group was found to be preferred. The C6 acetyl group was found to have more stable orientation when its C=O group was oriented outside the molecule, while C3 acetyl group was found to be iso-energetic for outside and inside orientations. A similarity in the VCD spectra of this molecule with that of (−)-(3S , 6S , 2 S )-22 led the authors to suggest the existence of several bands that reﬂect the absolute conﬁguration of molecular skeleton. 11.3.4.3. 3α,6β-Tropanediol (24), and Its Mono Esters. The absolute conﬁguration and conformational analyses for 24 (see Scheme 11.2) and four monoesters were reported, using experimental and theoretical VCD studies [46]. These monoesters, 6β-hydroxy-3α-senecioyloxytropane (25), 3α-hydroxy-6β-trigloyloxytropane (26), 3α-hydroxy-6β-senecioyloxytropane (27), and 3α-hydroxy-6β-angeloyloxytropane (28), were extracted from leaves of S . grahamii and S . pinnatus, but their optical rotation data were not provided to label them as (+)/(−) compounds. From experimental and theoretical VCD studies, the absolute conﬁgurations of investigated monoesters were deduced as (1R, 3R, 5S , 6R). For theoretical predictions, the initial conformers were screened with the Monte Carlo search method, for both axial and equatorial positions of the N–Me group. For (1R,3R,5S ,6R)-25, this resulted in ﬁnding 16 conformations with the axial N–Me group and 22 conformations with the N–Me equatorial group. These conformations were re-optimized at the B3LYP/6-31G* level to select 10 lowest-energy conformations within 2.5 kcal/mol energy difference. Of these conformations, two accounting for 72.8% population have an axial N–Me group and the remaining have an equatorial N–Me group. These conformers were re-optimized with a B3LYP functional and a DGDZVP basis set, and VCD predictions were undertaken at the same level. The two axial N–Me conformers have intramolecular hydrogen bonding between the hydroxyl group at C6 and the nitrogen atom of N–Me group. Solvation may have signiﬁcant inﬂuence on the experimental spectra, which was not incorporated in the calculations. This could be one reason for signiﬁcant differences between predicted and observed VCD spectra in the 1500- to 1100-cm−1 region. Nevertheless, VCD bands in 1100- to 950-cm−1 region showed similarities, and it was concluded that the 950- to 1100-cm−1 region is critical in assigning the absolute conﬁgurations for 3,6-tropanediol

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

derivatives. Based on the comparison in this region for (1R, 3R, 5S , 6R)-25 and experimental spectra, the absolute conﬁgurations of the investigated compound was suggested as (1R, 3R, 5S , 6R). For 26 and 27, since the O–H group at C6 is now esteriﬁed, hydrogen bonding with N–Me is not present; as a result, conformers with equatorial N–Me orientation (>80%) become more populated than axial N–Me conformers. The sample used for VCD experiments was identiﬁed via NMR as a 69:31 mixture of 26 and 27, and therefore the calculated VCD spectrum for this mixture was generated by combining individual calculated spectra in that ratio. A satisfactory agreement between the predicted and experimental VCD spectra was used to suggest the absolute conﬁgurations of both compounds as (1R,3R,5S ,6R).

11.3.5. Montanine-Type Alkaloids When hemanthamine (29) was reacted with SOCl2 , a compound (30) of the formula C17 H18 NO3 Cl was formed which was identiﬁed using 2D NMR to have montanine-type structure [47]. Relative stereochemistry of 30 was determined using ROESY spectrum, while the absolute conﬁguration of 30 was determined from combined analysis of experimental and theoretical VCD spectra. Monte Carlo conformational search revealed two low-energy conformations accounting for 99% population. The geometries of these two conformations were optimized ﬁrst at B3LYP/6-31G∗ level and subsequently at the B3LYP/DGDZVP level. VCD spectra were calculated for the optimized conformers at both B3LYP/DGDZVP and B3PW91/DGDZVP levels. The population-weighted predicted VCD spectra for (2S , 3S , 4aS , 11S )-30 were considered to be in good agreement with the experimental VCD spectra in CDCl3 (5.3 mg/150 μL) of (+)-30. Therefore the absolute conﬁguration was assigned as (+)-(2S , 3S , 4aS , 11S )-30.

11.3.6. Iridoids 11.3.6.1. Plumericin (31) and Isoplumericin (32) [48]. The relative conﬁgurations of 31 and its isomer, 32, were suggested [49] based on their chemical and spectroscopic properties. The X-ray diffraction data conﬁrmed the relative conﬁgurations assigned previously, but semiempirical analysis of experimental ECD spectra suggested [50] the opposite. The utilization of theoretical methods for interpreting the experimental VCD and ORD spectra resolved this controversy in favor of the originally suggested conﬁgurations. The conformational analysis at the MMFF94 level for 31 and 32 indicated four conformations each. These conformations were also supported by B3LYP/6-31G*, B3LYP/TZ2P, and B3PW91/TZ2P level optimizations. These conformations differ in the puckering angle of ring B and in the orientation of OCH3 group. In each case, two conformations account for 90% of the populations, so only two conformers are major contributors to the predicted spectra. The experimental VCD spectra of (+)-31 in CDCl3 (0.069 M) were considered to match well with the corresponding population-weighted predicted spectrum of (1R, 5S , 8S , 9S , 10S )-31, but not with that of (1S , 5R, 8R, 9R, 10R)-31. Therefore the absolute conﬁgurations were suggested as (1R, 5S , 8S , 9S , 10S )(+)-31 and (1R, 5S , 8S , 9S , 10S )-(+)-32. These assignments are in agreement with the original assignments [49] and were further supported by the analysis of predicted and experimental speciﬁc rotations. 11.3.6.2. Prismatomerin (33). In terms of chemical structures, 33 differs from 31 in a p-phenol group replacing the methyl group of 31 at the C14 position. The [α]D of 33 is −136 (EtOH), opposite in sign to that of 31. If it is assumed that the substitution

403

404

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

of phenol group for methyl group in 31 does not change the sign of optical rotation, then one would assign 33 the opposite conﬁguration to that of 31. But such assumptions can be dangerous because there is no way to reliably know the inﬂuence of chemical substitution on [α]D of chemically related molecules. For this reason, VCD spectroscopy was used [51] to determine the absolute conﬁguration of 33. Since the presence of O–H groups can result in inter- or intramolecular hydrogen bonding, which makes the theoretical analysis complex, the acetate derivative, prismatomerin acetate 34, was investigated. In addition to the four conformers, as were found for 31, the rotation of phenyl acetate group results in four more conformers. Thus a total of 16 conformers were found for 34. Twelve of these conformers accounted for 99% of the population, so populationweighted predicted VCD spectra were obtained from these 12 conformers. The predicted VCD spectra at the B3LYP/TZ2P and B3PW91/TZ2p levels for (1R, 5S , 8S , 9S , 10S )-34 were considered to agree well with the experimental VCD spectra of (−)-34, leading to the assignment (1R, 5S , 8S , 9S , 10S )-(−)-34. Since the conversion of phenol group to its acetate does not change the absolute conﬁguration, the absolute conﬁguration of 33 is also (1R, 5S , 8S , 9S , 10S )-(−). Note that naturally occurring 31 and 33 have the same absolute conﬁguration but oppositely signed [α]D values. Therefore it was suggested not to assign the absolute conﬁgurations based on the signs of [α]D values of chemically related molecules.

11.3.7. Meroditerpenoids 11.3.7.1. Sargaol Acetate (35) [52]. 35 is an optically active substance with positive optical rotation at 589 nm in chloroform solvent. To determine its absolute conﬁguration, a negative–positive couplet centered at 1200 cm−1 in the experimental VCD spectrum in CCl4 (5.9 mg/150 μL) was used. Theoretical VCD spectra were predicted for 35 with (R)-conﬁguration, using a B3LYP functional and a DGDZVP basis set. However, the tris-isoprenyl chain of this molecule introduces a large number of conformations. With a view to simplify the computational problem, the tris-isoprenyl chain was replaced with ethyl and isoprenyl, one substitution at a time. These modelmolecule predictions were considered to reproduce the major negative–positive feature found in the experimental spectrum, but weaker bands in the 1000- to 1100-cm−1 region were not. With an objective to improve the predictions, the authors used di-isoprenyl group substitution and claimed better agreement with experimental VCD spectrum on the lower-energy side of the spectrum. Finally, the authors investigated the complete structure without any model substitution. The conformer-averaged spectrum was no better than the model structure predictions, although all calculations correctly predicted the experimentally observed negative–positive couplet, suggesting the absolute conﬁguration (+)-(R)-35. 11.3.7.2. Isoepitaondiol Diacetate (36). To prevent possible epimerization of parent diol, its diacetate derivative (36), which was extracted from S.ﬂabelliforme, was used instead. Its structure was proposed based on various NMR spectral data for 36, and it was conﬁrmed using X-ray diffraction data. In light of this new structure, the reported structure of parent diol was corrected. The absolute conﬁguration of 36 could not be established from X-ray diffraction data, so VCD spectroscopy was used for this purpose [53]. Theoretical VCD predictions were obtained at the B3LYP/DGDZVP level. The Monte Carlo conformational search identiﬁed 13 low-energy conformers. This number was reduced to two, based on single-point energy calculations at the B3LYP/DGDZVP level.

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

The geometries of these two conformers were optimized and VCD calculations undertaken. The resemblance between experimental VCD spectrum in CDCl3 (6.6 mg/150 μL) and conformer-averaged predicted VCD spectrum in the 1700- to 1000-cm−1 region was used to propose the absolute conﬁguration for 36 sample as depicted in Scheme 11.2. 11.3.7.3. Stypotriol (37). The absolute conﬁguration of 37 was deduced from that of its triacetate derivative, 38. Theoretical VCD calculations for 38 were done at the B3LYP/DGDZVP level and compared to the experimental VCD spectra of 38. Based on this comparison, the absolute conﬁguration was concluded [54] to be (3S , 5R, 8R, 9R, 10S , 13S , 14S )-(−)-37.

11.3.8. Verticillane Diterpenoids 11.3.8.1. Verticilla-3E,7E-dien-12-ol (39). The absolute conﬁguration of 39 derived from x-ray analysis of its p-iodobenzoate derivative was at odds with that proposed earlier using the octant rule associated with ECD. To resolve this controversy experimental VCD spectrum of (+)-39 was measured in CCl4 (10 mg/200 μL) and compared to the predicted VCD spectrum [55]. For theoretical predictions, the structure of (1S , 11S , 12S )-39 was subjected to Monte Carlo conformational search, which identiﬁed six conformations. Further optimizations of geometries of these conformers at the B3LYP/6-31G*level indicated that three lowest-energy conformers accounted for 97% population. In these three conformers, the six-member ring is in chair conformation and cyclooctadodecane moiety in chair–chair–chair–chair conformation. All three conformers have the same hydrocarbon skeleton, which was veriﬁed with 1 H NMR coupling constants. The population-weighted predicted VCD spectrum at the B3LYP/6-31G∗ level was considered to have a good correspondence with the experimental VCD spectrum for (+)-39. The same level of agreement was also noted for the predicted VCD spectrum with a B3LYP/DGDZVP basis set leading to the suggested absolute conﬁguration, (+)-(1S , 11S , 12S )-39. This assignment is in accord with that derived from X-ray structure analysis of p-iodobenzoate derivative.

11.3.9. Sesquiterpenes 11.3.9.1. Quadrone (40) [56]. The total synthesis of quadrone enantiomers led to the determination of the absolute conﬁguration [57] of quadrone 40 as (1R, 2R, 5S , 8R, 11R)-(−). This assignment was independently conﬁrmed, using VCD, ECD, and ORD spectroscopic methods, in order to evaluate the applicability of these methods for structurally related sesquiterpenes, subersenone (41), subersanone (42), and seberosenol A acetate (43). The conformational analysis of 40 indicated the presence of three stable conformers. But the relative energies of these conformers at the B3LYP/6-31G∗ level indicated that only one conformer has predominant population. This predominant conformation has a cyclohexane ring in a chair conformation, a lactone ring in a boat conformation, and a cyclopentane ring in a non-planar conformation. The VCD spectrum predicted for this stable conformer with (1R, 2R, 5S , 8R, 11R) conﬁguration at the B3PW91/TZ2P level was considered to match well with the experimental (0.14 M and 0.24 M in CDCl3 ) VCD spectrum of (−)-40, which led to conﬁrming the previously established absolute conﬁguration as (1R, 2R, 5S , 8R, 11R)-(−)-40. This assignment was further conﬁrmed using predicted and experimental ECD spectra and also the [α]D values. The absolute conﬁgurations of related sesquiterpenes were suggested based

405

406

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

on the comparison between experimental and predicted [α]D values, as (1R, 2R, 8R, 11R)-(+)-41, (1R, 2R, 5S , 8R, 11R)-(−)-42, and (1R, 2R, 4R, 8R, 11R)-(−)-43, but it was suggested that a further analysis using VCD and ECD spectroscopies would be required for conﬁdent assignments. 11.3.9.2. 8-Epiisolippidiol-3-O-β-D-glucopyranoside (3β,8β,-Dihydroxy1α, 4β, 5α, 6β, 7α, 11β-guai-10(14)-ene-6,12-olide-3-O-β-D-glucopyranoside) (44). The structure of this compound was characterized on the basis of 1 H NMR spectra data (assuming α-conﬁguration of H7) and X-ray diffraction and ECD data of its dehydrogenated derivative. The known absolute stereochemistry of its dehydrogenated derivative did not favor the diastereomeric structure with opposite conﬁgurations of aglycone. VCD spectral analysis of this compound was undertaken [58] assuming that the absolute conﬁguration previously suggested [59] was correct. Conformational analysis involved the rotation of hydroxyl group at the C8 position, as well as rotation around the glycosidic bonds and hydroxymethyl group in the glucose moiety. The rotations around glycosidic bond were found to have major inﬂuence on energy. The orientation of hydroxyl hydrogen atom at C8 favored interaction with π -electrons of exomethylene double bond. The rotations of hydroxymethyl group of glucose moiety had smaller inﬂuence on energies as well as on spectra. A total of six conformers were investigated at the B3LYP/6-31G∗ level. The population-weighted predicted VCD spectra, after scaling the band positions, was considered to match the experimental VCD spectrum, and therefore the literature-suggested conﬁguration was considered to be correct. 11.3.9.3. Africanane and Lippifoliane. African-1(5)-ene-2,6-dione (45) and lippifoli-1(6)-en-5-one (46), both functionalized tricyclic systems, are derived from a shrub, Lippia integrifolia. X-ray diffraction analysis of a derivative, 4,10,11tribromo-10,11-seco-lippifoli-1(6)-en-5-one, along with a combined experimental and quantum chemical analysis of ECD spectra, was used previously to derive the absolute conﬁguration of 46 as (4R, 9S , 10R), while that of 45 was proposed based on the basis of its relationship to other natural products derived from Lippia integrifolia. An independent conﬁrmation of the absolute conﬁguration of 46 and determination of the absolute conﬁguration of 45 using VCD spectroscopy were undertaken [60]. The experimental VCD spectra of (+)-45 were obtained in CCl4 solvent (5.2 mg/100 μL). For theoretical VCD predictions, conformational analysis was undertaken using the Monte Carlo search method, which resulted in two conformations for 45 and four conformations for 46. Further geometry optimizations of these conformers at the B3LYP/6-31G∗ and B3LYP/DGDZVP levels indicated that only one conformer each accounts for > 90% population in these molecules. Additional support for these conformers was derived from a comparison with the X-ray structure of 4,10,11-tribromo-10,11-seco-lippifoli-1(6)-en-5-one and 1 H NMR coupling constants. For (4R, 9R, 10R)-45, VCD spectra were predicted at three different levels of theory (B3LYP/6-31G∗ , B3LYP/DGDZVP, and B3PW91/DGDZVP2) and all of them were considered to provide good correspondence with the experimental VCD spectra of (+)-45, thereby suggesting the possible conﬁguration as (+)-(4R, 9R, 10R)-45. For (4R, 9S , 10R)-46, VCD spectra were predicted at two different levels of theory (B3LYP/6-31G∗ and B3LYP/DGDZVP), and both of them were considered to provide good correspondence with the experimental VCD spectra of 46 [the sample studied was not identiﬁed with (+)/(−) optical rotation].

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

11.3.10. Halogenated Sesquiterpenes 11.3.10.1. Pacifenol-Related Compounds. The structure and absolute conﬁguration of pacifenol {2,7-dibromo-8-chloro-2,5,7,8,9,9a-hexahydro-5,8,10,10tetramethyl-6H-2,5a-methano-1-benzoxepin-5-ol} (47) (see Scheme 11.3), has been documented in the literature as (2R, 5R, 5aR, 7S , 8S , 9aS )-(−). When treated with sodium hydroxide, 47 yields dibrominated chamigrene (48). When 48 is treated with mchloroperbenzoic acid, it is transformed into a α-bromoketone (49) {(−)-(2R, 5R, 5aR, 8ζ , 9aS )−2,8-dibromo-2,5,9,9a-tetrahydro-5-hydroxy-5,8,10,10-tetramethyl-6H-2,5a-methano-1-benzoxepin-7(8H )-one}, whose absolute conﬁguration at the 8 position could be either (S ) or (R). Although the absolute conﬁguration of 49 could not be determined from X-ray diffraction data, the two diastereomers, (2R, 5R, 5aR, 8S , 9aS )-49 and (2R, 5R, 5aR, 8R, 9aS )-49, could be distinguished using combined experimental and theoretical VCD investigations [61]. The experimental VCD spectra were obtained in CDCl3 (8.5 mg/150 μL). For the theoretical studies, Monte Carlo conformational search, done separately for the two diastereomers, (2R, 5R, 5aR, 8S , 9aS )-49 and (2R, 5R, 5aR, 8R, 9aS )-49, indicated three conformations each. Further optimization of the geometries of these conformers at the B3LYP/DGDZVP level reduced the number of low-energy conformers to two each. Theoretical VCD spectra, obtained as population-weighted spectra of the conformers, for each of the two diasteromers, were compared to the experimental VCD spectrum. In addition to the visual comparison of VCD plots, experimental rotational strengths were plotted against predicted rotational strengths for each of the two diastereomers. These analyses suggested that the absolute conﬁguration of the reaction product 49 is (2R, 5R, 5aR, 8R, 9aS ). An interesting point to note is that in the 8R diastereomer, the cyclohexanone exists in a predominantly boat conformation, which is contrary to the notion of the chair form being more stable than the boat form. 11.3.10.2. Majapolene B (50) and acetylmajapolene B (51). These two brominated sesquiterpenes have moderate antibacterial activity against some marine bacteria. The presence of bromine atom in the chemical structure makes these compounds suitable for the X-ray determination of absolute conﬁguration, but it was not easy to obtain the good-quality crystals of these compounds. The experimental VCD spectra (∼0.1 M in CCl4 ) measured for (−)-50 and (−)-51, when analyzed [62] with corresponding theoretical VCD spectra, revealed their conﬁgurations to be (7S , 10S ). To arrive at this conclusion, conformational search was used to identify the low energy conformers, which were then optimized at the B3PW91/6-31G(d , p) level and their VCD spectra were calculated. The population-weighted predicted VCD spectrum for (7S , 10S ) conﬁguration was considered to compare well with the experimental VCD spectra of (−)-enantiomer. Therefore it was suggested that the conﬁgurations of the both sesquiterpenes studied are (−)-(7S ,10S ).

11.3.11. Endoperoxides 11.3.11.1. Acetylmajapolene A (52). The absolute conﬁguration of endoperoxides is usually derived by converting them to diol derivatives via reductive cleavage and using the established practice for determining the conﬁgurations of secondary alcohols. Such reactions may sometimes produce tertiary alcohols. These difﬁculties were avoided [63] for 52. This natural product, isolated from red alga Laurencia, contained a disatereomeric mixture, and these diastereomers could be separated on a CHIRALPAK column. The ﬁrst eluted diastereomer (A) and second eluted diastereomer (B) have the

407

408

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

same sign for optical rotation. The NMR spectra of the diastereomers are similar, so their stereochemical assignments were not possible. The absolute conﬁgurations of these diastereomers were assigned [63] using VCD spectral analysis. The experimental VCD spectra of the two diastereomers (0.15 M in CCl4 ) appeared similar except that the sign associated with a band at ∼1050 cm−1 is opposite for the two samples. A comparison of the experimental VCD spectra of the two diastereomers with predicted VCD spectra for (1R, 4R, 7S , 10S ) and (1S , 4S , 7S , 10S ) conﬁgurations indicated that the ﬁrst eluted diasteromer has the (1R, 4R, 7S , 10S ) conﬁguration and second eluted diasteremoer has the (1S , 4S , 7S , 10S ) conﬁguration. Absolute conﬁgurations at the 7 and 10 positions were selected as S based on those established for related metabolites 50 and 51.

11.3.12. Furochromones 11.3.12.1. 5-O-Methylvisamminol (53). The structure of 53 was obtained from X-ray diffraction analysis, while its absolute conﬁguration was derived from the experimental VCD spectrum in CDCl3 (5.3 mg in 150 μl) by comparing it with theoretical VCD predictions [64]. For the theoretical VCD spectra, initial conformational analysis included Monte Carlo search, which yielded 14 conformations. The geometries of these conformations were further optimized at the B3LYP/DGDZVP level and VCD spectra calculated at the same level. The population-weighted predicted VCD spectrum, at the B3LYP/DGDZVP level, for (S )-53 provided good correspondence with the experimental spectra of (+)-53, thereby establishing the conﬁguration as (+)-(S )-53. 11.3.12.2. 5-O-Methylvisamminol Derivatives. Two new dihydrofurochromones, (+)-4 -O-angeloyl-5-O-methylvisamminaol, (+)-54, and (+)-4 -O-senecioyl5-O-methylvisamminaol, (+)-55, were extracted from the roots of Prionosciadium thapsoides, a medicinal plant, found in Mexico. Reaction of (+)-53 with angeloyl and senecioyl chloride provided (+)-54 and (+)-55, and therefore the absolute conﬁgurations of these compounds were assigned as (S ) by chemical correlation to the parent molecule (+)-53. In the same chemical correlation approach the absolute conﬁguration of visamminol 56 was assigned [64] as (+)-(S )-56.

11.3.13. Eremophilanoids 11.3.13.1. 6-Hydroxyeuryopsin (57) and Its Acetyl Derivative 58 [65]. 57, isolated from Senecio toluccanus, and 58 have anti-feedant activity against certain insects. Experimental VCD spectrum for 57 was measured in CCl4 (5 mg/100 μL). The optical rotation sign was not provided to label this natural product as (+)/(−). For VCD spectral predictions, six conformers were identiﬁed via Monte Carlo search, and the geometries of these conformers were further optimized at the B3LYP/6-31G* level of theory. The energies of optimized structures indicated two conformers accounting for ∼95% population. The population-weighted predicted VCD spectrum from these two conformers with (4S , 5R, 6S ) conﬁguration was considered to be in good agreement with the experimental spectrum, except in the 1200-cm−1 region, which was attributed to intermolecular hydrogen bonding of the hydroxyl group. To eliminate this hydrogen bonding effect, the VCD spectrum of 58 was measured in CCl4 (5.7 mg/100 μL). The predicted VCD spectrum, following the conformational analysis similar to that of parent 6-hydroxy euryopsin, for the structures with (4S , 5R, 6S ) conﬁguration was considered to be in good agreement with the experimental spectrum including the 1200-cm−1 region.

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

Therefore the absolute conﬁguration of the natural product, 57, and of its acetylated derivative 58, were suggested [65] to be (4S , 5R, 6S ). Additional calculations at the B3PW91/DGDZVP level conﬁrmed this assignment. 11.3.13.2. Eremophilanolide (59). The chemical name [66] of 59 is 1αangeloyloxy-8β,10β-dihydroxyeremophil-7(11)-en-8α,12-olide. The experimental VCD spectra for this compound were measured in CCl4 (4.8 mg/100 μL). The theoretical VCD spectra were predicted in the same way as for 57, but only a single conformer was found to be stable. The predicted VCD spectra for the (1S , 4S , 5R, 6S , 8S , 10S ) conﬁguration, at both the B3LYP/6-31G∗ and B3PW91/DGDZVP levels of theory, were considered to be in good agreement with the experimental VCD spectrum. Hence the absolute conﬁguration of the natural product 59 was suggested [65] to be (1S , 4S , 5R, 6S , 8S , 10S ).

11.3.14. Eudesmanolides 11.3.14.1. 1-Hydroxy-15-acetoxyeudesm-11(13)-en-6,12-olide (60) [67]. This was a new compound isolated from Milkania. The molecular formula was deduced from a high-resolution mass spectrum, and its structure and relative conﬁgurations were deduced from 1 H, 13 C NMR, and coupling constants. The conformation in solid state, deduced from X-ray diffraction, was found to be essentially same as that obtained from molecular modeling. Monte Carlo conformational search identiﬁed 40 conformers, and this number was reduced to 8 on further optimization at the B3LYP/6-31G∗ level. These eight conformers were further optimized at the B3LYP/6-31G∗∗ level, and their VCD spectra were calculated. The population-weighted predicted VCD spectrum at the B3LYP/6-31G∗ level was compared to the experimental VCD spectrum, which was obtained at a concentration of 10 mg/20 μL CDCl3 . The predicted VCD spectrum for (1S , 4S , 5S , 6S , 7S , 10R)-60 was considered to be in good agreement with the experimental VCD spectrum of (+)-60. This assignment, (1S , 4S , 5S , 6S , 7S , 10R)-(+)-60, was considered [67] to be in agreement with those known for many eudesamanolides.

11.3.15. Presilphiperfolanes 11.3.15.1. 9-Epi-presilphiperfolan-1-ol, (61) [68]. A natural product constituent of the oil of Anemia tomentosa var. anthriscifolia was identiﬁed [69], from a comparison of its 13 C NMR chemical shifts with those of presilphiperfolan-1-ol (62), whose crystal structure was known, as (−)-epi-presilphiperfolan-1-ol (63). A further evaluation, based on X-ray diffraction data and VCD analysis, of the natural product led to a change in the identity of its structure. The experimental VCD spectra of the natural product were obtained in CCl4 (22.0 mg/150 μL). Using the relative conﬁguration and structure from X-ray diffraction data of the natural product, Monte Carlo conformational search was conducted for the natural product with (1S , 4S , 7R, 8R, 9S ) conﬁguration. This search yielded nine low-energy conformers and this number of conformers was reduced to six based on single-point energy calculations at the B3LYP/6-31G∗ level. The geometries of these six conformers were optimized at the B3LYP/DGDZVP level followed by VCD calculation. The population-weighted predicted VCD spectrum for (1S , 4S , 7R, 8R, 9S ) conﬁguration was considered to have a good correlation to the experimental VCD spectrum of the natural product, which led to the reassignment [68] of the identity of natural product, from (−)-63, to (−)-(1S , 4S , 7R, 8R, 9S )-61.

409

410

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

11.3.16. Longipinane Derivatives 11.3.16.1. 7,9-Diacetyloxylongipin-2-en-1-one (64) [70]. The absolute conﬁguration of this compound was established [71] as (4R, 5S , 7R, 9R, 10R, 11R), from a comparison of ECD spectra with its parent (without acetyl groups) and vulgarone B. This assignment was conﬁrmed using combined experimental and theoretical VCD spectral investigation. The experimental VCD spectra were obtained in CCl4 (7 mg/150 μL). Theoretical VCD spectra were obtained for (4R, 5S , 7R, 9R, 10R, 11R) conﬁguration. Geometries of two low-energy conformers identiﬁed via Monte Carlo conformational search were optimized at the B3LYP/6-31G∗ level. These optimized geometries were further optimized at the B3LYP/DGDZVP level, and VCD predictions were made at the B3LYP/DGDZVP and B3PW91/DGDZVP levels of theory. The predicted VCD spectra at both levels were considered to be in good accordance with the experimental VCD spectra. This observation was used to conﬁrm the literature assignment. VCD spectra for two additional diasteromers, (4R, 5S , 7R, 9S , 10R, 11R) and (4R, 5S , 7S , 9R, 10R, 11R), were also predicted to compare the sensitivity of VCD to epimeric stereoisomers. It was stated that signiﬁcant differences are evident in the 1250- to 950-cm−1 region for the C7 epimer and in the 1100- to 950-cm−1 region for the C9 epimer. 11.3.16.2. 7β –Angeloyloxy-8α-isovaleroyloxylongipin-2-en-1-one, (65) and 7β,8α-diacetyloxylongipin-2-en-1-one (66). The methanolic extract of the plant Stevia monardifolia yielded 65 along with two known compounds, 7β,8α–diangeloyloxy longipin-2-en-1-one and 7β,8α–diangeloyloxylongipinan-1-one. The mass spectral data, 1 H and 13 C NMR data of 65, along with crystal structure of 7β,8α–dihydroxylongipin-2-en-1-one, were used to determine the relative stereochemistry. The absolute conﬁguration was suggested [72] from the comparison of experimental and predicted VCD spectra. To obtain the predicted spectra, a Monte Carlo conformational search yielded 83 conformers. This number of conformers was reduced to 18 and 13 when optimized at the B3LYP/6-31G∗ level the B3LYP/DGDZVP level, respectively. The population-weighted VCD spectra of these 13 conformers at the B3LYP/DGDZVP level was then compared to the experimental VCD spectrum. The experimental VCD spectrum of (+)-65 was considered to compare well with that predicted for (4R, 5S , 7S , 8S , 10R, 11R)-65 at the B3LYP/DGDZVP level, leading to the assignment of natural product as (4R, 5S , 7S , 8S , 10R, 11R)-(+). The diacetyl derivative, 66, was used to further conﬁrm this assignment, where the experimental VCD spectrum of (+)-66 was considered to compare well with the population-weighted predicted VCD spectrum of (4R, 5S , 7S , 8S , 10R, 11R)-66.

11.3.17. Cruciferous Phytoalexins 11.3.17.1. Dioxibrassinin (67), and 3-Cyanomethyl-3-hydroxyoxindole (68). These compounds were synthesized [73] in racemic form and the (−)-enantiomer separated using chiral HPLC. There are no established spectra-structure correlations that can determine the absolute conﬁguration at C3 , containing the tertiary alcohol group, of oxindole ring. Therefore, (−)-67 and (−)-68 were subjected to VCD spectroscopic investigations. A model compound 3-hydroxy-3-methyloxindole, 69 was also investigated for the sake of facilitating spectral interpretations. Experimental VCD spectra were measured in DMSO-d6 (∼2 M). For obtaining the theoretical VCD predictions, several low energy conformers were selected from conformational analysis and their geometries re-optimized at B3LYP/6-31G(d , p) level. Although the

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

experimental VA spectra showed a simple pattern in the 1900- to 1400-cm−1 region, with three bands at 1730, 1620, and 1470 cm−1 , the associated VCD band signs differed among the three molecules. The 1730-cm−1 band originated from the C=O stretching vibration and 1620-cm−1 band originated from the aromatic ring stretching vibration. The population-weighted predicted VCD spectra were considered to compare well with the corresponding experimental spectra for (−)-(S )-67, (−)-(S )-68 and (−)-(S )-69. 11.3.17.2. Spirobrassinin (70), 1-Methoxyspirobrassinin (71), and 1-Methoxyspirobrassinin methyl ether (72). The absolute conﬁgurations of 71 and 72 were not known until they were determined using VCD spectroscopy [74]. These compounds were synthesized in racemic form, and the enantiomers separated using chiral HPLC. VCD spectra were measured in CDCl3 at a concentration of 0.15 M. Most of the VCD bands measured for (+)-71 were found to have signs opposite to those found for (S )-(−)-70. Therefore the absolute conﬁguration of 71 was established as (R)-(+). This assignment was supported by the observation of opposite signed ECD bands for (+)-71 and (S )-(−)-70. In the case of 72, such correlation could not be used due to the differences in their chemical structures. Therefore, theoretical VCD predictions were undertaken. The 11 conformations, identiﬁed using a conformational analysis program [17], were further optimized at B3LYP/6-31G(d , p) level, and four lowest-energy conformers were used to obtain theoretical VCD at the same level. The population-weighted predicted VCD spectrum for (2R, 3R)-72 was considered to provide a reasonable correlation to the experimental VCD spectrum of (−)-72, thereby assigning the conﬁguration of 72 as (2R, 3R)-(−). This conclusion was further supported by stereochemical transformation reactions of (−)-72 and (+)-72. 11.3.17.3. Brassicanal C (73) [75]. The chirality of this natural product arises from the sulﬁnate group attached at the C-2 position. The racemic form of 73 was synthesized and its enantiomers separated using chiral HPLC. The experimental VCD spectra were measured in CDCl3 (0.05 M). The geometries of 36 conformers of 73 identiﬁed by manually changing the dihedral angles were ﬁrst optimized at the B3LYP/631G(d ) level, and the resulting nonidentical conformations were further optimized at the B3LYP/6-311+G(2df , 2p) level. The four lowest-energy conformers obtained were used to calculate VCD. The population-weighted predicted VCD spectrum for (S )-73 was considered to compare reasonably well with the experimental VCD spectrum of (−)-73, resulting in the assignment of its absolute conﬁguration as (−)-(S ). This conclusion was further supported by comparing the experimental and theoretical predictions of optical rotation and ECD for (S )-(−)-73.

11.3.18. Furanones [76–78] Most of the naturally occurring furanones are often isolated as racemic compounds due to racemization of the enantiomers caused by keto–enol tautomerism. But it was possible to separate the enantiomers by HPLC. The absolute conﬁguration of 2,5-dimethyl-4hydroxy-3-(2H )-furanone, 74 (see Scheme 11.4), was difﬁcult to determine because of its rapid racemization through keto–enol tautomerism. This problem was overcome by carefully converting it to its methyl ether, 2,5-dimethyl-4-methoxy-3(2H )-furanone, 75, determining the absolute conﬁguration of the later and relating it back to the parent. The absolute conﬁguration of 75 was determined by comparing the experimental VCD spectrum (0.17 M in CCl4 ) with the predicted VCD spectrum at B3PW91/6-31G(d , p)

411

412

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

level. There was only one stable conformer for this molecule. The predicted spectrum for this conformer with (R)-conﬁguration was found to be identical to the experimental VCD spectrum of (+)-enantiomer, so the absolute conﬁguration of 75 was assigned as (+)-(R). Since careful reaction of (+)-74 with diazomethane yielded (+)-(R)-75, the absolute conﬁguration of the former was also deduced as (+)-(R). In the case of 4-acetoxy-2,5-dimethyl-3(2H )-furanone, 76, conformational analysis [17] yielded four conformations and optimization of their geometries at B3PW91/6-31G(d , p) level resulted in two low-energy conformers. The population-weighted predicted VCD spectrum for (R)-conﬁguration was considered to compare well with the experimental VCD spectrum (0.08M in CCl4 ) of (+)-enantiomer, suggesting the absolute conﬁguration as (R)-(+)-76. The absolute conﬁgurations of 5-ethyl-4-hydroxy-2-methyl-3(2H )-furanone (77) and 2ethyl-4-hydroxy-5-methyl-3(2H )-furanone (78) are also difﬁcult to determine because of their rapid racemization through keto–enol tautomerism. Just as in the case of 74, this problem was overcome by determining the absolute conﬁguration of their methyl ethers 79 and 80 using VCD spectroscopy and relating them back to the parent molecules. The absolute conﬁgurations determined in this manner were (R)-(+) for all of these compounds.

11.3.19. Furanocoumarins [79] Furanocoumarins contain a furan ring fused with coumarin. The furan ring may be fused in different ways, resulting in linear and bent structures. (+)-5,8-Dimethoxymarmesin, referred to as (+)-alternamin, a furanocoumarin isolated from the plant, Murraya alternans, was found to have antidote activity against snake venom. The chemical structure of alternamin, 81, was deduced from high-resolution mass spectra and NMR. While three different possibilities exist for fusing the furan ring (two linear (81-(a), 81-(b)) and one bent (81-(c)) structural forms), the NMR data suggested the linear form 81-(a), but the absolute conﬁguration was not known. The absolute conﬁguration of (+)-alternamin was established using VCD spectroscopy. The experimental VCD spectra measured in CDCl3 solvent were compared to the predicted VCD spectra for the three possible structural isomers, (a)–(c), with (S )-conﬁguration. In each case, conformational analysis [17] yielded multiple conformations. The low-energy conformers were subjected to geometry optimizations at B3PW91/6-311++G(d , p) level. The population-weighted predicted VCD spectrum for the (S )-conﬁguration of structure (a) was considered to yield the best agreement with the experimental VCD spectrum of (+)-81. Therefore the absolute conﬁguration of alternamin was assigned as (S )-(+)-81-(a).

11.3.20. Klaivanolide [80] Although the structure of 82 has been established from NMR spectra, its absolute conﬁguration was not. VCD spectroscopy was used [80] to establish the absolute conﬁguration of 82. The experimental IR and VCD spectra, in the 1900- to 800-cm−1 range, were obtained for (+)-82 in CDCl3 (0.03 M). Initial conformational analysis revealed 33 conformers. Further geometry optimization of these conformations using B3LYP functional and 6-31G∗ basis set yielded 24 stable conformations, of which seven have populations greater than 4%. Subsequent geometry optimizations with B3PW91 functional and TZ2P basis set identiﬁed ﬁve lowest energy conformers, all resulting from rotations around C7 -C8 , C8 -O9 and C5 -O18 bonds. The VA and VCD spectra were predicted for the ﬁve lowest–energy conformers of (7S )-82 using the B3PW91 functional and the TZ2P basis

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

set and the population-weighted spectra compared to the experimental spectra. The major features in the experimental VCD spectrum are one broad positive VCD band in the 1700 to 1800-cm−1 range (which is associated with C=O stretching vibrations) and a bisignate couplet with positive VCD at 1275 cm−1 and negative VCD at 1232 cm−1 (associated with C–O bond stretching vibrations). These features were considered to be reproduced well in the predicted VCD spectrum for (S )-82 with one notable difference. The predicted VCD spectrum in the 1700 to 1800-cm−1 range has resolved positive VCD bands, while the corresponding region in the experimental spectrum was broad and unresolved. Nevertheless, the satisfactory, yet qualitative, agreement between experimental VCD spectrum of (+)-82 and predicted VCD spectrum of (7S )-82 was used to suggest that the absolute conﬁguration of the studied natural product is (+)-(7S )-82.

11.3.21. Pheromones 11.3.21.1. 1-Acetoxymethyl-2,3,4,4-tetramethylcyclopentane (83). The relative conﬁguration of 83, a sex pheromone, was known from NMR spectra, but the absolute conﬁguration was not known until VCD spectral investigations were undertaken [81]. The experimental VCD spectrum (0.90 M in CDCl3 ) of (+)-83 was compared to the predicted VCD spectra for 83 with (1S , 2S , 3R)- and (1R, 2R, 3S )-conﬁgurations. A conformational search indicated 15 conformations, and population-weighted predicted VCD spectrum of (1S , 2S , 3R)-83 and experimentally measured VCD spectrum of (+)-83 were considered to be in good agreement leading to the absolute conﬁguration assignment as (1S , 2S , 3R)-(+)-83. 11.3.21.2. Frontalin (1,5-dimethyl-6,8-dioxabicyclo[3.2.1]octane), 84. The Southern pine beetle, Dendrooctoonus frontalis, uses (1S , 5R)-(−)-84 as the active pheromone. Since the absolute conﬁguration of this molecule was already known [82], this compound was useful to test the predictive capabilities of VCD. The experimental VCD spectra were measured for (+)-84, and theoretical spectra were calculated [83] for (1R,5S )-84. The conformer with a six-member ring in a chair conformation and a seven-membered ring in a boat conformation was found to be energetically favored. The VCD spectrum predicted at the B3LYP/6-31G∗ level for this stable conformation of (1R, 5S )-84, but not that of (1S , 5R)-84, was considered to be in good agreement with the experimental VCD spectrum observed for (+)-84. Also, the observed rotational strengths correlated well to the predicted rotational strengths of (1R, 5S ), but not to those of (1S , 5R). Thus VCD spectroscopy conﬁrmed the previously known conﬁguration of 84 as (1R, 5S )-(+).

11.3.22. Norlignan Norlignan is a class of phenylpropanoids. Hinokiresinol, belonging to this class, can have an E or Z double bond in its molecule. The Z -isomer is named nyasol, 85, and the E -isomer as hinokiresinol (sometimes as E -hinokiresinol), 86. There were some uncertainties regarding the absolute conﬁgurations of these compounds. To resolve these uncertainties, VCD spectroscopic investigation was undertaken [84] for 85. The experimental VCD spectra of (+)-85 were measured in DMSO-d6 solvent (5 mg/100 μL). Although the experimental VCD measurements were also done for KBR pellet, artifacts were suspected in the measured spectra. Conformational search and subsequent geometry optimizations at B3LYP/6-31G∗∗ level of theory identiﬁed eight lowest-energy conformers. Gas-phase predicted conformations did not differ signiﬁcantly from solution-phase

413

414

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

predicted conformations obtained with PCM. The population-weighted predicted VCD spectra in the gas phase at B3YP/6-31G∗∗ , and B3LYP/aug-cc-pVDZ levels of theory, and in the solution phase at B3LYP/6-31G*∗ (using PCM) were compared to the experimental spectra. The solvent absorption restricted the measurements to the 1800- to 1100-cm−1 region where only one dominant positive VCD band was found at 1510 cm−1 . This experimentally observed positive VCD band was considered to reproduce the corresponding band in the predicted spectrum for (S )-85 leading to the assignment of conﬁguration as (S )-(+). Based on this conclusion, and the chemical conversions, the absolute conﬁguration of 86 was suggested as (S )-(−).

11.3.23. Taxol Paclitaxel (Taxol) is a natural product derived in a very low yield from the Yew tree, Taxus brevifolia, and is used as an anticancer drug. The total synthesis of taxol involves complex reactions, and therefore commercial production was not economically feasible. For this reason, another precursor natural product, baccatin III, readily available in large quantities attracted much attention for the synthesis of paclitaxel. It was found that the VCD spectra of taxol and baccatin III (both obtained at 0.029 M in CDCl3 ) are quite similar [85]. The conformational analysis of baccatin III indicated 23 lowest-energy conformations, which upon further geometry optimizations yielded three lowest-energy conformers that account for 97% of the population. The population-weighted predicted VCD spectrum of baccatin III was considered to reproduce the experimental VCD spectrum quite well. Therefore the three lowest-energy conformers used for predicting the VCD spectrum were considered to be the predominant conformations of baccatin III in solution. The electron micrograph of the crystal structure of the paclitaxel–tubulin complex, which is considered to be a bioactive structure, indicated a conformation that is among the unstable conformations noted for baccatin III in solution. Based on this observation, it was suggested that a conformational change occurs in paclitaxel during its binding with tubulin.

11.3.24. Ginkgolides Ginkgo biloba is one of oldest living trees. Extract from the Ginkgo plant leaves and roots contains different ginkgolides: ginkgolide A (GA), ginkgolide B (GB), ginkgolide C (GC), ginkgolide J (GJ), and ginkgolide M (GM). This extract is often used to prevent the development of dementia or to improve focus. 11.3.24.1. Nature of Interactions Between Ginkgolides and Amyloid Peptide. The interactions with amyoid peptides were investigated [86] using amyloid Aβ(25–35) peptide, as a model for the full-length peptide. GA, GB, GC, and bilobalide (BB) and two ethers, GA-monoether and GA-diether, were investigated using VCD spectra. The experimental VA and VCD spectra of GA, GB, GC, BB, GA-monoether, and GA-diether were analyzed in conjunction with density functional theoretical predictions. The time-dependent experimental spectra of Aβ(25–35) peptide and the corresponding experimental spectra in the presence of ginkgolides indicated that the inﬂuence of ginkgolides in modulating the aggregation of Aβ(25–35) peptide is relatively minor. Such minor effects may have been due to the absence of a speciﬁc interaction with Aβ(25–35) peptide. It was suspected that the therapeutic effect of Ginkgo biloba extract may have been more complex.

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

11.3.24.2. Ginkgolide B. The experimental infrared absorption and VCD spectra of GB, in combination with quantum mechanical calculations, were used to suggest [87] the preferred conformations of GB in solution. The experimental data were obtained separately for GB in dimethyl sulfoxide-d6 (3.9 mg/20 μL), in CD3 CN (1.8 mg/50 μL) solvents, and as KBr pellet (0.5 mg in 250 mg KBr). Quantum chemical calculations were undertaken at the B3LYP/6-31G∗ level for GB and also for GA, GC, GJ, and GM to assess the possibility of distinguishing ginkgolides form their spectra. Although a number of diastereomers are possible for ginkgolides, two diastereomers with inversion at C1 (labeled GB-C1 i) and C10 (labeled GB-C10 i) were considered for GB. Two conformers within 7–8 kJ/mol were identiﬁed each for GB, GC, and GM. Only the ﬁrst lowest-energy conformer was used in further calculations, for GA, GJ, GBC1 i, and GBC10 i, while two lowest-energy conformers were used for GB, GC, and GM. Based on the comparison of predicted VCD spectra for different ginkgolides and assuming that the experimental spectra can be closely approximated by the corresponding predicted VCD spectra, it was suggested that VCD spectroscopy might allow discrimination between GB and other ginkgolides. Furthermore, the predicted VCD spectra for GB markedly differed from those of GBC1 i and GBC10 i, and this observation prompted the suggestion that the experimental VCD spectra might be able to provide discrimination among the diastereomers of GB. The structure of lowest-energy conformer of GB optimized with 6-31G∗ basis set agreed with its crystal structure. There are similarities between experimental VCD spectra of GB and theoretical VCD spectra predicted for the lowest-energy conformer of GB. But there are also differences between them, and these differences appeared to correspond to the predicted features for the second lowest-energy conformer. Based on this observation, it was suggested that the two lowest-energy conformers of GB are probably present in solution.

11.3.25. Peptides Although peptides represent a large class of molecules, a few peptides are also natural products. Among these natural products, pexiganan and cyclosporin have been studied using VCD spectroscopy. 11.3.25.1. Pexiganan. Pexiganan is an analogue of the magainin family of antimicrobial peptides and contains 22 amino acids (Gly-Ile-Gly-Lys-Phe-Leu-Lys-Lys-Ala-LysLys-Phe-Gly-Lys-Ala-Phe-Val-Lys-Ile-Leu-Lys-Lys). This peptide is present in the skin of the African clawed frog. Conformational analysis of pexiganan was conducted [88] using the experimental VCD studies in different solvents (triﬂuoroethanol, methanol, D2 O and DMSO-d6 ) and supplemented with IR and ECD measurements. All these spectroscopic measurements suggested that the pexiganan peptide has the tendency to adopt different structures in different environments: an α-helical conformation in TFE, a sheet-stabilized β-turn structure in methanol, a random coil with β-turn structure in D2 O, and a solvated β-turn structure in DMSO-d6 . 11.3.25.2. Cyclosporins. These are a group of cyclic peptides of 11 amino acids isolated from fungi. They differ from each other in the identity of one amino acid and contain an unnatural d-amino acid. To determine the solution state conformations of cyclosporin A (CsA), cyclosporin D (CsD), cyclosporin G (CsG), and cyclosporin H (CsH), VCD spectral investigations were undertaken [89]. The experimental IR and VCD spectra in the amide I region were measured for free cyclosporins (20 mM) and

415

416

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

their Mg2+ complexes (20 mM with 10-fold excess magnesium perchlorate) in CD3 CN. The spectra in the N–H stretching region for free cyclosporins were measured at 2 mM in CDCl3 . IR and VCD spectra were predicted for three different fragments, with side chains reduced to methyl groups and adding terminal methyl groups. The geometry optimizations and IR and VCD spectra were predicted at the BPW91/6-31G∗ level. Based on the comparison of experimental spectra with the predicted IR and VCD spectra of fragments, it was concluded that the conformers of CsA, CsD, and CsG in solution are similar to those found in crystal structures. The presence of additional conformers with decreased intramolecular hydrogen bonding was suggested for CsH in solution. Evidence for disruption of β-sheet structure for CsC and CsH in CDCl3 and for CsA in CD3 CN was suggested. The spectral alterations seen in the presence of Mg2+ ions indicated strong interactions between metal ions and cyclosporin.

11.3.26. Axially Chiral Natural Products 11.3.26.1. Dicurcuphenol B (87) and Dicurcuphenol C (88). The curcuphenol dimers contain both axial chirality (due to hindered rotation around the bond between two phenyl rings) and central chiraity (due to chiral centers in the alkyl side chain). 87 and 88 have the same alkyl side chains and differ only in their axial chirality. They both exhibit positive speciﬁc rotations at the sodium D line. The experimental VCD spectra were obtained at 0.23M for 87 and at 0.11 M for 88, both in CDCl3 . Some simpliﬁcations were made for determining the absolute conﬁguration using VCD spectroscopy [90]. (a) The VCD spectrum for structural analysis was generated by taking one-half of the experimental VCD spectral difference, (Dicurcuphenol B - Dicurcuphenol C)/2, with the assumption that this difference reﬂects the VCD spectrum due to axial chirality. (b) Since (S )-(+)-curcuphenol itself did not exhibit any VCD features in the 1600- to 1425-cm−1 region, and the VCD features seen in this region for Dicurcuphenol B and Dicurcuphenol C overlapped with each other, VCD features in this region are considered to be arising from axial chirality. (c) To avoid conformational mobility associated with an alkyl side chain, a model compound 89 that represents axial chirality but contains no side chain was used for predicting VCD spectra at the B3LYP/6-31G∗ level. The predicted VCD spectrum for the model compound with P chirality was considered to match well with the above-mentioned difference spectrum in the 1600- to 1425-cm−1 region. Therefore 87 and 88 were assigned P and M conﬁgurations, respectively. 11.3.26.2. Gossypol (90). 90 exists as an aldehyde in CDCl3 solvent and as an equilibrium mixture of tautomers in some other solvents [91]. Based on the ECD spectral interpretations using approximate theoretical methods, the absolute conﬁguration of gossypol enantiomers were suggested [92] in the literature to be (M )-(−) and (P )(+). This assignment has been conﬁrmed by VCD spectroscopy [93] and, in addition, predominant conformations of 90 have been identiﬁed. The experimental VCD spectra were measured for both enantiomers in CDCl3 solvent (5 mg of (+)-90 and 4 mg of (−)-90 each in 100 μL of CDCl3 ). For conformational analysis, at the B3LYP/6-31G∗ level, the rotational freedom associated with the aldehyde group and isopropyl group were considered. In addition, different intramolecular hydrogen bonding patterns among neighboring O–H groups and between O–H groups and oxygen of aldehyde group were considered. The population-weighted VCD spectrum, obtained from three low-energy conformers of (M )-90, was considered to correlate with the experimental VCD spectrum of (−)-90 and conﬁrmed the literature assignment.

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

11.3.26.3. Cephalochromin (91). Isolated from the culture broth of the endophytic fungus Pseudoanguillospora sp, 91 possesses the axial chirality of the atropisomeric naphthopyranone systems, in addition to central chirality elements as two tertiary stereogenic centers. This results in four diastereomers: (aS , 2R, 2 R), (aS , 2S , 2 S ), (aR, 2R, 2 R), and (aR, 2S , 2 S ). The axial chirality of (+)-91 was determined from ECD studies to be (aS ). Then two possibilities exist for the absolute conﬁguration of (+)-91: (aS , 2R, 2 R) and (aS , 2S , 2 S ). It was not possible to determine the nature of central chirality in the presence of overwhelming inﬂuence from axial chirality on its chiroptical properties. This problem was resolved [94] using VCD spectroscopy because vibrations attributable to naphthapyranone moieties (whose relative orientation determines the axial chirality) and stereogenic centers appear spatially separated at different vibrational frequencies. Based on the comparison between the experimental VCD spectra of (+)-91 and B3LYP/6-311G∗ predicted VCD spectra for (aS , 2R, 2 R)-91 and (aS , 2S , 2 S )-91, three VCD bands in the 1050- to 900-cm−1 region provided a means to distinguish the two diastereomers of 91, thereby allowing the determination of both types of chirality elements. The sign of the VCD band at 1038 cm−1 was used to label the absolute conﬁgurations of two stereoisomers of 91 as (+)589 -[(−VCD)1038cm−1 ]-(aS , 2R, 2 R) and (+)589 -[(+VCD)1038cm−1 ]-(aS , 2S , 2 S ).

11.4. CONCLUDING REMARKS VCD spectroscopy has evolved from a curiosity in 1970s into an accepted structural tool of the twenty-ﬁrst century for the determination of three-dimensional molecular structures of chiral molecules. A comparison of quantum chemical predictions of VCD with corresponding experimental observations is required for this purpose. In this process, however, VCD spectral calculations for all possible diastereomers, and all possible conformations in each case, must be undertaken. A conclusion on the absolute conﬁguration can be reached if the calculated VCD spectrum for only one of all possible absolute conﬁgurations matches the experimental VCD spectrum for an enantiomer of a given chiral compound. With the availability of commercial VCD instruments and increasingly advanced computational resources, the task of determining the absolute conﬁgurations using the combined experimental and quantum chemical VCD spectra is becoming a routine process.

ACKNOWLEDGMENTS I am grateful to Drs. Joseph-Nathan, K. Monde, and M. Urbanova for providing a list of their VCD publications on natural products. Funding from NSF (CHE-0804301) is gratefully acknowledged.

REFERENCES 1. N. Berova, K. Nakanishi, R. W. Woody, eds., Circular Dichroism: Principles and Applications, John Wiley & Sons, New York, 2000. 2. P. J. Stephens, J. Phys. Chem. 1985, 89 , 748–752; A. D. Buckingham, P. W. Fowler, P. A. Galwas, Chem. Phys. 1987, 112 , 1–14. 3. (a) Gaussian 09, Gaussian Inc., www.gaussian.com. (b) Turbomole, www.turbomole.com. (c) Amsterdam Density functional program; www.scm.com. (d) DALTON program; www.kjemi. uio.no/software/dalton.

417

418

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37.

P. J. Stephens, F. J. Devlin, J. J. Pan, Chirality 2008, 20 , 643–663. L. A. Naﬁe, Nat. Prod. Comm. 2008, 3 , 451–466. T. A. Keiderling, Nature 1986, 322 , 851–852. G. Shanmugam, P. L. Polavarapu J. Am. Chem. Soc. 2004, 126 , 10292–10295. L. A. Naﬁe, R. K. Dukor, Vibrational optical activity in chiral analysis, in Chiral Analysis, K. W. Busch and M. A. Busch, eds., Elsevier, Amsterdam, 2006. pp 505–544. J. R. Cheeseman, M. J. Frisch, F. J. Devlin, P. J. Stephens, Chem. Phys. Lett. 1996, 252 , 211–220. A. D. Becke, J. Chem. Phys. 1993, 98 , 5648–5652. R. Ditchﬁeld, W. J. Hehre, J. A. Pople, J. Chem. Phys. 1971, 54 , 724–728. J. Tomasi, B. Mennucci, R. Cammi, Chem. Rev . 2005, 105 , 2999–3094. P. L. Polavarapu, N. Jeirath, S. Walia, J. Phys. Chem. A 2009, 113 , 5423–5431. G. Scalmani and M. J. Frisch, J. Chem. Phys. 2010, 132 , 114110–114124. P. L. Polavarapu, G. Scalmani, E. K. Hawkins, C. Rizzo, N. Jeirath, I. Ibnusaud, D. Habel, D. S. Nair, S. Haleema, J. Nat. Prod . 2011, 74 , 321–328. Spartan ’08, Wavefunction, Inc., Irvine, CA; www.wavefun.com/products/spartan.html Conﬂex, CONFLEX Corporation, AIOS Meguro 6F, 2-15-19, Kami-Osaki, Shinagawa-ku, Tokyo 141-0021, Japan; www.conﬂex.us Hyperchem, Hypercube Inc., 1115 NW 4th street, Gainesville, FL 32601; www.hyper.com F. Mohammadi, N. G. J. Richards, W. C. Guida, R. Liskamp, M. Lipton, C. Cauﬁeld, G. Chang, T. Henderickson, W. C. Still, J. Comput. Chem. 1990, 11 , 440–467. (a) P. Zhang, P. L. Polavarapu, J. Phys. Chem. A 2007, 111 , 858–871; (b) J. He, P. L. Polavarapu, J. Chem. Theory Comput. 2005, 1 , 506–514; (c) P. L. Polavarapu, E. A. Donahue, G. Shanmugam, G. Scalmani, E. K. Hawkins, C. Rizzo, I. Ibnusaud, G. Thomas, D. Habel and D. Sebastian J. Phys. Chem A 2011, 115 , 5665–5673. A. Nakahashi, N. Miura, K. Monde, S. Tsukamoto, Bioorg. Med. Chem. Lett. 2009, 19 , 3027–3030. S. Abbate, G. Longhi, L. Ricard, C. Bertucci, C. Rosini, P. Salvadori, A. Moscowitz, J. Am. Chem. Soc. 1989, 111 , 836–840. E. Castiglioni, F. Lebon, G. Longhi, S. Abbate, Enantiomer 2002, 7 , 161–173. F. J. Devlin, P. J. Stephens, J. R. Cheeseman, M. J. Fritsch, J. Phys. Chem. A 1997, 101 , 9912–9924. F. J. Devlin, P. J. Stephens, J. R. Cheeseman, M. J. Fritsch, J. Am. Chem. Soc. 1996, 118 , 6327–6328. C. Guo, R. D. Shah, R. K. Dukor, T. B. Freedman, X. Cao, L. A. Naﬁe, Vibrational Spectrosc. 2006, 42 , 254–272. E. Castiglioni, F. Lebon, G. Longhi, S. Abbate, Enantiomer 2002, 7 , 161–173. M. Urbanova, V. Setnicka, K. Volka, Chirality 2000, 12 , 199–203. J. R. A. Moreno, F. P. Urena, J. J. L. Gonzalez, Vibrational Spectrosc. 2009, 51 , 318–325. J. R. A. Moreno, F. P. Urena, J. J. L. Gonzalez, Phys. Chem. Chem. Phys. 2009, 11 , 2459–2467. F. P. Urena, J. R. A. Moreno, J. J. L. Gonzalez, J. Phys. Chem. A 2008, 112 , 7887–7893. F. J. Devlin, P. J. Stephens, P. Besse, J. Org. Chem. 2005, 70 , 2980–2993. E. Burgue˜no-Tapia, L. G. Zepeda, P. Joseph-Nathan, Phytochem. 2010, 71 , 1158–1161. M. del H. Loandos, M. B. Villecco, E. Burgue˜no-Tapia, P. Joseph-Nathan, C. A. N. Catal´an, Nat. Prod. Comm. 2009, 4 , 1537–1545. (a) R. D. Singh, T. A. Keiderling, J. Am. Chem. Soc. 1981, 103 , 2387–2394; (b) P. Bour, J. McCann, H. Wieser, J. Phys. Chem. A 1998, 102 , 102–110. F. Maurer, H. Wieser, Proceedings of SPIE-The International Society for Optical Engineering 1992, 1575 (Int. Conf. Fourier Transform Spectrosc., 8th, 1991), 410–411. F. P. Urena, J. R. A. Moreno, J. J. L. Gonzalez, Tetrahed. Asym. 2009, 20 , 89–97.

D E T E R M I N AT I O N O F T H E S T R U C T U R E S O F C H I R A L N AT U R A L P R O D U C T S U S I N G V C D

38. E. Debie, L. Jaspers, P. Bultinck, W. Herrebout, B. van der Veken, Chem. Phys. Lett. 2008, 450 , 426–430. 39. E. Debie, L. Jaspers, P. Bultinck, W. Herrebout, B. van der Veken, Phys. Chem. Chem. Phys. 2008, 10 , 3498–3508. 40. V. P. Nicu, E. Debie, W. Herrebout, B. van der veken, P. Bultinck, E. J. Baerends, E. J. Chirality, 2010, 21, E287–E297. 41. T. Burgi, A. Vargaa, A. Baker, J. Chem. Soc. Perkin Trans. 2002, 2 , 1596–1601. 42. P. J. Stephens, J. J. Pan, F. J. Devlin, M. Urbanova, J. Hajicek, J. Org. Chem. 2007, 72 , 2508–2524. 43. P. J. Stephens, J. J. Pan, F. J. Devlin, M. Urbanova, O. Julinek, J. Hajicek, Chirality 2008, 20 , 454–470. 44. M. A. Mu˜noz, O. Mu˜noz, P. Joseph-Nathan, J. Nat. Prod . 2006, 69 , 1335–1340. 45. M. A. Mu˜noz, O. Mu˜noz, P. Joseph-Nathan, Chirality 2010, 22 , 234–241. 46. M. Reina, E. Burgue˜no-Tapia, M. A. Bucio, P. Joseph-Nathan, Phytochem. 2010, 71 , 810–815. 47. J. C. Cedr´on, A. Est´evez-Braun, A. G. Ravelo, D. Guti´errez, N. Flores, M. A. Bucio, N. P´erez-Hern´andez and P. Joseph-Nathan, Org. Lett. 2009, 11 , 1491–1494. 48. P. J. Stephens, J. J. Pan, F. J. Devlin, K. Krohn, T. Kurtan, J. Org. Chem. 2007, 72 , 3521–3536. 49. G. Albers-Schonberg, H. Schmid, Helv. Chim. Acta 1961, 44 , 1447–1473. 50. B. Elsasser, K. Krohn, M. N. Akhtar, U. Florke, S. F. Kouam, M. G. Kuigoua, B. T. Ngadgui, B. M. Abegaz, S. Antus, T. Kurtan, Chem Biodiversity 2005, 2 , 799–808. 51. P. J. Stephens, J. J. Pan, K. Krohn, J. Org. Chem. 2007, 72 , 7641–7649. 52. M. A. Mu˜noz, C. Areche, J. Rovirosa, A. San-Mart´ın, P. Joseph-Nathan, Heterocycles 2010, 81 , 625–635. 53. C. Areche, A. San-Mart´ın, J. Rovirosa, M. A. Mu˜noz, A. Hern´andez-Barrag´an, M. A. Bucio, P. Joseph-Nathan, J. Nat. Prod . 2010, 73 , 79–82. 54. M. A. Mu˜noz, C. Areche, A. San Mart´ın, J. Rovirosa, P. Joseph-Nathan, Nat. Prod. Comm. 2009, 4 , 1037–1040. 55. C. M. Cerda-Garc´ıa-Rojas, H. A. Garc´ıa-Guti´errez, J. D. Hern´andez-Hern´andez, L. U. Rom´anMar´ın, P. Joseph-Nathan, J. Nat. Prod . 2007, 70 , 1167–1172. 56. P. J. Stephens, D. M. McCann, F. J. Devlin, A. B. Smith, J. Nat. Prod . 2006, 69 , 1055–1064. 57. A. B. Smith, J. P. Konopelski, B. A. Wexler, P. A. Sprengeler, J. Am. Chem. Soc. 1991, 113 , 3533–3542. 58. Michalski, W. Kisiel, K. Michalska, V. Setnicka, and M. Urbanova, J. Mol. Struct. 2007, 871 , 67–72. 59. W. Kisiel, B. Barszcz, Phytochem. 1995, 39 , 1395. 60. C. M. Cerda-Garc´ıa-Rojas, C. A. N. Catal´an, A. C. Muro, P. Joseph-Nathan, J. Nat. Prod . 2008, 71 , 967–971. 61. M. A. Mu˜noz, C. Chamy, A. Carrasco, J. Rovirosa, A. San Mart´ın, P. Joseph-Nathan, Chirality 2009, 21 , E208–E214. 62. K. Monde, T. Taniguchi, N. Miura, C. S. Vairappan, M. Suzuki, Chirality 2006, 18 , 335–339. 63. K. Monde, T. Taniguchi, N. Miura, C. S. Vairappan, M. Suzuki, Tetrahed. Lett. 2006, 47 , 4389–4392. 64. J. M. Torres-Valencia, O. E. Ch´avez-R´ıos, C. M. Cerda-Garc´ıa-Rojas, E. Burgue˜no-Tapia and P. Joseph-Nathan, J. Nat. Prod . 2008, 71 , 1956–1960. 65. E. Burgue˜no-Tapia and P. Joseph-Nathan, Phytochem. 2008, 69 , 2251–2256. 66. E. Burgue˜no-Tapiaa, B. Hern´andez-Carlosb and P. Joseph-Nathan J. Mol. Struct. 2006, 825 , 115–123. 67. M. Krautmann, E. C. de Riscala, E. Burgue˜no-Tapia, Y. Mora-P´erez, C. A. N. Catal´an and P. Joseph-Nathan, J. Nat. Prod . 2007, 70 , 1173–1179.

419

420

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

68. P. Joseph-Nathan, S. G. Leit˜ao, S. C. Pinto, G. G. Leit˜ao, H. R. Bizzo, F. L. P. Costa, M. B. de Amorim, N. Martinez, E. Dellacassa, A. Hern´andez-Barrag´an and N. P´erezHern´andez, Tetrahedron Lett. 2010, 51 , 1963–1965. 69. S. C. Pinto, G. G. Leitao, H. R. Bizzo, N. Martinez, E. Dellacassa, F. Santos Jr., F. L. P, Costa, M. B. Amorim, S. G. Leitao, Tetrahedron Lett. 2009, 50 , 4785–4787. 70. C. M. Cerda-Garc´ıa-Rojas, E. Burgue˜no-Tapia, L. U. Rom´an-Mar´ın, J. D. Hern´andezHern´andez, T. Agull´o-Ortu˜no, A. Gonz´alez-Coloma and P. Joseph-Nathan, textitPlanta Med. (J. Med. Plant Res.), 2010, 76 , 297–302. 71. P. Joseph-Nathan, C. M. Cerda-Garcia-Rojas, R. E., del Rio, L. U. Roman-Marin, J. D. Hernandez, J Nat prod . 1986, 49 , 1053–1060. 72. R. E. Rojas-P´erez, E. Cedillo-Portugal, P. Joseph-Nathan and E. Burgue˜no-Tapia, Nat. Prod. Comm. 2009, 4 , 757–762. 73. K. Monde, T. Taniguchi, N. Miura, S. I. Nishimura, N. Harada, R. K. Dukor, L. A. Naﬁe, Tetrahedron Lett. 2003, 44 , 6017–6020. 74. K. Monde, T. Taniguchi, N. Miura, P. Kutschy, Z. Curillov´a, M. Pil´atov´a, J. Mojˇziˇs, Bioorg. Med. Chem. 2005, 13 , 5206–5212. 75. T. Taniguchi, K. Monde, K. Nakanishi, N. Berova, Org. Biomol. Chem. 2008, 6 , 4399–4405. 76. K. Monde, A. Nakahashi, N. Miura, Y. Yaguchi, D. Sugimoto, M. Emura, Chirality 2009, 21 , S110–S115. 77. M. Emura, Y. Yaguchi, A. Nakahashi, D. Sugimoto, N. Miura, K. Monde, J. Agric. Food Chem. 2009, 57 , 9909–9915. 78. Y. Yaguchi, A. Nakahashi, N. Miura, D. Sugimoto, K. Monde, M. Emura, Org. Lett. 2008, 10 , 4883–4885. 79. H. M. Min, M. Aye, T. Taniguchi, N. Miura, K. Monde, K. Ohzawa, T. Nikai, M. Niwa, Y. Takaya, Tetrahedron Lett. 2007, 48 , 6155–6158. 80. F. J. Devlin, P. J. Stephens, B. Figad`ere, Chirality 2009, 21 , E48–E53. 81. B. Figadere, F. J. Devlin, J. G. Millar, P. J. Stephens, Chem Comm. 2008, 9 , 1106–1108. 82. K. Mori, Tetrahedron 1975, 31 , 1381–1384. 83. C. S. Ashvar, P. J. Stephens, T. Eggimann and H. Wieser, Tetrahedron: Asymmetry 1998, 9 , 1107–1110. 84. P. R. Lassen, D. Mondrup Skytte, L. Hemmingsen, S. Feldbæk Nielsen, T. B. Freedman, L. A. Naﬁe, S. B. Christensen, J. Nat. Prod . 2005, 68, 1603–1609. 85. H. Izumi, A. Ogata, L. A. Naﬁe, R. K. Dukor, J. Org. Chem. 2008, 73 , 2367–2372. 86. J. He, A. G. Petrovic, S. V. Dzyuba, N. Berova, K. Nakanishi, P. L. Polavarapu, Spectrochim. Acta Part A 2008, 69 , 1213–1222. 87. N. H. Andersen, N. J. Christense, P. R. Lassen, T. B. Freedman, L. A. Naﬁe, K. Stromgaard, L. Hemmingsen, Chirality 2010, 22 , 217–223. 88. G. Shanmugam, P. L. Polavarapu, D. Gopinath, R. Jayakumar, Biopolymers: Peptide Science 2005, 80 , 636–642. 89. L. A. Bodack, L. A. Freedman, B. Z. Chowdhry, L. A. Naﬁe, Biopolymers 2004, 73 , 163–177. 90. R. H. Cichewicz, L. J. Clifford, P. R. Lassen, X. Cao, T. B. Freedman, L. A. Naﬁe, J. D. Deschamps, V. A. Kenyon, J. R. Flanary, T. R. Holman, P. Crews, Bioorg. Med. Chem. 2005, 13 , 5600–5612. 91. B. Brzezinski, K. Olejnik, S. Paszyc, J. Mol. Struct. 1990, 220 , 261–268. 92. L. Huang, Y. K. Si, G. Snatzke, D. K. Zheng, J. Zhou, Coll. Czech Chem Comm. 1988, 53 , 2664–2666. 93. T. B. Freedman, X. Cao, R. Oliveira, Q. B. Cass, L. A. Naﬁe Chirality 2003, 15 , 196–200. 94. P. L. Polavarapu, N. Jeirath, T. Kurt´an, G. Pescitelli, K. Krohn. Chirality 2009, 21 , E202–E207.

12 DETERMINATION OF MOLECULAR ABSOLUTE CONFIGURATION: GUIDELINES FOR SELECTING A SUITABLE CHIROPTICAL APPROACH1 Stefano Superchi, Carlo Rosini, Giuseppe Mazzeo, and Egidio Giorgio

12.1. INTRODUCTION The knowledge of the absolute conﬁguration (AC) of a chiral molecule is a fundamental prerequisite for the understanding and prediction of its interactions with chiral systems both at molecular and supramolecular level. The tight relationship between AC and bioactivity of chiral drugs and metabolites as well as the importance of chiral recognition in catalysis and materials is in fact well known [1]. The AC assignment is therefore an important problem commonly faced by researchers involved in medicinal chemistry, natural products chemistry, asymmetric catalysis, and materials chemistry. Many different methods to solve this fundamental problem are available to the chemist [2], but, to date, no one can be considered fully general. The classical chemical correlation procedure requires a long process, and often it involves complex stereo-controlled chemical reactions for transforming the compound under analysis into another having known AC. The anomalous X-ray scattering method [3], although being a very reliable method, presents some limitations, requiring the presence of “heavy” atoms on the molecule, crystalline products, and the availability of single crystals, as well as expensive and complex equipments. NMR and chiroptical spectroscopy can instead allow conﬁgurational assignments in solution. The NMR approach [4], requiring the derivatization of the molecule with a chiral auxiliary, can, however, be considered an indirect method and essentially empirical. In this framework the chiroptical spectroscopies like electronic circular dichroism (ECD), vibrational circular dichroism (VCD), and optical rotatory dispersion (ORD) can 1

Dedicated to the memory of Professor Carlo Rosini (1948–2010), an unforgettable friend and mentor, in recognition of his outstanding contribution to the development of chiroptical spectroscopy.

Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

421

422

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

provide a powerful tool for the AC assignment to chiral molecules [5]. A great advantage of these techniques is the possibility to achieve the conﬁgurational assignment in solution, then allowing to treat noncrystalline compounds, for which X–ray methods are not applicable. Moreover, the AC assignment is often obtained in a nonempirical fashion, thereby ensuring reliable results. Although ORD and ECD spectroscopies date back to the nineteenth century, in the last decades they had undergone a tremendous development from the experimental, instrumental, and theoretical-computational point of view. In this context the advent of computational approaches to chiroptical spectra interpretation has constituted the real breakthrough during the past few years [6] (see Chapters 22 and 23 in volume 1). This is because the quantum mechanical calculations of ORD and ECD have become much more accessible, and by that they have broadened the prospects for application of these already well-established techniques. Moreover, recent development of commercially available VCD spectrometers, along with implementation of quantum mechanical approaches for VCD calculation in computational packages, has extended the resources for molecular structural analysis with this new powerful technique [7] (see Chapters 23 and 24 in volume 1). The large number of methodologies and techniques now available can offer a wide choice, but, on the other hand, they may confuse nonspecialist in choosing the “right” method for his speciﬁc purposes. The aim of this chapter is then to provide a practical guide to nonspecialists in chiroptical spectroscopy, but who are involved in synthetic organic, medicinal, and natural product chemistry, on how to select the most suitable chiroptical techniques for absolute conﬁgurational assignment of a particular chiral substrate. For a quick and preliminary conﬁgurational assignment, the “ﬁrst” choice is often the “simplest” and most easily available one, which is not necessarily the most rigorous one. Of course, approaching a structural problem by more than a single technique often allows us to remove a possible ambiguity in the conﬁgurational assignment and, by that, to reach a more reliable result [5c, 8]. Accordingly, the scope and limitations of main chiroptical techniques employed for AC assignments will be brieﬂy reviewed. In the discussion we will provide a comparison of the different qualitative versus computational approaches. Finally, some guidelines and representative examples about how to make decision about the most suitable method will be provided, taking into account the chemical structure and other factors, such as substrate stability and available amount.

12.2. THE TECHNIQUES Whereas the previous chapters of this book review in detail the theoretical principles, instrumentation, experimental measurements, and applications of almost all types of chiroptical spectroscopies, here we will focus only on some of them, namely ORD, ECD, and VCD, which are today the most commonly used chiroptical techniques for AC assignments. The Raman optical activity (ROA) [9] (see Chapter 6 in volume 1) and NIR-VCD (see Chapter 10 in Volume 1) will not be discussed because, despite their high potential, these spectroscopies are not widely used yet for absolute conﬁgurational assignments. Optical Rotation/Optical Rotatory Dispersion 1. For experimental measurements a simple and less expensive polarimeter is commonly used in organic chemistry laboratories. By applying a set of ﬁlters, it can

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

2. 3. 4. 5. 6.

7.

record optical rotation (OR) at a few wavelengths.2 An ORD attachment to a common ECD spectropolarimeter provides a more advanced, continuous recording of OR in broader spectral region. Small quantities (a few milligrams) of sample are usually required. Qualitative methods for assignment of the AC from the sign of some speciﬁc ORD bands are available (i.e., octant rule) (vide infra). ORD data can be calculated by several commercially available packages [10]. Medium-size molecules can now be treated by computational approaches.3 The ORD computational simulation provides the sign and order of magnitude of one or a few OR values. Reproduction of the spectral trend (sign/position of some ORs) can be achieved. In this case, more reliable results are obtained when a plane ORD curve (i.e., far from resonant area and corresponding Cotton effects (CEs)) is calculated. According to some authors, a reliable AC assignment at single wavelength can be obtained only for OR values greater than 30–40 units [11], while the calculations of OR values at several frequencies in general is more reliable [12, 13]. It does not require the presence of chromophores, so even UV–vis transparent molecules can be treated.

Electronic Circular Dichroism 1. The instruments for ECD are more expensive than the common polarimeter, yet they are also available at affordable price and quite common in the practice of organic chemistry and biochemistry laboratories. 2. Very small samples are usually required: with substances having very high molar ε values, even micromolar solutions can be used [5]. 3. Qualitative spectrum/structure correlations (i.e., sector and chirality rules) are available, allowing its use also by experimental organic chemists who are not familiar with quantum mechanical calculations and spectroscopy. A drawback for qualitative approaches is that they are sometimes empirical and then not fully reliable. From this point of view, the exciton chirality method [14] (see Chapter 4 in this volume) represents one of the most reliable and versatile tools to be used. 4. ECD data of small and medium-size molecules can be simulated by several commercially available packages [10]. 5. Computational reproduction of (almost) entire spectrum in UV–vis region is possible: sign, position (λ), and intensity of some (1–10) CEs. For reliable results at least 30–50 excited states must be calculated. 6. The presence of UV–vis absorbing moieties (chromophores) is required.

2

For recording a full ORD spectrum it is necessary that an accessory (≥25,000 USD) be implemented on an ECD spectropolarimeter. 3 For instance, the single-wavelength (λ = 589 nm) TDDFT calculation of OR (B3LYP/6-31G*) for a single conformer of the molecule of benzylidene benzotricamphor [C51 H48 O3 (MW 708)] requires (including geometry optimization and frequencies calculations) 229 h with a desktop computer having CPU Intel® Pentium® D 3.20 GHz, 2 GB RAM [G. Mazzeo, E. Giorgio, C. Rosini, F. Fabris, E. Fregonese, U. Toniolo, O. De Lucchi, Chirality, 2009, 21 , E86–E97]. The same calculation with a newer and more powerful PC (Intel® Xeon® Quadcore E5420 at 2.50 GHz, 4 GB RAM) requires 29 h.

423

424

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Vibrational Circular Dichroism 1. The VCD spectra require a more expensive instrument and a good expertise about the factors that may affect the measurements quality. The most common experimental problem is, for example, the baseline stability, an artifact that can be avoided with dual photoelastic modulator (PEM) instruments [15]. 2. Larger amounts of samples are generally required: some authors suggest samples of 50–100 mg [16], but nowadays the minimum quantity is 5–10 mg. Often both enantiomers and the racemic mixture are needed. Alternatively, careful solvent subtraction is required. 3. VCD data can be simulated by several commercially available packages [10]. Medium-size molecules can now be treated computationally. Since the simulation of a VCD spectrum requires only the knowledge of the electronic ground state, these calculations often are simpler than that of ORD/ECD. The simulated VCD spectrum leads to simultaneous reproduction of many (20–50) CEs of different S/N ratios; and in principle, it provides a better opportunity for a safe matching with the experimental data, a necessary step for a safe conﬁgurational assignment. 4. It does not require the presence of UV–vis chromophores, therefore even transparent molecules can be treated. It must be recalled that both ORD and ECD rely on electronic properties, namely, on the differential refraction and absorption of circularly polarized UV–vis radiation, respectively. Moreover, both methods are also closely related to each other through Kramers–Kronig (KK) transforms [17], which allow us to convert one into the other.4 It follows that, in principle, both spectroscopies can provide the same type of information. From the experimental point of view, however, ECD is to be preferred over ORD. In fact, ECD bands occur only in correspondence of an optically active absorption band, while ORD spectrum results from the sum of contributions from all the chromophores of the molecule, even those in the far UV. It follows that in ORD it is difﬁcult to separate the contributions from different electronic transitions. Conversely, since in ECD overlaps occur only between very close absorptions, the spectral interpretation is much easier. However, an advantage of ORD over ECD is clearly seen when the compounds are devoid of chromophores absorbing at λ > 180 nm. These compounds are in fact ECD transparent, while still show an ORD curve. In such cases a practical alternative is provided by VCD, which, relying on vibrational transitions, does not require the presence of UV absorbing moieties on the molecule. In general, ECD methods are more sensitive, allowing microscale measurements. This is a relevant feature when dealing with natural products that are usually available in very small amounts. Moreover, these methods often allow us to reach a reliable conﬁgurational assignment simply by visual inspection of sign and position of speciﬁc diagnostic CEs. As already mentioned, the main limitation of ECD is the need of chromophores in order to get ECD signals intense enough to provide a reliable assignment. On the other hand, VCD allows us to treat also nonchromophoric molecules and gives rise to experimental/predicted spectral comparison on a much larger number of bands. 4

In theory, a complete transformation from ECD to ORD spectrum requires the knowledge of the full (λ from 0 to +∞) ECD spectrum. In practice, KK transform can be applied only to a part of the spectrum given that the accessible spectral window is usually limited to λ > 180 nm. In that case the resulting KK transformation of ECD into ORD takes into account only the contribution given by the section of the spectrum analyzed. The same happens when the reversed ORD to ECD transformation is done.

425

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

VCD

ECD

ORD

(aR) -Vanol B3LYP/6-311G(2d,2p)

(aR) -Vanol B3LYP/6-311G(2d,2p)

x3 (aR) -Vanol B3LYP/6-311G(2d,2p)

(+) -Vanol-expet

(+) -Vanol-experiment

1700

1500 1300 1100 Wavenumber

900 230

260

290 nm

320

(+) -Vanol-expet

350 350

450

550

650

nm

Figure 12.1. Calculated (top line) and experimental (bottom line) VCD, ECD, and ORD spectra of (+)-(aR)-VANOL. (Reprinted with permission from J. Org. Chem. 2009, 74, 5451–5457. Copyright 2009, American Chemical Society.)

Ph

OH

Ph

OH

(aR)-3,3'-diphenyl-[2,2'-binaphthalene]-1,1'-diol (VANOL)

Chart 12.1.

For interpretation of VCD spectra, the use of computational methods is, however, mandatory, making often the overall AC assignment more time-consuming, particularly in the case of conformationally ﬂexible substrates. In Figure 12.1, experimental and calculated VCD, ECD, and ORD spectra of (aR)3,3 -diphenyl-[2,2 -binaphthalene]-1,1 -diol (VANOL) (Chart 12.1) are compared [18]. There we can clearly see the higher complexity of the VCD spectrum, which allows a comparison on a much larger number of bands. An interesting case in which the use of VCD helps to solve a complex structural problem is when the molecule has a chromophoric chiral moiety that gives rise to very intense CEs in both ORD and ECD, and by that causes a signiﬁcant overlap with bands allied to other chiral moieties in the same molecule. The optical contributions, for example, of molecules containing both axial and central chirality are often difﬁcult to analyze. In such cases, each chirality element can be more conveniently identiﬁed and characterized by VCD spectroscopy, so that a safe assignment of AC can be reached [19]. The main operational features of the chiroptical techniques are summarized in Table 12.1.

426

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

TAB L E 12.1. The Main Operational Features of the Chiroptical Techniques Technique ORD ECD VCD

Instrumenta Cost (103 USD)

Chromophores Required

Sample Amount

Qualitative

Computationalb

≥ 25 ≥ 65 ≥ 100

No Yes No

1–10 mg > μg > 10 mg

Yes Yes No

Yes Yes Yes

a

The instruments costs are mean values referring to year 2010. ORD and VCD calculations require quantum mechanical approaches, while calculations of ECD can be performed also by “classical” (coupled oscillators) and semiempirical approaches. b

12.3. THE APPROACHES In general, the assignment of the molecular AC by chiroptical spectroscopy relies on the comparison between the experimental spectrum and predicted spectrum for a given enantiomer. If both spectra agree, then the AC of the molecule corresponds to the one chosen a priori for the spectral prediction. In order to predict the spectrum for a given enantiomer, its preferred conformation(s) must be also known. In fact, AC, conformation, and spectrum are tightly interrelated and knowing two of these information, the third can be obtained (Figure 12.2). For this reason, chiroptical spectroscopies are also employed as probes of the molecular conformation [20]. Therefore the knowledge of both the molecular conformation(s) and the mechanism relating the AC with the chiroptical response are two crucial aspects that have to be addressed. A conformational analysis of the molecule can be performed either experimentally, via NMR spectroscopy, or theoretically, by molecular modeling and calculations (molecular mechanics, semiempirical or quantum mechanical). In the ﬁrst case, however, only an averaged conformational situation is revealed, while a computational search can allow us to sort out any conformer of the molecule within a given energy window and to determine the relative conformers population. Regarding the spectral prediction, either a qualitative or a quantitative (computational) approach can be pointed out. In the ﬁrst case, empirical or nonempirical rules have to be considered, both allowing prediction of some diagnostic spectral features for a given enantiomer in a given conformation—that is, for a single geometry. This is the case of the well-known chirality rules that allow us to predict the wavelength position and sign of some diagnostic CEs in the ECD spectrum: octant rule [21], benzene [22] and diene [23] chirality rule, Mislow biphenyl rule [24], exciton chirality method [14]. The advantages of the qualitative approaches rely on their ease and less time required to reach a decision, not requiring a precise spectral prediction. In this case a visual Absolute Configuration

Figure 12.2. Relationship Conformation (NMR, Calculations, Molecular models)

ORD/ECD/VCD Spectrum (Experimental ↔ Predicted)

between molecular structural properties (conformation, AC) and spectrum.

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

inspection of the spectrum allows for the conﬁgurational assignment. Moreover, since by qualitative methods only some speciﬁc molecular fragments or chromophores of the molecule are taken into account, in such case complex molecules can be investigated as well. However, a major problem by the qualitative approaches arises when more than a single signiﬁcantly populated conformer is present and the predicted chiroptical response of some conformers is opposite in sign. In this case it is not possible to quantify the contribution of any single conformer to the overall experimental spectrum. A possible solution of this problem is either to (a) transform the compound under investigation into a more rigid derivative which exists as a single conformer or (b) apply a quantitative computational approach for conformational analysis and simulations of chiroptical properties. In fact, the quantitative approach provides a population and numerical values of the rotational strengths for any conformer, thus allowing the calculation of the net contribution by Boltzmann averaging over the populations. In some cases, however, if different conformers have chiroptical contributions of opposite sign, the weighted averaged value may become very small, and hence unreliable for any assignment. The quantitative (computational) approaches allow us to calculate the whole spectrum and therefore provide an estimate of the position, sign, and intensity of all bands allied to a speciﬁc structure and geometry. Many types of computational approaches have been described over the years, both quantum mechanical (semiempirical, ab initio) [6, 25]5 (see Chapters 22, 23, and 24 in Volume 1) and classical (coupled oscillators, DeVoe) [26, 27] (see Chapter 20 in Volume 1). The former provides more accurate answers whereas the latter, being computationally much less demanding, can be applied even to large systems. A complete computational treatment requires (a) conformational search and optimization, (b) assessment of the Boltzmann populations, (c) calculation of the chiroptical properties for each conformer, (d) Boltzmann averaging of the chiroptical properties, and (e) comparison with the experimental data. In some cases the solvent effect must be taken into account in the calculations [6e]. Therefore, there are two crucial computational steps for this approach: (a) the exact determination of the structure (geometry) and population of the main conformers and (b) the calculation of their chiroptical properties. In principle, ab initio methods could be applied for any kind of molecule within the reasonable size range. With these approaches, it is unnecessary to know a priori any property of the molecule, its fragments, and its chromophores (e.g., strength and direction of oscillators, group polarizabilities, etc.) or to have a suitable reference compound. However, as mentioned above, a very large size and a very high conformational ﬂexibility can be prohibitive for application of the ab initio methods. In fact, the larger the number of atoms, the higher the number of electronic or vibrational states and transitions to be computed [28]. Since by the computation of electronic transitions the number of electrons is taken into account, the presence of heavy atoms can obviously strongly affect the computation complexity. With ﬂexible molecules the calculation has to be repeated for each conformer and then a Boltzmann average has to be performed. It is therefore clear that the larger the number of accessible conformers, the more complex and time-consuming the spectral prediction. Often a chemical derivatization which provides a molecular “rigidiﬁcation” is advisable, leading to a decrease in the number of accessible 5

In ab initio electronic structure calculation the Hartree–Fock (HF) method was the ﬁrst widely used, followed by more accurate methods taking into account electron–electron repulsion correction like Density Functional Theory (DFT) and Coupled Cluster Theory (CC) [S.M. Bachrach, Computational Organic Chemistry, John Wiley & Sons, 2007]. For comparison of Time dependent DFT (TDDFT) and CC in chiroptical calculations see: T. D. Crawford, P. J. Stephens, Phys. Chem. A 2008, 112 , 1339–1345.

427

428

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

conformations that simplify the calculation [29, 30]. Alternatively, the problem of the high molecular mobility can be overcome by performing calculation of ECD spectra in the solid state and then taking into account the single conformation taken by X-ray analysis [31] (see Chapter 6 in this volume). Although computational methods can look as the ultimate solution for the AC assignment, their reliability depends on the right choice of many computational parameters and on the accuracy of the conformational analysis performed. For example, in a recent OR calculation study it has been shown that the use of different levels of theory for geometry optimization may lead not only to different conformers distribution, but also to minor differences in conformers’ geometry [32]. This study revealed in particular the critical effect of even subtle geometrical differences on the overall OR calculation result and hence on the correctness of conﬁgurational assignment. It moreover points out some warnings about the use of computational methods. In fact, in many cases, it is impossible to know a priori the right computational protocol and the best method, functional, basis set, and so on [6e]. Therefore, it is desirable that the results provided by one approach be validated by another independent method.

12.4. THE MOLECULAR STRUCTURE Of course, the choice of the suitable chiroptical method to determine the AC of a molecule depends on its structure. It mainly depends on the presence of one (or more) chromophore(s)6 or one (or more) functional group(s), which can allow the introduction of chromophoric moieties, and also on the molecular ﬂexibility (i.e., the number of conformers). As far as the molecular ﬂexibility is concerned, substrates can be roughly divided into rigid and ﬂexible molecules. To the rigid or relatively rigid belong these molecules that adopt only one or a few conformations, but only one has a predominant population. In contrast, the ﬂexible-type molecules exist as a mixture of many conformations, which renders the computational conformational analysis difﬁcult and time-consuming. In order to guide the researcher to the choice of the “right” method, a graphical guide as a ﬂowchart (Scheme 12.1) is provided. This ﬂowchart leads to the “right” choice, intended as the most easily available one or the simplest one. Of course, often the “right” choice cannot be the only one. Other alternative ways perhaps not as simple may also lead to a reliable answer as well. All in this line, more than one approach can often provide more reliable results [5c, 8]; therefore, if possible, the conﬁrmation of the ﬁrst assignment by an independent method is advisable. In Scheme 12.2 the ECD is given as the “ﬁrst” choice since it is one of the most available chiroptical approach and it allows also a simpler qualitative spectral analysis. Therefore, the ﬁrst selection to be made is on the presence of chromophores on the molecule. If chromophores are not present in the molecule and cannot be introduced by chemical reactions, then the molecule is UV–vis transparent and ECD spectroscopy cannot be applied. The absence of chromophores also often leads to low OR values at the D sodium line, therefore sometimes making unreliable even assignment by OR calculation. In this case the “right” choice is then the measurement of the VCD spectrum 6

An organic chromophore can be deﬁned as a part of the molecule containing a functional group or a combination of more functional groups displaying π electrons and responsible for electronic transition giving rise to absorptions in the UV–vis range.

429

Scheme 12.1.

ECD Computational approaches (DeVoe) (Quantum Mechanical)

NO

NO

More than two

Molecular structure

ECD Qualitative exciton coupling approaches (exciton chirality)

YES

Are there couplets in ECD spectrum?

YES

Is it rigid? Can be rigidified?

two

How many?

YES

Are there chromophores?

YES

one

YES

NO

Solid state ECD Quantum Mechanical approach

Is it possible to introduce a second chromophore?

Can they be introduced by chemical derivatization?

NO

ORD/ECD/VCD Quantum Mechanical approaches

NO

Use of Spectroscopic probes

ECD Qualitative approaches (sector/chirality rules)

YES

ORD/VCD Quantum Mechanical approaches

Is it rigid? Can be rigidified?

NO

430

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

and its comparison with a simulated one by quantum mechanical calculations. If a single chromophore with UV–vis absorption is present, the chiral molecule exhibits an ECD spectrum with one or more CEs of same or different sign. In such case the molecule can be treated either qualitatively by applying sector and chirality rules or by quantum mechanical calculation of ECD spectrum, depending on its molecular ﬂexibility. In the case of ﬂexible chromophoric molecules the conformational problem can be overcome by the quantum mechanical simulations of the ECD spectrum taken in the solid state then taking into account the single conformation presents in the crystal structure [31]. Alternatively, when the molecule possesses at least one suitable functional group, an auxiliary that works as a spectroscopic probe, giving rise in the ECD spectrum to CEs related to the AC, can be attached to the molecule. Such probes can be linked to the molecule by either noncovalent [33] or covalent [34] bonds, and the resulting CEs can be qualitatively interpreted by chirality rules. The presence of two chromophores (equal or different) allows the use of nonempirical exciton chirality method. An application of this method requires that the relative orientation of the chromophores and polarization of their interacting electric transition moments be known with certainty, and it also requires that a clear couplet feature (either degenerate or nondegenerate) be visible in the ECD spectrum. On the contrary, if the molecule is too ﬂexible and cannot be chemically rigidiﬁed, a computational approach cannot be avoided, being the only way to take into account the individual contribution of the several conformers. Also, molecules with more than two chromophores cannot be easily treated by the qualitative approaches and also by the exciton chirality method, therefore computations will be more desirable. The right choice of the method is better exempliﬁed by looking at some selected examples in which advantages, disadvantages, and limitations of the different approaches are discussed.

12.4.1. Chromophoric Molecules 12.4.1.1. Presence of Only One Chromophore. For compounds having a single chromophore (carbonyl, aryl, diene, enone, etc.), qualitative approaches, such as sector and helicity rules, have been developed. They link the sign of the ECD CEs allied to the main electronic transitions of the chromophore to the AC. These approaches, although widely employed, often present some ambiguity and can lead to incorrect assignments. For this reason, although qualitative methods are often the “ﬁrst” choice for their simplicity and rapidity, they sometimes need to be supported by more rigorous computational analysis. For several chromophores, belonging to either inherently chiral and achiral type [35], empirical and nonempirical rules relating the AC to the sign of corresponding ECD CEs have been developed. “Sector rules” have been described for treatment of symmetric, dissymmetrically perturbed chromophores, while “chirality rules” allow the treatment of inherently chiral chromophores [21–24] (see Chapters 2–4 in this volume). Some selected recent examples concerning the carbonyl and benzene chromophores are discussed herein, aiming to show some limitations and solution of pertinent problems. For more examples see other relevant chapters in this book. The Carbonyl Chromophore. The oldest and most popular sector rule is the so-called “octant rule,” allowing us to assign the AC of saturated ketones by the sign of the CE allied to the n → π ∗ transition of the carbonyl around 300 nm [20, 21, 36] (see Chapter 2 in this volume). Although it was originally proposed on empirical grounds [37], it was later supported by theoretical studies on the origin of the ketone n → π ∗ CE [21, 38].

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

In order to apply the octant rule for determining the AC of a ketone, the conformation of the latter must be known. A major problem then arises when the molecule shows an equilibrium of several conformations, since the octant rule allows to predict the sign of the n → π ∗ band for each conformer but not its intensity. For this reason the octant rule has usually been applied to cyclic, conformationally deﬁned ketones. Another problem of the octant rule concerns the priority assessment of substituents that fall in different, oppositely signed, octants. Bicyclo[3.3.1]nonanediones provide an interesting example, which highlights the problems and solutions of their AC assignments. By applying the octant rule to these bicyclic diketones, both carbonyl groups in the main conformations must be taken into account. The bicyclo[3.3.1]nonane-2,6-dione (1) (Chart 12.2) was obtained by kinetic resolution of the racemic mixture by baker’s yeast [39], and its AC was established by Berg and Butkus by applying the octant rule [40]. Taking the (1S ,5S ) enantiomer of 1 in its major chair–chair conformation and putting the C2 carbonyl into the octant, the situation depicted in Figure 12.3 is obtained. Here the only atoms not on nodal planes are C4, C9, C8, C7, and the oxygen. C4 and C9 cancel each other, while the oxygen and C7, being far from the carbonyl, give rise to a weak contribution. Therefore, the only effect to be taken into account is given by C8, which falls in a positive octant. Exactly the same situation is obtained when the C6 carbonyl is examined, given the C2 symmetry of the molecule. It follows that for (1S ,5S )-1 a positive CE allied to the n → π ∗ transition is expected in the ECD spectrum. The (+) enantiomer experimentally shows such positive CE, and therefore it has the (1S ,5S ) AC. In this case there is no signiﬁcant ambiguity in applying the octant rule. The molecule has a largely dominant conformation, and for both carbonyls the same sign of the ECD band is predicted. This assignment was conﬁrmed by chemical correlation [41] and, later, by Stephens et al. [42] by Time-Dependent Density Functional Theory (TDDFT) ECD and OR calculations. These authors performed a DFT conformational analysis of 1 at the B3LYP/6-31G* level, obtaining a 72:27 ratio of the chair–chair:chair–boat conformers. n → π ∗ C=O excitations and OR values were then calculated at the B3LYP/aug-ccpVDZ level for both conformers, and the ﬁnal values of ECD and OR were obtained by the Boltzmann average of the calculated values. In bicyclo[3.3.1]nonane-2,7-dione (2) (Chart 12.2), MMFF94 calculations give again the chair–chair conformation as the major one [43]. In this case, however, the two carbonyls are not equivalent and the two different situations depicted in Figure 12.4 are obtained. The location of the major conformer into octants, placing each C=O chromophore into the origin of the octants, leads to opposite signs of the n → π ∗ band (Figure 12.4) for the two carbonyls. When the C2 carbonyl group of the (1R,5S ) enantiomer is located into octants, only C4, C9, and C8 do not lie on nodal planes, while the C7 carbonyl lies very close to a nodal plane. Consequently, the major perturber is the axial C8, which falls in a positive

O

O

2 4

1

7 6

3 4

1

8

9 5

2

3

O

8

9

7

5 6

O (1S, 5S)-1

(1R, 5S)-2

Chart 12.2.

431

432

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

+

− 4

9

5

1 2

3

6

Figure 12.3. Location of (1S, 5S)-bicyclo[3.3.1]

8

−

+

7

nonane-2,6-dione (1) into octants placing the C2 carbonyl group at the origin of the octants.

octant. Therefore, for this enantiomer a positive CE allied to the carbonyl chromophore is predicted in the ECD spectrum. Interestingly, the placement of the C7 carbonyl group into the origin of the octants led to the prediction of the opposite CE and to the reverse AC for this enantiomer. In fact, on the right projection (Figure 12.4) the C1 and C2 atoms cancels the contribution of C5 and C4. Therefore, oxygen on C2 remains as the sole perturber since all the other atoms are on nodal planes. This projection predicts a negative sign of the CE due to the O perturber on the C2 atom, falling in a negative octant. The authors consider the latter effect weaker than the former and therefore predict a positive CE in the ECD spectrum for (1R,5S )-2. As a result, they assigned this AC to the (+)-enantiomer. Although this assignment was conﬁrmed by chemical correlation [43], obtaining 2 from 1 of known AC, it is clear that in this case the octant rule alone could not provide a reliable result. The AC of 2 was later assigned by means of chiroptical spectroscopy by Stephens et al. [42] through ab initio TDDFT calculations of ECD and OR. DFT B3LYP/6-31G* conformational analysis gave the chair–chair conformer as the only populated one, and TDDFT B3LYP/aug-cc-pVDZ calculations provided both n → π * ECD band and OR values, allowing to reach at the same (1R,5S )/(+) relationship for 2 in a more reliable way. Therefore, in this case, when uncertainty and ambiguity result from application of the octant rule, the use of a computational ab initio approach can provide more conﬁdent results. The Benzene Chromophore. The presence of an aromatic ring is one of the most common structural features in organic molecules. Therefore, many efforts have been devoted

+

− 9

5

4

+

−

3 2

4

1 3

2

6 8 7

−

+

−

5

9

1

6

7

8

+

Figure 12.4. Location of (1R, 5S)-bicyclo[3.3.1]nonane-2,7-dione (2) into octants placing the C2 and C7 carbonyl groups (left and right projections, respectively) at the origin of the octants.

433

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

in developing methods for determination of AC of chiral aryl substituted molecules. To this end, various semiempirical benzene sector rules have been proposed relating the sign of the ECD CEs allied to the benzene 1 La [44] and 1 Lb [22] transitions [45] with the AC of a phenyl- or benzyl-substituted stereogenic center. For substituted benzene compounds the benzene chirality rule [46] is instead operative (see Chapter 3 in this volume). The most widely used rule for the unsubstituted benzene chromophore is the “benzene sector rule” [22], which relates the 1 Lb ECD CE with the AC at the benzyl carbon. This rule has been successfully applied to many phenyl- and benzyl-substituted compounds, such as carbinamines, carbinamine salts, and carbinols, as well as for assignment to compounds of unknown AC [47]. It nevertheless failed with simple chiral alkyl benzenes like compounds 3–6 shown in Chart 12.3. For these compounds, all having (R) AC, the benzene sector rule predicts a negative CE allied to the 1 Lb transition. On the contrary, only 6 display a negative CE, while in all other cases positive CEs are observed in the 250- to 280-nm range. For these compounds an ab initio ECD calculation instead provided the right answer [48]. MMFF conformational search for each compound and DFT optimization at the B3LYP/6-31G(d) level was performed. The results showed three populated conformers for 3 and 5, ﬁve conformers for 4, and only a single conformation for 6. TDDFT-simulated ECD spectrum for each conformer at the B3LYP/TZVP level, followed by Boltzmann averaging, led to ﬁnely reproduced 1 Bb (below 200 nm) and 1 La (205–225 nm) ECD bands, while some disagreements were evident for the 1 Lb bands. This analysis clearly point out the complexity of ECD pertinent to phenyl chromophore. Not surprisingly, more reliable results were obtained after applications of computational approach. Other Chromophores. Several other semiempirical rules have been proposed for absolute conﬁgurational assignment such as sector rules for lactones [49], oxiranes [50], thioamides [51], β-lactams [52], and γ -lactams [53] as well as helicity rules for α, βunsaturated ketones [54, 55], dienes [23, 56], disulﬁdes [57], and biphenyls [24, 58] (see Chapter 2 in this volume). In general, AC assignment of these compounds can also be achieved by a theoretical treatment [13c, 59]. 12.4.1.2. Presence of Two Chromophores. Molecules that contain two identical chromophores (i.e., being “dimeric”) with electrically allowed transitions and which do not exchange electrons can be treated by the coupled oscillator or exciton chirality

3

4

5

N

N (–)–(R,R)–7

Chart 12.4.

6

Chart 12.3.

434

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

approach. This treatment ﬁnds its origin in the seminal papers of Kuhn [60], Kirkwood [61], and Mofﬁtt [62] on the coupled oscillator model, in both the classical and quantum mechanical formulation. Its ﬁrst application for determination of the AC of a small molecule, the natural alkaloid calycanthine, was reported by Mason [63]. A major breakthrough in this approach is however due to the introduction by Harada and Nakanishi in early 1970-ies of the so-called exciton chirality method [14] (see Chapter 4 in this volume). Because it is nonempirical approach, the method allows us to establish (safely and qualitatively) a correlation between ECD spectrum and molecular AC, avoiding any kind of calculations. Here, a few examples that point out some limitations of the method and the way to solve them are reported. Two Identical Chromophores. Some failures of the exciton chirality rule have been reported in the literature, but very often these defeats did not depend on the intrinsic nature of the model, but they simply derived from a wrong application of it. In many cases, wrong results are due to wrong placement of the transition dipoles. In fact, to correctly apply the exciton model, it is fundamental to know the right polarization direction of interacting dipoles. This aspect is clearly shown in the case of Tr¨oger’s base (7) (Chart 12.4), for which a (+)/(R, R) correlation was found by Mason et al. [64] by applying the exciton approach. More than 20 years later, Wilen et al. [65] found, by X-ray analysis, that the previous Mason’s assignment for 7 was wrong and that the (−)-enantiomer of 7 has instead (R, R) AC. This problem was later faced by Devlin and Stephens [66] by ab initio DFT calculation of VCD spectra. Equilibrium structures of 7 were calculated at the B3PW91/6-31G* and B3LYP/6-31G* levels ﬁnding a single C2 symmetric structure. Minor differences in bond lengths and angles were found using the two functionals. Also, mid-IR absorption and VCD spectra were calculated at the B3PW91/6-31G* level. The comparison of VCD rotational strengths calculated for the (R, R) AC showed excellent agreement with experimental intensities for (−)-7 deﬁning unambiguously the correspondence (R, R)(−)/(S , S )-(+). This assignment is in agreement with the one obtained by X-ray structure, but opposite to the Mason one deduced by the coupled oscillator approach, and it seems to reveal an intrinsic limitation of the exciton model. Actually, even a quantitative coupled oscillator DeVoe calculation [26] in which 7 was treated, similarly to Mason, as a system of two identical aniline chromophores, provided a wrong result [67]. More accurate analysis on the structure of the involved chromophore allowed to determine that in 7 a distorted aniline chromophore is actually present and that this distortion causes a signiﬁcant change in the polarization directions. Then repeating the DeVoe calculation with the correct oscillators direction and intensity the ECD spectrum of 7 was satisfactorily reproduced [67]. This demonstrates that the DeVoe coupled oscillator approach and the exciton chirality model work well, obviously, only if a correct description of the spectroscopic parameters is used. A main problem in applying the exciton chirality approach is given by the molecular mobility. In fact, when at room temperature the molecule displays more signiﬁcantly populated conformations, it is difﬁcult to determine the expected couplet feature by summing each conformer contribution. This problem can be overcome by transforming the original ﬂexible bis-chromophoric molecule in a rigid, conformationally deﬁned, derivative. This kind of approach has been followed to determine the AC of anti -1,2-diarylethane-1,2diols [68]. These acyclic diols were transformed into the corresponding 2,2-dimethyl1,3-dioxolanes, where the two hydroxyls form a ﬁve-membered ring. Since such ketals are conformationally rigid, the relative position of the two aryl moieties is ﬁxed and can

435

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

be easily determined. Futhermore, the sign of the CD exciton couplet reveals the chiral twist between the two interacting aryl chromophores, and by that the AC of starting diol. For example, in the dioxolanes of (1R, 2R)-1,2-di(4-substituted)phenylethane-1,2diols, both the longitudinally oriented aromatic transitions 1 Ba,b at 180–210 nm and 1 La at 210–240 nm form a positive chiral twist giving rise to a positive exciton couplet (Scheme 12.2), as experimentally observed. Often the positive couplet allied to the 1 La transition appears strongly unsymmetrical or only a single branch of the couplet is visible. This phenomenon is due to an overlap of the negative high-energy branch of such a couplet (centered at ∼220 nm) with the positive low-energy branch of the more intense positive couplet due to the 1 Ba,b transition (at ∼190 nm), giving rise to a reduction or even to a cancellation of the negative component of the 1 La couplet. By this method the absolute conﬁgurations of many anti -1,2-diarylethane-1,2-diols symmetrically and nonsymmetrically substituted with phenyl and naphthyl moieties have been assigned in a straightforward manner [68]. Two Different Chromophores. The exciton coupling approach also works when two different chromophores are present in the molecule and the directions of their excitonically coupled dipole transition moments are known. This is the case, for example, of optically active aryl alkyl sulfoxides 8–10 (Figure 12.5) in which the aryl and sulfoxide chromophores interact each other [69, 70]. The compounds 8–10 exemplify a type of structurally close compounds where more caution regarding their similarity will be advisable. In fact, some of them may possess a different degree of complexity if more than a single conformation is present and if more than two electric dipole moments have to be taken into account. In these compounds the absorption allied to the σ → σ ∗ transition of the S=O chromophore occurs at 210 nm, indicating that this chromophore is not conjugated with the aryl one, an effect due to the steric hindrance of the peri hydrogen. Therefore the sulfoxide and the naphthalene chromophore are reciprocally isolated, thus allowing us to apply the exciton coupling method. The transition dipoles to be taken into account are then the allowed σ → σ ∗ transition of the S=O chromophore polarized along the C1 → C2 direction of the C(1)(S=O)C(2) moiety and the long axis polarized 1 B transition of the naphthalene chromophore at 220 nm. Molecular mechanics calculations for (S )-1-naphthyl methyl sulfoxide (8) showed an E /Z equilibrium almost completely shifted toward the E conformer (Figure 12.5). Therefore the exciton chirality rule can be applied to this conformer where in (S )-8 the dipoles of the sulfoxide σ → σ ∗ and naphthalene 1 B transitions deﬁne a negative chirality, then giving rise to a negative couplet in the ECD spectrum (Figure 12.6a). The (−) enantiomer of 8 shows indeed a negative ECD coupling, thus leading to the (S )/(−) correlation [69].

HO

O

OH

O X

H

O O

X

X (R,R)

X

X (R,R)

H X POSITIVE chirality Positive Couplet in ECD

Scheme 12.2. Transformation of (R, R)-1,2-di(4-X-phenyl)-1,2-diols in the corresponding 2,2dimethyl-1,3-dioxolanes and application of the exciton chirality rule to the latter.

436

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Me – S O

–O

+

Me

+

S

(S)-8,Z

(S)-8,E Me + – S O

–O

Me

+

S

Me

Me

(S)-9,E

(S)-9,Z

Me – S O

–O

+

Me

+

S

Figure 12.5. Equilibrium between the (E) and (S)-10,E

(S)-10,Z

(a)

(Z) conformations of sulfoxides (S)-8–10.

(b)

Figure 12.6. Exciton chirality deﬁned by the allowed 1 B naphthalene and S=O transitions in (a) E-conformer of (S)-8 and (b) Z-conformer of (S)-9. (See insert for color representation of the ﬁgure.)

In the similar compound 1-(2-methylnaphtyl) methyl sulfoxide (9) contrary to what observed in 8 the Z conformer is the most stable one (Figure 12.5), having a 70:30 Z :E ratio [70]. In the Z conformation the sulfoxide σ → σ ∗ and naphthalene 1 B transitions deﬁne an opposite chirality in respect to the E one and therefore for (S )-9 a positive chirality is observed and then a positive couplet is expected in the ECD spectrum. This indicated an (S )-conﬁguration for the (−) enantiomer (Figure 12.6b). This qualitative analysis does not take into account the contribution of the minor conformer which may be not negligible, therefore for a reliable assignment a quantitative DeVoe calculation was performed taking both conformers as input geometries and averaging the results over the relative populations. It was found a good agreement between the calculated spectrum

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

for (S )-9 and the experimental one for the (−) enantiomer. Further conﬁrmation for AC was obtained by ab initio calculation of its VCD spectrum [71]. The third compound of this series, the 9-phenanthryl methyl sulfoxide (10), although apparently very similar to the previous ones, was actually more complex and required a different treatment [69]. In this compound the chromophores involved display many complex transition dipoles interactions. Therefore, even if only two chromophores are present, the two dipole treatment of the exciton chirality approach is not suitable to provide reliable results. The ECD spectrum of 10 shows a number of bands, no clear exciton couplets appear, and in correspondence to the strongly allowed band of the phenanthrene chromophore at about 250 nm, only a weak Cotton effect is measurable. All these facts hamper a simple application of the exciton chirality approach and therefore DeVoe calculations were undertaken. Molecular mechanics calculations afforded a 90/10 distribution of the E /Z conformers (Figure 12.5). To describe the phenanthrene chromophore transitions, a series of dipoles were placed at 250 nm (long-axis polarization), 240 nm (short-axis polarization), and 205 nm (short-axis polarization) as suggested by some CNDO/S-CI calculations. For the sulfoxide transition at 210 nm a single oscillator polarized along the CMe –CAr bond was used. A weighted average of the spectra of the two E and Z conformers afforded the calculated ECD spectrum which well reproduced the main features of the experimental one, providing the conﬁgurational correlation (−)/(S ). Exciton coupling between different chromophores also arises when, in order to apply the exciton chirality approach, a second chromophore is added by chemical reaction to a monochromophoric substrate. One example of this type is provided by the transformation of 1-arylethane-1,2-diols to the corresponding biphenyl boronates [72]. 1-Arylethane-1,2diols are acyclic diols having only one chromophoric group, the aryl one. Therefore in order to apply the exciton chirality approach, a second chromophore having well-deﬁned electrically allowed transitions and able to couple with the aryl one has to be introduced. Moreover, a great simpliﬁcation of the analysis is achieved if a rigid derivative is obtained. The diols were then transformed into the corresponding 4-biphenylboronates (Scheme 12.3), thus introducing a second chromophore and, at the same time, transforming acyclic diols into cyclic conformationally deﬁned derivatives. In biphenylboronates, due to sp 2 hybridation of the boron atom, no new stereogenic centers are formed and a strong UV absorption at ∼260 nm, allied to the biphenyl long-axis 1 La transition, is displayed. The aryl and the biphenyl transition dipoles are placed in a ﬁxed and rigid relative disposition whose chiral twist is determined only by the AC of the benzylic stereogenic center of the diol and revealed by the sign of the exciton couplet. In 1-arylethane-1,2diols having (R) AC at the benzylic carbon the aryl and the biphenyl moiety deﬁne, in the boronate, a negative chirality and then a negative CE is expected in correspondence to the 1 La biphenyl band at 260 nm (Scheme 12.3). To summarize, boronate approach can allow for the determination of AC at a benzylic stereogenic center of chiral diols, provided that there is no conformational ambiguity in the interpretation of the biphenyl exciton couplet at 260 nm. Many other examples of this type have been described in the literature [14], where the molecule under study contains only one chromophore and a functional group, such as hydroxyl and amino groups, which allows an introduction of a second chromophore. In this case a nondegenerate coupling occurs (see Chapter 4 in this volume). In a special case, if one of the chromophores absorbs below 200 nm (for example, the double bond

437

438

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

R OH HO

OH

Ar

R

O

B

O

OH

CHCl3, 4A MS

B

1L (260 a

nm)

1B

negative chirality

negative CE at 260 nm

Scheme 12.3. Transformation of (R)-1-arylethane-1,2-diols in the corresponding biphenylboronates and application of the exciton chirality rule to the latter.

in allylic alcohols or amines), only a single CE as the longer wavelength part of exciton couplet can be observed. Yet, even in such case the sign of this CE revealed the AC of substrate. 12.4.1.3. Presence of More than Two Chromophores. Systems with more than two chromophores cannot be solved by simple two dipole treatments of exciton chirality approach. The (−)-cupressoﬂavone (11) (Chart 12.5) is an example of such systems. The UV spectrum of (−)-11 exhibits two intense π –π ∗ bands at 324 and 273 nm. In correspondence of the ﬁrst UV band at 324 nm, positive and negative CEs at 362 and 326 nm, respectively, appear in the ECD spectrum. The second UV band at 273 nm is instead associated with a negative ECD shoulder around 300 nm and a positive CE at 267 nm. At ﬁrst the authors attempted an interpretation of these spectral features on the basis of the exciton chirality model. These CEs were then initially considered as two oppositely signed couplets deriving from exciton chirality deﬁned by two allowed transitions located on each ﬂavone monomer. The UV bands at 324 and 273 nm were in fact assigned to long-axis polarized transitions of p-methoxycinnamoyl and p-methoxybenzoyl chromophores, respectively. On the basis of such qualitative twodipole treatment, an aS conﬁguration was at ﬁrst assigned to biﬂavone (−)-11 [73]. In a subsequent study, Harada et al. [74] determined, by π -electron SCF-CI-DV MO calculations, energy position and rotational strengths of several π –π ∗ transitions of 11, thus obtaining its absorption and ECD spectra. Comparing the calculated and experimental ECD spectrum of (−)-11, they showed that the correct AC of (−)-11 is indeed (R).

OH

H3CO

O

O OCH3 OCH3

H3CO

O

OH

O

(–)–(R)–11

Chart 12.5.

439

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

These authors stated that the ﬁrst two-dipole treatment gave a wrong result because they did not take into account that in the 400- to 250-nm range there are several allowed electronic transitions with different polarizations [74]. Their interactions give rise to several couplings whose overlap does produce multiple bands of opposite sign; therefore the overall CD has a complicated proﬁle. Successively, the ECD spectrum of (−)-11 was reproduced by coupled oscillators DeVoe calculations [27b] demonstrating that also this simpliﬁed approach, which allow us to treat the interaction of several dipoles simultaneously, can afford the correct answer and constitutes a practical alternative to MO calculations.

12.4.2. Transparent (Nonchromophoric) Molecules 12.4.2.1. Presence of No Chromophores but One or Two Functional Groups. Molecules that do not absorb in the accessible UV–vis range are classiﬁed as nonchromophoric, transparent compounds and ECD analysis cannot be employed directly to them. However, in such case a suitable substrate with UV–vis absorption can be employed as auxiliary able to provide an ECD spectrum useful for determination of AC. Such an auxiliary can be considered as a chiroptical chromophoric “probe” for determination of molecular chirality. When the molecules are also endowed with a high molecular ﬂexibility, the derivatization with the probe should also reduce the conformational mobility of the molecule. For transparent molecules with two derivatizable functional groups (diols, aminoalcohols, diamines, etc.), the assignment of AC via ECD can be performed by double derivatization with suitable chromophores. A typical application of the exciton chirality rule is in fact the AC assignment of diols by transformation in the corresponding dibenzoates [75], a method extended to a number of transparent bifunctional compounds and to many natural products (see Chapter 4 in this volume). This approach is particularly straightforward when dealing with cyclic, conformationally deﬁned derivatives, but presents severe limitations with ﬂexible, conformationally mobile compounds, in which multiple accessible conformations (and then dipole orientations) must be taken into account. In the second case a practical solution can be provided by the transformation of the diol in a cyclic, conformationally deﬁned derivative. Following this concept, the AC assignment to aliphatic non-chromophoric diols has been approached by introduction of a ﬂexible bridged biphenyl moiety as a probe of the molecular chirality [34a, 76]. Alkyl and aryl substituted diols were transformed in the corresponding biphenyl dioxolanes (Scheme 12.4), thus obtaining a pair

OCH3 OCH3 OH OH

O P

R1 *

n

* R2

n = 0,1,2 R1, R2 = aryl, alkyl, H

O negative CE at 250 nm

R2 * n * R1

O M O

R2 * n * R1

positive CE at 250 nm

Scheme 12.4. Transformation of 1,n-diols in the corresponding biphenyldioxolanes. Equilibrium between P and M twisted dioxolanes.

440

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

of diastereoisomers having, respectively, P and M twist of the biphenyl moiety. In these compounds the low rotational barrier (∼14 kcal/mol) of the biphenyl allows, at room temperature, a thermodynamic equilibrium between the diastereoisomers. Therefore the most stable of them is also the major one. The mechanism for chirality transfer from the chiral diol to the biphenyl was clariﬁed, thus establishing a direct relationship between the diol AC and the preferred biphenyl torsion. Moreover, the twist of the biphenyl could be easily determined from the ECD spectrum where the biphenyl P torsion is allied to a negative Cotton effect related to the so-called A band of the biphenyl chromophore at 250 nm, while an M torsion gives rise to a positive A band at the same wavelength [24, 77]. Therefore, by simply looking at the sign of the A band in the CD spectrum, it is possible to identify the biphenyl torsion and then the AC of a particular diol. This method proved to be simple, straightforward, and general with mono- and di-substituted 1,2-, 1,3-, and 1,4-diols, both cyclic and acyclic. The conformational rigidiﬁcation of acyclic diols through transformation in cyclic chromophoric dioxolanes proved to be a useful tool also for simplifying the computational approaches for AC assignment [30]. By reacting ﬂexible and UV–vis transparent synand anti -1,2- and 1,3-diols with ﬂuorenone dimethyl acetal, the corresponding ketals were obtained (Scheme 12.5). These ketals are conformationally well-deﬁned (only one conformer in most cases). They exhibit medium to high OR values and ECD spectra with several (up to ﬁve) CEs in the 350- to 200-nm range, due to valence shell π → π ∗ transitions. These features have allowed simulation of the chiroptical properties of these compounds at the TDDFT/B3LYP/6-31G* level of theory. The simulated ECD spectra provided much better agreement with the corresponding experimental data than with the calculated OR values, where only a satisfactory agreement was found. The transformation of the diols in their ﬂuorenone ketals also turned out to be advantageous in performing VCD analysis for AC assignment, strongly reducing the number of conformers [78]. The assignment of AC to transparent monofunctional compounds by ECD is a really challenging task because in this case there is no possibility to obtain a bis-chromophoric system and then to apply the exciton chirality model. The solution of this problem was found by using “probes” of the molecular chirality—that is, chromophoric moieties which, linked covalently or not covalently to the substrate, give rise to CEs in the ECD spectrum which can be related to the AC of the molecule under investigation. Typical examples of noncovalently bound chirality probes are the “bis-porphyrin tweezers” that are applicable either to bifunctional compounds, such as diamines, amino alcohols, and amino acids, or to monofunctional secondary alcohols, primary and secondary amines, and carboxylic acids, all usually without useful for ECD studies UV–vis MeO

OH R1

*

OMe

OH n

* R2

n = 0,1; R1, R2 = aryl, alkyl, H several conformers LOW ORs; ECD transparent

R1

O

O

*

* R2 n

only one conformer OR values 40–50 degrees intense ECD spectrum

Scheme 12.5. Transformation of chiral 1,n-diols in the corresponding ketals with 9-ﬂuorenone. Conformational and chiroptical properties of the derivatives.

441

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

bands (Chapter 4 in this volume) [33]. In this approach, upon complexation between a chiral substrate as a guest and an achiral bis-metalloporphyrin as a host, a chirality transfer from the guest to the host takes place. The porphyrin moieties of the complex adopt mutually twisted orientations with one more preferred helicity. It was found that the preferred porphyrin helicity of the complex, either clockwise or anticlockwise, depends on the AC of the guest. The porphyrins’ electronic transitions give rise to exciton coupling and then to a typical couplet feature in the ECD spectrum in correspondence of the porphyrin Soret bands (380–420 nm). Therefore, from the sign of the couplet that reﬂects the porphyrin twist, the AC of the chiral guest can be determined (Figure 12.7). In the past few years, several approaches have been developed for predicting the sign of porphyrin based exciton couplet of the host-guest complex. The earliest one takes into account the relative steric size of the substituents at the stereogenic center estimated on the basis of conformational energy values. Although a qualitative steric-size model provides a simpler and faster assignment of the AC, in some cases it may provide ambiguous results if the steric parameters for some groups are lacking or when factors, other than a steric one, are involved in determining the most stable conformation of the

(R) (R) (R)

O

O

O NH2

O

NH

Zn

Zn +

H Zn

Zn H

O

O

O

N

O

NH2

O

O

O

1-2Zn

conjugate/tweezer complex 15 N

N

10

20

Zn

+ 15'

N

N

N

N

5

10'

20'

Zn

Zn Z n

Zn Zn

N

N

O

−

5'

O

Favored conformation I 1-2Zn O

O

Predicted positive CD exciton couplet

Unfavored conformation II Predicted negative CD exciton couplet

Figure 12.7. Host–guest complexation between I-2Zn tweezers and a conjugate prepared from (+)-isomenthol as a representative of chiral secondary alcohols. The preferred positive helicity between the two porphyrins in the complex, deﬁned as the projection angle between 5–15 and 50–150 directions, is governed by the AC of the isomenthol; in such a case the complex exhibits in solution a positive ECD exciton couplet (not shown). (Reprinted with permission from Chem. Commun. 2009, 5958–5980. Copyright 2009, The Royal Society of Chemistry.)

442

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

host–guest complex. The recently developed approach instead relies on more accurate computational conformational analysis by MC/OPLS-2005 [34, 79] and provides not only reliable results about AC, but also a deeper insight on the mechanism of this host–guest complexation process. Another approach for the assignment of AC to transparent monofunctional molecules is given by the use of covalently bonded biphenyl probes. This approach has been employed speciﬁcally for conﬁgurational analysis of 2-substituted chiral carboxylic acids [34b]. In this approach the chiral acids are converted in the corresponding biphenyl amides, whose ﬂexible biphenyl moiety acts as a “probe” of the acid chirality giving rise to CEs related to the acid AC. In fact, in these derivatives, the stereogenic center of the acid induces a preferential torsion of the biphenyl moiety, detectable by the sign of the biphenyl A band (at 250 nm) in the ECD spectrum [24, 77] (Scheme 12.6). The mechanism of transfer of chirality from the acid stereogenic center to the biphenyl moiety has been analyzed and understood in detail, deﬁning two different mechanisms operative in amides derived from 2-alkyl and 2-aryl substituted acids, respectively. Therefore, for both classes of compounds, it has been possible to deﬁne a model that allows us to predict, for a given acid AC, the preferred twist of the biphenyl moiety and, subsequently, the sign of the A band at 250 nm in the ECD spectrum, related to the biphenyl torsion. Following this protocol, to establish the AC of a 2-substituted chiral acid, it is simply necessary to prepare its biphenyl amide, to record the amide ECD spectrum, and to look at the sign of the A band. From the sign of such a band the torsion of the biphenyl can be deduced and then the acid AC. The main advantages of this approach are the reliability and simplicity of the AC assignment [34b]. In fact, in most of the cases, no conformational analysis is needed and the correlation between the ECD spectrum of the biphenylamide and the acid AC can be established while taking into account only the size of the substituents on the stereogenic center. When the steric size values are unknown or when some ambiguities arise upon the relative size of the substituents, then a simple molecular mechanics conformational analysis can allow a reliable spectrum–structure correlation. A variety of substrates, with different structures and multiple functionalities, has been investigated, in particular this approach displayed its full validity with alkyl and aryl substituted acids, with α-hydroxyacids and α-aminoacids [34b]. 12.4.2.2. Presence of No Chromophores and No Functional Groups. Saturated hydrocarbons are virtually UV-vis transparent, therefore for these compounds ECD and hence qualitative methods based on this technique cannot be applied. These compounds can then be treated either by transforming them into chromophoric derivatives, without affecting the molecular chirality, or by applying OR and VCD calculations [11c].

R2 R3 + R1

COOH

O

O NH

N

P

R1 negative CE at 250 nm

R3 R2

M

N R1

R3 R2

positive CE at 250 nm

Scheme 12.6. Transformation of 2-substituted carboxylic acids in the corresponding biphenylamides. Equilibrium between P and M twisted amides.

443

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

O H H

H

(S)-PHTP (S)-PHTP

(S)-2-keto-PHTP

Scheme 12.7.

(a) stereochemistry of (S)-anti-trans-anti-trans-anti-trans-perhydrophenylene (PHTP) (12). (b) Transformation of (S)-(12) in (S)-13.

An interesting approach to saturated hydrocarbons is provided by the case of perhydrotriphenylene (PHTP) (Scheme 12.7). In 1967 Farina and Audisio described the preparation of optically active anti-trans-anti-trans-anti-trans-perhydrophenylene (12) isomer [80]. The racemic compound 12 was transformed in the corresponding carboxylic acid by radical acylation with oxalyl chloride, followed by hydrolysis. The racemic acid of 12 was then resolved into enantiomers through transformation in its dehydro-abietylamine salt. Recovery of the acid followed by decarboxylation provided optically active 12. Originally, the AC of (−)-12 was assigned as (R) by the Brewster OR sign prediction method [81] Subsequently, the same authors converted an optically active acid of 12 to the 2keto derivative 13 and applied the octant rule to this ketone [82]. The enantiomer (+)-13 exhibited a positive band at 290 nm in ORD, therefore supporting (S ) conﬁguration for (+)-13 and for (+)-12 from which it derived [83]. Yet, the AC assignment of 12 could not be considered as fully reliable because of the empirical nature of the Brewster method and also the uncertain identiﬁcation of the structure of 13, and hence the octant rule assignment. Therefore it was more recently reinvestigated by Stephens et al. [84] by ab initio DFT calculations of VCD and ORD of 12. Monte Carlo conformational search by MMFF94 force ﬁeld, followed by optimization using DFT at the B3LYP/6-31G* level, found that the only conformation of 12 signiﬁcantly populated at room temperature is the one where all four cyclohexane rings adopt a chair conformation. The simulation of mid-IR and VCD spectra of (S )-12 by using DFT at the B3PW91/TZ2P level showed a very good agreement with the experimental spectra for the (+) enantiomer and, by that, conﬁrmed the previous assignment. The TDDFT calculation of OR of 12 at several wavelengths also was performed by TDDFT at B3LYP/aug-cc-pVDZ level. Qualitatively, calculated and experimental rotations were in agreement in sign and in dispersion, supporting once more the (S )/(+) relationship for 12.

12.5. CONCLUSIONS Certainly, selection of most suitable chiroptical method for the purpose of safe absolute conﬁgurational assignment depends on many factors which, ﬁrst of all, include equipment availability and speciﬁc chemical structure. The most important structural features of the molecule under investigation that must be taken into account are (a) the presence of chromophores, (b) the presence of functional groups that allow an introduction of chromophores if desirable, and (c) the level of substrate conformational ﬂexibility. For transparent nonfunctionalizable molecules, VCD and ORD computational predictions are the method of choice, while for other compounds also ECD spectroscopy can be used.

444

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The advantages of the latter are its higher sensitivity and the possibility of achieving conﬁgurational assignments by qualitative methods, in principle easier and quicker than the computational ones. For the nonspecialist the qualitative methods, when applicable, often still remain the ﬁrst choice. When the critical factors governing application of nonempirical exciton chirality method are taken into account, and no conformational ambiguity exists, the method can provide a straightforward and fast assignment of AC of a wide variety of chiral substrates even at a microscale level. In general, the computational approaches for AC analysis are becoming much more reliable and affordable, due to the fast methodological and technological advancement, although still some limitations given by the molecular size and conformational mobility exist. There is no doubt that the recent advance in chiroptical methods will open new opportunities for selection of the most suitable method for conﬁgurational assignment, or even new ways to apply routinely a few suitable methods instead of one with unparallel efﬁciency.

REFERENCES 1. (a) T. A. Baillie, K. M. Schultz, in The Impact of Stereochemistry on Drug Development and Use, H. Aboul-Enein, and I. W. Wainer, eds, John Wiley & Sons, New York, 1997, Chapter 2, pp. 21–43. (b) Chirality in Agrochemicals, N. Kurihara and J. Miyamoto, eds., John Wiley & Sons, New York, 1998. (c) Chirality in Industry, A. N. Collins, G. N. Sheldrake, and J. Crosby, eds., John Wiley & Sons, New York, 1997. 2. E. L. Eliel, S. H. Wilen, L. N. Mander, Stereochemistry of Organic Compounds, John Wiley & Sons, New York, 1994. 3. (a) H. Duddeck, in Houben Weyl Stereoselective Synthesis, G. Helmchen, R. W. Hoffmann, J. Mulzer, and E. Schaumann, eds., Thieme, Stuttgart, 1996, Vol. 1, p. 293. (b) H. D. Flack, G. Bernardinelli, Chirality 2008, 20 , 681–690. 4. (a) M. Seco, E. Qui˜noa, R. Riguera, Chem. Rev . 2004, 104 , 17–117. (b) T. Wenzel, Discrimination of Chiral Compounds Using NMR Spectroscopy, Wiley-Interscience, Hoboken, NJ, 2007. 5. (a) Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000. (b) N. Berova, L. Di Bari, G. Pescitelli, Chem. Soc. Rev . 2007, 36 , 914–931. (c) P. Polavarapu, Chirality, 2008, 20 , 664–672. 6. (a) M. Pecul, K. Ruud, Adv. Quantum. Chem. 2006, 50 , 185–227. (b) T. D. Crawford, Theor. Chem. Acc. 2006, 115 , 227–245. (c) P. L. Polavarapu, Chem. Record 2007, 7 , 125–136. (d) T. D. Crawford, M. C. Tam, M. L. Abrams, J. Phys. Chem. A 2007, 111 , 12057–12068. (e) J. Autschbach, Chirality 2009, 21 , E116–E152. 7. (a) L. A. Naﬁe, T. B. Freedman, Vibrational optical activity, in Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000; Chapter 4, pp. 97–132. (b) P. J. Stephens, F. J. Devlin, Chirality 2000, 12 , 172–179. (c) P. L. Polavarapu, Int. J. Quant. Chem. 2006, 106 , 1809–1814. (d) P. J. Stephens, F. J. Devlin, J.-J. Pan, Chirality 2008, 20 , 643–663. 8. P. J. Stephens, D. M. McCann, F. J. Devlin, J. R. Cheeseman, M. J. Frisch, J. Am. Chem. Soc. 2004, 126 , 7514–7521. 9. (a) L. D. Barron, L. Hecht, Vibrational optical activity: From fundamentals to biochemical applications, in Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, Chapter 23. (b) L. D. Barron, Molecular Light Scattering and Optical Activity, 2nd ed., Cambridge University Press, Cambridge, 2004. 10. (a) GAUSSIAN09, Revision A.02, Gaussian, Inc., Wallingford CT, 2009, www.gaussian.com. (b) DALTON, Release 2.0 (2005), www.daltonprogram.org. (c) TURBOMOLE, Version

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

11.

12.

13. 14.

15. 16. 17. 18. 19. 20. 21.

22.

23.

24.

25.

26. 27. 28. 29.

6.0, COSMOlogic GmbH & Co. KG, www.turbomole-gmbh.com. (d) ADF2009.01, SCM, Theoretical Chemistry, Vrije Universiteit, Amsterdam, The Netherlands, www.scm.com. (a) D. M. McCann, P. J. Stephens, J. R. Cheeseman, J. Org. Chem. 2004, 69 , 8709–8717. (b) P. J. Stephens, D. M. McCann, Chirality 2005, 17 (Suppl.) S52–S64. (c) D. M. McCann, P. J. Stephens, J. Org. Chem. 2006, 71 , 6074–6098. (a) E. Giorgio, C. Minichino, R. G. Viglione, R. Zanasi, C. Rosini, J. Org. Chem. 2003, 68 , 5186–5192. (b) E. Giorgio, R. G. Viglione, R. Zanasi, C. Rosini, J. Am. Chem. Soc. 2004, 126 , 12968–12976. J. Autschbach, L. Jensen, G. C. Schatz, Y. C. E. Tse, M. Krykunov, J. Phys. Chem. A 2006, 110 , 2461–2473. (a) N. Harada, K. Nakanishi, Circular Dichroic Spectroscopy: Exciton Coupling in Organic Stereochemistry, University Science Books, Mill Valley, CA, 1983. (b) N. Berova, K. Nakanishi, Exciton chirality method: Principles and applications, in Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, R. W. Woody, eds., Wiley-VCH, New York, 2000, Chapter 12, pp. 337–382. L. A. Naﬁe, H. Bujus, A. Rilling, X. Cao, R. K. Dukor, Appl. Spectr. 2004, 58 , 647–654. T. Furo, T. Mori, T. Wada, Y. Inoue, J. Am. Chem. Soc. 2005, 127 , 8242–8243. (a) H. A. Kramers, Nature 1926, 117 , 775–778. (b) R. D. L. Kronig, J. Opt. Soc. Am. 1926, 12 , 547–557. P. L. Polavarapu, A. G. Petrovic, S. E. Vick, W. D. Wulff, H. Ren, Z. Ding, R. J. Staples, J. Org. Chem. 2009, 74 , 5451–5457. P. L. Polavarapu, N. Jeirath, T. Kurtan, G. Pescitelli, K. Krohn, Chirality 2009, 21 , E202–E207. D. A. Lightner, J. E. Gurst, Organic Conformational Analysis and Stereochemistry from Circular Dichroism Spectroscopy, Wiley-VCH, New York, 2000. D. A. Lightner, The Octant Rule, in Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, Chapter 10, pp. 261–303. (a) H. E. Smith, Chem. Rev . 1998, 98 , 1709–1740. (b) H. E. Smith, Applications of the benzene sector and benzene chirality rules, in Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, Chapter 14, pp. 397–429. J. Gawronsky, Circular dichroism and chirality of dienes, in Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, Chapter 11, pp. 305–335. (a) E. Bunnenberg, C. Djerassi, K. Mislow, A. Moscowitz, J. Am. Chem. Soc. 1962, 84 , 2823–2826. (b) K. Mislow, E. Bunnenberg, R. Records, K. Wellman, C. Djerassi, J. Am. Chem. Soc. 1963, 85 , 1342–1349. J. Sandstr¨om, Determination of conformations and absolute conﬁgurations by semiempirical calculation of CD spectra, in Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, Chapter 16, pp. 459–490. (a) H. DeVoe, J. Chem. Phys. 1964, 41 , 393–400. (b) H. DeVoe, J. Chem. Phys. 1965, 43 , 3199–3208. (a) C. Rosini, M. Zandomeneghi, P. Salvadori, Tetrahedron: Asymmetry 1993, 4 , 545–554. (b) S. Superchi, E. Giorgio, C. Rosini, Chirality 2004, 16 , 422–451. Today molecules having M.W. up to 1000 u.m.a. can be computationally treated with common desktop PC. F. J. Devlin, P. J. Stephens, P. Besse, J. Org. Chem. 2005, 70 , 2980–2993.

445

446

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

30. S. Tartaglia, D. Padula, P. Scafato, L. Chiummiento, C. Rosini, J. Org. Chem. 2008, 73 , 4865–4873. 31. G. Pescitelli, T. Kurt´an, U. Fl¨orke, K. Krohn, Chirality 2009, 21 , E181–E201. 32. G. Mazzeo, E. Giorgio, R. Zanasi, N. Berova, C. Rosini, J. Org. Chem. 2010, 75 , 4600–4603. 33. N. Berova, G. Pescitelli, A. G. Petrovic, G. Proni, Chem. Commun. 2009, 5958–5980. 34. (a) S. Superchi, D. Casarini, A. Laurita, A. Bavoso, C. Rosini, Angew. Chem. Int. Ed. Engl . 2001, 40 , 451–454. (b) S. Superchi, R. Bisaccia, D. Casarini, A. Laurita, C. Rosini, J. Am. Chem. Soc. 2006, 128 , 6893–6902. 35. A. Moskowitz, Tetrahedron 1961, 13 , 48–56. 36. C. Djerassi, Optical Rotatory Dispersion, McGraw-Hill, New York, 1960. 37. (a) W. Mofﬁtt, A. Moscowitz, J. Chem. Phys. 1959, 30 , 648–660. (b) W. Mofﬁtt, R. B. Woodward, A. Moscowitz, W. Klyne, C. Djerassi, J. Am. Chem. Soc. 1961, 83 , 4013–4018. 38. T. D. Bouman, D. A. Lightner, J. Am. Chem. Soc. 1976, 98 , 3145–3154. 39. G. Hoffmann, R. Wiartalla, Tetrahedron Lett. 1982, 23 , 3887–3888. 40. U. Berg, E. J. Butkus, Chem. Res., Synop. 1993, 116–117. 41. H. Gerlach, Helv. Chim. Acta 1978, 61 , 2773–2776. 42. P. J. Stephens, D. M. McCann, E. Butkus, S. Stoncius, J. R. Cheeseman, M. J. Frisch, J. Org. Chem. 2004, 69 , 1948–1958. 43. E. Butkus, S. Stoncius, A. Zilinskas, Chirality 2001, 13 , 694–698. 44. G. Snatzke, P. C. Ho, Tetrahedron 1971, 27 , 3645–3653 45. H. H. Jaff´e, M. Orchin, Theory and Applications of Ultraviolet Spectroscopy, John Wiley & Sons, New York, 1962, p. 242. 46. S. T. Pickard, H. E. Smith, J. Am. Chem. Soc. 1990, 112 , 5741–5747. 47. (a) M. Dawn, H. E. Smith, Chirality, 1993, 5 , 20–23. (b) A. Rumbero, I. Borreguero, J. V. Sinisterra, A. R. Alcantara, Tetrahedron 1999, 55 , 14947–14960. 48. G. Pescitelli, L. Di Bari, A. M. Caporusso, P. Salvadori, Chirality 2008, 20 , 393–399. 49. W. Klyne, P. M. Scopes, The carboxyl and related chromophores, in Fundamental aspects and recent developments in optical rotatory dispersion and circular dichroism, F. Ciardelli, P. Salvadori, eds., Heyden, London, 1973, pp 126–147. 50. (a) A. Gedanken, K. Hintzer, V. J. Schurig, Chem. Soc. Chem. Commun. 1984, 1615–1616; A. Rodger, J. Am. Chem. Soc. 1988, 110 , 5941–5945. (b) A. Gedanken, Chiroptical properties of alcohols, ethers and peroxides, in The Chemistry of Hydroxyl Ether and Peroxide Groups, Vol. 2, S. Patai, ed., John Wiley & Sons, New York, 1993. 51. M. Milewska, M. Gdaniec, H. Maluszynska, T. Polonski, Tetrahedron: Asymmetry 1998, 9 , 3011–3023. 52. H. Rehling, H. Jensen, Tetrahedron Lett. 1972, 27 , 2793–2796. 53. J. Frelek, I. Panﬁl, Z. Urbanczyk-Lipkowska, M. Chmielewski, J. Org. Chem. 1999, 64 , 6126–6134. 54. J. Gawronski, Tetrahedron 1982, 38 , 3–26. 55. J. Gawronski in The Chemistry of Enones, Eds. S. Patai, Z. Rappoport, John Wiley & Sons, New York, 1989, pp. 55–105. 56. (a) M. Duraisamy, H. M. Walborsky, J. Am. Chem. Soc. 1983, 105 , 3252–3264. (b) M. Duraisamy, H. M. Walborsky, Ibid . 3264–3269. (c) M. Duraisamy, H. M. Walborsky, Ibid . 3270–3273. (d) H. M. Walborsky, S. M. Reddy, J. H. Brewster, J. Org. Chem. 1988, 53 , 4832–4846. 57. (a) L. A. Neubert, M. Carmack, J. Am. Chem. Soc. 1974, 96 , 943–945. (b) R. W. Woody, Tetrahedron 1973, 29 , 1273–1283. 58. J. Gawronski, P. Grycz, M. Kwit, U. Rychlewska, Chem. Eur. J . 2002, 18 , 4210–4215.

D E T E R M I N AT I O N O F M O L E C U L A R A B S O L U T E C O N F I G U R AT I O N

59. See, for example: P. J. Stephens, D. M. McCann, F. J. Devlin, T. C. Flood, E. Butkus, S. Stoncius, J. R. Cheeseman, J. Org. Chem. 2005, 70 , 3903–3913. 60. W. Kuhn, Trans. Faraday Soc. 1930, 46 , 293–308. 61. J. G. Kirkwood, J. Chem. Phys. 1937, 5 , 479–491. 62. W. Mofﬁtt, J. Chem. Phys. 1956, 25 , 467–478. 63. (a) S. F. Mason, Proc. Chem. Soc. 1962, 362–363. (b) S. F. Mason, G. W. Vane, J. Chem. Soc. (B) 1966, 370–374. 64. S. F. Mason, G. W. Vane, R. J. Wells, J. S. Whitehurst, J. Chem. Soc. (B) 1967, 553–556. 65. S. H. Wilen, J. Z. Qi, P. G. Williard, J. Org. Chem. 1991, 56 , 485–487. 66. F. J. Devlin, P. J. Stephens, Chem. Comm. 1999, 361–362. 67. E. Giorgio, PhD Thesis Universit`a della Basilicata (Italy), 2004. 68. (a) C. Rosini, S. Scamuzzi, G. Uccello-Barretta, P. Salvadori, J. Org. Chem. 1994, 59 , 7395–7400. (b) C. Rosini, S. Scamuzzi, M. Pisani Focati, P. Salvadori, J. Org. Chem. 1995, 60 , 8289–8293. (c) C. Rosini, G. P. Spada, G. Proni, S. Masiero, S. Scamuzzi, J. Am. Chem. Soc. 1997, 109 , 505–512. (d) M. I. Donnoli, P. Scafato, S. Superchi, C. Rosini, Chirality 2001, 13 , 258–265. 69. C. Rosini, M. I. Donnoli, S. Superchi, Chem. Eur. J . 2001, 7 , 72–79. 70. M. I. Donnoli, E. Giorgio, S. Superchi, C. Rosini, Org. Biomol. Chem. 2003, 1 , 3444–3449. 71. P. J. Stephens, A. Aamouche, F. J. Devlin, S. Superchi, M. I. Donnoli, C. Rosini, J. Org. Chem. 2001, 66 , 3671–3677. 72. (a) S. Superchi, M. I. Donnoli, C. Rosini, Org. Lett. 1999, 1 , 2093–2096. (b) S. Superchi, D. Casarini, C. Summa, C. Rosini, J. Org. Chem. 2004, 69 , 1685–1694. 73. M. Parveen, N. U. Khan, B. Achari, P. K. Dutta, Abstracts of Papers, 3rd International Symposium on Flavonoides in Biology and Medicine, Singapore, November 13–17, 1989; Abstract, p. 22. 74. N. Harada, H. Ono, H. Uda, M. Parveen, K. Nizam ud Din, B. Achari, P. K. Dutta, J. Am. Chem. Soc. 1992, 114 , 7687–7692. 75. (a) H. Uzawan, Y. Nishida, H. Ohrui, H. Meguro, J. Org. Chem. 1990, 55 , 116–122. (b) N. Harada, A. Saito, H. Ono, J. Gawronski, K. Gawronska, T. Sugioka, H. Uda, T. Kuriki, J. Am. Chem. Soc. 1991, 113 , 3842–3850. (c) N. Harada, A. Saito, H. Ono, S. Murai, H-Y. Li, J. Gawronski, K. Gawronska, T. Sugioka, H. Uda, Enantiomer 1996, 1 , 119–138. 76. P. Scafato, S. Superchi, Chirality 2010, 22 (1E), E3–E10. 77. P. Rashidi-Ranjbar, J. Sandstr¨om, J. Mol. Struct. 1991, 246 , 25–32. R. Isaksson, P. RashidiRanjbar, J. Sandstr¨om, J. Chem. Soc. Perkin Trans. 1991, 1147–1152. 78. G. Longhi, S. Abbate, P. Scafato, C. Rosini, Phys. Chem. Chem. Phys. 2010, 12 , 4725–4732. 79. (a) A. G. Petrovic, Y. Chen, G. Pescitelli, N. Berova, G. Proni, Chirality, 2010, 22 , 129–139. (c) Y. Chen, A. G. Petrovic, M. Roje, G. Pescitelli, Y. Yang, M. Kayser, N. Berova, G. Proni, Chirality 2010, 22 , 140–152. 80. M. Farina, G. Audisio, Tetrahedron Lett. 1967, 14 , 1285–1288. 81. J. H. Brewster, J. Am. Chem. Soc. 1959, 81 , 5475–5483. 82. M. Farina, G. Audisio, Tetrahedron 1970, 26 , 1827–1837. 83. M. Farina, G. Audisio, Tetrahedron 1970, 26 , 1839–1844. 84. P. J. J. Stephens, F. J. Devlin, S. Sch¨urch, J. Hulliger, Theor. Chem. Acc. 2008, 119 , 19–28.

447

PART III INORGANIC STEREOCHEMISTRY

13 APPLICATIONS OF ELECTRONIC CIRCULAR DICHROISM TO INORGANIC STEREOCHEMISTRY Sumio Kaizaki

13.1. INTRODUCTION Inorganic stereochemistry has made great progress associated with advances in chiroptical methods. The discovery of chiral structures in coordination compounds has not only played historical roles at the origin of coordination chemistry in the beginning of modern chemistry, but has also provided important clues in expanding it toward asymmetric catalysis, bioinorganic chemistry, and supramolecular chemistry in recent years. The ﬁrst application of chiroptical spectra to inorganic stereochemistry was made by the measurement of optical rotation (OR) after the syntheses of chiral transition metal complexes by the “father of coordination chemistry” Alfred Werner in the early twentieth century [1], although this technique had been used for organic chemistry. A ﬁrm proof of the octahedral coordination theory by synthesis of a chiral “carbon-free” compound is the well-known tetranuclear hexol-type complex [Co{(μ-OH)2 Co(NH3 )4 }]6+ , which was resolved into enantiomers in an epoch-making experiment [2]. The separated complexes were shown to possess equal and oppositely signed optical rotations at the NaD line (589 nm), conﬁrming their mirror-image structure. However, the signs of OR are not suitable for the determination of absolute chiral structures. Though the more reliable wavelength-dependent OR through an absorbing region, called optical rotary dispersion (ORD), was measuring chiral differences associated with the whole absorption region of metal complexes, it is now common and more useful to measure circular dichroism, or absorption differences between right- and left-circularly polarized light, through an absorption band, which is a technique equivalent to ORD [3]. Cotton’s ﬁrst observation (see Chapter 1 in this volume) of optical rotation measurements through an absorption band, along with interpretation in terms of differential Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

451

452

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

absorption of the circularly polarized light, was performed on solutions of L-tartrate complexes of chromium(III) and copper(II) [4]. This indicates relatively easier application of electronic circular dichroism (ECD) measurements and the merit of most transition metal complexes, which possess moderately intense d –d absorption bands in the more accessible visible and near-infrared region, in contrast to most organic compounds in which CD measurements are limited to the ultraviolet region. The availability of commercial instrumentation in the 1960s led to numerous investigations aimed at the determination of chiral inorganic stereochemistry through the measurement of ECD, as well as a better understanding of UV–vis spectra and electronic structure, leading to advances in the chiroptical theory for metal complexes [5–10]. On the basis of the fundamental information on the structure–spectra relation, recent prime interests are in application-oriented subjects such as asymmetric catalysis and chiral interactions between metal complexes and biomolecules such as proteins and DNA. As described below, however, there are often exceptions and complications in the development of reliable rules relating CD sign patterns to absolute conﬁguration or conformation, with need of conﬁrmation by NMR spectra or X-ray analysis of suitably grown single crystals and/or DFT theoretical calculations. In this chapter, the application of electronic circular dichroism (ECD) to inorganic stereochemistry will be reviewed and is limited to coordination compounds. ECD will mostly originate from the ligand-centered and charge-transfer transitions, as well as the metal-centered d –d or f –f transitions. Special attention will be given to the structure–spectra relation for a variety of chiral structures of coordination compounds, with the emphasis on current experimental and theoretical techniques.

13.2. ECD IN THE D–D TRANSITIONS OF TRIS- OR BIS-BIDENTATE TRANSITION METAL COMPLEXES The relationships between chiral structures and ECD for inert cobalt(III) and chromium(III) complexes with d 6 and d 3 conﬁgurations have been the most often studied, since stable chiral complexes can be synthesized, and the d –d bands have been very well characterized on the basis of the ligand-ﬁeld theory. The standard CD criterion to determine the chiral structure for the tris- or bis-bidentate Co(III) complexes and the related metal complexes is the CD spectrum in the ﬁrst d –d ligand-ﬁeld transition of tris(ethylenediamine) cobalt(III) ion, (+)589 -[Co(en)3 ]3+ , with ﬁve-membered diamine rings. The reasons to choose this Co(III) complex as the CD criterion are twofold. For one reason, the absolute conﬁguration of this complex was ﬁrst determined to be the so-called D conﬁguration in (+)589 -[Co(en)3 ]2 Cl6 ·NaCl·6H2 O crystal with anomalous X-ray diffraction techniques in 1955 by Saito et al. [11]. The original stereochemical descriptor D was later renamed to be , according to the helicity of the complex or the IUPAC skew-line convention [12], as shown in Figure 13.1. For the other reason, since the ﬁrst d –d ligand ﬁeld 1 T1 ←1 A1 transition of the Co(III) complexes is electric dipole-forbidden, but magnetic dipole-allowed, the CD intensity or dissymmetry factor g is relatively large compared with the second electric dipole-forbidden and magnetic dipole-forbidden d –d ligand ﬁeld 1 T2 ←1 A1 transition. As shown in Figure 13.2, -(+)-[Co(en)3 ]3+ gives a major positive and then a minor negative CD component from the low-frequency side in solution for the electric dipole-forbidden and magnetic dipole-allowed d –d 1 T1 ←1 A1 transition around 21,000 cm−1 . These correspond to the trigonal split states, 1 E and 1 A2 , which originate from the 1 T1 state in octahedral

A P P L I C AT I O N S O F E L E C T R O N I C C I R C U L A R D I C H R O I S M T O I N O R G A N I C S T E R E O C H E M I S T RY

Figure 13.1. Absolute conﬁguration of (+)589 -[Co(en)3 ]2 Cl6 ·NaCl·6H2 O.

Figure 13.2. Vis–UV and CD spectrum of -(+)-[Co(en)3 ]3+ in water (broken line) and the axial single-crystal CD spectrum (solid line) of -(+)-{[Co(en)3 ]Cl3 }2 ·NaCl·6H2 O. The assignments to the octahedral (1 T1g ) and trigonal (1 E and 1 A2 ) states are represented.

ﬁeld as shown in Figure 13.2 [13]. Uniaxial single-crystal polarized CD measurements of (+)589 -[Co(en)3 ]2 Cl6 ·NaCl·6H2 O reveal that the CD sign for the 1 E component is positive and hence the major positive band observed in solution is due to the 1 E component (Figure 13.2). The crystal CD intensity is found to be an order of magnitude larger than that of the solution CD [14]. On the other hand, the apparent trigonal splitting estimated from the solution CD is 3000 cm−1 , in contrast to very small splitting of 0–70 cm−1 from the single-crystal measurements. These facts indicate that the observation of weaker CD intensity and the apparently large trigonal splitting in the solution CD results from a mutual cancellation between oppositely signed 1 E and 1 A2 CD components with similar intensities. The net CD sign is that of the 1 E component.

453

454

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The CD component rule may also be applied in situations involving ligand-to-metal charge-transfer (CT) transitions in the UV region. In this application, negative and positive CD peaks in the CT region around 40,000–50,000 cm−1 and 50,000–60,000 cm−1 , respectively, are associated with the conﬁguration, and positive and negative ones are assigned to a conﬁguration. The validity of the extension of this rule to the CT region is reasonable if one considers that the major 1 E(1 T1 ) transition “borrows” CD intensity from the lower frequency 1 E(CT) rather than from the higher frequency 1 A2 (CT) on the basis of the assigment to the 1 E(CT) from the single-crystal CD (Figure 13.2) [14]. One exceptional case showing a positive sign results from the large contribution from the conformation of the ptn (2,4-diaminopentane) ligand in the six-membered diamine chelate, -(+)546 -[Co(R,R-ptn)3 ]3+ [7]. An empirical rule for the absolute conﬁgurations of tris- and bis-bidentate chelate complexes of the cis-[CoX2 (en)2 ] type was proposed from the experimental results that a positive and negative sign of the major 1 E CD component (of trigonal parentage) indicated a and complex, respectively [15]. It is important to note that some exceptions for the diamine complexes were found to be the strained ﬁve-membered diamine chelate complex (δδδ)-[Co(R,R-cptn)3 ]3+ (cptn = trans1,2-cyclopentanediamine), the six-membered diamine chelate complex, -[Co(tn)3 ]3+ (tn = trimethylendiamine) and -[Co(R,R-ptn)3 ]3+ (R,R-ptn = 2R, 4R-pentanediamine), and the seven-membered diamine chelate complex -[Co(tmd)3 ]3+ (tmd = 1, 4diaminobutane), which display a major negative CD in the spin-allowed 1 T1 ←1 A1 transition. For the (δδδ)-[Co(R,R-cptn)3 ]3+ complex, the conformational contribution ε(δδδ) from the three RR-cptn ligands exceeds the conﬁgurational ε() contribution as found experimentally from the CD for the mixed RR- and SS -cptn complexes, so the apparent solution CD intensity for the lower frequency 1 E component is smaller than that for the 1 A2 one [16]. Thus, the empirical rule still holds with respect to the conﬁgurational CD. The results for the -[Co(tn)3 ]3+ complex are due to the oppositely (negative) signed 1 E CD component for the -conﬁguration, probably because of the distortion from a regular octahedron: the chelate N–Co–N bite angle α(N–Co–N) > 90◦ or the radial (azimuthal) expansion (φ > 60◦ ), as compared to the cases for the -ﬁve-membered diamine chelate complexes with the smaller bite angle (α(N–Co–N) < 90◦ ) and azimuthal contraction (φ < 60◦ ) giving a positive major 1 E CD as theoretically predicted (vide infra) [17, 18]. However, these exceptions clearly are problematic if one is relying upon ECD measurements to determine absolute conﬁguration. The same empirical rule has been shown to be applicable for tris- or bis-diamine Cr(III) and Ni(II) complexes; that is, the conﬁguration gives a positive major CD band in the electric dipole-forbidden and magnetic dipole-allowed d –d transition of 4 T2 ← 4A2 for Cr(III) and 3 T2 ← 3A2 for Ni(II) [19]. The solution CD spectra of (+)-[Cr(en)3 ]3+ gives only one positive component (Figure 13.3), which is assigned to one (4 E(4 T2 )) of the trigonal split states by the single-crystal polarized CD measurement [20, 21]. By this CD component rule, using the signs of the 4 E(4 T2 ) component, or the trigonal parentage for the lower symmetry complexes, the absolute conﬁgurations of many tris- or bis-bidentate complexes were determined. For Cr(III) complexes, it has been found that characteristic weak but sharp CD patterns in the spin-forbidden transitions within the t2g subshell of Cr(III) complexes yield important information on the CD behavior in the spin-allowed as well as spin-forbidden transitions and are correlated with the absolute conﬁguration. In the

A P P L I C AT I O N S O F E L E C T R O N I C C I R C U L A R D I C H R O I S M T O I N O R G A N I C S T E R E O C H E M I S T RY

455

A(+)589−[Cr(en)3]3+ 2

−0.5 AB log e

log e 2 Δe

−1

1

1

× 100

CD 0 (2A,E) Eb (2A,Ea) 2 2 E A2 2E

−1

(4A1) 4E

Eg 2T1g

15

16

Figure 13.3. CD spectra in the spin-forbidden and

4

2

T2g

20 σ (103 cm−1)

spin-allowed transitions of -(+)-[Cr(en)3 ]3+ in

25

water. The assignments to the doublet and quartet states are given in terms of the single-group and/or double-group representations.

room-temperature solution CD spectra in the spin-forbidden 2 E, 2 Tl ← 4A2 d –d transitions of -(+)-[Cr(en)3 ]3+ , three sharp peaks, (+), (−) and (+), are observed, as shown in Figure 13.3. The CD sign of the lowest-frequency 2 E component is the same as that of the major 4 E CD component [22–24]. This empirical correlation also holds for the circularly polarized luminescence (CPL) of (+)-[Cr(en)3 ]3+ [25] and the chiral ––[EuIII CrIII (L2)3 ]6+ complex (see Figure 13.10) [26] in the 2 E– 4 A2 transitions, as well as the CD of a number of mixed-ligand Cr(III) complexes of tris-chelate type [22–24]. This is elucidated on the basis of the theoretical approaches for the rotational strengths between the spin-forbidden and the spin-allowed transitions in terms of the intensity borrowing mechanism through the spin–orbit coupling between the quartet-doublet states [23, 24]. That is, the rotational strength of the 2 E(2 Eg ) component is predicted to be proportional to the net rotational strength R(4 T2 ) = 2 R(4 E) + R(4 A1 ) : R(2 E(2 Eg )) = 32k [R(4 E) + R(4 A1 )](k = ζ /18(E (4 T2g ) − E (2 ), where ζ is the spin–orbit coupling constant and E (4 T2g ) − E (2 ) is the energy interval between 4 T2g and 2 . The CD sign of the 2 E(2 Eg ) component should be the same as that of the 4 E one with |R(4 A1 )| < |R(4 E)|. The other sharp CD peaks were assigned to the 2 A2 (2 T1g ) and 2 E(2 T1g ) from the lower-frequency side (Figure 13.3). The CD signs of the remaining higher-frequency 2 A2 (2 T1g ) and 2 E(2 T1g ) states are negative and positive as predicted in view of R(2 A2 (2 T1g )) = 2k [R(4 E) + 4R(4 A1 )] and R(2 E(2 T1g ))) = 24k [5R(4 E) + 2R(4 A1 )] for the negative 4 A1 and positive 4 E CD components. Some exceptions to the 2 E CD sign rule are observed for (+)546 -[Cr(acac or acaX)2 (en)]+ (acac = acetylacetonate; acaX = 3-halogenoacetylacetonate) and (−)D -[Cr(bgH)3 ]3+ (bgH = biguanide(NH{C(NH2 )NH}2 )) complexes [23]. In these cases, since the 2 E component is strongly split into the 2A(2 Eg ) and E (2 Eg ) levels, where denotes the double-group irreducible representation, owing to the large trigonal

456

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

splitting (K ) of the 4 T2 state, the two components on the lower-frequency side are assigned to the 2A(2 Eg ) and E (2 Eg ), of which the observed signs, (−) and (+), for the -isomer agree with the predicted signs (R(2A(2 Eg )) = 4k [3R(4 E) + 6R(4 A1 )] and (R(E (2 Eg )) = 4k [5R(4 E) + 2R(4 A1 )]) with R(4 E) > 0 and R(4 A1 ) < 0 and the energy ordering between the 2A(2 Eg ) and E (2 Eg ) levels.

13.2.1. Theoretical studies on the structure-spectra relationships There have been a number of qualitative or quantitative attempts to develop a theoretical basis for CD component rules for -[Co(en)3 ]3+ and related complexes by using the crystal-ﬁeld theory with metal d –p orbital mixing [17] and the molecular orbital models (ligand-ﬁeld theory) with mixing between the metal d and ligand orbitals [14, 27–31]. These efforts have primarily been concerned with developing possible chiral spectrastructure relationships or so-called “sector rules” with the emphasis on the source of d –d electric dipole-forbidden transition intensity. Mason and Seal [27] have calculated more rigorously the rotational strengths in the d –d transitions for various kinds of diamine Co(III) complexes by using the so-called dynamic-coupling ligand-polarization model where the electric dipole transition moment results from the allowed intraligand transitions, which is complementary to the crystal-ﬁeld theory. This afforded quantitatively the net rotational strengths, R(1 T1 ), as well as the rotational strengths R(1 E) and R(1 A2 ) for the 1 E and 1 A2 components. These calculations, based on X-ray crystallographic structural data, include the polarizabilities of each XHn (X = C and N) group in the amine ligands together with the allowed higher-order hexadecapole moments in the d –d transitions of the cobalt ion. The calculated signs and magnitudes for the rotational strengths are in fairly good agreement with the observed ones, but still not consistent with the net rotational strength R(1 T1 ) for the -[Co(tmd)3 ]3+ and (λλλ)-[Co(R, R-ptn)3 ]3+ complexes. Later, two ab initio (see Chapter 22 in Volume 1) calculations of the CD of [Co(en)3 ]3+ have been reported to reproduce the CD pattern. Both theoretical studies suggested that the electric dipole transition moments stem from N–H and N–C σ –σ * intra-ligand transitions, as in the dynamic coupling model [8, 32, 33]. This result may help explain the violation of the empirical CD rule for [Co(tn)3 ]3+ . Very recently, time-dependent density functional theory (TD-DFT) calculations have been applied to bis-diamine as well as tris-diamine or tris(π -conjugated bidentate chelate) Co(III) or Rh(III) complexes in the entire experimental spectral region, including the charge-transfer transitions, by Autschbach and Ziegler [34, 35]. This ab initio theoretical method reproduces the experimental CD spectra not only in the d –d transitions, but also in the charge-transfer region. For the Rh(III) complexes, there is a good ﬁt in energy and signed magnitude between the observed and calculated CD, but for the Co(III) complexes, energy shifts in the CT transitions are needed to ﬁt the observed spectra. The agreement between the experimental and theoretical CT energy is improved by a theoretical estimation of the solvent effect for +3 charged Co(III) complexes. The TD-DFT calculations of the hypothetical -[M(NH3 )6 ]3+ show that the CD is dependent on two geometrical deformations from a regular octahedron. One is the azimuthal distortion that allows the even d (eg ) orbital to mix with the odd σ ligand orbitals, resulting in nonzero rotational strengths in the magnetic dipole-allowed d –d transitions. The model calculations predict a positive 1 E CD for the azimuthal contraction with φ < 60◦ and a negative 1 E CD for the azimuthal expansion with φ > 60◦ for the -conﬁguration (vide supra). The other is the polar distortion, giving rise to trigonal splitting or governing the signs of the trigonal splitting parameter K = 2/3{E (E) − E (A)}. It has been demonstrated in these calculations

A P P L I C AT I O N S O F E L E C T R O N I C C I R C U L A R D I C H R O I S M T O I N O R G A N I C S T E R E O C H E M I S T RY

that K > 0 for the polar elongation (s/h < 1.22) and K < 0 for the polar compression (s/h > 1.22). (h denotes the distance between opposite faces, and s is the length of the triangular side in the octahedron.) This simple geometrical consideration accounts for the exceptional case of the major CD component rule for [Co(tn)3 ]3+ . The TD-DFT calculations are also performed for the corresponding paramagnetic Cr(III) complexes with an open-shell ground-state conﬁguration for the tris(en) and the tris(acetylacetonato) or tris(oxalato) complexes with π -conjugated bidentate chelates [36]. It appears that the description of the relative contributions of the d –d and charge-transfer transitions to the CD spectrum is also supported by the DFT calculations. This DFT analysis provides a new reliable tool in developing an understanding of the relationship between CD spectra and absolute conﬁguration of metal complexes. Theoretical simulations of transition metal complexes is discussed in more detail in Chapter 22 in volume 1.

13.3. POLYNUCLEAR COMPLEXES WITH CONFIGURATIONAL CHIRALITY 13.3.1. Dinuclear Complexes

e (mol−1dm3cm−1)

The dihydroxy-bridged dinuclear complex [Cr2 (μ-OH)2 (S , S -chxn)4 ]4+ (S , S chxn = (1S , 2S )-1,2-trans-cyclohexanediamine) is stereospeciﬁcally formed to take – absolute conﬁguration [37]. This assignment is in agreement with that based on the major positive CD signs in the ﬁrst d –d transition. As shown in Figure 13.4, a negative CD peak with a sharp half-band width (ν1/2 = 740 cm−1 ) near 285 nm or 35,100 cm−1 is observed, corresponding to the weak absorption shoulder near 285 nm with ε = 70 M−1 cm−1 . Since the half-band width is small and the transition energy (35,100 cm−1 ) is nearly equal to the sum of 15,000 cm−1 and 20,000 cm−1 , respectively, of the 2 2 E, T1 -4 A2 and 2 T2 −4 A2 transition, this CD peak is due to the double excitation from the lowest singlet (S = 0) level in the ground antiferromagnetic spin-coupled 4 A2 , 4 A2

200

AB

× 15

100 0

×2 N 300

400

500

600

Δe (mol−1dm3cm−1)

N +2

CD

N ×4

+1 300

N H O

700

× 25

400

M

M

O H N Λ

N N

N Λ

0 −1

× 10

500 λ (nm)

600

700

−2

Figure 13.4. UV–vis (upper) and CD spectra (lower) of -[Cr(OH)(S,S-chxn)2 (H2 O)]2+ (broken line) and -[Cr2 (OH)2 (S,S-chxn)4 ]4+ (solid line) in water. The dihydroxo-bridged dinuclear structure is shown on the right-hand side.

457

458

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

state to the excited 2 E, 2 T2 or 2 T1 , 2 T2 singlet level. Thereby, the transitions become formally singlet–singlet spin-allowed, otherwise quartet–doublet spin-forbidden, by the exchange-coupling mechanism for the antiferromagntic interaction (2J = −16 cm−1 ) [37]. This CD peak is a kind of dimer band that intensiﬁes with lowering temperature, giving a probe for dinuclear structures. Another interesting example of binuclear octahedral-based chirality is the chiral polyoxometalate (+)589 -[Co2 Mo10 O34 (OH)4 ]6− , where the CD measurement has conﬁrmed the same D2 symmetry [38] as the binuclear complexes discussed above, revealed by X-ray analysis [39]. (+)589 -[Co2 Mo10 O34 (OH)4 ]6− is used as a dopant to form polypyrrole ﬁlms in conducting polymers with the magnetotransport properties in terms of a magnetochiral anisotropy effect [40]. Though the temperature-dependent conductivity of the chiral polymer is different from that of the racemic polymer, the CD of the chiral polymer shows that macroasymmetry is not induced in the chiral polyoxometalate dopant.

13.3.2. Tetranuclear Complexes of Hexol Type Shimura et al. [41–44] succeeded in completely resolving the diastereomers (diamine = en, of tetranuclear complexes [Co{(μ-OH)2 Co(diamine)2 }3 ]6+ meso-R, S -butanediamine, R-propylenediamine(R-pn), and (1R, 2R)-1,2-transcyclohexanediamine(R, R-chxn)), which are analogous to the well-known chiral tetranuclear hexol complex [Co{(μ-OH)2 Co(NH3 )4 }3 ]6+ with and enantiomers [2]. The structures for these diamine complexes consist of the eight diastereomers ()/(), ()/(), ()/(), and ()/() as shown in Figure 13.5. The absolute conﬁgurations of the metal-centered conﬁgurations around the Co(O–O)3 moiety were determined on the basis of the CD component rule around 16,500 cm−1 . From the CD signs for the cis-[Co(diamine)2 (H2 O)2 ]3+ generated by acid decomposition of the diastereomers, the absolute conﬁgurations of the peripheral Co(O–O)(en)2 moiety could be assigned. The CD spectra of [Co{(μ-OH)2 Co(en)2 }3 ]6+ diastereomers were not found to follow simple additivity for the two main CD contributions due to the central Co(O–O)3 and the peripheral Co(O–O)(N–N)2 moieties. On the other hand, from the CD behavior of the corresponding chiral R-pn or R, R-chxn complexes, it was found that the vicinal contribution of the R-pn and R, R-chxn complexes is due to the asymmetric carbon(s) with R-conﬁguration but not to the λ-ring conformation, in view of the CD intensity being 2ε (R-pn) = ε (R, R-chxn). Such a dominant vicinal contribution of asymmetric carbons is the ﬁrst case for octahedral six-coordinate complexes, unlike the mononuclear complexes with chiral organic ligands. CD spectra of another type of chiral hexol structures have been measured in Anderson-type heteropoly acids, telluratocobaltate(III) [Co4 Te3 O18 (en)3 ]6− , periodatocobaltate(III) [Co4 I3 O18 (en)3 ]4− , and diastereomers of [Co4 I3 O18 (l-ala)3 ]3− [45]. On the basis of the CD signs in the d –d transitions, the absolute conﬁguration of (+) [Co4 Te3 O18 (en)3 ]6− and (−)-[Co4 I3 O18 (en)3 ]3− are ( )( )3 and ( )( )3 , respectively, where , and concern the absolute conﬁgurations around the central Co(MO6 )3 , Co(CoO4 en)3 , and the peripheral Co(en)(MO6 )2 , respectively. The CD intensities of the heteropoly acids are weaker than those of the hexol-type Co(III) complexes. This may be due to the insertion of three MO6 octahedra between the peripheral Co(III) octahedra of the hexol-like Co4 moiety [44]. Recently, Sato and co-workers [46] succeeded in preparing and resolving into eight enantiomers of another hexol type of so-called

A P P L I C AT I O N S O F E L E C T R O N I C C I R C U L A R D I C H R O I S M T O I N O R G A N I C S T E R E O C H E M I S T RY

(a)

459

(b)

Figure 13.5. Four diastereomers of (c)

(d)

[Co{(OH)2 Co(en)2 }3 ]6+ : (a) (), (b) (), (c) (), (d) ().

star-burst tetranuclear tetraacetylethanate(taet)-bridged Ru(III) complexes with acetylacetonate, -[{-Ru(III)(acac)2 (taet)}3 Ru(III)]. Additivity between the CD contributions from the conﬁgurational chirality of the central and peripheral moieties is found in contrast to the nonadditivity for [Co{(μ-OH)2 Co(en)2 }3 ]6+ diastereomers mentioned above. This CD behavior is used as a diagnostic tool to discriminate among the diastereomeric ()/(), ()/(), ()/(), ()/() enantiomers.

13.4. ECD IN THE 4f –4f TRANSITIONS Though a number of experimental studies on ECD in the 4f –4f transitions attempted to ﬁnd the CD-sensitive bands and/or the relationship between the CD signs and the chirality of ligands [47–51], the structure–spectra relationship including the absolute conﬁgurations around lanthanide ions could not be examined in detail in contrast to transition-metal complexes. This is because lanthanide complexes in solution are too labile to ﬁx the chiral structures and because the 4f –4f transition intensities are too weak for reliable ECD measurements, which require very high concentrations and long cell pathlengths, and are less understood in theory [52–54]. A nonracemic ground state can be generated by the so-called Pfeiffer effect on addition of a chiral compound [55]. However, this method cannot be used to provide the structure–spectra relationship, since the absolute conﬁguration cannot be deﬁnitely identiﬁed. On the other hand, there are several examples of conﬁgurationally chiral complexes that are stereospeciﬁcally formed with chiral ligands, but recently a limited number of CD data for emissive lanthanide(III) complexes, such as Eu(III), Tb(III), or Yb(III) complexes, makes it feasible to provide

460

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

the structural probe in connection to circularly polarized luminescence (CPL) [10, 56, 57] (see Chapter 3 in volume 1). Kaizaki and co-workers [58, 59] reported ECD for two series of lanthanide complexes with conﬁgurational chirality. Chiral (-)-[(acac)2 Cr(ox)Ln(HBpz3 )2 ] was prepared from a mixture by a chiral interaction between -[Cr(acac)2 (ox)]− and racemic [Ln(HBpz3 )2 ]+ (HBpz− 3 = hydrotris(pyrazol-1-yl)borate), resulting in stereospeciﬁc complex formation. The retention of the absolute conﬁguration around SAPR-8-[Ln(ox)(HBpz3 )2 ]+ (SAPR-8 = square antiprism eight-coordinate) in the Cr(ox)Ln assembly upon combination with inert -[(acac)2 Cr(ox)]− was conﬁrmed by X-ray analysis as well as by the large intensity of 4f –4f CD of the Ln(III) moiety in solution. The CD in the 4f –4f transitions of these structurally well-deﬁned complexes enables us to propose a criterion to determine the absolute conﬁguration, though they are limited to the NIR region in view of overlapping with the d –d bands of the Cr(III) moiety in the UV–vis region. By comparing CD patterns of the (-)-Cr(ox)Ln complexes, an empirical criterion for the relation between the 4f –4f CD signs and the absolute conﬁgurations around the Ln(HBpz3 )2 (ox) moiety was given: a positive sign of the major CD band in the 4 I9/2 → 4F3/2 (Nd), 6 H5/2 → 6F11/2 (Sm), 6 H15/2 → 6F7/2 (Dy), 5 I → 5H (Ho), 3 H →3 H (Tm) transitions for which Richardson’s classiﬁcations [54] 8 5 6 4 are type 5 and RII for the rotational strength () and DIII for the dissymmetry factor(g), with S = 0, L ≥ 0, 2 ≤ J ≤ 6(J = 0 = J ). The 4 I15/2 → 4I11/2 transition of Er is an exception for this relation. On the other hand, much invaluable information on the spectra–structure relation in the CD of the --Cr(ox)Ln complexes could be provided by a series of ECD spectra in the vis–NIR region for the Cs[Ln((+)-hfbc)4 ] (Cs-Ln) complexes, where (+)-hfbc is heptaﬂuorobutyryl-(+)-camphorate, taking a -SAPR-8-(llll ) conﬁguration (l between sites in different squares) with four helically bladed propellers, as shown on the right side of Figure 13.6 [60, 61]. The conﬁgurational chirality of Cs[Ln((+)-hfbc)4 ] is retained by the intramolecular interaction between the ﬂuorocarbon and cesium ion, as revealed by the exciton CD spectra (vide infra). The Cs–Yb complex gives a strong positive CD peak in the 2 F7/2 → 2F5/2 transition with g = +0.318, much larger than that (+0.06) of the Cr(ox)Yb complex and the reported values (−0.14) for -[Yb(DOTMA)][56] (see Chapter 11 in volume 1). The CD peak in the 2 F7/2 → 2F5/2 transition gives a suitable criterion for the absolute conﬁguration of the Yb complexes: a positive sign for the conﬁguration, because this transition is CD-sensitive in view of the transition type 1 with S = 0, L = 0, J = 0, 1(J = 0 = J ) and the transition properties belong to the RI and DII class for which the CD and g values are larger, respectively, than those of RII and DIII [54]. For the 6H 6 5/2 → H7/2 transition for the Cs–Sm complex with the same type 1 and the same DII class as that for the Yb complexes, however, this criterion could not be adopted. That is, two large negative CD peaks are observed around 1100 cm−1 with g = −0.01 to −0.005 for the absolute conﬁguration of the Cs–Sm complex, not in accordance with that for the Yb complexes [62]. The aforementioned criterion for the Cr(ox)Ln complexes can be applied to the other Cs–Ln complexes. There are two exceptions, for the Cs–Sm complex (6 H5/2 → 6 F11/2 ) and the Cs–Er complex(4 I15/2 → 4I11/2 ), which give a negative sign for the conﬁguration. Therefore, this criterion is not suitable for the Cs–Ln complexes. So far, Richardson’s classiﬁcation or selection rules could not provide a criterion of the absolute conﬁguration of the lanthanide(III) complexes.

A P P L I C AT I O N S O F E L E C T R O N I C C I R C U L A R D I C H R O I S M T O I N O R G A N I C S T E R E O C H E M I S T RY

0.05 0

Δε

−0.05 Nd

−0.15

ν (10 −0.25 14.00

3 cm−1)

16.00

18.00

20.00

0.3 MI

Δε

0 −0.3 −0.6 −0.9 20.00

Ln Ho Δ-SAPR-8-C4(llll)-M[Ln(+)-(hfbc)4] with an encapsulated alkali metal ion

ν (103 cm−1) 21.00

22.00

23.00

24.00

25.00

0.04

Δε

0 −0.04 −0.08 −0.12

ν (103 cm−1)

Er

−0.16 12.00 14.00 16.00 18.00 20.00 22.00 24.00

Figure 13.6. CD spectra in the hypersensitive 4f –4f transitions of -Cs[Ln((+)-hfbc)4 ] in CHCl3 (left) and the proposed structure in solution (right). (See insert for color representation of the ﬁgure.)

Close comparison among the ECD in the UV–vis to NIR region of the -Cs–Ln complexes reveals that the CD signs in the hypersensitive transitions with environmental sensitivity are correlated to the absolute conﬁgurations of the Cs–Ln complexes [62], though theoretically the hypersensitive transitions belonging to RII or DIII class of type 5 for the S = 0 or type 11 for the S = 0 with L ≥ 0, 2 ≤ J ≤ 6(J = 0 = J ) may not be the most favorable chiroptical probe as claimed by Richardson [54]. That is, as shown in Figure 13.6, the CD components are negative in the 4 I9/2 → 4G5/2 (Cs–Nd), the 5 I8 → 5G6 (Cs–Ho), and the 4 I15/2 → 2H11/2 (Cs–Er) transitions. For other hypersensitive transitions, the -Cs–Eu complex gives a negative 5 D0 – 7 F2 CPL peak at 16,300 cm−1 [63] as well as a negative 5 D2 – 7 F0 CD shoulder near 21,700 cm−1 on the intense negative peak at 24,600 cm−1 which may be assigned to the singlet–triplet π –π * transition of (+)-hfbc in view of the intensity and position as well as the band width [62]. For -Na3 [Eu(ODA)3 ]·2NaClO4 ·6H2 O(ODA = oxydiacetate) with trigonal D3 symmetry, a negative 5 D2 – 7 F0 CD peak was clearly observed at 21,500 cm−1 [64], and the CPL (see Chapter 3 in Volume 1) in the 5 D0 – 7 F2 emission showed a negative major component at 16,200 cm−1 [57, 65]. --[EuIII CrIII (L2)3 ]6+ with distorted trigonal D3 symmetry around Eu(III) gave a negative 5 D0 – 7 F2 CPL peak at 21,500 cm−1 [26].

461

462

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

This relation of the CD and CPL signs with the absolute conﬁguration is in contrast to the case of the CD-sensitive 5 D0 – 7 F1 magnetic dipole-allowed transition of which the CPL and/or CD sign for the -Cs–Eu and --[EuIII CrIII (L2)3 ] is opposite to that for -Na3 [Eu(ODA)3 ]·2NaClO4 ·6H2 O. It is noted that these CD or CPL peaks in the hypersensitive transitions are composed of dominant negative component(s) and are not disturbed by neighboring positive peak(s) and are thus without mutual cancellation. Therefore, these hypersensitive CD or CPL components could be an appropriate criterion of the chiral structure–spectra relation for SAPR-8(llll ) or trigonal D3 lanthanide(III) complexes, even though there are some differences in ligand electronic structure and coordination polyhedra among these Eu(III) complexes [57].

13.5. EXCITON ECD IN THE INTRALIGAND TRANSITIONS 13.5.1. Tris- and Bis-Chelate Complexes For tris- and bis-chelated octahedral six-coordinate complexes having conjugated aromatic bidentate ligands, a large CD couplet can be observed in the UV region, as predicted (Figure 13.7) [66] in contrast to the corresponding mono-chelated complexes that give a single CD component with a positive sign for the -conﬁguration and vice versa. These CD bands are considered to be due to exciton splitting (see Chapter 4 in this volume) and the zero-order rotational strengths are induced from the helical charge displacement by the coulombic coupling of the two or three long-axis π –π * transitions λ phen

bpy

[M(phen)3]

[M(phen)2X2]

[M(bpy)(phen)X2]

[M(phen)2(bpy)]

[M(bpy) 2(phen)]

Figure 13.7. Predicted exciton CD patterns of bpy and phen complexes.

A P P L I C AT I O N S O F E L E C T R O N I C C I R C U L A R D I C H R O I S M T O I N O R G A N I C S T E R E O C H E M I S T RY

of the identical ligands. For tris-chelate complexes with D3 symmetry, there are E and A2 couplings for which the rotational strengths are positive and negative, respectively, going from lower to higher frequency for the conﬁguration, on the basis of the simple exciton approximation. This pattern of oppositely signed CD bands is called a couplet and is designated as a positive couplet if the low-frequency band is positive, as in this case, and as a negative couplet if the low-frequency band is negative. The situation is the same for bis-chelated complexes as for the tris complexes. On the basis of this exciton model, the absolute conﬁgurations can be determined for tris- and bis-(chelated) complexes with 2,2 -bipyridine (bpy), 1,10-phenanthroline (phen), acetylacetonate (acac), 1,2-benzenediolate or catecholate (cat), and related aromatic ligands. The exciton CD for an octahedral six-coordinate metal complex with a closed-shell conﬁguration was observed for -(+)-[Si(acac)3 ]+ by Larsen et al. [67], -(+)-[As(cat)3 ]+ by Mason and Mason [68], and -(+)-[Si(bpy or phen)3 ]+ by Yoshikawa et al. [69] and were shown to give a positive couplet in the pure π –π * transition of the acac, cat, or phen ligand. This assignment was conﬁrmed by the X-ray analysis of -(+)-[As(cat)3 ]− by Ito et al. [70]. Even for the transition-metal complexes with an open-shell conﬁguration, such as [M(bpy)3 ]n+ and [M(phen)3 ]n+ (M = Fe2+,3+ ; Ru2+,3+ ; Os2+,3+ ), where there is mixing with the d –d and CT transitions, exciton couplets have been observed for the intraligand π –π * transitions. In particular, most Co(III) and Cr(III) complexes exhibit agreement between the absolute conﬁgurations based on the d –d CD and the exciton CD and/or the X-ray structure determinations [71]. For -[Co(bpy)3 ]3+ , -[Ni(bpy)3 ]2+ , and [Ni(phen)3 ]2+ , however, the absolute conﬁguration based on the exciton CD analysis is in agreement with the X-ray crystal structure, but the empirical d –d CD gives the incorrect result [71, 72]. The reverse situation is encountered for the tris-biguanide Co(III) complex. -(−)D -[CoIII (bgH)3 ]3+ (bgH = NH{C(NH2 )NH}2 ), for which the major CD rule gives the same absolute conﬁguration as the X-ray analysis, exhibits a UV CD couplet opposite to the exciton CD prediction, unlike the corresponding -(−)D -[CrIII (bgH)3 ]3+ . This may be due to uncertain assignment of the intraligand transition arising from large mixing of the d –π orbital with the bgH orbital in the Co(III) complex [73, 74]. Recent time-dependent density functional theory (TD DFT) studies showed that the exciton CD for [M(phen)3 ]2+ (M = Fe, Ru, Os) complexes [75, 76] and related complexes [77] are reproduced in good agreement with the experimental pattern. For nonidentical mixed-ligand complexes, such as -(−)589 -[Cr(ox)(bpy)(phen)]+ , the CD pattern would be two positive peaks if there were no coupling between the nonidentical π –π * transitions of bpy and phen, since a positive single CD component in the intraligand π –π* transition is observed for mono-phen or bpy complexes such as -[Co(en)2 (phen)]3+ [78] and --[Cr2 (L-tart)2 H(bpy or phen)2 ]− [79]. However, this mixed-ligand complex gives a positive couplet in the region of the bpy and phen transitions [80] as predicted by the exciton theory as in Figure 13.7. This type of nondegenerate exciton CD may also be applied with success to more complicated nonidentical couplings in [M(bpy)2 (phen)]n+ or [M(bpy)(phen)2 ]n+ . These complexes exhibit the expected characteristic exciton CD patterns with three components: (+),(+),(−) from lower to higher frequency for the -isomers, as theoretically predicted (Figure 13.7) [71, 81]. The chiral tetrahedral (T-4) and distorted square-planar four-coordinate (SP-4) metal complexes which are stereospeciﬁcally formed with bis-(−)-α-pinene-bpy-type ligands give exciton CD spectra [82]. The absolute conﬁgurations of the Ag(I), Pd(II), and Zn(II) complexes predicted by the exciton CD model nicely agree with those determined by the X-ray analysis.

463

464

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

13.5.2. (β-Diketonato) Lanthanide(III) Complex Several achiral tris-β-diketonato lanthanide(III) complexes have been known as a chiroptical probe for speciﬁc recognition of chiral amino acids, amino alcohols, and diols since the pioneering work of Nakanishi and Dillon [83]. The CD method is based on the CD couplet near 300 nm which is useful to determine the chiral conﬁguration of organic compounds functioning as chiral bidentate ligands [84, 85]. The signs of the CD couplets depend on the conformation of the bidentate chelates, with the bulky groups in the equatorial positions. That is, R-bidentate ligands with the anticlockwise (λ) conformation give a positive couplet and vice versa. Assuming the helical bladed arrangement around the central tetrakis-chelated Ln ion with the R-bidentate ligand, it is supposed to take a absolute conﬁguration on the basis of the exciton theory. However, it is difﬁcult to assume the exciton CD bands in the β-diketonate intraligand transition region, since SAPR-8(ssss) or -(ssll ) conﬁgurations are more stable than the SAPR-8-(llll ) (Figure 13.8), in view of more favorable chelation at sites within the same square planes due to steric hindrance, as found for most tetrakis(β-diketonato) lanthanide(III) complexes. The exciton CD spectra of the β-diketonate Ln complexes are actually observed for tetrakis-3-heptaﬂuorobutyryl-(+)-camphorate Ln(III) complexes with an encapsulated alkali metal ion, M[Ln((+)-hfbc)4 ]is (Figure 13.6) [60, 61]. Since complexes in this group are stereospeciﬁcally formed with chiral -SAPR-(C4 (llll )) conﬁgurations with the aid of CF· · ·M+ intramolecular interaction, as shown in Figure 13.9, a negative CD couplet in the π –π * transition of the β-diketonate chromophore is safely assigned to exciton CD, leading to the -SAPR-8 absolute conﬁguration (Figure 13.6) [60, 61]. It is noted that the exciton CD intensities (ε) depend on the alkali metal ion size, Cs+ > Rb+ > K+ > Na+ , reﬂecting the variation of twist angles around the C4 axis in the helically four-bladed C4 (llll ) chiral conﬁguration. The absolute conﬁguration for the Cs–Ln(III) complexes is supported by other chiroptical methods: CPL shows the largest g value for the Cs–Eu [63], VCD [86] and 4f –4f CD (vide supra).

Figure 13.8. Possible conﬁgurations of a square antiprism (SAPR)-8 complex. (a) llll and (b) ssss means chelation in the sites within different and the same

(a)

square, respectively.

(b)

200

Δε

100 0 −100 −200 250

Figure 13.9. Exciton CD spectra of 300 λ (nm)

350

M[La((+)-hfbc)4 ] in CHCl3 . M: Cs (red), Rb (green), K (blue), Na (black). (See insert for color representation of the ﬁgure.)

A P P L I C AT I O N S O F E L E C T R O N I C C I R C U L A R D I C H R O I S M T O I N O R G A N I C S T E R E O C H E M I S T RY

internuclear (net effect) intranuclear (Ln end) intranuclear (Cr end)

Δε (M−1cm−1)

50

CrIII

LnIII 0

N

N

−50

N

N

N

N

L2 275

300

375 325 350 Wavelength/nm

400

425

O N

Figure 13.10. Right: Structure of the ligand L2(below) and -[LnIII CrIII (L2)3 ]6+ (above). Left: Schematic vertical lines summering the dominant coupling effects in the CD spectra of -[LnIII CrIII (L2)3 ]6+ . The black line corresponds to the CD spectrum of -[GdCr(III)(L2)3 ]6+ in CH3 CN. (See insert for color representation of the ﬁgure.)

In some polynuclear complexes, including helicates, exciton CD is observed in the intraligand π –π * transition for each metal center, but for many of these species the CD depends on the bridging units connecting the individual chromophores, leading to anomalous CD patterns [87]. Recently, the CD spectra observed in the π –π * transitions of the dichlorido-bridged dinuclear complexes, -[(L1)2 Co(μ-Cl)2 Co(L1)2 ]2+ (L1 = chiral tetradentate Schiff base chelate with diimine chromophores) [87] and -[LnCr(L2)3 ]6+ (Figure 13.10) [26], show an exciton CD pattern opposite to that expected for the absolute conﬁguration from the X-ray analysis. This has been elucidated on the basis of semiempirical (ZINDO) calculations by considering the internuclear exciton coupling [87, 88]. The other dinuclear complexes -[(bpy)2 Ru(bpm)Ru(bpy)2 ]2+ (bpm = 2, 2 bipyrimidine) and the related trinuclear complexes show weak exciton CD bands with the correct signs. Such anomalous CD intensity as compared with that of the corresponding analogous mononuclear complexes can be reproduced with the oppositely signed internuclear exciton CD contributions by the ZINDO calculations, as schematically depicted in Figure 13.10. This theoretical approach provides a potential tool to determine the absolute conﬁguration on the basis of the exciton CD bands for polynuclear complexes, especially playing a crucial role for systems exhibiting the incorrect CD signs, opposite to those of the mononuclear complexes, leading to the correct determination of the absolute conﬁgurations. In summary, even though the implicit approximations used in the application of exciton CD to determine absolute conﬁguration ignore important complications such as the exciton splitting order, d -orbital mixing, and vibronic coupling, it has been a remarkably successful method for determination of the absolute conﬁguration with careful consideration of internuclear exciton contributions for polynuclear complexes.

13.6. ATROPISOMERISM AND ECD Atropisomers with rotamer conformations of coordinated monodentate ligands were found for the diastereomeric trans-[CoCl2 py4 ](hydrogen L-dibenzoyltartrate) compound in a

465

466

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

solid powder by CD [89] and later conﬁrmed by X-ray analysis [90]. Other examples are found for the conglomerates in space group P 21 21 21 of isomorphous nitrito-κ-N K[Co(NO2 )4 (NH3 )2 ] and NH4 [Co(NO2 )4 (NH3 )2 ] [91], as well as mer-[Co(NO2 )3 (NH3 )3 ] [92] by neutron diffraction studies, and the CD spectra were measured later [93]. For these pyridine and nitrito-κ-N complexes, the racemization of the atropisomers of monodentate lignds are too fast in solution to permit measurement of the solution CD spectra.

13.6.1. Nitrito-κ-O Complexes Such lability in the pyridine and nitrito-κ-N complexes could be reduced by chiral peripheral ligands near the monodentate ligands. This was observed for a dianionotetramine Co(III) complex with a chiral tetradentate ligand. The CD spectra in the ﬁrst ligand ﬁeld d –d band region of trans-R, R-[Co(N3 )2 (3,2,3-tet)]+ (3, 2, 3-tet(NH2 (CH2 )3 NH(CH2 )2 NH(CH2 )3 NH2 ) = chiral tetradentate linear tetramine ligand with δ conformation of the central ethylene backbone) gave enantiomeric patterns in H2 O and DMSO [94]. This suggests chirality due to rotational conformations of axial monodentate aniono ligands. However, such rotational or atropisomeric chirality cannot be evaluated by ECD because the d –d CD is susceptibile to chiral contributions from the ring conformation or geometrical structures of complexes. On the other hand, CD in the intraligand transitions of monodentate ligands could provide chiral rotational information on the monodentate ligand itself. Monodentate nitrito-κ-O Cr(III) complexes provide examples. Kaizaki [95] observed unique vibronic CD spectra with about 1000 cm−1 progression due to the NO stretching vibrations in the intraligand n –π * transition of the nitrito-κ-O ligands from 24,000 to 30,000 cm−1 for trans-[Cr(ONO)2 (R-pn)2 ]+ and trans-[Cr(ONO)2 (R,R-ptn)2 ]+ , as shown in Figure 13.11. The considerable solvent-dependent vibronic CD intensity enhancement reveals stereoselective solvation to the amino equatorial NH protons in increasing order of the bulkiness as well as the donor number of the solvents. The intraand intermolecular interactions of nitrito ligands with diamine chelates and/or the solvent molecules determine the most probable location of the nitrito ligands at equilibrium near amino nitrogen atoms in their predominant chiral rotamer conformations, which are sensitive to the N–Cr–N diamine chelate angles and/or the crowding around the chelate or the amino protons, causing the CD sign inversion between the R-pn and R, R-ptn complexes and very weak vibronic CD for -(+)450 ε -cis-[Cr(ONO)2 (en)2 ]+ as shown in Figure 13.11.

13.6.2. Guanine Derivative–Platinum(II) Complexes Analogous rotational isomers or atropisomers to the above nitrito complexes are found in square-planar four-coordinate SP-4 platinum(II) complexes. CD spectra of a number of model guanine platinum(II) complexes have been studied in order to reveal the selective binding to target DNA of platinum anticancer drugs that form a critical lesion by interor intrastrand cross-linking with two adjacent guanine bases or an adenine and a guanine base, respectively. A simple model complex, [Pt(R, R-dach)(9-EtG)2 ] (R, R-dach or R, R-chxn = R, R-cyclohexanediamine; 9-EtG = 9-ethylguanine), showed two CD couplets centered at 280 nm and 230 nm in the intraligand π –π * transitions of the 9-EtG, and the corresponding S , S -dach complex gives the enantiomeric CD pattern [96]. This clearly indicates a transmission of chirality of the dach ligand to the coordinated cis-guanine

A P P L I C AT I O N S O F E L E C T R O N I C C I R C U L A R D I C H R O I S M T O I N O R G A N I C S T E R E O C H E M I S T RY

467

2 AB

0.8 1.5 0.6 log ε

Δε (mol−1 dm3 cm−1)

1

0.4 (+) 0.2

1

0 CD

0.2 (−) 0.4

Figure 13.11. UV–vis(AB) and CD spectra 18

20

22

24

26

28

30

σ (103 cm−1)

of (+)-cis-[Cr(ONO)2 (en)2 ]+ (— —), trans-[Cr(ONO)2 (R-pn)2 ]+ (— — —) and trans-[Cr(ONO)2 (R, R-ptn)2 ]+ (· · ·).

bases through the diamine protons. Though the CD inversion was assumed to result from a change of the tilting direction of the guanine bases relative to the coordination plane [96], Natile et al. [97] elucidated more convincingly the CD results by recent studies on more rigid diamine complexes with [Pt{(S , R, R, S or R, S , S , R)-Me2 dab}G2 ] (Me2 dab = N , N -dimethyl-2,3-diaminobutane with four asymmetric centers at the N, C chelate ring atoms). A correlation is demonstrated between the CD signs and the two chiral conformers among three possible ones depicted in Figure 13.12: the major chiral head-to-tail (HT) forms, in which the guanine bases have their six-membered rings located on the opposite sides, and a head-to-head (HH) conformer where the guanine bases have their six-membered rings located on the same side [98]. The CD couplets are interpreted on the basis of the exciton coupling theory, which predicts the sign inversion from HT to HT or vice versa. As shown in Figure 13.13, the guanine derivative complexes [Pt(diamine)(3 -GMP)2 ] (diamine = (NH3 )2 , en or tn: 3 -GMP = guanosine3 -monophosphate) give -type CD with the HT atropisomer, whereas the CD patterns of the corresponding 5 -GMP (guanosine-5 -monophosphate) complexes are the reverse of those of the 3 -GMP complexes. The same CD behavior is observed for the R, R- and S , S -dach complexes. This indicates that the chirality of the D-ribose in the GPM plays a more important role than that of the diamine in forming the HT or HT conformation,

N

N

N

N

N

N

Pt

Pt

Pt

ΔHT

HH

ΔHT

Figure 13.12. Schematic representation of HH(head-to-head), HT (head-tail), and HT (headto-tail) atropisomers for cis-[PtA2 X2 ] complexes.

468

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

a

8

a

b c

Δε (M−1cm−1)

Δε (M−1cm−1)

5

0 −5

b

4

c 0 −4 −8

250

300

250

nm

300 nm

Figure 13.13. CD spectra of cis-[Pt(amine)2 (n -GMP)2 ] at pH 7. (Left) 3 -GMP. (Right) 5 -GMP. Amine = (a) (NH3 )2 , (b) tn(trimethylediamine, (c) en(ethylenediamine).

through an internucleotide hydrogen bond between a phosphate group and the N1H of the neighboring GMP. These intramolecular interactions are sensitive to the bite angles of amine ligands, in view of the weaker CD intensity for both the (en) complexes than that for the (NH3 )2 and (tn) complexes (Figure 13.13) [98]. Other than monodentate complexes, there have been a number of atropisomers of metal complexes with bidentate chelates. Examples are the optical resolution and CD spectra of dinuclear atropisomeric (3,4-diacetyl-2,5-hexanedionato)bis[(2,2 , 2 triaminotriethylamine)cobalt(III), [{Co(tren)}2 tae], and the related mononuclear 3-aryl2,4-pentanedionato complexes by Nakano et al. [99]. It is noted that the CD in the intraligand or charge-transfer transition region is comparable in intensity to the d –d CD for the dinuclear complex, whereas much larger CD intensities are observed in the d –d CD for the mononuclear complexes than for [Co(acac)(en)2 ]+ . This may depend on whether the major chirality originates from the relative conﬁguration of the tren moiety, but not directly from the β-diketonate rings. Recently, McCormick and Wang [100] reported the atropisomers of a polypyridyl N , N -chelate induced by zinc(II) and a chiral carboxylate, Zn(R- or S -O2 CCH(Br)CHMe2 )2 , where the zinc(II) ion functions as a mediator to facilitate the recognition event between the atropisomeric ligand and the chiral Zn(II) carboxylate, according to CD measurements.

13.7. SUMMARY Application of ECD to inorganic stereochemistry is going to become even more essential as chiral-selective structural probes, as asymmetric catalysts, and as key components of biomolecular processes. It will become increasingly important for researchers to explore the chiral structure–spectra relations that will make it possible to understand the various mechanisms and to design new chiral substances. Since there are still some problems in establishing reliable rules relating CD signs to absolute conﬁguration or conformation, the development of synthetic methods for chiral complexes, the progress of ECD techniques, and the advances in theoretical computation of chiroptical properties in recent years have certainly resulted in newer, more consistent ways to understand the inorganic stereochemistry in both ﬂexible and rigid systems.

A P P L I C AT I O N S O F E L E C T R O N I C C I R C U L A R D I C H R O I S M T O I N O R G A N I C S T E R E O C H E M I S T RY

REFERENCES 1. A. Werner, V. L. King, E. Scholze, B er. Dtsch. Chem. Ges. 1911, 44 , 1887–1898. 2. A. Werner, Ber. Dtsch. Chem. Ges. 1914, 47 , 3087–3094. 3. S. F. Mason, Molecular Optical Activity and the Chiral Discriminations, Cambridge University Press, Cambridge, 1982. 4. A. Cotton, Ann. Chim. Phys. 1896, 8 , 347–432. 5. J. Fujita, Y. Shimura, Optical rotary dispersion and circular dichroism, in Spectroscopy and Structures of Metal Chelate Compounds, K. Nakamoto and P. McCarthy, eds., John Wiley & Sons, New York, 1968, pp. 156–215. 6. R. D. Peacock, B. Stewart, Coord. Chem. Rev . 1982, 46 , 129–157. 7. R. Kuroda, Y. Saito, Circular dichroism of inorganic complexes: Interpretation and applications, in Circular Dichroism: Principles and Applications, N. Berova, K. Nakanishi, and R. W. Woody, eds., John Wiley & Sons, New York, 2000, pp. 563–599. 8. R. D. Peacock, B. Stewart, Chiral molecules spectroscopy, in Comprehensive Coordination Chemistry II: From Biology to Nanotechnology, Vol. 2. Fundamentals: Physical Methods, Theoretical Analysis and Case Studies, A. B. P. Lever, ed., Elsevier Pergamon, Amsterdam, 2004, pp. 65–81. 9. H. Amouri, M. Gruselle, Chirality in Transition Metal Chemistry: Molecules, Supramolecular Assemblies and Materials. John Wiley & Sons, Chichester, UK, 2008. 10. J. P. Riehl, S. Kaizaki, Absolute chiral structures of inorganic compounds, in Physical Inorganic Chemistry, Vol. 1. Principles, Methods, and Models, A. Bakac, ed., John Wiley & Sons. Hoboken, NJ, 2010, pp. 143–197. 11. Y. Saito, K. Nakatsu, M. Shiro, H. Kuroya, Acta Crystallogr. 1955. 8 , 729–730. 12. N. G. Connelly, T. Damhus, R. M. Hartshorn, A. T. Hutton, eds., The Red Book— Nomenclature of Inorganic Chemistry: IUPAC Recommendations, 2005, The Royal Society of Chemistry, Cambridge, 2005. 13. A. J. McCaffery, S. F. Mason, Mol. Phys. 1963, 6 , 359–371. 14. S. F. Mason, R. H. Seal, Chem. Commun. 1975, 331–333. 15. A. J. McCaffery, S. F. Mason, and B. J. Norman, J. Chem. Soc. 1965, 5094–5107. 16. M. Kunimatsu, H. Kanno, M. Kojima, K. Kashiwabara, J. Fujita, Bull. Chem. Soc. Jpn. 1980. 53 , 1571–1576. 17. T. S. Piper, A. G. Karipides, Mol. Phys. 1962. 5 , 475–483. 18. R. R. Judkins, D. J. Royer, Inorg. Chem. 1974, 13 , 945–950. 19. D. F. Harding, S. F. Mason, B. J. Peart, J. Chem. Soc. Dalton Trans. 1973, 955–960. 20. S. F. Mason, Fundamental Aspects and Recent Developments in Optical Rotatory Dispersion and Circular Dichroism, F. Ciardelli and P. Salvadori, eds., Heyden & Son, 1973, 21. U. Geiser, H. U. G¨udel, Inorg. Chem. 1981. 20 , 3013–3019. 22. S. Kaizaki, J. Hidaka, Y. Shimura, Bull. Chem. Soc. Jpn. 1970. 43 , 1100–1109. 23. S. Kaizaki, J. Hidaka, Y. Shimura, Inorg. Chem. 1973, 12 , 135–142. 24. S. Kaizaki, M. Ito, N. Nishimura, Y. Matsushita, Inorg Chem. 1985. 24 , 2080–2088. 25. G. L. Hilmes, H. G. Brittain, F. S. Richardson, Inorg. Chem. 1977. 16 , 528–533. 26. M. Cantuel, G. Bernardinelli, G. Muller, J. P. Riehl, C. Piguet, Inorg. Chem. 2004, 43 , 1840–1849. 27. S. F. Mason, R. H. Seal, Mol. Phys. 1976, 31 , 755–775. 28. S. F. Mason, Mol. Phys. 1979, 37 , 843–853. 29. F. S. Richardson, Chem. Rev . 1979, 79 , 17–36. 30. R. S. Schreiner, P. J. Hauser, Inorg. Chem. 1974. 13 , 2185–2192.

469

470

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

31. R. W. Strickland, F. S. Richardson, Inorg. Chem. 1973. 12 , 1025–1036. 32. M. C. Ernst, D. J. Royer, Inorg. Chem. 1993, 32 , 1226–1232. 33. B. Stewart, R. D. Peakock, L. Alagna, T. Prosperi, S. Turchini, J. Goulon, A. Rogalev, C. Goulon-Ginet, J. Am. Chem. Soc. 1999, 121 , 10233–10234. 34. J. Autschbach, F. E. Jorge, T. Ziegler, Inorg. Chem. 2003, 42 , 2867–2877. 35. F. E. Jorge, J. Autschbach, T. Ziegler, Inorg. Chem. 2003, 42 , 8902–8910. 36. J. Fan, M. Seth, J. Autschbach, T. Ziegler, Inorg. Chem. 2008, 47 , 11656–11668. 37. S. Kaizaki, N. Azuma, A. Fuyuhiro, Inorg. Chim. Acta, 1998 274 , 210–214. 38. T. Ama, J. Hidaka, Y. Shimura, Bull. Chem. Soc. Jpn. 1970, 43 , 2654. 39. H. T. Evans, Jr., J. S. Showell, J. Am. Chem. Soc. 1969. 91 , 6881–6882. 40. E. Coronado, S. Curreli, C. Gimenez-Saiz, C. J. Gomez-Garcia, J. Roth, Synthetic Metals, 2005, 154 , 241–244. 41. T. Kudo, Y. Shimura, Bull. Chem. Soc. Jpn. 1979, 52, 1648–1652. 42. T. Kudo, Y. Shimura, Bull. Chem. Soc. Jpn. 1979, 52, 3553–3558. 43. T. Kudo, Y. Shimura, Bull. Chem. Soc,. Jpn. 1980, 53 , 1588–1594. 44. Y. Shimura, Rev. Inorg. Chem. 1984. 6 , 149–193. 45. Y. Hosokawa, Y. Shimura, Bull. Chem. Soc. Jpn. 1986, 59 , 2913–2914. 46. N. Fujimoto, Y. Mori, A. Yamagishi, H. Sato, Chem. Commun. 2010, 46 , 5473–5475. 47. S. Misumi, S. Kida, T. Isobe, Spectrochim. Acta A, 1968, 24 , 271–275. 48. S. Misumi, S. Kida, T. Isobe, Y. Nishida, H. Furuta, Bull. Chem. Soc., Jpn. 1969, 42 . 3433–3438. 49. L. I. Katzin, Inorg. Chem. 1968, 7 , 1183–1191. 50. L. I. Katzin, Inorg. Chem. 1969, 8 , 1649–1654. 51. B. Norden, I. Grenthe, Acta Chem. Scand . 1972, 26 , 407–409. 52. J. D. Saxe, T. R. Faulkner, F. S. Richardson, J. Chem. Phys. 1982, 76 , 1607–1623. 53. F. S. Richardson, T. R. Faulkner, J. Chem. Phys. 1982, 76 , 1595–1606. 54. F. S. Richardson, Inorg. Chem. 1980, 19 , 2806–2812. 55. N. C¸oruh, G. L. Hilmes, J. P. Riehl, Inorg. Chem. 1988, 27 , 3647–3651. 56. L. Di Bari, P. Salvadori, Coord. Chem. Rev . 2005, 249 , 2854–2879. 57. P. Gawryszewska, J. Legendziewicz, Z. Ciunik, N. Esfandiari, G. Muller, C. Piguet, M. Canteul, J. P. Riehl, Chirality 2006, 18 , 406–412. 58. M. A. Subhan, T. Suzuki, S. Kaizaki, J. Chem. Soc., Dalton Trans. 2002, 1416–1422. 59. M. A. Subhan, T. Suzuki, S. Kaizaki, J. Chem. Soc., Dalton Trans. 2001 , 492–497. 60. D. Shirotani, T. Suzuki, S. Kaizaki, Inorg. Chem. 2006, 45 , 6111–6113. 61. D. Shirotani, T. Suzuki, Y. Yamanari, S. Kaizaki, J. Alloys Compd . 2008, 451 , 325–328. 62. D. Shirotani, H. Sato, K. Yamanari, S. Kaizaki, Unpublished results, 2010. 63. J. L. Lunkely, D. Shirotani, K. Yamanari, S. Kaizaki, G. Muller, J. Am. Chem. Soc. 2008, 130 , 13814–13815. 64. A. C. Sen, M. Chowdhury, J. Chem. Soc., Faraday Trans., 2 , 1981, 77 , 1393–1300. 65. J. P. Morley, J. D. Saxe, F. S. Richardson, Mol. Phys. 1982, 47 , 379–406. 66. B. Bosnich, Acc. Chem. Res. 1969, 2 , 266–273. 67. E. Larsen, S. F. Mason, G. H. Searle, Acta Chim. Scand . 1966, 20 , 191–196. 68. J. Mason, S. F. Mason, Tetrahed . 1967, 23 , 1919–1927. 69. Y. Ohmori, M. Namba, Y. Kuroda, M. Kojima, Y. Yoshikawa, Inorg. Chem. 1992, 31 , 2299–2300. 70. T. Ito, A. Kobayashi, F. Marumo, Y. Saito, Inorg. Nucl. Chem. Lett. 1971, 7 , 1097–1102. 71. S. F. Mason, Inorg. Chim. Acta Rev . 1968, 89–109.

A P P L I C AT I O N S O F E L E C T R O N I C C I R C U L A R D I C H R O I S M T O I N O R G A N I C S T E R E O C H E M I S T RY

72. Y. Ohashi, K. Yanagi, Y. Mitsuhashi, K, Nagata, Y. Kaizu, Y. Sasada, H. Kobayashi, J. Am. Chem. Soc. 1979, 101 , 4739–4740. 73. G. R. Brubaker, L. E. Webb, J. Am. Chem. Soc. 1969, 91 , 7199–7200. 74. K. Michelsen, Acta Chem. Scand . 1965, 19 , 1175–1179. 75. A. Le Guennic, W. Hieringer, A. Autschbach, J. Phys. Chem. A, 2005, 109 , 4836–4846. 76. J. Fan, J. Autschbach, T. Ziegler, Inorg. Chem. 2010, 49 , 1355–1362. 77. F. J. Coughlin, K. D. Oyler, R. A. Pascal, S. Bernhard, Inorg. Chem. 2008, 47 , 974–979. 78. J. Hidaka, B. Douglas, Inorg. Chem. 1964, 3 , 1180–1184. 79. S. Kaizaki, J. Hidaka, Y. Shimura, Bull. Chem. Soc. Jpn. 1969, 42 , 988–994. 80. S. Kaizaki, J. Hidaka, Y. Shimura, Bull. Chem. Soc. Jpn. 1970, 43 , 3024. 81. B. Bosnich, Inorg. Chem. 1968, 7 , 2379–2386. 82. O. Mamula, Alex von Zelwsky, T. Bark, H. Stoeckli-Evans, A. Neels, G. Bernardinelli, Chem. Eur. J . 2000, 6 , 3575–3585. 83. K. Nakanishi, J. Dillon, J. Am. Chem. Soc. 1971, 93 , 4058–4060. 84. L. Di Bari, M. Lelli, G. Pintacuda, P. Salvadori, Chirality 2002, 14 , 262–273. 85. H. Tsukube, S. Shinoda, H. Tamaki, Coord. Chem. Rev . 2002, 226 , 227–234. 86. H. Sato, D. Shirotani, K. Yamanari, S. Kaizaki, Inorg. Chem. 2010. 49 , 356–358. 87. S. G. Telfer,, N. Tajima, R. Kuroda, J. Am. Chem. Soc. 2004, 126 , 1408–1418. 88. S. G. Telfer, N. Tajima, R. Kuroda, M. Cantuel, C. Piguet, Inorg. Chem. 2004, 43 , 5302–53210. 89. S. Utsuno, J. Am. Chem. Soc. 1982, 104 , 5846–5847. 90. N. C. Payne, M. E. Thachuk, Chirality 1989, 1 , 284–292. 91. I. Bernal, Inorg. Chim. Acta 1985, 96 , 99–110. 92. I. Bernal, Inorg. Chim. Acta 1986, 115 , 25–29. 93. S. Utsuno, A. Suganuma, Y. Yoshikawa, Inorg. Chim. Acta 1996, 244 , 1–2. 94. B. Bosnich, J. M. Harrowﬁeld, Inorg. Chem. 1975, 14 , 828–836. 95. S. Kaizaki, Bull. Chem. Soc. Jpn. 1983, 56 , 3625–3631. 96. M. Gullotti, G. Pacchioni, A. Pasini, R. Ugo, Inorg. Chem. 1982, 21 , 2006–2014. 97. M. Benedetti, L. G. Marzilli, G. Natile, Chem. Eur. J . 2005, 11 , 5302–5310. 98. G. Natile, L. G. Marzilli, Coord. Chem. Rev . 2006, 250 , 1315–1331. 99. Y. Nakano, Y. Yoshikawa, H. Kondo, J. Am. Chem. Soc. 1986, 108 , 7630–7632. 100. T. M. McCormick, S. Wang, Inorg. Chem. 2008, 47 , 10017–10024.

471

PART IV BIOMOLECULES

14 ELECTRONIC CIRCULAR DICHROISM OF PROTEINS Robert W. Woody

14.1. INTRODUCTION Circular dichroism is extensively used in the characterization of protein structure and folding because of its high conformational sensitivity. In this chapter, we will discuss electronic CD almost solely, so we will use the common abbreviation CD to refer to electronic CD and use VCD when we refer to vibrational CD. CD provides a sensitive and convenient method for determining the secondary structure of proteins. As such, it is one of the principal methods for the initial characterization of proteins and is an invaluable tool for the investigation of protein folding. CD is widely used to validate the conformational integrity of mutant proteins and to assess their conformational stability relative to the wild-type protein. CD also has important applications in detecting and quantitating ligand binding to proteins. Additional chiroptical methods for studying protein conformation are discussed in other chapters of this volume: VCD (Chapter 22) and Raman optical activity (Chapter 23). Related topics of interest to readers of this chapter include: the CD of peptides (Chapter 15), peptidomimetics (Chapter 16), nucleic acids (Chapter 17), peptide nucleic acids (Chapter 18), protein–nucleic acid interactions (Chapter 19), drug and natural products binding to nucleic acids (Chapter 20), ligand binding to serum proteins (Chapter 21), and glycoconjugates (Chapter 24). Theoretical methods for predicting protein CD are discussed in Chapter 20 of volume 1. The CD spectrum of a protein is conveniently divided into three regions, each of which is dominated by different types of chromophores and provides different kinds of information. The far UV ranges from 250 nm to the lower limit of measurements, ∼170 nm in water, and ∼120 nm in ﬁlms. The dominant chromophore is the amide group, and the CD is largely determined by the secondary structure of the protein. In Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

475

476

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

the near UV from 250 to 300 nm, protein CD is dominated by aromatic side chains and provides information about tertiary structure. In the proximal near UV, visible, and near IR, from 300 nm to ∼1000 nm, CD is observed only in proteins that have chromophoric prosthetic groups or other bound chromophores with absorption bands in this region. Thus this region is useful for monitoring ligand binding and ligand conformation. Recent reviews of protein CD include those of Sreerama and Woody [1, 2], Kelly and Price [3, 4], Martin and Schilstra [5], and Wallace [6]. The present chapter will focus on developments of the past decade.

14.2. THE PEPTIDE BACKBONE The far-UV CD of proteins is largely determined by the peptide backbone contributions, which arise from the peptide chromophore. Various types of secondary structure have characteristic CD patterns, which makes the far-UV CD of proteins useful for secondary structure analysis.

14.2.1. The Amide Chromophore The absorption spectra of amides have been studied extensively in the gas phase [7, 8], in solution [9], and in crystals [10]. In condensed phases, where Rydberg transitions are suppressed, only three excited states are observed down to ∼130 nm in simple amides: the n –π ∗ transition at ∼220 nm and the ﬁrst (NV1 ) and second (NV2 ) π –π ∗ transitions at ∼190 and 140 nm, respectively. The n –π ∗ transition is analogous to that in other carbonyl chromophores. It is symmetry-allowed in the Cs symmetry of amides, but it is weak in absorption (εmax ∼100 M−1 cm−1 ). A large magnetic dipole transition moment (∼1BM) is directed along the carbonyl bond. The energy of the n –π ∗ transition depends on the extent and strength of hydrogen bonding. In nonpolar solvents, the amide n –π ∗ transition is at ∼230 nm, whereas in strongly hydrogen-bonding solvents, λmax ∼210 nm. The n –π ∗ transitions in tertiary amides are red-shifted by 5–10 nm relative to those in secondary amides [9]. The 190-nm (NV1 ) π –π ∗ transition has a moderate intensity (εmax ∼9000 M−1 cm−1 with a transition moment magnitude |μ|∼3 D). It occurs at 185–190 nm in secondary amides and near 200 nm in tertiary amides. The transition is polarized approximately along the C–N bond direction [10]. The NV2 transition has been observed in crystals near 140 nm, is less intense than the NV1 , and its polarization is approximately orthogonal to that of the NV1 . Ab initio calculations for amides as isolated molecules [11–13] and in cyclohexane and water [14] have been reported. These ab initio results are in good agreement with experiment.

14.2.2. General Aspects of Protein CD Isolated amide chromophores are achiral and therefore have no CD. Interactions of the n –π ∗ and π –π ∗ transitions in polypeptides are largely responsible for the far-UV CD of proteins. These interactions are of three types: (1) coupled electric dipole transition moments on the amide groups, μ–μ coupling, of which the exciton coupling between NV1 transition moments is the most important [15] (Chapters 20 in volume 1 and 4 in this volume); (2) coupling of the electric dipole transition moment on one peptide and

477

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

the magnetic dipole transition moment on another (μ–m coupling) [16]; (3) mixing of NV1 (NV2 ) and n –π ∗ excited states within a peptide group induced by the electric ﬁeld of the polypeptide (one-electron or static-ﬁeld coupling) [17]. These effects are all taken into account in Tinoco’s ﬁrst-order perturbation theory [18] and the matrix method of Schellman and co-workers [19]. The ﬁrst mechanism (μ–μ coupling) is considered in DeVoe theory [20] and in Applequist’s atom dipole interaction model [21]. These models are described in Chapter 20 in volume 1.

14.2.3. CD of Secondary Structural Elements We brieﬂy describe the CD spectra of the major types of secondary structure that appear in proteins. More extensive information may be found in Chapter 15 in this volume. 14.2.3.1. α-Helix. Pauling and Corey’s [22] α-helix is the major element of secondary structure in many proteins and accounts for about one-third of the residues in globular proteins [23]. The α-helix was the ﬁrst secondary structure to be characterized by CD [24, 25], with measurements on poly(Glu), poly(Lys), and poly(GluOMe). The α-helix CD spectrum was found to be largely insensitive to the side chains and solvent [26], so long as the α-helix was maintained. The spectrum of α-helical poly(Glu), shown in Figure 14.1, has three bands above 180 nm. The negative band at 222 nm is assigned to the n –π ∗ transition [27, 28]. The negative band at ∼207 nm and the positive band at ∼190 nm are both attributed to the π –π ∗ (NV1 ) transition, resulting from exciton coupling [15] among the π –π ∗ transition moments. The 207-nm band is polarized along the helix axis, and the 190 nm band is polarized in the plane perpendicular to the axis. In addition to these well-characterized bands, the α-helix CD spectrum has a positive band near 175 nm, which appears as a shoulder on the 190-nm band; a negative band

ππ* (perp)

30

Δε (M−1 cm−1)

20

10

0

–10 ππ* (par) –20

180

200

nπ* 220

240

λ (nm)

Figure 14.1. CD spectra of model polypeptide secondary structures. α-helix (

), poly(Glu) in water, pH. 4.5 [29]. The band assignments are indicated by labels adjacent to the extrema. β-sheet (– – –), poly(Lys-Leu), 0.5 M NaF, pH 7 [62]. Unordered conformation (· · · · · ·), poly(Glu) in water,

pH 8 [29].

478

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

near 160 nm; and a positive band below 140 nm [29]. The latter band is attributable to the NV2 transition, but the other two bands have no counterpart in simple amides. They probably result from charge-transfer transitions between neighboring amides. This assignment is consistent with ab initio calculations on small peptides [30–33]. 14.2.3.2. 310 -Helix. The 310 -helix occurs less frequently in proteins (∼3% of residues) [23]. Its CD spectrum has been characterized experimentally by Toniolo et al. [34, 35]. The long-wavelength negative bands are dramatically weaker than those of the α-helix and especially that of the n –π ∗ transition, which is only a shoulder in the 310 helix spectrum. The ratio ε222 /ε204 is ∼0.2–0.3, compared to ∼1.0 for the α-helix. Based upon theoretical calculations, a value for this ratio signiﬁcantly below unity was proposed [36] as a criterion for 310 -helix formation. Although this criterion has been strongly questioned [37], the work of Toniolo et al. supports it. The 310 -helix appears to have a positive band near 190 nm, but its magnitude is uncertain. 14.2.3.3. β-Sheet. The CD spectrum of a model β-sheet is shown in Figure 14.1. A negative band near 217 nm, a positive band near 195 nm, and a negative band near 175 nm are characteristic of the β-sheet conformation [38, 39]. The 217-nm band is assigned to the n –π ∗ transition, and the 195- and 175-nm bands are attributable to exciton splitting of the π –π ∗ transition. A positive band at 168 nm is observed [40] in the spectrum of concanavalin A, a β-rich protein, and is attributed to a charge-transfer transition. In contrast to the α-helix, β-sheet CD spectra show much more variation in absolute magnitude and in relative magnitudes of the bands. This is probably a reﬂection of the inherent variability of β-sheets. β-sheets may be parallel, antiparallel, or mixed; the extent of twisting varies from weak to strong; they vary in the number and length of strands; and they may contain defects such as β-bulges [41]. Calculations [42] suggest that the CD of parallel and antiparallel β-sheets are similar, and this is supported by the observation [43] that the CD spectrum of pectate lyase C, a protein rich in parallel β-sheet, is similar to that of poly(Lys), which is an antiparallel β-sheet. 14.2.3.4. β-Turns. β-Turns are an important element of secondary structure in proteins. Woody [44] predicted the CD spectrum for various types of β-turns described by Venkatachalam [45]. He found several CD patterns, but the two most common types of β-turns, types I and II, were predicted to have a β-sheet-like CD spectrum, with bands red-shifted by 5–10 nm (a class B spectrum). For one type of β-turn, expected for the sequence D-X-L-Pro that adopts a type II β-turn (the prime indicates a mirror image form), an α-helix-like CD pattern (class C) was predicted. Experimental data on cyclic peptides [46–49] have supported these predictions for the type II and II turns, but indicate an α-helix-like spectrum for type I turns, in contrast to the predictions. The observed spectra for type I and II β-turns in cyclic peptide models are shown in Figure 14.2. 14.2.3.5. Unordered Polypeptides. Unordered polypeptides have been modeled by the charged homopolypeptides poly(Glu) and poly(Lys) at neutral pH. These polypeptides have a relatively weak positive CD band at ∼217 nm and a strong negative band at ∼197 nm, as shown in Figure 14.1. Tiffany and Krimm [50, 51] pointed out that this spectrum bears a strong resemblance to that of poly(Pro)II, the left-handed threefold helical form adopted by poly(Pro) in water. On this basis, they suggested that poly(Glu) and

479

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

15

Δε (M−1cm−1)

10

5

0

–5

–10

200

220

240

λ (nm)

Figure 14.2. CD spectra of model β-turns [49]. Type I turn (• • •), cyclo(L-Ala-Aha) (Aha = ε-aminohexanoic acid). Type II turn (

), cyclo(L-Ala-D-Ala-Aha).

poly(Lys) are not truly unordered but have a signiﬁcant amount of the poly(Pro)II (PII ) conformation. This proposal was not widely accepted at the time, but is now supported by a large body of experimental and theoretical evidence [52, 53]. Although the positive 217-nm band of the PII or unordered conformation is assigned to the n –π ∗ transition, the strong negative band at 200 nm and the absence of any positive counterpart are not consistent with a dominant exciton effect in the π π ∗ region. Woody [54] has shown that mixing of the NV1 π –π ∗ transition with transitions in the deep UV occurs in the PII conformation and is responsible for suppressing the exciton contributions and generating the strong negative CD band near 200 nm.

14.2.4. Secondary Structural Analysis of Proteins One of the most important and widely used applications of CD spectroscopy is the analysis of the secondary structure of proteins. The CD spectra of proteins frequently show the characteristics of their dominant secondary structure, as illustrated in Figure 14.3, which shows the CD spectra of hemoglobin, an α-rich protein; concanavalin A, a βrich protein; and α-synuclein, a predominantly disordered protein. However, because the α-helix spectrum is much more intense than the β-sheet spectrum, proteins with signiﬁcant amounts of both secondary structures generally show the qualitative features of the α-helix. Figure 14.4 shows the CD spectra of ribonuclease A (21% α, 33% β) and subtilisin (30% α and 18% β), which both have α-like features. Also shown in Figure 14.4 is the CD spectrum of elastase (11% α, 34% β, and 34% unordered), the spectrum of which resembles that of unordered polypeptides. Clearly, protein CD spectra contain important information about secondary structure, but extracting this information is not a simple task. Methods of secondary structure analysis have been reviewed extensively [2, 55–61].

480

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

14

Δε (M−1 cm−1)

9

4

–1

–6

170

180

190

200

210 220 λ (nm)

230

240

250

), an α-rich protein; concanavalin A (• • •), a β-rich protein; α-synuclein (ooo), an intrinsically disordered protein. The myoglobin and

Figure 14.3. CD spectra of three proteins: myoglobin (

concanavalin A spectra are from the CDPro database [83] and were a private communication from W. C. Johnson. The α-synuclein spectrum is from reference 210.

6

Δε (M−1 cm−1)

4

2

0

–2

–4 170

180

190

200

210 220 λ (nm)

230

240

250

), subtilisin (• • •), and elastase (ooo). The spectra are from the CDPro database [83] and were a private communication from W. C. Johnson.

Figure 14.4. CD spectra of ribonuclease A (

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

A basic assumption in secondary structure analysis by CD is that the spectrum is a linear combination of component secondary structure spectra: fi Bi λ , (14.1) Cλ = λ

where Cλ is the observed CD at wavelength λ, fi is the fraction of secondary structure i , and Bi λ is the CD of secondary structure i at wavelength λ. Additional implicit assumptions are that the contributions are additive and that the effects of tertiary structure, aromatic side chains, and variations in secondary structure geometry and segment length are negligible. Although these assumptions are not always valid, current methods minimize their effects by using ﬂexible methods in the analysis. Initially, spectra of the secondary structure types, Bi λ , were taken from model systems as in Figure 14.1 [38, 62]. However, the model systems have long and regular helices and strands and thus differ signiﬁcantly from the relatively short and irregular secondary structure elements that occur in globular proteins. Consequently, later methods have derived, explicitly or implicitly, the secondary structure CD spectra from the CD spectra of proteins with structures known from X-ray diffraction (reference proteins). The secondary structure spectra and the secondary structure fractions have been derived by many methods, using simple least squares [63–66], ridge regression [67, 68], singular value decomposition [69–72], neural networks [73–77], principal component factor analysis [78], convex constraint analysis [79], partial least squares [80, 81], and support vector machines [80]. The importance of ﬂexibility in secondary structure analysis by CD was recognized by Manavalan and Johnson [70] and implemented in their Variable Selection Method. Flexibility permits the program to adapt the set of reference proteins to best describe the CD spectrum of the protein under analysis and thus helps overcome limitations imposed by non-peptide CD contributions and variations in helix and strand geometry and length. Flexibility has also been introduced by ridge regression (CONTIN) [67], the locally linearized model [71], the self-consistent method (SELCON) [72], CDsstr [82], and neural network methods [73–77]. The databases of CD spectra for structurally characterized proteins have undergone signiﬁcant development in recent years. The CDPro website [83] (a list of website addresses and contacts for software useful in the analysis and prediction of protein CD spectra is provided in Section 14.6) offers 10 sets of CD spectra for various wavelength ranges (from 178–260 nm to 190–240 nm) and permits analysis for various combinations of secondary structure. Some sets include denatured proteins, and these are well-suited for analyzing intrinsically disordered proteins and protein folding/unfolding transitions [84]. Other sets contain integral membrane proteins and are useful for the analysis of such proteins [85]. A basis set containing 50 proteins (RaSP50) has been constructed by Oberg et al. [86]. The proteins were selected to cover all known protein folding types and the full range of α-helix and β-sheet contents. The authors acquired the CD spectra from 185–260 nm for each protein and analyzed the reported crystal structure data for secondary structure by DSSP [87]. Increasing availability of synchrotron light sources for CD spectroscopy [6] has stimulated the development of protein CD basis sets that extend into the vacuum UV. Matsuo et al. [88, 89] reported a database for 31 proteins that extends to 160 nm. Lees et al. [90] described two databases, SP175 and SP170, with short-wavelength cutoffs of 175 nm (72 proteins) and 170 nm (39 proteins), respectively. Like the RaSP50 database

481

482

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

[86], the proteins were selected to provide a broad representation of protein folds and secondary structure contents. The SP175 data set is available on the DICHROWEB server [91, 92] and in the Protein Circular Dichroism Data Bank [93, 94]. Although publications describing methods of secondary structure analysis report performance statistics, most have reported comparisons with only a few other methods. There have been few broader comparisons of methods. Greenﬁeld [58] compared CONTIN [67], VARSLC [70], SELCON [72], and K2D [74], as well as several older methods that lacked ﬂexibility. Greenﬁeld recommended SELCON, CONTIN, and K2D for analysis of globular protein structure. Lees et al. [80] compared SELMAT3 (their version [90] of SELCON3 [83]); PLS, partial least squares; SIMPLS, simultaneous partial least squares; PCR, principal component regression; NN, neural network; SIMPL-NN, simultaneous partial least squares neural network; and SVM, support vector machine methods. Cross-validation was performed with the SP175 database [90]; that is, each of the 72 proteins was analyzed using the remaining proteins as the data base. Two performance measures were used: the Pearson product-moment correlation coefﬁcient [95], r, which is 1 for perfect agreement between calculated (CD) and observed (X-ray) structure fractions and is 0 for random agreement; and δ, the root-mean-square deviation between calculated and observed fractions. In all tests, the performance of the various methods was quite similar for all methods. For example, in a three-state model (α-helix, β-sheet, other), r for the α-helix content varied from 0.957 to 0.971, and δ ranged from 0.052 to 0.063. No one method was best for the all three structural types, but the SIMPL-NN model performed best for the β-sheet and other, whereas the PLS method did best for the α-helix. Tests were also performed for the six-state model used by Sreerama et al. [96] to calculate the numbers of α-helix and β-strand segments. Again, the range of performance parameters was narrow for each structural type. In this test, only two methods (SELMAT3 and PLS) were compared. SELMAT3 performed best for two structural types, and PLS worked best for four types. Secondary structure analysis by the methods described above requires an accurate protein concentration because the per-residue ε or [θ ] is used. The effects of errors in the protein concentration were discussed by Hennessey and Johnson [69], who showed that a 5–10% error in concentration propagated as proportional changes in secondary structure contents. Therefore, it is important to use spectrophotometrically determined concentrations and not ones based upon less accurate methods such as dye binding. For some samples (e.g., with protein ﬁlms), it is impossible to obtain accurate protein concentrations. McPhie [97–99] and Goormaghtigh and co-workers [100, 101] have developed methods that are independent of protein concentration. While not as accurate as the methods previously described, they are sufﬁciently accurate to be useful in cases in which accurate protein concentrations are unavailable. McPhie’s method [97–99] utilizes the Kuhn anisotropy ratio, g, which is the ratio ε/ε. He measured g-factor spectra from 190nm to 240 nm for a basis set of 32 proteins and used these in a constrained least squares or ridge regression analysis to derive secondary structures. The analyses for α-helix and β-sheet were satisfactory with r ∼0.85 and 0.75, respectively, and rms deviations of ∼0.1. Results for turns and unordered structure were inferior, with r∼0.4. In the method of Raussens et al. [100] (note also the erratum [101]), the observed spectrum is ﬁrst normalized to an ellipticity of 1 at 207 nm, a wavelength selected to minimize the deviations of the α-helix contents of the set of 50 reference proteins [86]

483

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

from X-ray values. Then, for each type of secondary structure, a model of the form fi = a1 + a2 Eλ1 + a3 Eλ1 2 ,

(14.2)

where Eλ1 is the normalized ellipticity of the test protein at wavelength λ1 , is used. The wavelength λ1 and the coefﬁcients a1 , a2 , and a3 were optimized to give the best ﬁt to the secondary structure contents of type i in the set of reference proteins. For the α-helix and β-sheet, this model was augmented by adding linear and quadratic terms for a second wavelength. For the turns and unordered conformation, the additional terms were unnecessary. This method gives standard deviations of ∼10–12% for the α-helix and β-sheet, comparable to those from the method of McPhie [97–99]. It is possible to extract information from far-UV CD spectra beyond the content of secondary structures. The number of α-helices and β-strands in a protein can be estimated [96, 102]. The basis for this lies in the fact that amide groups at the ends of α-helical or β-strand segments differ in their environment and hence their CD from those in the middle of the segments. Sreerama and Woody assumed that nα and nβ residues per helix and strand, respectively, are affected by being terminal residues. For each helix in a reference protein, nα residues were assigned as “distorted,” hD , and the remainder were considered as “regular,” hR , and correspondingly for each beta strand, giving βD and βR . Analysis of the reference proteins showed that the values of nα and nβ that gave the best values for helix and strand numbers are nα = 4 and nβ = 2. With these choices, the numbers of helices and strands can be estimated with an uncertainty of ∼ ±3. Protein secondary structure can also be derived from IR absorption [103–105], Raman [106–108], VCD (Chapter 21 in this volume) and Raman optical activity (Chapter 22 in this volume). Methods for combined analysis of CD and IR absorption spectra have been developed [109–113]. VCD and CD spectra have also been combined [111, 114]. In general, the combined analysis of CD and vibrational spectroscopy, IR absorption, or VCD has been found to improve the results over the individual methods. CD gives superior results for the α-helix whereas vibrational spectroscopy provides better results for β-sheets. Thus, the two types of spectroscopy complement each other.

14.3. PROTEIN SIDE CHAINS Side-chain chromophores in proteins include the three aromatic side chains (Phe, Tyr, and Trp) and the disulﬁde group of cystine. These chromophores have absorption bands in the near UV and are responsible for the near-UV CD of proteins. They also have higher-energy transitions, in the far UV, which are much more intense than the near-UV bands. In the far-UV, side-chain CD is usually obscured by the much more numerous amide groups but, in some cases, especially strong aromatic CD combined with weak backbone CD permits the detection and characterization of side-chain CD bands in the far UV. Side-chain CD has been reviewed [115–118].

14.3.1. Near-UV CD The near-UV CD of proteins is dominated by aromatic side-chain contributions. The CD of these chromophores is determined by coupling with the peptide backbone and with other aromatic groups. Globular proteins with a well-deﬁned tertiary structure generally exhibit a near-UV CD spectrum, often with some distinct ﬁne structure. Thus, the nearUV CD spectrum is sensitive to the protein’s tertiary structure. Conformational changes

484

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

or ligand binding that affect the geometry or environment of one or more aromatic side chains can therefore perturb the near-UV CD spectrum. Because of the relatively small numbers of aromatic chromophores, such perturbations may be more easily detectable than those of the far-UV CD, which require a signiﬁcant fraction of the residues in the protein to be perturbed. Each of the three aromatic amino acids gives rise to absorption and CD bands with characteristic wavelengths and ﬁne structure. These features frequently permit assignment of speciﬁc features in the near-UV CD spectrum to one type of aromatic side chain. Site-directed mutagenesis can provide deﬁnitive assignment to a speciﬁc residue in the sequence. A chromophoric residue can be mutated to a weaker chromophore or a nonchromophoric residue. The difference CD spectrum of (wild-type − mutant) provides the CD contribution of the mutated residue to the wild-type spectrum, if the mutation has not signiﬁcantly altered the conformation. Such studies have been performed for a substantial number of aromatic residues in several proteins: barnase [119], carbonic anhydrase II [120], dihydrofolate reductase [121], gene protein 5 from bacteriophage fd [122], pancreatic trypsin inhibitor [123], human tissue factor [124], and ribonuclease A [125]. For a number of these systems, theoretical calculations have given good agreement with the sign and approximate magnitude of the individual aromatic CD contributions [118, 123, 125–127]. Disulﬁdes also may contribute signiﬁcantly to protein CD. Their near-UV CD generally consists of a relatively weak and broad band with a maximum near 260 nm. The disulﬁde peak is generally obscured by aromatic CD bands, but a long-wavelength tail may extend to 300 nm or above.

14.3.2. Far-UV CD Aromatic CD bands in the far-UV are most readily observed in β-rich proteins with relatively weak backbone CD—for example, immunoglobulin folds, lectins, and snake toxins. Positive bands near 230 nm are generally due to aromatic side chains because, except for the PII conformation, none of the common secondary structures have positive CD in this region, and the long-wavelength positive band of the PII conformation is relatively weak. Proteins that exhibit positive CD bands near 230 nm include avidin [128], gene 5 protein from ﬁlamentous phages [122, 129], and cobrotoxin [130]. For the most favorable backbone and side-chain conformations, coupling between the La transition of Tyr or Phe and the amide transitions of nearest-neighbor peptide groups favors positive rotational strength in the La band [131]. However, there appears to be no strong bias toward positive rotational strengths for the 225-nm Bb band of Trp [132]. It is interesting to note that all of the aromatic amino acids in the form of their N -acetyl, N -methylamides exhibit positive La (Tyr, Phe) and Bb (Trp) CD bands [133]. Aromatic side chains are frequently found in clusters in proteins [134], so exciton couplets involving aromatic side chains might be expected. The dependence of the exciton couplet strength on the square of εmax (Chapter 4) excludes observable exciton couplets in the near-UV. However, the intense Bb band of Trp (εmax ∼35, 000 M−1 cm−1 ) can give rise to strong couplets in Trp–Trp pairs and, through coupling with the nearly degenerate La band of Tyr (εmax ∼10, 000), in Trp–Tyr pairs. Trp–Trp exciton couplets have been reported in chymotrypsin and chymotrypsinogen [126, 135], dihydrofolate reductase [121, 136], and hemoglobin variant III from Chironomus thummi thummi [137]. A Trp–Tyr exciton couplet has been established by site-directed mutagenesis in PagP, a bacterial outer membrane enzyme that transfers fatty acyl groups [138]. The coat protein

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

of bacteriophage fd has an exciton couplet-like feature superimposed on an α-helix CD pattern [139], with a negative lobe at 223 nm and a positive lobe at 210 nm. This may represent a Trp (Bb )-Phe (La ) exciton couplet.

14.4. EXTRINSIC CHROMOPHORES For proteins that lack prosthetic groups, the electronic CD spectrum effectively ends at 300 nm. Therefore, the CD bands of chromophoric prosthetic groups, metal ions, substrates and substrate analogues, and inhibitors can be studied at wavelengths above 300 nm without interference from intrinsic protein groups. CD bands from such chromophores are called extrinsic CD bands. Thus, CD in the very near UV and visible regions is useful for studying protein–ligand interactions. In most cases, the free ligands are achiral or, if chiral, are ﬂexible and therefore have only weak CD when free. Thus, the extrinsic CD is characteristic of the bound ligand and its interaction with the protein. The CD of protein–ligand complexes has been reviewed [140–143], and Chapter 21 also provides many examples. Extrinsic CD bands can arise from three mechanisms: (1) inherent chirality of the bound chromophore, (2) coupling of transitions on the chromophore with transitions in the peptide and aromatic groups of the protein, and (3) mixing of transitions within the chromophore caused by the static ﬁeld of the protein. Of these mechanisms, the most readily analyzed case is (1) because it depends only on the geometry of the ligand, whereas the other two mechanisms involve interactions with the protein. Many extrinsic chromophores that are achiral when free are actually unresolvable mixtures of enantiomeric or nearly enantiomeric conformers. Binding to a protein selects one of these conformers. Examples include 11-cis retinal [144], bilirubin [145], and di- and triphenyl methyl dyes [146–148].

14.5. APPLICATIONS 14.5.1. Protein Folding CD, NMR, and ﬂuorescence are the major biophysical tools in the study of protein folding. The use of CD in monitoring the kinetics of protein folding has been reviewed [149, 150]. CD secondary structural analysis provides information about the native and the unfolded states of the protein, the two end points of the folding process. CD has also yielded important information about a key type of intermediate in the folding process, the molten globule (MG) in which the protein largely retains its native secondary structure but lacks a deﬁned tertiary structure. MGs were ﬁrst recognized as partially unfolded proteins observed at low pH (for reviews, see references 151 and 152). The protein αlactalbumin is a widely studied example. Far- and near-UV CD spectra of α-lactalbumin are shown in Figure 14.5 [153]. In the far-UV CD, spectra 1 and 2 are for the native protein at neutral pH in the presence and absence of Ca2+ , respectively; spectrum 3 is taken at pH 2 and is that of the MG; spectra 4 and 5 are taken at higher temperatures of 41◦ C and 78◦ C, respectively; and spectrum 6 is that of the fully unfolded protein in 6 M guanidinium chloride (GuCl). The squares and circles will be discussed in the next paragraph. The MG spectrum differs from those of the native protein and the protein that is unfolded by heat or GuCl, but it resembles the native spectrum more than those of the unfolded forms. In the near UV, spectra 1–3 are as for the far UV; spectrum 4 is that

485

486

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

3

0

4 5

[θ] (deg.cm2.dmol–1)

[θ] x10–3 (deg.cm2.dmol–1)

0

–5 5 1

–10

5

4

–15

200

2

–100

–200

–300

1 2

3

220 240 Wavelength (nm)

260

(a)

240

280 300 260 Wavelength (nm)

320

(b)

Figure 14.5. CD spectra of the native, unfolded, and molten globule (A-state) of α-lactalbumin in the far-UV (a) and near-UV regions (b). The open circles and squares show the CD values obtained by extrapolating to zero time of the refolding curves. In both panels, spectra 1 and 2 are for the holo and the apo forms, respectively, in the native state; spectrum 3 is for the A state. In (a), spectra 4 and 5 are of the thermally unfolded protein at 41◦ C and 78◦ C, respectively; spectrum 6 is that of the unfolded state in GuCl. In (b), spectrum 4 is that of the thermally unfolded protein at 62.5◦ C; spectrum 5 is that of the GuCl-unfolded state. (Reprinted with permission from Kuwajima et al. [153], © 1985 American Chemical Society.)

of thermally unfolded protein at 62.5◦ C; and spectrum 5 is that of the GuCl-unfolded protein. We see that the MG has a near-UV CD spectrum comparable to those of the unfolded forms and qualitatively different from that of the native form. These data lead to the conclusion that the MG of α-lactalbumin has secondary structure that resembles the native protein but lacks well-deﬁned tertiary structure. The open symbols in the CD spectra of α-lactalbumin (Figure 14.5) are derived from kinetic folding experiments [153]. Stopped-ﬂow experiments monitored by CD were used to study the refolding of the GuCl-unfolded protein on dilution in neutral buffer. Extrapolation of the signal observed at each wavelength to zero time generated the points shown in Figure 14.5. That these points nearly coincide with the equilibrium MG spectrum (curve 3) implies that during the dead time of the experiment (3 s in the data shown, but subsequent measurements with a 15-ms dead time [154] demonstrate that the intermediate forms on the millisecond timescale) the protein folds to a conformation with substantial secondary structure but no well-deﬁned tertiary structure. In addition to the similarity of the CD spectrum of this rapidly formed intermediate to that of the equilibrium MG, the unfolding of the kinetic intermediate by GuCl, monitored by CD, coincides with that of the equilibrium MG [154]. Kinetic bursts corresponding to the formation of MG-like intermediates have been observed for many proteins (reviewed in reference 155), and such species are now generally accepted as key intermediates in the folding of most proteins [152]. Another aspect of protein folding in which CD has had a major impact is the recognition and characterization of intrinsically disordered proteins (IDPs) [156, 157]. Recently,

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

it has been established that many proteins, especially from eukaryotes, have extensive regions that are disordered in the native state, in some cases extending throughout the protein [156]. Many lines of evidence, including far-UV CD, now support this onceheretical view. IDPs have far-UV CD spectra characteristic of unordered polypeptide chains (Section 14.2.3.5). These proteins usually fold into well-deﬁned structures upon binding to their targets—other proteins, nucleic acids, or other ligands. These coupled folding/binding reactions are conveniently followed by CD. The CD of IDPs generally becomes more negative near 220 nm and less negative near 200 nm with increasing temperature. These changes are nearly linear in temperature. This behavior has sometimes been interpreted as indicating “temperature-induced formation of secondary structure” [158]. Although the direction of the observed changes is consistent with α-helix and/or β-sheet formation, it is also consistent with a melting of PII -helix [157]. Recently, a combined CD, NMR, and small-angle X-ray scattering study of several IDPs has provided strong evidence for this latter interpretation [159].

14.5.2. Membrane Proteins Membrane proteins constitute about one-third of the proteins coded for in the human genome. CD spectroscopy should be especially useful in characterizing the secondary structure of these proteins because both X-ray diffraction and NMR are difﬁcult to apply to these systems—they are difﬁcult to crystallize and their size poses problems for NMR methods. Despite this, relatively few CD studies of membrane proteins have been reported. Membrane protein CD is known to be subject to artifacts produced by inhomogeneous samples and by differential scattering of left- and right-circularly polarized light [160]. Both of these problems can be avoided by solubilizing the protein in nonionic detergents, thereby generating a molecularly dispersed solution. It has been claimed that the lipid environment of membrane proteins leads to wavelength shifts in their CD spectra that make soluble proteins unsuitable for analyzing the CD spectra of membrane proteins. Wallace et al. [161] analyzed the CD spectra of eight membrane proteins and reported that the use of soluble proteins in the analysis gave inaccurate results. They attributed this to the problem of lipid-induced wavelength shifts and called for the development of a basis set of membrane proteins. Sreerama and Woody [85] developed a basis set of membrane proteins, based upon data of Park et al. [162], who reported CD spectra for 30 membrane proteins. At the time of the work of Park et al., structures were available for only four of the proteins, and one of these was of low resolution. Twelve years later, high-resolution structures were available for nearly half (13) of the proteins, providing an adequate set of reference proteins. Sreerama and Woody found that secondary structure analysis of the membrane proteins using a soluble protein basis set gave acceptable results, only slightly inferior to those obtained using the membrane protein basis set. The best results were obtained with a combined basis set. Sreerama and Woody pointed out that the data of Park et al. do not support the idea [161] of signiﬁcant wavelength shifts in membrane protein CD. The wavelength maximum of the positive π –π ∗ CD band for α-rich membrane proteins ranges from 192 to 196 nm, whereas that for α-rich soluble proteins ranges from 192 to 195 nm. Thus, a special set of reference proteins is not required for analyzing the secondary structure of membrane proteins. Turk et al. [163] used the combined soluble + membrane protein basis set [85] to analyze the secondary structure of a bacterial Na+ /galactose co-transporter. They found 85% α-helix, in good agreement with sequence analysis and topological analysis that

487

488

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

indicated 14 trans-membrane helices. Turk et al. also analyzed the CD of a fusion protein between the co-transporter and green ﬂuorescent protein (GFP), a β-rich protein with only one turn of α-helix. They obtained 60% α-helix which, corrected for the presence of the β-rich GFP component, is equivalent to the helix content of the co-transporter. Matsuo et al. [164] have used vacuum-UV CD to study the conformation of α1 glycoprotein (AGP) in solution and bound to liposomes. Spectra of the protein component were obtained by subtraction of the glycan spectrum, which contributes 10–20% of the CD at 193 nm. Matsuo et al. [89] used their vacuum-UV CD data for 31 reference proteins to determine the helix content and the number of helices and strands. Native AGP was found to contain 14% α-helix in three segments and 38% β-sheet in 10–11 strands. On binding to liposomes, the α-helix content increased to 50% in 7–8 segments and the β-sheet content decreased to 3–6% in two strands. Matsuo et al. [165] combined this information with a sequence-based secondary structure prediction method using a neural network and thus predicted the locations of the α-helical and β-strand segments in the amino acid sequence for the native and liposome-bound protein. The bacterial membrane protein phospholipid, lipid A palmitoyltransferase (PagP), is a β-barrel protein that transfers acyl chains from phospholipids in the outer leaﬂet of the outer membrane to a lipid A precursor to form lipid A. It has a strong preference for C16 , palmitoyl, chains. The NMR [166] and X-ray [167] structures of PagP reveal that the central cavity of the β-barrel is closed at the bottom by a glycyl (Gly88) residue, thus deﬁning the preferred length of the acyl chain that can be transferred. Bishop and coworkers [138] probed the nature of this “molecular ruler” by mutating the Gly at the bottom of the cavity to Ala, Cys, S-methyl Cys, and Met. As the length of the replacement side chain increases, one would expect the length of the preferred acyl chain to decrease by one methylene group for each additional atom in the side chain. This was found to be the case for all mutants except Gly88 → Cys, in which case C14 and C15 acyl chains were transferred with equal facility; that is, the acyl chain speciﬁcity is lost. Concomitant studies by CD showed that the wild-type protein and all mutants except Gly88 → Cys show a positive CD band at about 232 nm that was suspected to arise from an aromatic side chain. Theoretical calculations [138] based upon the crystal structure [167] predicted a strong positive couplet centered at about 228 nm associated with Tyr26 and Trp66, the side chains of which are near each other and near Gly88 at the bottom of the pocket. This prediction was veriﬁed by site-directed mutagenesis, which demonstrated that the positive 232-nm band disappears on mutation of either Tyr26 (Tyr26 → Phe) or Trp66 (Trp66 → Phe), whereas the negative CD band at 218 nm decreases. Thus, both theory and experiment show the existence of a Trp–Tyr exciton couplet (see Chapter 4 in this volume) in PagP, and the presence of this couplet correlates exactly with the speciﬁcity of acyl-chain transfer. The CD at 232 nm provides an excellent probe for the intact tertiary structure of PagP and is useful for monitoring protein stability in the wild-type and mutant proteins. The cause of the breakdown of the couplet and of acyl chain-length speciﬁcity was elucidated later [168], again with help from the exciton CD. In the original study [138], wild-type and mutant proteins were refolded at pH 8, the pH optimum for enzyme activity. When the Gly88 → Cys mutant was refolded at lower pH, down to pH 5, the exciton couplet was observed and strong acyl chain-length speciﬁcity for C14 was observed at pH7. Using the exciton intensity at 232 nm, PagP refolding at various pHs was shown to follow the titration of Cys88, and a pKa of 7.5 was obtained, which is about 2 pH units below the normal pKa of a Cys–SH group (∼9.5). Above the pKa , the charged thiolate group at the bottom of the cavity perturbs both the Tyr26–Trp66 exciton coupling and

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

the operation of the molecular ruler. In this system, exciton CD permits the determination of the pKa of a speciﬁc ionizable group in PagP.

14.5.3. Heme Proteins Heme proteins have been extensively studied by CD, and there has been great interest in the extrinsic bands of the heme [169–171]. The origin of the heme CD bands was elucidated by Hsu and Woody [172], who calculated the CD of the Soret and other heme bands for myoglobin and hemoglobin, using the available crystal structures. They were able to account for the sign and approximate magnitude of the Soret and visible bands and of the UV bands near 350 and 260 nm by the coupled oscillator interactions of the heme transitions with the π –π ∗ transitions of the aromatic side chains. The contributions of coupling with peptide transitions were found to be small. This interpretation was strengthened by calculations on hemoglobin from an insect [173, 174] and a lamprey [175], which showed that the same mechanism accounts for the CD bands in these proteins, with Soret bands opposite in sign to those of mammalian hemoglobins. A recent study by Dartigalongue and Hache [176] utilized Applequist’s method [177] to calculate the Soret CD of MbCO and Mb in the native form and in postulated intermediates following photolysis of the CO from MbCO. Their results for the native form are consistent with those of Hsu and Woody [172]. They found that although the net contribution of the proximal His to the Soret CD is small, in accord with Hsu and Woody, the orientation of the proximal His has a dramatic effect on the Soret CD spectrum. Rotation of the proximal His by 30◦ is predicted to lead to nearly an order of magnitude decrease in the Soret rotational strength. A potential role for inherent chirality in heme protein CD has been suggested by the observation that heme isomers in myoglobin (Mb) have very different Soret CD bands. The isomers differ by a 180◦ rotation about an axis through the α- and γ -methine carbons. NMR studies of Lamar et al. [178] showed that native MbCN consists of a 9:1 mixture of isomers A and B. MbCO freshly reconstituted from heme and globin consists of a 50:50 A:B mixture and its Soret CD has only about 50% of the magnitude of the equilibrium species [179]. Extrapolation of the linear relationship between the Soret CD intensity and the fraction of isomer B to zero and to one showed that the Soret band of the dominant isomer, A, is strong and positive, whereas that of the minor isomer, B, is weakly negative [180]. The model of Hsu and Woody [172] does not predict such a dramatic change upon 180◦ rotation of a planar heme about an in-plane axis. A nonplanar heme, however, might explain the results because the 180◦ rotation could reverse the sign of the intrinsic heme contributions while the aromatic coupling contributions undergo only minor changes. MD simulations for the major isomer, combined with CD calculations, indicate that the heme–aromatic coupling is the largest single contributor (∼40 %) and that intrinsic contributions of heme nonplanarity and heme–peptide contributions each contribute ∼30% of the Soret CD intensity [181]. The heme in cytochrome c is substantially distorted from planarity by its covalent attachment to the protein [182], and this distortion is likely to be preserved in the heme undecapeptide produced by proteolytic cleavage. A combination of molecular dynamics and CD calculations on the heme undecapeptide provided a good description of the Soret CD, which showed a signiﬁcant contribution from heme nonplanarity [183]. Near-UV CD has been useful for spectroscopically monitoring the allosteric transition in hemoglobin (Hb). Perutz et al. [184] reported that deoxyHb has a distinct negative band at 287 nm, whereas the CD of HbO2 is weak from 280 to 300 nm. They proposed

489

490

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

the negative band as a marker for the R → T transition, where R is the relaxed quaternary structure characteristic of liganded Hb, such as HbO2 , and T is the taut conformer found in deoxyHb. Perutz et al. proposed that the 287-nm band in deoxyHb arises from two aromatic side chains in the α1 β2 interface that is critical in the R → T transition: Tyr42 in the α-chain (α42 Tyr) and Trp37 in the β-chain (β37 Trp). These aromatic side chains are H-bonded in T but not in R, and this might account for the difference in CD. There are two additional aromatic residues in the α1 β2 interface, α140 Tyr and α145 Tyr. Nagai and co-workers [185] have explored the contributions of the four aromatic side chains to the difference CD of the R → T transition, using mutants in which each of the residues is replaced by a nonchromophoric residue. The residues identiﬁed by Perutz et al. [184] contributed 4% (α42 Tyr) and 18% (β37 Trp), whereas α140 Tyr and α145 Tyr contributed 32% and 27%, respectively. Jin et al. [185] suggested that α140 Tyr and α145 Tyr may be more important in the difference CD spectrum because they shift from solvent-exposed positions in the R conformer to hydrophobic pockets in the T conformer. By contrast, although α42 Tyr and β37 Trp change their H-bonding states, their hydrophobic environments do not undergo a change in the R → T transition. In another study, Li et al. [186] compared the near-UV CD of isolated α- and βsubunits of Hb to demonstrate that the change in CD at 287 nm is not solely due to quaternary structural differences. About 50% of the difference CD is associated with changes in tertiary structure of the individual subunits induced by oxygen binding. Applications of CD spectroscopy to the study of cytochrome c have been reviewed [187]. Far-UV CD spectra of cytochrome c down to 175 nm have been analyzed [187] by DichroWeb [92, 188] to characterize the secondary structure of the protein in four states (II to V, in order of increasing pH). In state III, which prevails at physiological pH, the content of α-helix, β-sheet, and β-turns was found to agree well with the X-ray structure [189]. State II, at pH 3, appears to be a molten globule state, with both α-helix and β-turn contents showing small decreases relative to the native form. The two high-pH forms, IV and V, showed signiﬁcant decreases in secondary structure relative to state III, but even state V had 31% α-helix and 20% β-turn, suggesting that the pH-induced conformational changes are predominantly at the tertiary structure level. Schweitzer-Stenner and co-workers [190, 191] have used the visible CD spectrum of cytochrome c, associated with the Soret and visible bands, to measure the electrostatic ﬁeld created by the protein at the heme iron. The Soret and visible bands are each associated with two nearly degenerate transitions that are split by the electrostatic ﬁeld (Stark effect) arising from the protein. This splitting manifests itself in the CD spectrum as couplets. Simulation of the CD spectra permitted determination of the ﬁeld strength for both oxidized and reduced cytochrome c [191]. Oxidized cytochrome c has absorption and CD bands at 695 nm, attributed to charge transfer from heme → Fe3+ [192] or Met80 S → Fe3+ [193]. This band has several subbands, probably arising from different protein conformers. The 695-nm band disappears upon opening of the heme crevice during unfolding. The temperature dependence of the sub-bands was analyzed, providing thermodynamic data for the unfolding of the various conformers [194].

14.5.4. Retinal Proteins The CD of rhodopsin has long been of interest because of its potential for providing information about the chirality of the retinal chain. A major concern is the question of

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

whether the extrinsic CD of rhodopsin is dominated by inherent chirality [195–200] or by coupling with protein transitions [201–203]. Recently, Pescitelli et al. [204] have reported theoretical calculations that strongly support inherent chirality as the dominant mechanism. They used time-dependent density functional theory (see Chapter 22 in volume 1) to calculate the CD of the bound chromophore, using geometries from several crystal structures, and to generate the transition parameters necessary to calculate the contributions of coupling (see Chapter 20 in volume 1) with aromatic and peptide groups in the protein. The two contributions were of opposite signs for both the α (500 nm) and β (340 nm) bands. For the α band, the inherent chirality contribution was twice as large as that from coupling, and for the β band the inherent contribution was much more dominant. Thus, in this case, inherent chirality is dominant, but this may not be true in other retinal proteins (e.g., bacteriorhodopsin).

14.6. COMPUTER RESOURCES Computer programs for performing secondary structure analysis and for simulating protein CD spectra are available at the websites or through the contacts listed below. CDPro [83] SELCON3 [83], CDsstr [82], CONTINLL [67, 83] Contact: [email protected] or [email protected] Internet: http://lamar.colostate.edu/∼sreeram/CDPro CDsstr [82] CD analysis and XTLsstr [205] crystal structure analysis http://biochem.science.oregonstate.edu/dr-johnsons-softwar-downloadinstructions CONTIN [67] ridge regression Internet: http://s-provencher.com/pages/contin-cd.shtml K2D2 [77] neural network Internet : http://www.ogic.ca/projects/k2d2 CCA[79] convex constraint analysis Internet : http://www.chem.elte.hu/departments/jimre/ DICHROWEB [91, 92] web-based secondary structure analysis Internet: http://www.cryst.bbk.ac.uk/cdweb/html/home.html DICHROPROT [206] linear regression [38], SELCON2 [72], SELCON3 [83], CONTIN [67], VARSLC [70] Internet: http://dicroprot-pbil.ibcp.fr PROTEIN CD DATA BANK [93, 94] deposition of and access to protein CD data Internet: http://pcddb.cryst.bbk.ac.uk/home.php DICHROCALC [207] web-based protein CD calculation Internet: http://comp.chem.nottingham.ac.uk/dichrocalc MATMAC [208] Tinoco [18] and matrix method [19] calculations of protein CD Contact: joerg.ﬂ[email protected] PROTEIN [209] matrix method [19] calculations of protein CD Contact: [email protected]

491

492

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

REFERENCES 1. N. Sreerama, R. W. Woody, Circular dichroism of peptides and proteins, in Circular Dichroism: Principles and Applications, N. Berova, K. Nakanishi, and R. W. Woody, eds., John Wiley & Sons, New York, 2000, pp. 601–620. 2. N. Sreerama, R. W. Woody, Meth. Enzymol . 2004, 383 , 318–351. 3. S. M. Kelly, N. C. Price, Biochim. Biophys. Acta Protein Struct. Mol. Enzymol . 1997, 1338 , 161–185. 4. S. M. Kelly, T. J. Jess, N. C. Price, Biochim. Biophys. Acta Proteins Proteom. 2005, 1751 , 119–139. 5. S. R. Martin, M. J. Schilstra, Meth. Cell Biol . 2008, 84 , 263–293. 6. B. A. Wallace, Q. Rev. Biophys. 2009, 42 , 317–370. 7. M. B. Robin, Amides, in Higher Excited States of Polyatomic Molecules, Vol. 2, Academic Press, New York, 1974, pp. 122–146. 8. M. B. Robin, Amides, in Higher Excited States of Polyatomic Molecules, Vol. 3, Academic Press, Orlando, FL, 1985, pp. pp. 303–308. 9. E. B. Nielsen, J. A. Schellman, J. Phys. Chem. 1967, 71 , 2297–2304. 10. L. B. Clark, J. Am. Chem. Soc., 1995 117 , 7974–7986. 11. L. Serrano-Andres, M. P. F¨ulscher, J. Am. Chem. Soc. 1996, 118 , 12190–12199. 12. L. Serrano-Andres, M. P. F¨ulscher, J. Am. Chem. Soc. 1996, 118 , 12200–12206. 13. J. D. Hirst, D. M. Hirst, C. L. Brooks, J. Phys. Chem. A 1997, 101 , 4821–4827. 14. N. A. Besley, J. D. Hirst, J. Phys. Chem. A 1998, 102 , 10791–10797. 15. W. Mofﬁtt, J. Chem. Phys. 1956, 25 , 467–478. 16. J. A. Schellman, Acc. Chem. Res. 1968, 1 , 144–151. 17. E. U. Condon, W. Altar, H. Eyring, J. Chem. Phys. 1937, 5 , 753–775. 18. I. Tinoco, Jr., Adv. Chem. Phys. 1962, 4 , 113–160. 19. P. M. Bayley, E. B. Nielsen, J. A. Schellman, J. Phys. Chem. 1969, 73 , 228–243. 20. H. DeVoe, J. Chem. Phys. 1965, 43 , 3199–3208. 21. J. Applequist, J. Chem. Phys. 1979, 71 , 4332–4338. 22. L. Pauling, R. B. Corey, H. R. Branson, Proc. Natl. Acad. Sci. USA 1951, 37 , 205–211. 23. D. J. Barlow, J. M. Thornton, J. Mol. Biol . 1988, 201 , 601–619. 24. G. Holzwarth, P. Doty, W. B. Gratzer, J. Am. Chem. Soc. 1962, 84 , 3194–3196. 25. G. Holzwarth, P. Doty, J. Am. Chem. Soc. 1965, 87 , 218–228. 26. R. W. Woody, J. Polym. Sci., Part D: Macromol. Rev . 1977, 12 , 181–320. 27. J. A. Schellman, P. Oriel, J. Chem. Phys. 1962, 37 , 2114–2124. 28. R. W. Woody, I. Tinoco, Jr., J. Chem. Phys. 1967, 46 , 4927–4945. 29. W. C. Johnson, Jr., I. Tinoco, Jr., J. Am. Chem. Soc. 1972, 94 , 4389–4390. 30. L. Serrano-Andres, M. P. F¨ulscher, J. Am. Chem. Soc. 1998, 120 , 10912–10920. 31. L. Serrano-Andres, M. P. F¨ulscher, J. Phys. Chem. B 2001, 105 , 9323–9330. 32. M. T. Oakley, J. D. Hirst, J. Am. Chem. Soc. 2006, 128 , 12414–12415. 33. B. M. Bulheller, A. J. Miles, B. A. Wallace, J. D. Hirst, J. Phys. Chem. B 2008, 112 , 1866–1874. 34. C. Toniolo, A. Polese, F. Formaggio, M. Crisma, J. Kamphuis, J. Am. Chem. Soc. 1996, 118 , 2744–2745. 35. C. Toniolo, F. Formaggio, S. Tognon, Q. B. Broxterman, B. Kaptein, R. Huang, V. Setnicka, T. A. Keiderling, I. H. McColl, L. Hecht, L. D. Barron, Biopolymers 2004, 75 , 32–45. 36. M. C. Manning, R. W. Woody, Biopolymers 1991, 31 , 569–586.

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

37. 38. 39. 40. 41. 42. 43. 44.

45. 46. 47. 48. 49.

50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61.

62. 63. 64. 65. 66. 67. 68. 69. 70. 71.

T. S. Sudha, E. K. S. Vijayakumar, P. Balaram, Int. J. Pept. Protein Res. 1983, 22 , 464–468. N. J. Greenﬁeld, G. D. Fasman, Biochemistry, 1969, 8 , 4108–4116. S. Brahms, J. Brahms, G. Spach, A. Brack, Proc. Natl. Acad. Sci. USA 1977, 74 , 3208–3212. J. G. Lees, B. A. Wallace, Spectroscopy Int. J ., 2002, 16 , 121–125. J. S. Richardson, E. D. Getzoff, D. C. Richardson, Proc. Natl. Acad. Sci. USA 1978, 75 , 2574–2578. R. W. Woody, Biopolymers 1969, 8 , 669–683. D. E. Kamen, Y. Griko, R. W. Woody, Biochemistry 2000, 39 , 15932–15943. R. W. Woody, Studies of theoretical circular dichroism of polypeptides: Contributions of β turns, in Peptides, Polypeptides, and Proteins, E. R. Blout, F. A. Bovey, M. Goodman, and N. Lotan, eds., New York, John Wiley & Sons, 1974, pp. 338–350. C. M. Venkatachalam, Biopolymers 1968, 6 , 1425–1436. C. A. Bush, S. K. Sarkar, K. D. Kopple, Biochemistry 1978, 17 , 4951–4954. V. S. Ananthanarayanan, N. Shyamasundar, Biochem. Biophys. Res. Commun. 1981, 102 , 295–301. L. M. Gierasch, C. M. Deber, V. Madison, C. H. Niu, E. R. Blout, Biochemistry 1981, 20 , 4730–4738. J. Bandekar, D. J. Evans, S. Krimm, S. J. Leach, S. Lee, J. R. McQuie, E. Minasian, G. Nemethy, M. S. Pottle, H. A. Scheraga, E. R. Stimson, R. W. Woody, Int. J. Pept. Protein Res. 1982, 19 , 187–205. M. L. Tiffany, S. Krimm, Biopolymers 1968, 6 , 1379–1382. M. L. Tiffany, S. Krimm, Biopolymers 1969, 8 , 347–359. R. W. Woody, Adv. Biophys. Chem. 1992, 2 , 37–79. Z. S. Shi, R. W. Woody, N. R. Kallenbach, Adv. Prot. Chem. 2002, 62 , 163–240. R. W. Woody, J. Am. Chem. Soc. 2009, 131 , 8234–8245. J. T. Yang, C. S. C. Wu, H. M. Martinez, Meth. Enzymol . 1986, 130 , 208–269. W. C. Johnson, Annu. Rev. Biophys. Biophys. Chem. 1988, 17 , 145–166. W. C. Johnson, Proteins Struct. Funct. Genet. 1990, 7 , 205–214. N. J. Greenﬁeld, Anal. Biochem. 1996, 235 , 1–10. N. J. Greenﬁeld, Meth. Enzymol . 2004, 383 , 282–317. N. J. Greenﬁeld, Nat. Protoc. 2006, 1 , 2876–2890. S. Y. Venyaminov, Determination of protein secondary structure, in Circular Dichroism and the Conformation of Biomolecules, G. D. Fasman, ed., Plenum Press, New York, 1996, pp. 69–107. S. Brahms, J. Brahms, J. Mol. Biol . 1980, 138 , 149–178. V. P. Saxena, D. B. Wetlaufer, Proc. Natl. Acad. Sci. USA 1971, 68 , 969–972. Y. H. Chen, J. T. Yang, Biochem. Biophys. Res. Commun, 1971, 44 , 1285–1291. Y. H. Chen, J. T. Yang, H. M. Martinez, Biochemistry 1972, 11 , 4120–4131. I. A. Bolotina, V. O. Chekhov, V. Y. Lugauskas, A. V. Finkelshtein, O. B. Ptitsyn, Mol. Biol . (English translation of Molekul. Biol .) 1980, 14 , 701–709. S. W. Provencher, J. Gl¨ockner, Biochemistry 1981, 20 , 33–37. V. V. Shubin, M. L. Khazin, T. B. Eﬁmovskaya, Mol. Biol . (English translation of Molekul. Biol .) 1990, 24 , 165–176. J. P. Hennessey, W. C. Johnson, Jr., Biochemistry 1981, 20 , 1085–1094. P. Manavalan, W. C. Johnson, Jr., Anal. Biochem. 1987, 167 , 76–85. I. H. M. van Stokkum, H. J. W. Spoelder, M. Bloemendal, R. van Grondelle, F. C. A. Groen, Anal. Biochem. 1990, 191 , 110–118.

493

494

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113.

N. Sreerama, R. W. Woody, Anal. Biochem. 1993, 209 , 32–44. G. Bohm, R. Muhr, R. Jaenicke, Prot. Eng. 1992, 5 , 191–195. M. A. Andrade, P. Chacon, J. J. Merelo, F. Moran, Protein Eng. 1993, 6 , 383–390. N. Sreerama, R. W. Woody, J. Mol. Biol . 1994, 242 , 497–507. P. Unneberg, J. J. Merelo, P. Chacon, F. Moran, Proteins Struct. Funct. Genet. 2001, 42 , 460–470. C. Perez-Iratxeta, M. A. Andrade-Navarro, BMC Struct. Biol . 2008, 8 , 25–29. P. Pancoska, T. A. Keiderling, Biochemistry 1991, 30 , 6885–6895. A. Perczel, M. Hollosi, G. Tusnady, G. D. Fasman, Protein Eng. 1991, 4 , 669–679. J. G. Lees, A. J. Miles, R. W. Janes, B. A. Wallace, BMC Bioinform. 2006, 7 , 507–516. S. Navea, R. Tauler, E. Goormaghtigh, A. de Juan, Proteins Struct. Funct. Bioinform. 2006, 63 , 527–541. W. C. Johnson, Proteins Struct. Funct. Genet. 1999, 35 , 307–312. N. Sreerama, R. W. Woody, Anal. Biochem. 2000, 287 , 252–260. N. Sreerama, S. Y. Venyaminov, R. W. Woody, Anal. Biochem. 2000, 287 , 243–251. N. Sreerama, R. W. Woody, Protein Sci . 2004, 13 , 100–112. K. A. Oberg, J. M. Ruysschaert, E. Goormaghtigh, Protein Sci . 2003, 12 , 2015–2031. W. Kabsch, C. Sander, Biopolymers 1983, 22 , 2577–2637. K. Matsuo, R. Yonehara, K. Gekko, J. Biochem. 2004, 135 , 405–411. K. Matsuo, R. Yonehara, K. Gekko, J. Biochem. 2005, 138 , 79–88. J. G. Lees, A. J. Miles, F. Wien, B. A. Wallace, Bioinformatics 2006, 22 , 1955–1962. A. Lobley, L. Whitmore, B. A. Wallace, Bioinformatics 2002, 18 , 211–212. L. Whitmore, B. A. Wallace, Nucl. Acids Res. 2004, 32 , W668–W673. B. A. Wallace, L. Whitmore, R. W. Janes, Proteins Struct. Funct. Bioinform. 2006, 62 , 1–3. L. Whitmore, R. W. Janes, B. A. Wallace, Chirality 2006, 18 , 426–429. W. H. Press, B. P. Flannery, S. A. Teukolsky, W. T. Vetterling, Numerical Recipes: The Art of Scientiﬁc Computing, Cambridge University Press, Cambridge, England, 1989. N. Sreerama, S. Y. Venyaminov, R. W. Woody, Protein Sci . 1999, 8 , 370–380. P. McPhie, Anal. Biochem. 2001, 293 , 109–119. P. McPhie, Biopolymers 2004, 75 , 140–147. P. McPhie, Anal. Biochem. 2008, 375 , 379–381. V. Raussens, J. M. Ruysschaert, E. Goormaghtigh, Anal. Biochem. 2003, 319 , 114–121. V. Raussens, J. M. Ruysschaert, E. Goormaghtigh, Anal. Biochem. 2006, 359 , 150–150. P. Pancoska, V. Janota, T. A. Keiderling, Anal. Biochem. 1999, 267 , 72–83. N. N. Kalnin, I. A. Baikalov, S. Y. Venyaminov, Biopolymers 1990, 30 , 1273–1280. W. K. Surewicz, H. H. Mantsch, D. Chapman, Biochemistry 1993, 32 , 389–394. A. Barth, C. Zscherp, Q. Rev. Biophys. 2002, 35 , 369–430. R. W. Williams, J. Mol. Biol . 1983, 166 , 581–603. R. W. Williams, Meth. Enzymol . 1986, 130 , 311–331. B. M. Bussian, C. Sander, Biochemistry 1989, 28 , 4271–4277. R. W. Sarver, W. C. Krueger, Anal. Biochem. 1991, 199 , 61–67. R. Pribic, Anal. Biochem. 1994, 223 , 26–34. V. Baumruk, P. Pancoska, T. A. Keiderling, J. Mol. Biol . 1996, 259 , 774–791. K. A. Oberg, J. M. Ruysschaert, E. Goormaghtigh, Eur. J. Biochem. 2004, 271 , 2937–2948. E. Goormaghtigh, R. Gasper, A. Benard, A. Goldsztein, V. Raussens, Biochim. Biophys. Acta Proteins Proteom. 2009, 1794 , 1332–1343.

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

114. P. Pancoska, E. Bitto, V. Janota, M. Urbanova, V. P. Gupta, T. A. Keiderling, Protein Sci . 1995, 4 , 1384–1401. 115. E. H. Strickland, CRC Crit. Rev. Biochem. 1974, 2 , 113–175. 116. P. C. Kahn, Meth. Enzymol . 1979, 61 , 339–378. 117. R. W. Woody, A. K. Dunker, Aromatic and cystine side-chain circular dichroism in proteins, in Circular Dichroism and Conformational Analysis of Biomolecules, G. D. Fasman, ed., Plenum Press, New York, 1996, pp. 109–157. 118. R. W. Woody, Aromatic side-chain contributions to protein circular dichroism, in Methods in Protein Structure and Stability Analysis, V. N. Uversky and E. Permyakov, eds., Nova Science Publishers, Hauppauge, NY, 2007, pp. 291–344. 119. S. Vuilleumier, J. Sancho, R. Loewenthal, A. R. Fersht, Biochemistry 1993, 32 , 10303–10313. 120. P. O. Freskg˚ard, L. G. M˚artensson, P. Jonasson, B. H. Jonsson, U. Carlsson, Biochemistry 1994, 33 , 14281–14288. 121. E. Ohmae, Y. Sasaki, K. Gekko, J. Biochem. 2001, 130 , 439–447. 122. T. M. Thompson, B. L. Mark, C. W. Gray, T. C. Terwilliger, N. Sreerama, R. W. Woody, D. M. Gray, Biochemistry 1998, 37 , 7463–7477. 123. N. Sreerama, M. C. Manning, M. E. Powers, J. X. Zhang, D. P. Goldenberg, R. W. Woody, Biochemistry 1999, 38 , 10814–10822. 124. D. Andersson, U. Carlsson, P. O. Freskg˚ard, Eur. J. Biochem. 2001, 268 , 1118–1128. 125. A. Y. M. Woody, R. W. Woody, Biopolymers 2003, 72 , 500–513. 126. I. B. Grishina, R. W. Woody, Faraday Disc. 1994, 99 , 245–262. 127. D. M. Rogers, J. D. Hirst, Biochemistry 2004, 43 , 11092–11102. 128. N. M. Green, M. D. Melamed, Biochem. J . 1966, 100 , 614–621. 129. L. A. Day, Biochemistry 1973, 12 , 5329–5339. 130. C. C. Yang, C. C. Chang, K. Hayashi, T. Suzuki, K. Ikeda, K. Hamaguchi, Biochim. Biophys. Acta 1968, 168 , 373–376. 131. R. W. Woody, Biopolymers 1978, 17 , 1451–1467. 132. R. W. Woody, Eur. Biophys. J . 1994, 23 , 253–262. 133. M. Shiraki, Sci. Pap. Coll. Gen. Educ. Univ. Tokyo 1969, 19 , 151–173. 134. S. K. Burley, G. A. Petsko, Science 1985, 229 , 23–28. 135. C. R. Cantor, S. N. Timasheff, Optical spectroscopy of proteins, in The Proteins, 3rd ed., H. Neurath and R. L. Hill, eds., Academic Press, New York, 1982, pp. 145–306. 136. K. Kuwajima, E. P. Garvey, B. E. Finn, C. R. Matthews, S. Sugai, Biochemistry 1991, 30 , 7693–7703. 137. A. Wollmer, Konformationsanalyse von Proteinen mit Hilfe des Circulardichroismus und der optischen Rotationsdispersion. Habilitation thesis, RWTH Aachen, 1972. 138. M. A. Khan, C. Neale, C. Michaux, R. Pomes, G. G. Prive, R. W. Woody, R. E. Bishop, Biochemistry 2007, 46 , 4565–4579. 139. G. E. Arnold, L. A. Day, A. K. Dunker, Biochemistry 1992, 31 , 7948–7956. 140. G. Blauer, Struct. Bonding (Berlin) 1974, 18 , 69–129. 141. N. J. Greenﬁeld, CRC Crit. Rev. Biochem. 1975, 3 , 71–110. 142. F. Zsila, Z. Bikadi, I. Fitos, M. Simonyi, Curr. Drug Discov. Technol . 2004, 1 , 133–153. 143. A. Rodger, R. Marrington, D. Roper, S. Windsor, Meth. Mol. Biol . 2005, 305 , 343–363. 144. B. Honig, T. G. Ebrey, Annu. Rev. Biophys. Bioeng. 1974, 3 , 151–177. 145. G. Blauer, G. Wagniere, J. Am. Chem. Soc. 1975, 97 , 1949–1954. 146. J. S. Wicken, R. W. Woody, Biochemistry 1973, 12 , 3459–3466. 147. J. S. Wicken, R. W. Woody, Arch. Biochem. Biophys. 1974, 162 , 12–16.

495

496

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

148. 149. 150. 151. 152. 153. 154. 155. 156. 157.

158. 159. 160. 161. 162. 163. 164. 165. 166. 167. 168. 169. 170. 171.

172. 173. 174. 175.

176. 177. 178. 179. 180.

J. F. Towell, R. W. Woody, Biochemistry 1980, 19 , 4231–4237. K. Kuwajima, Meth. Mol. Biol . 1995, 40 , 115–135. N. J. Greenﬁeld, Nat. Protoc. 2006, 1 , 2891–2899. O. B. Ptitsyn, Adv. Protein Chem. 1995, 47 , 83–229. M. Arai, K. Kuwajima, Adv. Protein Chem. 2000, 53 , 209–282. K. Kuwajima, Y. Hiraoka, M. Ikeguchi, S. Sugai, Biochemistry 1985, 24 , 874–881. M. Arai, K. Kuwajima, Fold. Design 1996, 1 , 275–287. R. W. Woody, Meth. Enzymol . 2004, 380 , 242–285. P. Tompa, Structure and Function of Intrinsically Disordered Proteins, CRC Press, Boca Raton, FL, 2010. R. W. Woody, Circular dichroism of intrinsically disordered proteins, in Instrumental Analysis of Intrinsically Disordered Proteins: Assessing Structure and Conformation, V. N. Uversky and S. Longhi, ed., John Wiley & Sons, Hoboken, NJ, 2009, pp. 303–321. V. N. Uversky, J. R. Gillespie, A. L. Fink, Proteins Struct. Funct. Genet. 2000, 41 , 415–427. M. Kjaergaard, A.-B. Nørholm, R. Hendus-Altenburger, S. F. Pedersen, F. M. Poulsen, B. B. Kragelund, Protein Sci . 2010, 19 , 1555–1564. D. W. Urry, T. H. Ji, Arch. Biochem. Biophys. 1968, 128 , 802–807. B. A. Wallace, J. G. Lees, A. J. W. Orry, A. Lobley, R. W. Janes, Protein Sci . 2003, 12 , 875–884. K. Park, A. Perczel, G. D. Fasman, Prot. Sci . 1992, 1 , 1032–1049. E. Turk, O. K. Gasymov, S. Lanza, J. Horwitz, E. M. Wright, Biochemistry 2006, 45 , 1470–1479. K. Matsuo, H. Namatame, M. Taniguchi, K. Gekko, Biochemistry 2009, 48 , 9103–9111. K. Matsuo, H. Watanabe, K. Gekko, Proteins-Struct. Funct. Bioinform., 2008, 73 , 104–112. P. M. Hwang, W. Y. Choy, E. I. Lo, L. Chen, J. D. Forman-Kay, C. R. H. Raetz, G. G. Prive, R. E. Bishop, L. E. Kay, Proc. Natl. Acad. Sci. USA 2002, 99 , 13560–13565. V. E. Ahn, E. I. Lo, C. K. Engel, L. Chen, P. M. Hwang, L. E. Kay, R. E. Bishop, G. G. Prive, EMBO J . 2004, 23 , 2931–2941. M. A. Khan, J. Moktar, P. J. Mott, R. E. Bishop, Biochemistry 2010, 49 , 2368–2379. Y. P. Myer, Meth. Enzymol . 1978, 54 , 249–284. Y. P. Myer, A. Pande, Circular dichroism studies of hemoproteins and heme models, in The Porphyrins, D. Dolphin, ed., Academic press, New York, 1978, pp. 271–322. R. Woody, Circular dichroism probes of hemoglobin structure, in Biochemical and Clinical Aspects of Hemoglobin Abnormalities, W. S. Caughey, ed., Academic press, New York, 1978, pp. 279–298. M. C. Hsu, R. W. Woody, J. Am. Chem. Soc. 1971, 93 , 3515–3525. J. Fleischhauer, A. Wollmer, Z. Naturforsch. 1972, B 27 , 530–532. W. Strassburger, H. Thiele, W. Steigemann, E. Weber, J. Fleischhauer, A. Wollmer, HoppeSeylers Z. Physiolog. Chem. 1978, 359 , 1155–1156. R. W. Woody, Contributions of aromatic amino acid residues to the optical activity of protein–ligand complexes, in Protein–Ligand Interactions, H. Sund and G. Blauer, eds., Walter de Gruyter, Berlin, 1975, pp. 60–77. T. Dartigalongue, F. Hache, J. Chem. Phys. 2005, 123 , 184901-1–184901-9 J. Applequist, K. R. Sundberg, M. L. Olson, L. C. Weiss, J. Chem. Phys. 1979, 70 , 1240–1246. G. N. Lamar, H. Toi, R. Krishnamoorthi, J. Am. Chem. Soc. 1984, 106 , 6395–6401. H. S. Aojula, M. T. Wilson, A. Drake, Biochem. J . 1986, 237 , 613–616. H. S. Aojula, M. T. Wilson, G. R. Moore, D. J. Williamson, Biochem. J . 1988, 250 , 853–858.

ELECTRONIC CIRCULAR DICHROISM OF PROTEINS

181. C. Kieﬂ, N. Sreerama, R. Haddad, L. S. Sun, W. Jentzen, Y. Lu, Y. Qiu, J. A. Shelnutt, R. W. Woody, J. Am. Chem. Soc. 2002, 124 , 3385–3394. 182. J. D. Hobbs, J. A. Shelnutt, J. Protein Chem. 1995, 14 , 19–25. 183. G. Blauer, N. Sreerama, R. W. Woody, Biochemistry 1993, 32 , 6674–6679. 184. M. F. Perutz, J. E. Ladner, S. R. Simon, C. Ho, Biochemistry 1974, 13 , 2163–2173. 185. Y. Jin, H. Sakurai, Y. Nagai, M. Nagai, Biopolymers 2004, 74 , 60–63. 186. R. Li, Y. Nagai, M. Nagai, J. Inorg. Biochem. 2000, 82 , 93–101. 187. R. Schweitzer-Stenner, A. Hagarman, D. Verbaro, J. B. Soffer, Meth. Enzymol . 2009, 466 , 109–153. 188. L. Whitmore, B. A. Wallace, Biopolymers 2008, 89 , 392–400. 189. G. V. Louie, G. D. Brayer, J. Mol. Biol . 1990, 214 , 527–555. 190. I. Dragomir, A. Hagarman, C. Wallace, R. Schweitzer-Stenner, Biophys. J . 2007, 92 , 989–998. 191. R. Schweitzer-Stenner, J. Phys. Chem. B 2008, 112 , 10358–10366. 192. W. A. Eaton, R. M. Hochstrasser, J. Chem. Phys. 1967, 46 , 2533–2539. 193. J. McKnight, M. R. Cheesman, A. J. Thomson, J. S. Miles, A. W. Munro, Eur. J. Biochem. 1993, 213 , 683–687. 194. R. Schweitzer-Stenner, R. Shah, A. Hagarman, I. Dragomir, J. Phys. Chem. B 2007, 111 , 9603–9607. 195. M. J. Burke, D. C. Pratt, T. R. Faulkner, A. Moscowitz, Exp. Eye Res. 1973, 17 , 557–572. 196. B. Honig, P. Kahn, T. G. Ebrey, Biochemistry 1973, 12 , 1637–1643. 197. H. Kakitani, T. Kakitani, S. Yomosa, J. Phys. Soc. Japan 1977, 42 , 996–1004. 198. V. Buss, K. Kolster, F. Terstegen, R. Vahrenhorst, Angew. Chem. Intl. Ed ., 1998, 37 , 1893–1895. 199. Q. Tan, J. Lou, B. Borhan, E. Karnaukhova, N. Berova, K. Nakanishi, Angew. Chem. Intl. Ed . 1997, 36 , 2089–2093. 200. Y. Fujimoto, N. Fishkin, G. Pescitelli, J. Decatur, N. Berova, K. Nakanishi, J. Am. Chem. Soc. 2002, 124 , 7294–7302. 201. A. S. Waggoner, L. Stryer, Biochemistry 1971, 10 , 3250–3254. 202. A. Kropf, B. P. Whittenberger, S. P. Goff, A. S. Waggoner, Exp. Eye Res. 1973, 17 , 591–606. 203. E. M. Johnston, R. Zand, Biochemistry 1973, 12 , 4637–4643. 204. G. Pescitelli, N. Sreerama, P. Salvadori, K. Nakanishi, N. Berova, R. W. Woody, J. Am. Chem. Soc. 2008, 130 , 6170–6181. 205. S. M. King, W. C. Johnson, Proteins Struct. Funct. Genet. 1999, 35 , 313–320. 206. G. Deleage, C. Geourjon, Comput. Appl. Biosci . 1993, 9 , 197–199. 207. B. M. Bulheller, J. D. Hirst, Bioinformatics 2009, 25 , 539–540. 208. G. Kurapkat, P. Kruger, A. Wollmer, J. Fleischhauer, B. Kramer, E. Zobel, A. Koslowski, H. Botterweck, R. W. Woody, Biopolymers 1997, 41 , 267–287. 209. R. W. Woody, N. Sreerama, J. Chem. Phys. 1999, 111 , 2844–2845. 210. V. N. Uversky, J. Li, A. L. Fink, J. Biol. Chem. 2001, 276 , 10737–10744.

497

15 ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES Claudio Toniolo, Fernando Formaggio, and Robert W. Woody

15.1. INTRODUCTION Peptides are of great importance in biology for their many activities: hormones, intracellular signals, neurotransmitters, immunogens, antibiotics, toxins, and so on. The biological signiﬁcance of peptides makes them important for the pharmaceutical industry and biotechnology. Isolation and synthesis of natural peptides and generation of a wide range of analogues are increasingly relevant activities in these ﬁelds. Knowledge of the conformation of a peptide is essential for understanding its mechanism of action. Electronic circular dichroism (hereafter the common abbreviation CD will be used for this spectroscopy) has been an invaluable tool for the conformational analysis of peptides since the introduction of commercial instrumentation in the early 1960s, replacing its dispersive counterpart, optical rotatory dispersion. The unique stereochemical sensitivity of a chirospectroscopic method such as CD guarantees it an important role in solution studies, along with NMR, IR absorption, and Raman. Over the past two decades, CD has been joined by two methods that combine the advantages of chiral sensitivity with the richness of vibrational spectroscopy: vibrational circular dichroism and Raman optical activity. The applications of these methods to peptide systems are described in Chapters 22 and 23 of this volume, respectively. Readers of this chapter may also ﬁnd the chapters on the CD of peptidomimetics (Chapter 16, this volume), the CD of proteins (Chapter 14, this volume), and the independent systems theory of CD (Chapter 20, Volume 1) of interest. Of the many general reviews of the applications of CD to protein and peptide systems, those of Sears and Beychok [1], Johnson [2, 3], Woody [4] and Kelly et al. [5] may be found especially useful. Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

499

500

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

15.2. α-HELIX The α-helix is well deﬁned, since it corresponds to a narrow region of the Ramachandran (φ, ψ) map [6], and it is the most thoroughly investigated type of polypeptide secondary structure. It was ﬁrst proposed by Pauling et al. [7] in 1951 and accounts for approximately one-third of the residues in globular proteins [8]. For a detailed discussion of the 3D-structural parameters of this helix, the reader should refer to Section 15.3 dealing with the closely related 310 -helix. The α-helix has an intense and characteristic CD spectrum [9] (Figure 15.1). The negative band at ∼222 nm is attributable to the n → π ∗ transition of the peptide group [10, 11]. The peptide π → π ∗ transition gives rise to the other two bands, through exciton splitting (see Chapter 4, this volume) [11–13]: the negative band at ∼208 nm (parallel exciton band) and the positive band at ∼190 nm (perpendicular exciton band). The α-helix CD spectrum shows only minor solvent effects. The case of poly(lAla) in 1,1,1,3,3,3-hexaﬂuoroisopropanol (HFIP) is an apparent exception [14]. In HFIP, poly(l-Ala) has a CD spectrum that is reduced about twofold in amplitude, maxima that are blue-shifted by several nanometers, and an n → π ∗ band that is diminished to a shoulder, relative to the normal α-helix spectrum. Other helix-forming polypeptides in HFIP exhibit normal CD spectra. Substantial differences are also seen in the UV

80

[θ] x 10–3 (deg x cm2 x dmol–1)

40

0

–40

Figure 15.1. CD spectrum of poly(L-Glu) in 180

200

220 Wavelength (nm)

240

260

aqueous solution (pH 4.3). (Redrawn from reference 9.)

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

and IR absorption spectra of poly(l-Ala) in HFIP, compared to those of other α-helical polypeptides. Parrish and Blout [14] proposed that poly(l-Ala) in HFIP adopts a modiﬁed α-helix conformation in which the peptide planes, nearly parallel to the helix axis in the Pauling and Corey structure [7], are tilted so that the carbonyl oxygens can H-bond to the solvent, while retaining the intramolecular 1 ← 5 H-bonding. The small size of l-Ala makes this conformation, known as the αII conformation [15], feasible, in contrast to other amino acids. Monitoring helix-coil equilibria [16–19] is one of the most common applications of CD, which is the method of choice because of the large difference between the spectrum of the α-helix and that of the unordered conformation (see Section 15.7). For this and other applications, it is important to quantitate α-helix content. Figure 15.2 shows a family of CD spectra generated by melting of α-helix. A sharp isodichroic point near 203 nm is characteristic of helix-coil equilibria and implies a two-state equilibrium. Helix-coil equilibria are commonly monitored by measuring the CD at 222 nm, where the difference between helix and unordered CD is at a maximum and the signal-to-noise ratio is favorable. If only the α-helix and unordered forms are present, as is implied by an isodichroic point, the fraction of helix, fα can be calculated: fα = ([θ ]222,obs − [θ ]222,u )/([θ ]222,α − [θ ]222,u ) where [θ ]222,obs is the observed 222-nm ellipticity, and [θ ]222,α and [θ ]222,u are the ellipticities of the α-helix and unordered forms, respectively. Chen and Yang [20] proposed values of −32, 640 and −2340 deg cm2 dmol−1 for the α-helix and unordered forms, respectively, so fα = −([θ ]222,obs + 2340)/(30,300) This equation is useful for obtaining approximate values of fα , but precise work requires more reﬁned parameters. Section 15.7 should be consulted for a discussion of the work of Park et al. [21] on this topic. For peptides, the dependence of the CD upon helix length becomes signiﬁcant. Amide groups at the ends of a helix have a different environment and contribute differently to the CD than those in the middle of the helix, giving rise to end effects. Yang and co-workers [22, 23] proposed an empirical equation for the helix-length dependence of CD: [θn ]λ = [θ∞ ]λ (n − kn )/n where [θn ]λ is the mean residue ellipticity at wavelength λ of a helix with n residues; [θ∞ ]λ is the mean residue ellipticity of an inﬁnite helix at the same wavelength; and kn is a constant that corrects for end effects. The constant kn can be thought of as the number of residues effectively missing from the helix. Its value will depend on the CD band studied. Chen et al. [23] showed that this equation ﬁts the theoretical helix-length dependence predicted in two different studies [11, 24]. The helix-length dependence has also been formulated in terms of the number of amides, r, which has a sounder theoretical basis than the number of residues: [θr ]λ = [θ∞ ]λ (r − kr )/r where [θr ]λ is the ellipticity per amide at wavelength λ and kr is the correction for the number of missing amides. For an unblocked peptide, r = n − 1, so kr = kn − 1; for

501

502

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

60

[θ] x 10–3 (deg x cm2 x dmol–1)

40

20

0 70°C 30°C 15°C

–20

1°C

Figure 15.2. CD spectra of an Nα -acetylated (Ac) 17-mer peptide amide based on L-Ala, L-Glu,

–40 200

220 Wavelength (nm)

240

and L-Lys in aqueous solution (pH 7) as a function of heating from 1◦ C to 70◦ C. (Redrawn from reference 18.)

an N α -acyl peptide or a C α -amidated peptide, r = n and kr = kn ; for an N α -acyl, C α amidated peptide, r = n + 1, and kr = kn + 1. Gans et al. [25] used the latter formulation to ﬁt the theoretical CD results of Manning and Woody [26] and experimental data for peptides with r = 14–22 under conditions where they were expected to be nearly 100% helical. The experimental data gave kr = 4.7 and [θ∞ ]222 = −40,000 deg cm2 dmol−1 . This value of [θ∞ ]222 is consistent with the data for synthetic polypeptides in a variety of solvents [27]. The CD of the α-helix also depends on temperature, although not so strongly as that of the unordered conformation. Baldwin and co-workers [18, 28] reported a linear temperature dependence of [θ∞ ]222 = −42,500 + 125t deg cm2 dmol−1 (t = ◦ C) for lAla-rich peptides, assuming kr = 3. The values of the slope and intercept vary somewhat with the peptide sequence. CD, along with 220 MHz (in 1969!) 1 H NMR, was instrumental in detecting the onset of the helix conformation at the heptamer level in terminally-protected l-Glu(OEt) (where OEt is ethoxy) homo-oligopeptides in structure-supporting organic solvents [29]. Remarkably, in this work TFE (2,2,2-triﬂuoroethanol), a well-known UV-transparent and helicogenic solvent, was used in CD for the ﬁrst time. Following the same approach (CD technique, TFE as the solvent), high amounts of α-helical conformation were reported

503

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

11

12–15

60 10 9

[θ] x 10–3 (deg x cm2 x dmol–1)

40

8 20 7 6 1 0 5 6 –20

15

–40 180

200

220 Wavelength (nm)

240

Figure 15.3. CD spectra of Boc-(L-Met)n -NHPEG (n = 1–15 ; Boc, tert-butyloxycarbonyl) in TFE solution. (Redrawn from reference 30.)

for monodisperse (l-Met)n (Figure 15.3) and [l- Lys(Z)]n , where Z is benzyloxycarbonyl, homo-oligopeptides to the 15-mer level [30, 31]. These two peptide series were solubilized by covalently linking a polyethylene glycol (PEG) chain to their C-terminus. The threshold for helix stability results from the entropic barrier for helix nucleation [32], which can be overcome by providing a nucleus to initiate the helix. Kemp and co-workers [33–35] designed a series of nuclei, RHelOH, where HelOH = (1S , 4S , 7S , 10S )-2-oxo-9-thia-3,12-diazatricyclo[8.2.1.03.7 ]tridecane-4carboxylic acid and R = Ac, Acβ-l-Asp. These conformationally restricted templates can adopt a conformation in which three carbonyl groups are oriented like the three carbonyls at the N-terminus of an α-helix and thus provide a nucleus for helix formation. A CD study [36] of l-Ala–rich peptides of sequence AcHel–(l-Ala–l-Ala–l-Ala–lAla–l-Lys)n –l-Ala–l-Ala–NH2 (n = 1–5) gave unusually large [θ ]222 values: −50,00 deg cm2 dmol−1 at 0◦ C in 16 mol% ethylene glycol/water, pH 1, 0.2 M NaClO4 , and −60,000 cm2 dmol−1 at −20◦ C. The 208-nm band is not enhanced under these conditions (data are not available for the 190-nm band). A different kind of helix nucleus was introduced by Bierzynski and co-workers [37], who found that when the Ca2+ -binding loop of calmodulin binds La3+ ions, the four peptide groups at the C-terminus adopt an α-helical conformation, and additional residues extend this helix. This ﬁnding permitted the measurement of the CD of α-helices

504

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

with 4, 8, and 11 peptides [38]. The important result is that the α-helix CD pattern persists even when the helix contains only four amides, a single turn, although the amplitude is diminished and the shape is altered. Earlier theoretical studies [11, 13, 24, 26] predicted that the 208-nm negative band would be detectable only for helices of at least 10 residues. However, calculations [38, 39] using the experimental π → π ∗ transition dipole moment direction [40] predict a 208-nm band discernable even at the dipeptide level. The induction or augmentation of α-helix can also be accomplished by a strongly helicogenic Cα -tetrasubstituted α-amino acid placed at the N-terminus [41]. According to a CD analysis, Ac-(S )-Iva (Iva, isovaline) is able to increase the α-helix content of an already partially folded 13-mer peptide by about 7% in aqueous solution. Signiﬁcantly, here the intensity of the 222-nm band is not unusual; that is, it is lower than that at 206 nm. Kemp and co-workers [42–44] discovered that some l-Ala-rich peptides show anomalous temperature dependence of the n → π ∗ CD at low temperatures. In such peptides, [θ ]222 values as large as −65,000 deg cm2 dmol−1 have been observed at −50◦ C [42]. Peptides exhibiting this behavior can be recognized by a dual-wavelength parametric test in which [θ ]222 is plotted against [θ ]208 for a family of spectra generated, for example, by varying temperature. If the shape of the spectrum of the helix is independent of temperature, this plot will be a straight line with a slope of 1.1–1.3. For such peptides, helix content can be analyzed using a limiting ellipticity at 222 nm of −40,000 deg cm2 dmol−1 as described above. However, for a number of l-Ala-rich peptides studied by Kemp’s group, the dual wavelength plot curves upward, corresponding to enhanced [θ ]222 relative to [θ ]208 as temperature decreases. The temperature and length dependence of the CD of such systems has been characterized [44]. What is the origin of the remarkable enhancement of the n → π ∗ band in these peptides? This anomalously strong CD reﬂects a conformational modiﬁcation of the αhelix, perhaps a conformation more like the Pauling and co-workers [7] helix, which is predicted [9–11] to have a more intense n → π ∗ band than the Barlow and Thornton [8] helix characteristic of proteins. Lending some credence to this is the observation [43] of an Hα –HN coupling constant that indicates φ = −50◦ , close to that for the Pauling and co-workers helix (−47◦ ) and differing from that for the Barlow and Thornton helix (−62◦ ). Woody [45] suggested that the high amplitude might arise from a strong electrostatic mixing of the n → π ∗ and π → π ∗ transitions in peptide groups perturbed by the proximity of a charged side-chain l-Lys ammonium group [46]. However, the observation of an intensiﬁed n → π ∗ band in pure (l-Ala)n sequences [43], as well as a natural protein [47] and other peptides [48], excludes this explanation. Dang and Hirst [49] have suggested a shortening of the H-bonds in the helix caused by the low temperatures and the co-solvents. They calculated a [θ ]max of ∼ − 40, 000 deg cm2 ˚ increasing in magnitude to −50, 000 deg dmol−1 for the normal H-bond length of 3.0 A, ˚ cm2 dmol−1 for an H-bond length of 2.7 A.

15.3. 310 -HELIX The 310 -helix (or, more properly, the 3.010 -helix) (Figure 15.4b) is an alternative helical conformation that has been observed in many peptide and protein structures [50–56]. It is more tightly wound than the α-helix (or 3.613 -helix) (Figure 15.4a). The backbone torsion angles of the right-handed 310 -helix (ideally φ = −60◦ , ψ = −30◦ ) are within

505

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

(a)

(b)

1 ← 5 trans, trans

1← 4 trans (III)

3.613-helix (α-helix)

3.010-helix

Figure 15.4. (a) A 10-residue, right-handed 3.613 -helix (α-helix) and its building block, one of the 1 ← 5 trans, trans intramolecularly H-bonded peptide conformations (also termed α-bend or C13 conformation). (b) A 10-residue, right-handed 3.010 -helix and its building block, one of the 1 ← 4 trans intramolecularly H-bonded peptide conformations (also termed type-III β-turn or C10 conformation). Intramolecular C=O· · ·H–N H-bonds are represented by dashed lines, O atoms are larger and gray, and N and H atoms are white.

the same region of the conformational map as those of the α-helix (ideally φ = −55◦ , ψ = −45◦ ). However, the intramolecular C=O · · · H–N H-bonding schemes are significantly different in the two helices, being of the 1 ← 4 type (trans C10 -form or type-III β-turn) in the 310 -helix, while of the 1 ← 5 type (trans –trans C13 -form or α-bend) in the α-helix. A long polypeptide 310 -helix formed by Cα -monosubstituted α-amino acid residues is less stable than the α-helix, its van der Waals energy is less favorable (it has several close, although not forbidden, short contacts), and the H-bond geometry is not optimal. Thus, for many years it was considered unlikely that long stretches of 310 -helix would be observed. However, there is no disallowed region of the conformational (φ, ψ) space separating these two regularly folded secondary structures. Therefore, the α-helix may be gradually transformed into a 310 -helix (and vice versa) maintaining a near-helical conformation of the chain throughout. Furthermore, if one of the conformations should turn out to be impossible (say, as a result of side-chain interactions), the main chain may slip into the other conformation. In fact, the 310 -helix appears to derive its relevance mainly from its proximity in the conformational energy map to the more stable α-helix. Thus, the role of the 310 -helix as an important intermediate in the mechanism of folding of α-helical proteins may be envisaged [54]. According to a statistical analysis from X-ray diffraction structures of globular proteins [8], the majority of 310 -helices are short (the mean length is 3.3 residues, that is, approximately one turn of helix) and a signiﬁcant percentage of 310 -helices occur as an N- or C-terminal extension to an α-helix. However, a few relatively long

506

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

NH HN

H3C H2C

H3C H3C

CH3

H2C

CH3

HN

CO

HN

CO

Aib

(S)-Iva

H2C

CH3 CH3

HN CO (S)-(αMe)Nva

H3C HC

N H

CH3

HN CO (S)-(αMe)Val

HN

CO

(S)-ATANP

Figure 15.5. Chemical formulae of the C α -methylated α-amino acid residues Aib, (S)-Iva, (S)-(αMe)Nva, (S)-(αMe)Val, and (S)-ATANP discussed in this section.

(7–12 residues) 310 -helical segments were also detected [57]. Conformational energy calculations demonstrated that Aib (α-aminoisobutyric acid or C α,α -dimethylglycine, Figure 15.5), the prototype of achiral, C α -tetrasubstituted α-amino acids, can effectively promote the onset of 310 -helices in short (or relatively short) peptides due to steric interactions involving the gem-dimethyl groups linked to the α-carbon (Thorpe–Ingold effect) [50–53, 58]. Since 1978, by taking advantage of the incredibly high crystallinity of peptides rich in Cα -tetrasubstituted residues, hundreds of X-ray diffraction structures of 310 -helical peptides were solved. The two longest 310 -helical sequences known to date are the homo-deca- and undecapeptides -(Aib)10,11 -. The minimal main-chain length required for a peptide heavily based on Aib residues to form an α-helix in the crystal state corresponds to seven residues (interestingly, about 13 amino acids are required to induce an α-helix in the solid state for a peptide with protein residues only). By contrast, there is no critical main-chain length dependence for 310 -helix formation; that is, incipient 310 -helices are formed at the lowest possible level (an N α -acylated tripeptide). An N α - and C-blocked -(Aib–l-Ala)3 - peptide gives a regular 310 -helix, but an -(Aib–lAla)4 - peptide gives a predominant α-helix. In peptides of eight or more residues, the α-helix is preferred over the 310 -helix if the percentage of Aib residues does not exceed 50%. However, one or two 310 -helical residues may be observed at either end of the αhelical stretch (the short bits of 310 -helix tighten up the ends of the α-helix by moving the related peptide groups nearer to the axis). This conformational preference was also found in many members of the family of peptaibiotics (membrane-active, Aib-rich, naturally occurring peptides exhibiting antibiotic activity) [59]. The average number of α-helical residues in undeca- and longer peptides is seven (two turns). The average parameters for 310 - and α-helices observed in studies at atomic resolution are listed in Table 15.1 [50]. The number of residues per turn (3.24) is intermediate between those of the theoretical 310 -helix and the theoretical/experimental α-(3.613 -)helix. In a perfect 310 -helix,

TAB L E 15.1. Average Parameters for Right-Handed 310 - and α-Helices [50] Parameter ϕ ψ N· · ·O=C H-bond angle Rotation (per residue) Axial translation (per residue) Residues per turn Pitch

310 -Helix

α-Helix

−57◦ −30◦ 128◦ 111◦ ˚ 1.94 A 3.24 ˚ 6.29 A

−63◦ −42◦ 156◦ 99◦ ˚ 1.56 A 3.63 ˚ 5.67 A

507

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

the side chains on successive turns are exactly eclipsed since there is an integer number of residues per turn. However, the experimentally observed, slightly fractional number of residues per turn does not line up side chains, thereby inducing a slightly staggered, energetically more favorable, disposition. This property may have some implications if a relatively long, amphiphilic 310 -helix needs to be built. On the contrary, the α-helix, with its largely fractional number of amino acids per turn, requires two turns (seven residues or a “heptad” repeat) to position two side chains exactly one on top of the other on the same helical face. Distinct hydrophobic/hydrophilic faces, in turn, are of paramount importance for a correct construction of peptide α- and 310 -helix coiled coils. Finally, it is particularly worth mentioning that a fully developed, stable 310 -helix in solution requires only about eight Cα -tetrasubstituted α-amino acid residues, but this ﬁgure is considerably higher (≈20 amino acids) for an α-helix based on protein amino acids under the same experimental conditions. In 1996, the CD spectrum of Ac-[(S )-(αMe)Val]8 -OtBu [(αMe)Val, C α methylvaline; OtBu, tert-butoxy] was monitored in TFE solution and proposed as the standard pattern exhibited by a right-handed 310 -helix [60, 61] (Figure 15.6). This proposal was substantiated by independent X-ray diffraction and 2D-NMR analyses. In particular, this spectrum is characterized by a negative Cotton effect at 207 nm accompanied by a negative shoulder centered near 222 nm. The ellipticity ratio R = [θ ]222 /[θ ]207 is 0.4. It was gratifying to note that these experimental ﬁndings were

2

[θ]R x 10–3 (deg x cm2 x dmol–1)

0

–2

–4

–6

–8

Figure 15.6. CD spectrum of the 310 -helical –10 180

200

220 Wavelength (nm)

240

260

homo-octapeptide Ac–[(S)-(αMe)Val]8 –OtBu in TFE solution. (Redrawn from references 60 and 61.)

508

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

40

[θ] x 10–3 (deg x cm2 x dmol–1)

20

0

–20

–40

Figure 15.7. Calculated CD spectrum for a 160

180

200

220

Wavelength (nm)

240

260

48-residue 310 -helix peptide with ϕ = −60◦ ,

ψ = −30◦ . (Redrawn from reference 26.)

in good agreement with the theoretical curves published a few years before by Manning and Woody [26] (Figure 15.7). Additional CD calculations of 310 -helix were reported by Bode and Applequist [62], but their method does not treat the n → π ∗ transition. Moreover, the ellipticity at about 195 nm is positive, albeit only slightly. It is relevant to stress the point that the overall shape of this spectrum closely resembles that of an unordered peptide. However, the two spectra can be readily distinguished from the position of the π → π ∗ band (well above 200 nm for the 310 -helix). Whether the α- and 310 -helices can be distinguished by CD has been a long-standing question [11]. In this connection, it should be noted that although the R value can be useful to discriminate between a 310 -helix and an α-helix (R ≈ 1) when used in combination with data from other physicochemical techniques, an uncritical interpretation based solely on this CD parameter as a diagnostic can be risky. Indeed, the R value can be affected by the chemical structure (e.g., C α -methylation) of the peptide building blocks, by environmental (solvent, temperature) changes, and 3D-structural factors, such as incomplete α-helix formation, coexistence of diastereoisomeric (right- and left-handed) helices, and intramolecular C=O· · ·H–N H-bond lengths (including those of bifurcated H-bonds) [49, 63–66]. Finally, it was emphasized that the 222 nm n → π ∗ CD band may deserve a theoretical reexamination [43].

509

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

40

[θ]T x 10–3 (deg x cm2 x dmol–1)

20

0 MeOH 0

–20

MeOH 1

–40

Figure 15.8. CD spectra of the

HFIP 0

homo-heptapeptide alkylamide Ac-[(S)-(αMe)Val]7 -NHiPr, where NHiPr is

HFIP 1

isopropylamino, in methanol (MeOH) and HFIP solutions. Repeated cycles of 310 -helix/α-helix conversion were performed, the order of

HFIP 2 –60 200

220 Wavelength (nm)

240

solvent switches being HFIP 0, MeOH 0, HFIP 1, MeOH 1, HFIP 2. (Redrawn from reference 67.)

The difference in length between the more elongated peptide 310 -helix and the ˚ more compact α-helix is about 0.4 A/residue. This property makes the 310 -/α-helix reversible conversion very promising as a molecular switching tool between the N- and C-terminal functions of a peptide backbone. Using homo-peptides of various main-chain length (for the N α -acetyl, isopropyl amide homo-heptamer, see Figure 15.8), all based on the strongly helicogenic, C α -tetrasubstituted α-amino acid (S )-(αMe)Val, two of us (F.F. and C.T.) have shown that a well-deﬁned, solvent-controlled, reversible 310 -/αhelix transition takes place even in a homo-oligomer as short as a terminally blocked hexapeptide [67–69]. Homo-peptide sequences blocked as a urethane or an acetamide at the N-terminus and as a methyl ester or an N -alkylamide at the C-terminus are all appropriate. Analogous molecular spring properties of related Aib/(αMe)Val peptides were also assessed by CD [70–73]. The next step of the research in this area was the ﬁrst CD characterization of a water-soluble 310 -helical peptide [74–78]. To this goal, two terminally blocked heptapeptides were prepared, each containing: (1) two residues of ATANP, 2-amino-3[1(1,4,7-triazacyclononane)]-(S )-propanoic acid (Figure 15.5), a chiral Cα -trisubstituted α-amino acid with excellent water-solubilizing properties; and (2) ﬁve residues of

510

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Cα -tetrasubstituted α-amino acids, either the achiral Aib or the chiral (S )-Iva (Figure 15.5) and (S )-(αMe)Nva [(αMe)Nva, C α -methyl norvaline; Figure 15.5]. The CD patterns of the Aib- and Iva-containing peptides 1 and 2, respectively, in water are similar (Figure 15.9) and are characterized by a negative band at 202 nm and a negative shoulder at ∼222 nm. In both cases the R ratio is 0.20. Not surprisingly, the negative ellipticity values of peptide 2 almost double those of peptide 1. We believe that this observation is strictly related to the presence in peptide 2 of ﬁve chiral (S )-Iva residues as opposed to ﬁve achiral Aib residues in peptide 1. More speciﬁcally, it is our contention that this experimental ﬁnding is mainly associated with the concomitant occurrence of a non-negligible amount of left-handed 310 -helix in the conformational equilibrium mixtures of peptide 1, characterized by less than 30% of chiral (S ) residues. Indeed, there is no evidence in the literature for a different ability of the structurally related Aib and Iva residues to support a helical conformation. If the CD spectrum of peptide 2 in water is compared with that of the (αMe)Val homo-octapeptide in TFE solution, then: (i) The shapes of the curves are very close, despite an ∼5-nm blue shift for peptide 2 in water. (ii) Although both R values would be in the range expected for a 310 -helix, that of peptide 2 is appreciably reduced as a consequence of a concomitant higher intensity of the negative π → π ∗ band and a lower intensity of the negative n → π ∗ band. Thus, on the basis of this chirospectroscopic analysis, we are conﬁdent that both peptides 1 and 2 would be folded in the 310 -helical structure in water. The CD curves of the two peptides are only marginally modiﬁed by variations of pH, temperature, or concentration. Finally, it is worth mentioning that recently Basu and co-workers [66] have discussed the CD pattern of a partially 310 -helical, water-soluble, terminally blocked Aib–l-Ala–l-Lys 12-mer peptide. In a sequential peptide the alternation of an l-Pro residue, which disrupts the conventional H-bonding schemes observed in helices (lacking the N–H donor group), and a helix-forming residue such as Aib may give rise to a novel helical structure, called the β-bend ribbon spiral [79, 80]. This structure may be considered a variant of the 310 -helix, having approximately the same helical fold of the peptide chain and being stabilized by intramolecular C=O· · ·H–N H-bonds of the β-turn type. The complete characterization of this peptide conformation, which may be of relevance in the development of models for peptaibiotics and for the numerous (l-Pro–X)n (with X = l–Pro) segments found in globular and ﬁbrous proteins, was achieved by X-ray diffraction analyses of terminally blocked (l-Pro–Aib)n sequential peptides (Figure 15.10). The repeating -l-Pro–Aib- dipeptide units show on the average the sequence of backbone torsion angles (φi , ψi and φi +1 , ψi +1 ) − 78◦ , −10◦ and −54◦ , −40◦ . The mean heli˚ and p = 6.29 A. ˚ Typically, the value for the cal parameters are n = 3.43, d = 1.94 A, -l-Pro–Aib- ω torsion angles deviate signiﬁcantly (| ω| > 10◦ ) from the ideal trans 180◦ value. The required energy for this structural change is partially regained by the formation of acceptable intramolecular C=O· · ·H–N H-bonds. By investigating the terminally blocked -(l-Pro–Aib)4 - octapeptide, we were able to determine the diagnostic CD signature for this unusual ordered peptide conformation [81] (Figure 15.11), which adds physical evidence for the similarity of the β-bend ribbon structure in solution to that of the 310 -helix. The small differences observed between the spectra of these two conformations may be in part attributed to the high percentage (50%) of tertiary amide (Xxx–l-Pro) chromophores present in the former. Also, according to the Xray diffraction analyses, some of the l-Pro–Aib amide bonds in these peptides are signiﬁcantly nonplanar. This phenomenon can lead to reasonably marked CD consequences, at least in the amide n → π ∗ region, that are independent of inter-residue coupling.

511

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

0

[θ]R x 10–3 (deg x cm2 x dmol–1)

1

–3

–6

2

–9

Figure 15.9. CD spectra of the 310 -helical peptides Ac–Aib–(S)-ATANP–(Aib)2 –(S)-

200

220 Wavelength (nm)

240

ATANP-(Aib)2 -OMe (1), where OMe is methoxy, and Ac-(S)-Iva-(S)-ATANP-[(S)-Iva]2 -(S)-ATANP[(S)-Iva]2 -OMe (2) in water. (Redrawn from reference 75.)

Figure 15.10. X-Ray diffraction structure of the β-bend ribbon spiral forming heptapeptide pBrBz-Aib–(L-Pro–Aib)3 –OMe, where pBrBz is para-bromobenzoyl. The three intramolecular Hbonds of the β-turn (C10 ) type are indicated by dashed lines, O atoms are larger and gray, N and H atoms are white. (Redrawn from reference 80.)

512

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

30

[θ]T x 10–3 (deg x cm2 x dmol–1)

0

–30

–60

–90

Figure 15.11. CD spectrum of the β-bend –120 200

220 Wavelength (nm)

240

260

ribbon spiral forming, sequential hexapeptide Ac–(L-Pro-Aib)3 –OMe in TFE solution. (Redrawn from reference 81.)

15.4. β-SHEETS The pleated-sheet β-structure is the second most common ordered secondary structure in peptides and proteins. Both types (parallel- or antiparallel-chain disposition) of βstructures were ﬁrst proposed by Pauling and Corey in 1951 [82]. The antiparallel-chain β-structure is more widespread than its parallel-chain counterpart. In the former, the directionality of the interstrand H-bonds is optimal, but this can be counterbalanced by side-chain–side-chain interactions and chain topology considerations. In the parallelchain β-structure, the side-chain Cβ atoms of the residues in register and in adjacent ˚ (Figure 15.12). In contrast, strands repeat themselves regularly at a distance of 4.5 A in the antiparallel-chain β-structure these separations strictly alternate between a longer ˚ and a shorter distance (3.5 A). ˚ This short distance in the antiparalleldistance (5.7 A) chain β-structure may be responsible for the statistical observation [83] that sterically hindered amino acids (e.g., the β-branched Val and Ile residues) cannot be easily accommodated and therefore strongly prefer the parallel-chain β-structure. The backbone φ, ψ torsion angles for the two types of pleated-sheet β-structures are partially extended and quite close to each other (φ ≈ −140◦ ; ψ ≈ 135◦ ). The differences of their φ, ψ torsion angles (by 30–45◦ only) from those of the fully extended structure (2.05 -helix;

513

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

NNN

NCN

4.5 Å

4.5 Å

5.7 Å

7.0 Å

3.5 Å

6.9 Å

CCC CN C (a)

(b)

Figure 15.12. Representations of the parallel (a) and antiparallel (b) chain dispositions of the β-sheet structure. Relevant inter- and intrastrand distances are reported.

see Chapter 16, this volume) explain their slightly wavy appearance. Particularly effective β-structure-forming residues are those with sterically demanding, β-branched side chains (Val, Ile, and Thr) and those which can form side-chain to main-chain intra-strand H-bonds (Ser, Thr, Cys) [84]. While Pro is a β-structure breaker, one of the preferred structures for Gly is the antiparallel-chain β-structure. The pleated-sheet β-structure can be either of the intra- or the intermolecular type. The antiparallel-chain, intramolecular type is also termed β-hairpin (two strands and one β-turn). If the strands and turns are multiple, then the resulting structure is termed “cross-β” or “β-meander.” The β-sheet structures may deviate substantially [85] from the nearly planar structures originally proposed [82]. Indeed, both antiparallel- and parallel-chain sheets display an offset of one strand with respect to its neighbor. The overall twist for sheets comprised of l-amino acids is typically left-handed. The β-structure, due to its poor solubility, is responsible for a variety of neurodegenerative “conformational diseases,” such as those characterized by ﬁbril formation and amyloid deposits [86]. This last property of the β-sheet conformation has also made its CD characterization rather difﬁcult. CD spectra for three polypeptides in the β-sheet conformation are shown in Figure 15.13 [87–89]. The β-sheet CD spectrum is characterized by a negative n → π ∗ band near 217 nm and a pair of exciton-split π → π ∗ bands, positive near 195 nm and negative near 175 nm (the latter not shown in Figure 15.13). In contrast to the α-helix, β-sheet CD spectra show much greater variability. This is understandable, in view of the variability of β-sheets, which can be parallel or antiparallel and vary in length and number of strands, as well as in the extent of twist [85]. Is it possible to distinguish parallel and antiparallel β-sheets by CD? It was proposed [90, 91] in the late 1970s that the two types of β-sheet could be distinguished by their crossover wavelength, λco , that is, the wavelength at which the CD passes through zero between the 195-nm positive band and the 175-nm negative band. A λco ∼ 178 nm was reported for ﬁlms of Boc-(l-Ala)7 -OMe, an antiparallel β-sheet according to IR absorption [92, 93]. For ﬁlms of Boc-(l-Val)7 -OMe, a parallel β-sheet by IR absorption criteria, λco ∼ 192 nm. Films of Boc–(Xxx)7 –OMe {Xxx = l-Leu, l-Cha (β-cyclohexylalanine), l-Nva (norvaline), l-Nle (norleucine)} exhibited intermediate λco values, suggesting that

514

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

40

[θ] x 10–3 (deg x cm2 x dmol–1)

c

a 20

b

0 b

c

Figure 15.13. CD spectra of polypeptide models forming a β-sheet conformation: (a) poly(L-Lys), pH 11.1, 15 min at 52◦ C followed by cooling to 25◦ C; (b) poly(L-Lys) in 1% SDS

a –20 200

220 Wavelength (nm)

240

mixture; (c) copoly(L-Lys– L-Leu) in 0.1 M NaF, pH 7. (Redrawn from reference 45.)

peptides of these amino acids formed mixed β-sheets. However, conformational energy calculations [94, 95] predicted that β-sheets with small, linear side chains, (e.g., l-Ala), are only slightly twisted, whereas sheets with side chains branched at Cβ (e.g., l-Val), have a strong twist. Calculations of the CD for slightly twisted and strongly twisted β-sheets showed that the extent of twist affects the CD spectrum more than the parallel versus antiparallel character. Thus, the λco parameter depends on the degree of twist as well as the relative chain orientation and probably cannot be used as a reliable marker of the latter. In fact, theory indicates that the CD of parallel and antiparallel β-sheets are similar. The protein pectate lyase C (pelC) has more than 30% parallel β-sheet, no antiparallel sheet, and very little α-helix, and the sheet is only slightly twisted [96]. The CD spectrum of pelC [97] strongly resembles that of poly(l-Lys) in the β-form (Figure 15.13) in peak positions and relative amplitudes. Poly(l-Lys) β-sheet is antiparallel by IR absorption criteria and, because of the linear side chains, is expected to have a small twist. These observations support the conclusion that CD differences between parallel and antiparallel β-sheets with comparable twists are minor. There is much interest in peptides that model β-sheets. Intramolecular β-sheets are of greatest interest because β-sheet-to-coil transitions [98] in such peptides can provide thermodynamic parameters describing the effects of amino acid composition and sequence

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

on stability, analogous to the numerous studies of the α-helix. Until recently, β-sheet models have been intermolecular [90, 91, 99, 100], but two-stranded (β-hairpins) and three-stranded β-sheets are now being studied [101]. Evidence for β-hairpins in several model peptides has come from NMR, but the CD spectra show no clear proof for β-structure [102–104]. Contributions of aromatic side chains and β-turns may obscure the β-hairpin signature. Alternatively, NOEs can detect contacts characteristic of the β-hairpin that are present in one part of the molecule while the rest of the molecule is not in the hairpin conformation. By contrast, CD detection requires that all or nearly all of the molecule be in the hairpin. Detection of such “nascent” 3D-structures by NMR and not by CD has been documented for α-helices [105]. Peptides expected to form β-hairpins that show distinct negative bands in the 215 to 220-nm region have been reported [102, 106, 107]. For such peptides, the content of hairpins has been estimated by both NMR and CD, with reasonable agreement. A model hexadecapeptide was reported to be 55% β-hairpin (CD), 41–47% (NOE intensities), and 47% (chemical shifts) [107]. A peptide designed [108] to form a three-stranded β-sheet (betanova) has been investigated intensively. NMR supported the presence of the β-sheet, and it was claimed that CD also corroborated the β-hairpin structure, although no spectra were published. CD at 217 nm and ﬂuorescence indicated a cooperative thermal transition, which was interpreted as a two-state transition. The UV resonance Raman spectrum of betanova was consistent with β-structure but showed no indication of a cooperative thermal transition [109]. Boyden and Asher [109] presented the CD spectrum, which was consistent with an unordered conformation, rather than β-structure, even at 0◦ C. This ﬁnding was rationalized [108, 109] as resulting from aromatic side-chain contributions (betanova has one Tyr and one Trp out of 20 residues), but this explanation seems inconsistent with resonance Raman evidence for a molten globule [110] in which the side chains are not in a well-deﬁned conformation. The Serrano group [111] reexamined betanova and, using NMR chemical shifts rather than NOEs, revised their estimates of β-sheet content sharply downward, from ∼80% [108] to ∼10%. Keiderling and co-workers [112] reported a detailed study of betanova. Using CD, IR absorption, and F¨orster resonance energy transfer (FRET), they showed that betanova is predominantly unordered, even at 5◦ C. Both CD and IR absorption gave an estimated β-sheet content of 20–26%. FRET between a donor and acceptor pair at opposite ends ˚ at 5◦ C, which barely increased to 46 of the peptide gave an end-to-end distance of 45 A ◦ ˚ A at 80 C. By contrast, the NMR structure reported by Kortemme et al. [108] gave this ˚ Kuznetsov et al. [112] also studied another three-stranded β-sheet condistance as 21 A. D D struct, P– P, designed by Schenck and Gellman [113] to include two type-II β-turns, stabilized by -d-Pro–Gly- sequences. The original publication showed that the CD spectrum, NOEs, and chemical shifts are all consistent with the targeted structure. Kuznetsov et al. [112] found 42–59% β-sheet by CD and IR absorption. FRET measurements gave ˚ in good agreement with the 30–35 A ˚ estimated from an end-to-end distance of 31 A, structural models of the three-stranded β-sheet. In contrast to the α-helix, there is no generally accepted method to calculate the βsheet content from CD spectra. One suggestion [114] uses the difference [θ ]195 − [θ ]217 taking a value of 50–55 × 103 deg cm2 dmol−1 , derived from the data in Figure 15.13, to represent 100%. As long as the sheet is not strongly twisted, this is a reasonable method, but for Boc–(l-Val)7 –OMe in TFE this difference is ∼104 × 103 deg cm2 dmol−1 [115]. Solubility is a major problem with peptides that form β-sheet. PEGs were attached to the N-terminus of peptides to improve solubility [30,116–119]. Several series of

515

516

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

PEG-based homo-oligopeptides with hydrophobic side chains were examined by CD. The prevailing conformation is the β-sheet structure. However, an interesting ﬁnding is that an α-helix to β-sheet transition is observed in the longest (l-Met)n oligomers upon deblocking of the N-terminal protection under acidic conditions [30]. Also, the β-sheet structure of these oligopeptides can be disrupted by incorporation of a guest l-Pro residue [119]. Remarkably, these results, and the analogous ones for oligopeptides not linked to PEG, using Aib, l-Pro, or a d-residue [120], anticipated those recently reported for the important ﬁeld of amyloid-forming peptides [121–125]. Decapeptides with the sequences H–(l-Val)5 –(l-Ala)2 –(l-Val)3 –NH–PEGM and H–(l-Ala)5 –(l-Val)3 –(l-Ala)2 –NH–PEGM, where PEGM is polyethylene glycol monomethyl ether (Mr ≈ 5 × 103 ) were studied by CD in water and TFE [126]. In both solvents, H–(l-Ala)5 –(l-Val)3 –(l-Ala)2 –NH–PEGM gave CD spectra with a characteristic β-sheet pattern (negative maxima at 216 nm in water and 215 nm in TFE; positive maximum at ∼190 nm in water). By contrast, H–(l-Val)5 –(l-Ala)2 –(lVal)3 –NH–PEGM exhibited a CD spectrum consistent with an unordered conformation in water and a mixture of unordered and α-helix conformations in TFE. The critical main-chain length for forming a β-sheet was extensively investigated. The ﬁrst monodisperse homo-oligopeptide series examined was (l-Ile)n [99]. In TFE, using CD the critical main-chain length for full development of the β-sheet was found at the heptamer level (Figure 15.14). However, the secondary structure of this oligomer can be disrupted by dilution or addition of a more polar solvent. Subsequently, a large variety of peptide series based on hydrophobic α-amino acids was studied in organic solvents and found to adopt the β-sheet structure [127–129]. Interestingly, the CD spectrum of the β-sheet structure of a terminally-protected, all-d-(Ala)7 -homo-oligomer was found to be the mirror image of that recorded for its enantiomeric peptide [130]. Dilution, heating, and co-solvent addition experiments clearly indicated that the rank order of stability for the homo-heptapeptides examined is l-Ile > l-Val > l-Cys(Me) (S -methyl cysteine) > l-Ala > l-Nva > l-Nle > l-Met [129]. To the same end, peptides of the type H–l-Val–Xxx–(l-Val)3 –NH–PEGM, where Xxx = l-Val, l-Leu, l-Ile [126], and (Xxx–Yyy)n –Zzz–NH–PEGM [131], where Xxx is a hydrophilic amino acid, Yyy is a hydrophobic amino acid, or vice versa, and Zzz is Gly or l-Ser, were prepared and studied. At or above the critical main-chain length, the CD spectra are characteristic of β-sheets. The critical main-chain lengths ranged from 6 to more than 11 residues, depending on the sequence and the solvent. Generally, the bulkier the side chain, the shorter the critical main-chain length. Interestingly, IR absorption criteria indicate that all of the peptides with the host–guest block sequence adopt an antiparallel β-sheet conformation, in contrast to the demonstration [91, 92] that l-Val and l-Ile homo-oligomers assume a parallel β-sheet conformation beyond the critical main-chain length. The preference [126] for antiparallel sheets in the PEGMprotected peptides was attributed to avoidance of steric clashes that would arise between the protecting groups in a parallel β-sheet.

15.5. β-TURNS AND γ-TURNS In 1968 Venkatachalam [132] and Geddes et al. [133], independently, proposed three types (trans I–III) of the 1 ← 4 intramolecularly H-bonded, folded peptide conformation (also called C10 form or β-turn), where there is an H-bond between the C=O group of residue 1 and the N–H group of residue 4 (Figure 15.15). The type-I β-turn has

517

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

40 7 8

30

[θ] x 10–3 (deg x cm2 x dmol–1)

20

10

2 3 4

0

5 6

–10

7 8

–20

Figure 15.14. CD spectra of the Boc–(L-Ile)n –OMe (n = 2–8) homo-oligopeptides in TFE solution (peptide

–30 200

220 Wavelength (nm)

240

concentration: ∼2 mg/ml). (Redrawn from reference 99.)

Figure 15.15. Four types of the 1 ← 4 intramolecularly H-bonded (β-turn) peptide conformation. Intramolecular C=O· · ·H–N H-bonds are represented by dashed lines, O atoms are larger and gray, and N and H atoms are white.

518

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

φ2 = −60◦ , ψ2 = −30◦ , and φ3 = −90◦ , ψ3 = 0◦ , while the type-II β-turn has φ2 = −60◦ , ψ2 = 120◦ and φ3 = 80◦ , ψ3 = 0◦ . The two nonhelical turns are interrelated by a rotation of 180◦ at the second peptide moiety. The type-III β-turn has φ2,3 = −60◦ and ψ2,3 = −30◦ , thus forming one turn of the 310 -helix. Obviously, turns of type-I’, II’, and III’ also exist, where the prime superscript indicates that the given turn is the mirror image of the corresponding unprimed one, except of course for the positions of the Cβ and other atoms of the side chains. For the H-bond to remain intact, only certain sidechain groups may be accommodated at Cα (2) and Cα (3). All three types of the trans C10 folded conformation were found in crystals of synthetic and naturally occurring, linear and cyclic oligopeptides according to X-ray diffraction analyses. In globular proteins, the average frequency of trans β-turns is as high as 30%. The most frequently occurring residues in β-turns are l-Pro and l-Ser in the second position, and l-Asn, l-Asp, and Gly in the third position. β-Turns in proteins are often stabilized by antiparallel β-sheets (forming β-hairpins). All three types of β-turns were authenticated in solution as well. An additional type of β-turn (VIa) is that having the central amide group in the cis conformation (ω = 0◦ ) [134] (Figure 15.5). Its characteristic backbone torsion angles are φ2 = −60◦ , ψ2 = 120◦ , and φ3 = −90◦ , ψ3 = 0◦ . Model building studies indicate that, because of the dimension of the pseudo-ring generated by the 1 ← 4 intramolecular Hbond, only the cis peptide structure with both C=O and N–H bonds of the central amide group pointing outwards exists. The occurrence of this structure, although relatively rare, was demonstrated in the crystal state and in solution as well. γ -Turns (also called C7 forms or 1 ← 3 intramolecularly H-bonded, folded peptide conformations) (Figure 15.16) are far less abundant in peptides and proteins than β-turns. In γ -turns, a large change in direction occurs over two amides and a single Cα atom [135, 136]. γ -Turns are pseudo-ring structures that are stabilized by an intramolecular H-bond between the H(3) and O(1) atoms. The trans amide groups lie in two planes, which make an angle of about 115◦ . When R in-NH-CHR-CO- is not an H atom, two different conformers (equatorial and axial) can exist, which are represented on the usual Ramachandran φ, ψ map by two centrosymmetric points, the coordinates of which [for an (l)-residue] are φ = −75◦ , ψ = 65◦ (for the equatorial or inverse γ -turn form) and φ = +75◦ , ψ = −65◦ (for the axial or classical γ -turn form). While the H-bond is strongly bent, it has a normal H· · ·O distance and it still makes a sizable contribution to the stabilization of the folded structure. Actually, some variations of the ω value (| ω| ≈ 10◦ ) are required for the stabilization of the 1 ← 3 intramolecularly H-bonded peptide conformations. However, the small energy of torsion rotation is more than compensated for by the energy of the H-bond [137]. X-ray diffraction analyses and several spectroscopic techniques unambiguously demonstrated the existence of these two types

Figure 15.16. The axial (γ -turn) and equatorial (inverse γ -turn) 1 ← 3 intramolecularly H-bonded peptide conformations. Intramolecular C=O· · ·H–N H-bonds are represented by dashed lines, O atoms are larger and gray, and N and H atoms are white.

519

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

of γ -turn conformations. All types of β- and γ -turns were discussed in detail in several review articles [138–142]. Unlike the α-helix and β-sheet, there is no unique CD signature for β-turns because of the range of conformers included in this structural category. The CD was predicted [143] for each of 15 β-turn conformations and variants thereof [132, 134]. Various spectral patterns were predicted, but the predominant one had a weak negative n → π ∗ band at 220–230 nm and had strong π → π ∗ bands, positive at 200–210 nm and negative at 180–190 nm (Figure 15.17). This pattern, called a class B spectrum, resembles that of a β-sheet, but is red-shifted by ∼10 nm. Certain types of β-turns have a relatively high probability of giving other spectral types. For example, some type-II -turns have an α-helix-like (class C) spectrum. Such turns are favored by a heterochiral -d–l- sequence, such as the -d-Phe–l-Pro- sequence in gramicidin S. The prediction that type-II β-turns have a class B spectrum has been veriﬁed by CD studies of cyclic peptides with these types of turns [87, 144–146]. However, a cyclic peptide with a type-I turn was observed [146] to have an α-helix-like CD spectrum (class C), rather than the expected [143] class B spectrum. Cyclic peptides with the -d-Xxx–l-Pro- sequence have been found to have a class C spectrum [145–149], in accordance with predictions. Similarly, an -l-Pro–d-Xxx- sequence requires a type-II variant that is predicted to have a relatively high probability of exhibiting a class C

spectrum (a left-handed α-helix spectrum), and such a spectrum is observed for peptides with the -l-Pro-d-Ala- sequence [150].

30

[θ] x 10–3 (deg x cm2 x dmol–1)

20

10

0

–10

–20

–30

Figure 15.17. Calculated CD spectrum for 180

200

220

Wavelength (nm)

240

Venkatachalam’s type-II β-turn (class B spectrum). (Redrawn from reference 143.)

520

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The most deﬁnitive study of β-turn CD is that of Bandekar et al. [151], who studied cyclic tripeptides of the type cyclo(l-Ala–Xxx–Aha), where Aha is 6-aminohexanoic acid. With Xxx = l-Ala, a type-I turn is obtained, and the CD spectrum is α-helix-like, in agreement with Gierasch et al. [146]. When Xxx = d-Ala, a type-II turn is required and the CD spectrum is of class B. Thus, of the two most common types of turns, type-I is expected to give a class C spectrum, whereas type-II will show a class B spectrum. The result for the type-II turn agrees with the predictions of Woody [143], but that for the type-I turn does not. In addition, calculations for the low-energy conformers [151] gave results in conﬂict with experiment. However, Sathyanarayanan and Applequist [152] applied the classical dipole interaction model and obtained qualitative agreement with experiment for these two β-turn models (Figure 15.18). Studies utilizing CD, NMR, and molecular dynamics (MD) simulations have examined 14 cyclic and linear peptides [153–156]. The fractions of type-I and type-II β-turns were determined from measured NOEs and MD simulations. Convex constraint analysis [153] was used to obtain a set of basis CD pectra for type-I, type-II, and type-II β-turns [155, 156], which are shown in Figure 15.19. The CD analysis gave good agreement with the NMR and MD results. Spectra 1 and 4 are both attributed to type-I turns. They have the same sign pattern and are of class C, in accord with expectations [151], but spectrum

[θ] x 10–3 (deg x cm2 x dmol–1)

III

I II

Figure 15.18. Calculated CD spectra for Ac–(L-Ala)2 –NHMe (-NHMe, methylamino) in type-I, II, and III β-turn conformations. It should be noted that, as the atom dipole interaction model of reference 152 does not include the

160

180

200 Wavelength (nm)

220

240

n → π ∗ transition, the 220-nm region is not reproduced reliably. (Redrawn from reference 152.)

521

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

[θ] x 10−3 (deg x cm2 x dmol−1)

40

1

2

20

0 4 –20

–40

Figure 15.19. The four component CD curves for

3 180

200 220 Wavelength (nm)

240

convex constraint analysis of CD spectra from cyclic β-turn model peptides. (Redrawn from reference 155.)

4 differs strongly in relative amplitudes from that of an α-helix. In fact, it resembles the spectrum of a 310 -helix [60, 61, 75]. The authors suggested that spectrum 1 is that of a “strained” type-I turn and spectrum 4 is that of an unstrained turn. However, the type-I spectrum is like those seen with cyclic hexapeptides [146] and tripeptides [151], and at least the hexapeptides do not appear to be strained. It seems more likely that spectrum 1 is that of a canonical type-I turn and spectrum 4 is that of a type-III turn—that is, one turn of a 310 -helix. Spectrum 2 in Figure 15.19 is related to type-II β-turns, is of class C , and differs from the class B spectrum expected for type-II turn in that the n → π ∗ band is weakly positive, rather than negative. Only two peptides in the sample showed predominantly type-II turns, and these have the turns at the -l-Pro-d-Ser- and -l-Val-d-Ser- sequences. The former is analogous to the peptides studied by Ananthanarayanan and Shyamasundar [150], which has a class C spectrum, and the steric bulk of the Val may favor a similar type-II variant for the latter peptide. Thus, these peptides may be type-II by NMR criteria, but differ enough from the canonical type-II turns to exhibit different CD spectra. Component 3 is a class C spectrum and correlates with type-II turns. It is therefore consistent with expectations. Extensive CD studies were performed on selected synthetic sequences of tropoelastin [157–159]. This ﬁbrous protein is known to fold into a spiral structure heavily based on -l-Pro–Gly- type-II β-turns [160]. Not unexpectedly, the spectra resemble those of the type-II turn. The CD curve of a peptide as short as a terminally blocked dipeptide amide, Ac–[d-(αMe)Pro]2 –NHi Pr [(αMe)Pro, C α -methyl proline; NHi Pr, isopropylamide], in acetonitrile solution was reported and proposed as the reference spectrum for a type-III’ β-turn [161] (Figure 15.20). This dipeptide is unique in that it is exclusively based on the conformationally very restricted (αMe)Pro residue. As a result, it is rigidly folded in this structure in the crystal state and 100% folded in this same conformation in CDCl3 , according to an X-ray diffraction, NMR, and FT-IR absorption investigation. The CD properties of γ -turns were not well characterized. In the original deﬁnition of the γ -turn [135], the ϕ, ψ angles at the Cα -atom located at the central residue of the turn are (68◦ , −61◦ ), which are near those of the C7 ax conformation—that is, in a region that is unfavorable for an l-amino acid residue. The inverse γ -turn, with

522

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

[θ]T x 10–3 (deg x cm2 x dmol–1)

30

20

10

0

Figure 15.20. CD spectrum of the terminally blocked, type-III β-turn forming,

200

220 Wavelength (nm)

240

Ac[D-(αMe)Pro]2 –NHiPr dipeptide amide in acetonitrile solution. (Redrawn from reference 161.)

the corresponding torsion angles of (−68◦ , 61◦ ), is near the C7 eq conformation and is accessible to all l-amino acid residues. Therefore, the inverse γ -turn is the only form that is likely to be encountered in most peptides, except for a Gly or a d-amino acid central residue. It was reported [162] that the inverse γ -turn is a dominant conformer in Ac-l-Pro-NHMe in carbon tetrachloride and is populated to a signiﬁcant extent in chloroform. CD measurements, performed in the latter solvent down to 220 nm, revealed a strong, negative band in the n → π ∗ region (Figure 15.21). Thus, an inverse γ -turn is expected to give rise to a strong, negative CD band at 220–230 nm and a γ -turn should produce a large, positive n → π ∗ band. A similar CD curve was calculated for the pentapeptide Ac–(Gly)2 –l-Ala–(Gly)2 –NH2 with the l-Ala residue in the inverse γ -turn conformation [163] (Figure 15.22). For reviews covering in-depth the CD properties of β- and γ -turn peptides, see references 45, 141, 143, and 164–166.

15.6. POLY(L-PRO) HELICES AND COLLAGEN TRIPLE HELIX Amide bonds are usually found in the trans conformation (ω = 180◦ ) in linear peptides, whereas cis amide bonds (ω = 0◦ ) are observed in constrained situations such as those

523

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

0

[θ] x 10–3 (deg x cm2 x dmol–1)

–10

–20

–30

–40

–50

Figure 15.21. CD spectrum of the inverse 200

220

240

Wavelength (nm)

260

γ -turn forming Ac– L-Pro–NHMe in chloroform solution. (Redrawn from reference 162.)

occurring in cyclic peptides, particularly if they are characterized by a small ring size. A trans amide bond in a secondary amide is energetically more stable than a cis amide bond by ∼2 kcal/mol, which explains the overwhelming occurrence in linear peptides of the trans bonding in the crystal state and its large preponderance in solution. However, the energy difference between the two conformations markedly decreases in tertiary amides. Thus, not surprisingly, cis peptide bonds reported in the literature for linear peptides in almost all cases involve a tertiary amide in an -Xxx–Yyy- sequence where Yyy is an l-Pro or an N -alkylated (Sar, l-MeAla, peptoid unit, etc.) residue. Pro–Pro bonds generally adopt the trans conformation, but, in some instances, particularly when the peptide is short or the sequence is syndiotactic (l–d or d–l), a cis conformation does occur [167]. Interestingly, the homopolymer poly(l-Pro) is dimorphic in that, under appropriate experimental conditions, the “all-trans” peptide bond conformation (type-II, PII ) may exhibit a transition (mutarotation) to the “all-cis” peptide bond conformation (type-I, PI ) [168–175] (Figure15.23). The PI and PII helices show quite similar ϕ, ψ values (semi -extended conformations). The ϕ, ψ values are about −65◦ , 155◦ . The classical PII conformation is a threefold helix, which, with appropriate side-chain replacements (e.g., an -OH function at position 4 of the ring), may be endowed with an amphiphilic character. The transition from PI to PII implies a remarkable increase in the long dimension of the 3D-structure (Figure 15.23). As for the ϕ, ψ torsion angles in l-Pro-based peptides, the former is almost invariant (∼ − 65◦ ) owing to the restrictions imposed by the ﬁve-membered pyrrolidine ring

524

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

4

[θ] x 10–3 (deg x cm2 x dmol–1)

0

–4

–8

Figure 15.22. Calculated CD spectra for 180

200

220

Wavelength (nm)

240

Ac-(Gly)2 -L-Ala-(Gly)2 -NH2 with L-Ala in the inverse γ -turn conformation. (Redrawn from reference 163.)

(Ni to Ci α cyclization). The ψ values accessible to l-Pro residues correspond either to the semi -extended region mentioned above (trans conformation) or to the 310 -/αhelical region (cis conformation). This latter 3D-structure, usually stabilized by 1 ← 4 (or 1 ← 5) intramolecular C=O· · ·H–N H-bonds, cannot be formed by l-Pro residues only. Pro residues, in fact, can exclusively occur at the ﬁrst two (three) positions of a 310 (α-) helix, respectively, because any following l-Pro residue would act as a helix breaker in that it lacks the H-bonding donor (N–H) fuctionality. The presence of a single l-Pro in the middle of a helix normally induces a “kink” in the structure. In the large majority of published examples, the semi -extended conformation is that adopted by l-Pro, indicating that this residue has an intrinsic propensity to be in the PI or PII structure. This ﬁnding is especially veriﬁed in the longest homo-oligo(l-Pro)n and is related to unfavorable steric interactions originating between the δ-carbon of an l-Proi residue and the β-carbon of an l-Proi −1 residue if both are folded in an α-/310 -helical conformation. As for the role of the Pro pyrrolidine ring, there is a weak correlation between the type of ring puckering and the backbone φ angle [176, 177], with rings in the up (χ1 < 0◦ ) conformation preferring less negative φ(∼ − 60◦ ) and in the down (χ1 > 0◦ ) conformation preferring more negative φ (∼ − 70◦ ) values. Moreover, the ﬂexibility of the pyrrolidine ring is expected to expand the range of φ, ψ values available to the l-Pro residue, leading to a minimization of unfavorable short-range interactions.

525

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

Figure 15.23. A -(Pro)10 - stretch in the PI poly(L-Pro) (a) and PII poly(L-Pro) (b) conformations. O atoms are larger and gray, N

(a)

(b)

atoms are white.

In summary, the available 3D-structural evidence strongly supports the view that short oligo(l-Pro)n spacers and templates cannot be uncritically viewed as “rigid rods,” as usually considered, but rather that the cis trans (ω torsion angle) and cis trans

(ψ torsion angle) equilibria, typical of the Pro residue, may severely hamper reliable conclusions from the experimental data [178–189]. Interestingly, calculations have shown that even long (l-Pro)n peptides may be quite ﬂexible with a deﬁned tendency to fold in a “worm-like” chain characterized by multiple bends [190].

526

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Collagen, the most widespread ﬁbrous protein, is known to be a triple-helical coiled coil in which each of the three strands has a left-handed PII poly(l-Pro) helical conformation. The three strands wrap parallel and in register around a common helical axis forming a right-handed superstructure and are held together by interchain H-bonds. The closepacked nature of the triple helix strictly requires a Gly residue in every third position. These conditions are achieved by the repetitive “consensus” triplet -(Gly-l-Pro-Xxx)n - in which Xxx is any amino acid that can adopt the semi -extended conformation {including an additional l-Pro or an l-Hyp [4(R)-hydroxy-(S )-proline] residue} [191–195]. Due to its elongated nature, no intrachain C=O· · ·H–N H-bonds are possible for this structure, even if secondary amides do occur in the sequence. A stable collagen-type triple helix requires at least ﬁve triplets. l-Pro residues can be replaced by acyclic, N -alkylated analogues with the l-conﬁguration [196]. The PI conformation is found only in the solid state and in solvents of low polarity—for example, higher alcohols (1-propanol). PII is the conformation authenticated in water and in halogenated alcohols (e.g., TFE) [197]. The CD spectra of both forms are shown in Figure 15.24. The PII spectrum is very nonconservative and characterized by a weak, positive n → π ∗ band at 226 nm and a strong, negative π → π ∗ band at 206 nm. An additional, much weaker negative band (not shown in Figure 15.24) is observed in the vicinity of 165 nm. 60

I

[θ] × 10–3 (deg × cm2 × dmol–1)

30

0

–30

II

–60

Figure 15.24. CD spectra of PI (I) and PII (II) 180

200

220 Wavelength (nm)

240

poly(L-Pro) in 1-propanol solution and in water, respectively. (Redrawn from reference 197.)

527

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

Since poly(l-Pro) contains tertiary amides, the transition energies for the n → π ∗ and π → π ∗ transitions are shifted to longer wavelengths relative to secondary amides. In general, both transitions are red-shifted by ∼10 nm. Also, the ground-state charge distribution is different for tertiary amides. The major difference is that secondary amides contain a highly polar N–H group, with nitrogen being negatively charged and hydrogen positively charged. In Xxx-Pro tertiary amides the hydrogen is replaced with a methylene group, which carries only a small positive total charge, mirrored by a decrease in the negative charge on the amide nitrogen. Nothing about the PII -helix requires the Pro ring; and, in fact, any of the natural amino acids can ﬁt into the PII -helix. This was recognized with the work of Tiffany and Krimm [198, 199], who pointed out the strong resemblance of the CD spectra of ionized poly(l-Glu) and poly(l-Lys) to that of poly(l-Pro). On this basis, they proposed that ionized poly(l-Glu) and poly(l-Lys), used at the time as models of the unordered conformation, must have signiﬁcant amounts of the PII conformation. This hypothesis is now supported by a large body of evidence, as discussed in Section 15.7. Short stretches of the PII -helix are found in some globular proteins [173], some of which are completely devoid of Pro. Quantum mechanical calculations have not succeeded in accounting for the CD spectrum of poly(l-Pro) in the PII conformation [200], although the classically based dipole interaction model gives reasonable results [201]. Calculations for poly(l-Ala) in the PII conformation reproduce the weak positive n → π ∗ band and the strong negative band near 200 nm, as shown in Figure 15.25. However, calculations using the PII poly(l-Pro) conformation from X-ray ﬁber diffraction [202] give poor agreement with experiment (Figure 15.25). It is not clear whether this result is due to differences between solution and solid-state structures (φ,ψ differences and conformational heterogeneity) or to the parameters used to treat the Pro side chains. In contrast to PII poly(l-Pro), PI poly(l-Pro) displays a more conservative CD spectrum [197] (Figure 15.24). It is unusual in that it has a strong, positive band at longer

[θ] × 10–3 (deg × cm2 × dmol–1)

20

10

0

–10

–20

–30 170

180

190

200 210 220 Wavelength (nm)

230

240

250

Figure 15.25. Calculated CD spectra [200] for PII poly(L-Ala)20 with (φ, ψ) = (−60◦ , +160◦ ) ( and the poly(L-Pro)20 (– – –) structure of Sasisekharan [202].

)

528

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

wavelengths (∼215 nm) with a slightly weaker, negative counterpart at shorter wavelengths (∼195 nm). The n → π ∗ transition is presumably responsible for the very weak, negative band near 235 nm. Calculated CD spectra [26] for chains of 10 residues in length agree well with experiments (Figure 15.26). The positions of the two strong features are accurately predicted. However, the calculations underestimate the intensity of the positive 215 nm band and they predict a weak, negative band for the n → π ∗ transition. As opposed to peptides forming α-helices and β-sheets, the CD spectra of (l-Pro)n homo-oligomers are found to exhibit the qualitative features of PII poly(l-Pro)n even for n = 3, although with substantially reduced amplitude and shifted wavelengths [203, 204]. The intrinsic helix-length dependence of the PII -helix CD is not qualitatively different from that of the α-helix but, for the latter, short helices are stable only if nucleated [37, 43]. According to experimental CD analyses, the terminal unprotected (l-Pro)n oligomers show a solvent-dependent conformational change, beginning at the trimer level [205] (Figure 15.27). The intensities of the Cotton effects increase with main-chain length and reach the characteristic values of the polymer when n ≈ 20. Rather surprisingly, the corresponding N-protected oligomers do not show mutarotation on change of the solvent (only the all-trans PII -helix is formed by these peptides). It is also interesting to note that oligo (l-Pro)n peptides, when dissolved in 1-propanol, adopt ﬁrst a PII -helix, which

[θ] x 10–3 (deg x cm2 x dmol–1)

40

20

0

–20

180

200

220 Wavelength (nm)

240

260

Figure 15.26. Calculated CD spectrum for PI poly(L-Pro)10 . (Redrawn from reference 26.)

529

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

60

37 0

40

[θ] x 10–3 (deg x cm2 x dmol–1)

[θ] x 10–3 (deg x cm2 x dmol–1)

18 12 8 20 5 0

2 6 12

–20

–10

2 3

–20

4 5

–30 15 –40

40 31

37

–50

–40 200

220 Wavelength (nm)

240

180

200 220 Wavelength (nm)

240

Figure 15.27. CD spectra of terminally unprotected homo-oligo (L-Pro)n (n = 2–40): (left) in a 9:1 1-propanol/water mixture (PI helix); (right) in aqueous solution (PII helix). (Redrawn from reference 205.)

slowly interconverts to the PI -helix typically found in this low-polarity solvent [206] (Figure 15.28). Many other valuable and stimulating review articles, as well as theoretical and experimental papers, were published on the CD properties of oligo and poly(l-Pro) helices (including those generated by the related ring-substituted and ring-contracted Pro analogues). A large selection [24, 26, 45, 165, 207–247] is provided in the list of references. Because each of the three strands of the collagen triple helix is based on the PII poly(l-Pro) conformation, it is not surprising that the experimental CD spectrum of this self-associated ﬁbrous protein would be qualitatively similar to that of its monomeric building block [199, 248]. In particular, the CD bands of collagen, though closely related in shape to those of PII poly(l-Pro), have larger intensity values. The experimentally observed overall blue shift of the CD spectrum of collagen relative to that of PII poly(lPro) is paralleled by calculations and depends on the ratio of secondary versus tertiary peptide bonds [24]. In natural collagen, the weak, positive band is seen at 220 nm and the negative band at 197 nm [248] (Figure 15.29). Partially denaturated collagen was found to give CD spectra with lower intensity, red-shifted crossover point, and higher 197 nm/220 nm ellipticity ratio. A ratio of 8.5 between these two bands was recommended as a sensitive measure of the purity of a natural collagen sample. Numerous CD studies of oligomeric and polymeric (Gly-l-Pro–Xxx) peptide triplets, capable of triple helix formation, were performed [196, 242, 249–253]. Figure 15.30 shows the CD spectrum of a polymeric peptide triplet, poly(Gly–l-Pro–l-Nva), in ethylene glycol solution. In monodisperse (Gly–l-Pro–Xxx) oligomers, the stability of the triple helix conformation is governed by peptide main-chain length and concentration, as well as by temperature, the nature of the Nα -blocking group, and the solvent. The central l-Pro of the triplet can

530

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

14 days

10

3 days

[θ] x 10–3 (deg x cm2 x dmol–1)

2 days 1 day

0

–10

2 min

–20

Figure 15.28. Time-dependent evolution of 200

220 Wavelength (nm)

240

the CD spectrum of (L-Pro)13 in 1-propanol. (Redrawn from reference 206.)

be replaced by an acyclic, N -alkylated residue. The ratio of 220 nm/197 nm ellipticities, the inverse of the ratio mentioned above, was found to be useful in establishing the onset of the triple helix in solution. The critical value of 0.15 is typically achieved at room temperature in water for oligopeptides as long as at least ﬁve triplets.

15.7. UNORDERED CONFORMATION Most small peptides exist in aqueous solution as ensembles of conformers, rather than having a well-deﬁned conformation. Such peptides are called unordered and have a characteristic CD spectrum [88, 199, 254, 255] with a strong negative band near 197 nm and a weak band approximately at 220 nm (Figure 15.31). The latter band may be a negative shoulder on the short-wavelength band. The distribution of the ensemble in (φ, ψ) space depends upon the peptide sequence, the solvent, and the temperature. The PII -helix (see Section 15.6) has been demonstrated to be an important component of the ensemble in many (perhaps most) peptides [208, 209, 211]. The importance of the PII conformation was ﬁrst recognized by Tiffany and Krimm [199, 255, 256], who pointed out the remarkable resemblance of the CD spectra of ionized poly(l-Glu) and poly(l-Lys) to the poly(l-Pro) spectrum (Figure 15.32) (the ∼10−nm difference in peak wavelengths results from secondary versus tertiary amides). On the basis of this resemblance, Tiffany and Krimm proposed that “unordered” poly(l-Glu) and poly(l-Lys) must

531

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

[θ] × 10–3 (deg × cm2 × dmol–1)

0

–20

–40

–60

–80

Figure 15.29. CD spectrum of calfskin 180

200

220

Wavelength (nm)

240

collagen (10 mM phosphate buffer, pH 3.5; 15◦ C). (Redrawn from reference 248.)

have a signiﬁcant amount of local order in the form of short stretches of PII helix, interspersed with less ordered conformations. Drake et al. [257] showed that the CD spectrum of charged poly(l-Lys) exhibits a sharp isodichroic point over a nearly 200◦ temperature range, demonstrating a two-state equilibrium. The spectrum of the low-temperature form is that of a PII helix, whereas that of the high-temperature form has a substantially diminished negative band near 197 nm and a weak negative shoulder near 220 nm. The high-temperature limit must be that of a truly unordered polypeptide with a broad ensemble of conformers from PII , α-helix and β-sheet regions of the Ramachandran map. The vibrational circular dichroism (VCD) spectra of charged poly(l-Glu) and poly(l-Lys) also closely resemble that of poly(l-Pro) [258, 259], providing further convincing proof for Tiffany and Krimm’s proposal. This evidence has been reviewed [208, 209, 211]. Recent studies of small oligopeptides suggest that the conformational distribution in such oligopeptides is much more homogeneous than previously thought. SchweitzerStenner and co-workers [260–265] have used CD in conjunction with IR absorption, polarized Raman, and VCD to analyze the conformation of di- and tripeptides and found sequence-dependent variations in the (φ, ψ) angles of the central residue, but a conformation in the PII region was detected to be prevalent in most tripeptides investigated. These studies also suggest a surprisingly narrow range of conformers [265]. The same conclusion has been achieved for longer l-Ala-rich peptides [266, 267]. For example, a

532

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

10

[θ] x 10–3 (deg x cm2 x dmol–1)

0

–10

–20

–30

–40

Figure 15.30. CD spectrum of 200

220 Wavelength (nm)

240

poly(Gly– L-Pro– L-Nva) [average P (triplet) ≈ 120] at 20◦ C in ethylene glycol. (Redrawn from reference 250.)

combination of NMR and molecular dynamics simulations indicate ∼80–85% PII conformation at each of the α-carbons of (l-Ala)n (n = 3 − 7) [267]. A PII -like conformation appears to prevail even at the level of N -acetyl amino acids [268]. These compounds have a very weak and positive n → π ∗ band near 220 nm and a strong negative π → π ∗ band between 195 and 200 nm (Figure 15.33). By contrast, amino acid amides have only a broad and weak positive CD in the 180 to 200-nm region, attributable to a single perturbed amide chromophore. An l-Ala-rich heptadecapeptide with l-Pro at the central residue has been studied by Park et al. [21]. The CD of this peptide is typical of unordered peptides and shows a positive band near 217 nm at low temperatures, indicative of substantial PII content. The temperature-dependence of the CD of this peptide in 8 M guanidinium chloride and the CD data of Drake et al. [257] for poly(l-Lys) were ﬁt to a two-state thermodynamic model with temperature-independent CD for each component and a temperature-independent

H. The ﬁt gave [θ ]222 = +9580 deg cm2 dmol−1 for the low-temperature form, PII , and [θ ]222 = −5560 deg cm2 dmol−1 for the high-temperature form, the truly unordered peptide. These data provide a measure of the PII content of this peptide, which can be applied, with due reservations, to others: fPII = ([θ ]222 + 5560)/15,140

533

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

[θ]T x 10–3 (deg x cm2 x dmol–1)

0

c

–10

b –20

–30

Figure 15.31. CD spectra of poly(L-Lys) in aqueous solution at pH 5.7 (a), the 20-mer S-peptide of ribonuclease A in water at 25◦ C (b), and the Ac-/-NH2 terminally blocked, 11-mer XAO (where X = α, β-diaminobutyric

a –40

200

220

240

Wavelength (nm)

acid or Dap, A = alanine, O = ornithine) in aqueous solution (pH 7.0) at 55◦ C (c). (Redrawn from references 88, 277, and 266, respectively.)

Bienkiewicz et al. [269] used the data of Park et al. [21] for the short-wavelength negative band to provide another method for PII estimation: fPII = ([θ ]200 + 9100)/(−36,600) The CD of unordered peptides generally shows a nearly linear dependence on temperature in the range usually investigated [28, 270]. The slope and intercept of the linear relationship need to be evaluated for each peptide and solvent condition. The implications of the PII component of the unordered peptides for the analysis of helix–coil equilibria were examined by Park et al. [21]. Helix–coil transitions are commonly monitored by CD, generally at 222 nm, and have been assumed to be two-state equilibria between helix and the unordered conformation. The [θ ]222 of the unordered form has frequently been taken to be zero. Park et al. [21] pointed out that we are really dealing with a three-state equilibrium between α-helix, PII -helix, and truly unordered conformations. They proposed that the PII + unordered conformation should be modeled by a peptide with the same sequence as the peptide under investigation, but with the central residue replaced by l-Pro. In the 1970s, the concept of the peptide chromophore parameter was introduced and deﬁned as two amino acid residues, l-Xxx–l-Xxx, forming the peptide bond [271, 272].

534

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

10

[θ] x 10−3 (deg x cm2 x dmol−1)

0 –10 –20 –30 –40 –50 –60 150

160

170

180

190

200

210

220

230

240

250

Wavelength (nm)

Figure 15.32. CD spectra of poly(L-Glu) (• • •) at pH 8 (coil) and poly(L-Pro) (o o o). (Redrawn from references 247 and 248, respectively.)

[θ] x 10−3 (deg x cm2 x dmol−1)

10

0

–10

–20

Figure 15.33. CD spectra of L-Ala –30 180

190

200

210 220 230 Wavelength (nm)

240

250

260

derivatives. Ac– L-Ala–OH (— —), H– L-Ala–NH2 (–·–·–), and Ac-L-Ala-NH2 (– – –). (Redrawn from reference 268.)

As a result, in a CD study the total molar ellipticity of an unordered polypeptide chain is represented at any wavelength by the equation [θ ]T = i ni [θ ]R , where ni is the number of peptide chromophores of type i and [θ ]R is the intrinsic (internal or terminal) peptide chromophore ellipticity. To calculate the most relevant l-Xxx–l-Xxx internal [θ ]R values for each homo-dipeptide sequence, the total molar ellipticity values of N- and C- protected homo-tripeptides were subtracted from those of

535

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

[θ]R x 10–3 (deg x cm2 x dmol–1)

0

0 a b

e

c

–10

–10

d

Figure 15.34. CD spectra of the -L-Ala– L-Ala–20

–20

200

220

240

Wavelength (nm)

200

220

240

(a), -L-Leu– L-Leu- (b), -L-Nva– L-Nva- (c) where Nva is norvaline, -L-Val– L-Val- (d), and -L-Ile– L-Ile- (e) internal peptide chromophores in HFIP (1,1,1,3,3,3-hexaﬂuoroisopropanol) solution at room temperature. (Redrawn from reference 272.)

Wavelength (nm)

[θ]R x 10–3 (deg x cm2 x dmol–1)

0

–10

a

–20

b

Figure 15.35. CD spectra of the -L-Ala– L-Alainternal peptide chromophore in water at 65◦ C (a) (redrawn from reference 272); and calculated

–30

200

220 Wavelength (nm)

240

from -(Gly)2 –(L-Ala)2 –(Gly)2 - in water (b) (Redrawn from reference 271.)

the corresponding homo-tetrapeptides. The [θ ]R values thus obtained represent rigorously only those of the corresponding homo-polymers in an unordered conformation. In particular, it was found that: (i) In alcoholic solution the CD spectra of the internal peptide chromophores of aliphatic hydrocarbon-containing peptides, each of the l-conﬁguration, show a weak, negative Cotton effect at about 222 nm accompanied by a signiﬁcantly more intense, negative Cotton effect just below 200 nm (Figure 15.34). (ii) Within the

536

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

ROH family of solvents, the effect of solvent polarity on the internal [θ ]R values is of minor signiﬁcance. (iii) The corresponding CD spectrum for the l-Ala–l-Ala internal peptide chromophore in aqueous solution at 65◦ C, where the population of higher energy conformers is increased by heating, is similar (Figure 15.35) to that in alcohols. It is worth pointing out that the CD spectra for the l-Lys–l-Lys and l-Glu–l–Glu internal peptide chromophores show a weak, positive dichroism at about 215 nm, but only in the pH region where their side chains are ionized. Not surprisingly, this band is barely discernible in the case of the l-Glu(OMe)–l-Glu(OMe) internal peptide chromophore [273]. It is evident that these data on the internal peptide chromophores ﬁt beautifully with all of the conclusions described above for the CD properties of the unordered peptide conformation. Deﬁnitive conﬁrmation of these conclusions, particularly in aqueous solution, can be extracted from the results published by Goodman, Toniolo, Mutter, and their co-workers on the CD spectra of a variety of monodisperse homo-oligopeptide series [29, 99, 116, 120, 274–276] and from the related data on two classical peptides [163, 264, 266, 277–281]. Furthermore, selected stimulating readings on this exciting topic may be found in references 210, 217, 268, and 282–292.

REFERENCES 1. D. W. Sears, S. Beychok, Physical Principles and Techniques of Protein Chemistry, Vol. C, S. J. Leach, ed., Academic Press, New York, 1973, pp. 445–593. 2. W. C. Johnson, Jr., Meth. Biochem. Anal . 1985, 31 , 61–163. 3. W. C. Johnson, Jr., Annu. Rev. Biophys. Chem. 1988, 17 , 145–166. 4. R. W. Woody, Meth. Enzymol . 1995, 246 , 34–71. 5. S. M. Kelly, T. J. Jess, N. C. Price, Biochim. Biophys. Acta-Proteins, Proteom. 2005, 1751 , 119–139. 6. G. N. Ramachandran, V. Sasisekharan, Adv. Protein Chem. 1968, 23 , 283–438. 7. L. Pauling, R. B. Corey, H. R. Branson, Proc. Natl. Acad. Sci. USA 1951, 37 , 205–211. 8. D. J. Barlow, J. M. Thornton, J. Mol. Biol . 1988, 201 , 601–619. 9. G. Holzwarth, P. Doty, J. Am. Chem. Soc. 1965, 87 , 218–228. 10. J. A. Schellman, P. Oriel, J. Chem. Phys. 1962, 37 , 2114–2124. 11. R. W. Woody, I. Tinoco, Jr., J. Chem. Phys. 1967, 46 , 4927–4945. 12. W. J. Mofﬁt, J. Chem. Phys. 1956, 25 , 467–478. 13. I. Tinoco, Jr., R. W. Woody, D. F. Bradley, J. Chem. Phys. 1963, 38 , 1317–1325. 14. J. R. Parrish, Jr., E. R. Blout, Biopolymers 1971, 10 , 1491–1512. 15. G. N´emethy, D. C. Phillips, S. J. Leach, H. A. Scheraga, Nature 1967, 214 , 363–365. 16. J. M. Scholtz, R. L. Baldwin, Annu. Rev. Biophys.Biomol. Struct. 1992, 21 , 95–118. 17. C. A. Rohl, R. L. Baldwin, Biochemistry, 1997, 36 , 8435–8442. 18. S. Marqusee, R. L. Baldwin, Proc. Natl. Acad. Sci. USA 1987, 84 , 8898–8902. 19. N. R. Kallenback, E. J. Spek, Meth. Enzymol . 1998, 295 , 26–41. 20. Y.-H. Chen, J. T. Yang, Biochem. Biophys. Res. Commun. 1971, 44 , 1285–1291. 21. S. H. Park, W. Shalongo, E. Stellwagen, Protein Sci . 1997, 6 , 1694–1700. 22. Y.-H. Chen, J. T. Yang, H. M. Martinez, Biochemistry 1972, 11 , 4120–4131. 23. Y.-H. Chen, J. T. Yang, K. H. Chau, Biochemistry 1974, 13 , 3350–3359. 24. V. Madison, J. Schellman, Biopolymers 1972, 11 , 1041–1076.

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

25. P. J. Gans, P. C. Lyu, M. C. Manning, R. W. Woody, N. R. Kallenbach, Biopolymers 1991, 31 , 1605–1614. 26. M. C. Manning, R. W. Woody, Biopolymers 1991, 31 , 569–586. 27. R. W. Woody, J. Polym. Sci., part D: Macromol. Rev . 1977, 12 , 181–320. 28. C. A. Rohl, A. Chakrabartty, R. L. Baldwin, Protein Sci . 1996, 5 , 2623–2637. 29. M. Goodman, A. S. Verdini, C. Toniolo, W. D. Phillips, F. A. Bovey, Proc. Natl. Acad. Sci. USA 1969, 64 , 444–450. 30. C. Toniolo, G. M. Bonora, S. Salardi, M. Mutter, Macromolecules 1979, 12 , 620–625. 31. C. Toniolo, G. M. Bonora, H. Anzinger, M. Mutter, Macromolecules 1983, 16 , 147–149. 32. J. A. Schellman, Compt. Rend. Trav. Lab. Carlsberg, Ser. Chim. 1955, 29 , 230–259. 33. D. S. Kemp, J. H. Rothman, Tetrahedron Lett. 1995, 36 , 4019–4022. 34. D. S. Kemp, J. H. Rothman, Tetrahedron Lett. 1995, 36 , 4023–4026. 35. W. Maison, E. Arce, P. Renold, R. J. Kennedy, D. S. Kemp, J. Am. Chem. Soc. 2001, 123 , 10245–10254. 36. P. Wallimann, R. J. Kennedy, D. S. Kemp, Angew. Chem. Int. Edit. 1999, 38 , 1290–1292. 37. M. Siedlicka, G. Goch, A. Ejchart, H. Sticht, A. Bierzynski, Proc. Natl. Acad. Sci. USA 1999, 96 , 903–908. 38. D.-H. Chin, R. W. Woody, C. A. Rohl, R. L. Baldwin, Proc. Natl. Acad. Sci. USA 2002, 99 , 15416–15421. 39. R. W. Woody, A. Koslowski, Biophys. Chem. 2002, 101 , 535–551. 40. L. B. Clark, J. Am. Chem. Soc. 1995, 117 , 7974–7986. 41. M. Gobbo, L. Biondi, F. Filira, F. Formaggio, M. Crisma, R. Rocchi, C. Toniolo, Q. B. Broxterman, J. Kamphuis, Lett. Pept. Sci . 1998, 5 , 105–107. 42. P. Wallimann, R. J. Kennedy, J. S. Miller, W. Shalongo, D. S. Kemp, J. Am. Chem. Soc. 2003, 125 , 1203–1220. 43. B. Heitmann, G. E. Job, R. J. Kennedy, S. M. Walker, D. S. Kemp, J. Am. Chem. Soc. 2005, 127 , 1690–1704. 44. G. E. Job, R. J. Kennedy, B. Heitmann, J. S. Miller, S. M. Walker, D. S. Kemp, J. Am. Chem. Soc. 2006, 128 , 8227–8233. 45. R. W. Woody, Houben-Weyl: Methods of Organic Chemistry, Vol. E22b, M. Goodman, A. Felix, L. Moroder, C. Toniolo, eds., Thieme, Stuttgart, Germany, 2003, pp. 739–765. 46. K. Groebke, P. Renold, K. Y. Tsang, T. J. Allen, K. F. McClure, D. S. Kemp, Proc. Natl. Acad. Sci. USA 1996, 93 , 4025–4029. 47. C. B. Marshall, A. Chakrabartty, P. L. Davies, J. Biol. Chem. 2005, 280 , 17920–17929. 48. R. S. Farmer, K. L. Kiick, Biomacromolecules 2005, 6 , 1531–1539. 49. Z. Dang, J. D. Hirst, Angew. Chem. Int. Edit. 2001, 40 , 3619–3621. 50. C. Toniolo, E. Benedetti, Trends Biochem. Sci . 1991, 16 , 350–353. 51. C. Toniolo, M. Crisma, F. Formaggio, C. Peggion, Q. B. Broxterman, B. Kaptein, J. Incl. Phenom. Macrocycl. Chem. 2005, 51 , 121–136. 52. M. Crisma, F. Formaggio, A. Moretto, C. Toniolo, Biopolymers (Pept. Sci.) 2006, 84 , 3–12. 53. I. L. Karle, P. Balaram, Biochemistry 1990, 29 , 6747–6756. 54. K. A. Bolin, G. L. Millhauser, Acc. Chem. Res. 1999, 32 , 1027–1033. 55. T. Ohyama, A. Hiroki, K. Kobayashi, R. Katakai, Peptide Chemistry 1996 , C. Kitada, ed., Protein Research Foundation, Osaka, Japan, 1997, pp. 365–368. 56. A. Freiberg, M. P. Machner, W. Pfeil, W.-D. Schubert, D. W. Heinz, R. Seckler, J. Mol. Biol . 2004, 337 , 453–461. 57. A. V. Kajava, B. Kobe, Protein Sci . 2002, 11 , 1082–1090.

537

538

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

58. C. Toniolo, M. Crisma, F. Formaggio, C. Peggion, Biopolymers (Pept. Sci.). 2001, 60 , 396–419. 59. C. Toniolo, H. Br¨uckner, Peptaibiotics: Fungal Peptides Containing α-Dialkyl α-Amino Acids, Wiley-VCH, Weinheim, Germany, 2009. 60. C. Toniolo, A. Polese, F. Formaggio, M. Crisma, J. Kamphuis, J. Am. Chem. Soc. 1996, 118 , 2744–2745. 61. A. Polese, F. Formaggio, M. Crisma, G. Valle, C. Toniolo, G. M. Bonora, Q. B. Broxterman, J. Kamphuis, Chem. Eur. J . 1996, 2 , 1104–1111. 62. K. A. Bode, J. Applequist, J. Phys. Chem. 1996, 100 , 17825–17834. 63. T. S. Sudha, E. K. S. Vijayakumar, P. Balaram, Int. J. Pept. Protein Res. 1983, 22 , 464–468. 64. S. Aravinda, S. Datta, N. Shamala, P. Balaram, Angew. Chem. Int. Edit. 2004, 43 , 6728–6731. 65. N. H. Andersen, Z. Liu, K. S. Prickett, FEBS Lett. 1996, 399 , 47–52. 66. R. Banerjee, S. Chattopadhyay, G. Basu, Proteins 2009, 76 , 184–200. 67. A. Moretto, F. Formaggio, B. Kaptein, Q. B. Broxterman, L. Wu, T. A. Keiderling, C. Toniolo, Biopolymers (Pept. Sci.) 2008, 90 , 567–574. 68. A. Moretto, M. Crisma, F. Formaggio, B. Kaptein, Q. B. Broxterman, T. A. Keiderling, C. Toniolo, Biopolymers (Pept. Sci.) 2007, 88 , 233–238. 69. G. Yoder, A. Polese, R. A. G. D. Silva, F. Formaggio, M. Crisma, Q. B. Broxterman, J. Kamphuis, C. Toniolo, T. A. Keiderling, J. Am. Chem. Soc. 1997, 117 , 10278–10285. 70. F. Donald, G. Hungerford, D. J. S. Birch, B. D. Moore, J. Chem. Soc., Chem. Commun. 1995, 313–314. 71. G. Hungerford, M. Martinez-Insua, D. J. S. Birch, B. D. Moore, Angew. Chem. Int. Edit. Engl . 1996, 35 , 326–329. 72. P. Pengo, L. Pasquato, S. Moro, A. Brigo, F. Fogolari, Q. B. Broxterman, B. Kaptein, P. Scrimin, Angew. Chem. Int. Edit. 2003, 42 , 3388–3392. 73. M. Bellanda, S. Mammi, S. Geremia, N. Demitri, L. Randaccio, Q. B. Broxterman, B. Kaptein, P. Pengo, L. Pasquato, P. Scrimin, Chem. Eur. J . 2007, 13 , 407–416. 74. B. Sacc`a, F. Formaggio, M. Crisma, C. Toniolo, R. Gennaro, Gazz. Chim. Ital . 1997, 127 , 495–500. 75. F. Formaggio, M. Crisma, P. Rossi, P. Scrimin, B. Kaptein, Q. B. Broxterman, C. Toniolo, Chem. Eur. J . 2000, 6 , 4498–4504. ˇ 76. C. Toniolo, F. Formaggio, S. Tognon, Q. B. Broxterman, B. Kaptein, R. Huang, V. Setnicka, T. A. Keiderling, I. H. Mc Coll, L. Hecht, L. D. Barron, Biopolymers 2004, 75 , 32–45. 77. F. Formaggio, C. Peggion, M. Crisma, B. Kaptein, Q. B. Broxterman, J.-P. Mazaleyrat, M. Wakselman, C. Toniolo, Chirality 2004, 16 , 388–397. 78. E. Longo, A. Moretto, F. Formaggio, C. Toniolo, Chirality 2011, 23 , 756–760. 79. I. L. Karle, J. Flippen-Andersen, M. Sukumar, P. Balaram, Proc. Natl. Acad. Sci. USA 1987, 84 , 5087–5091. 80. B. Di Blasio, V. Pavone, M. Saviano, A. Lombardi, F. Nastri, C. Pedone, E. Benedetti, M. Crisma, M. Anzolin, C. Toniolo, J. Am. Chem. Soc. 1992, 114 , 6273–6278. 81. G. Yoder, T. A. Keiderling, F. Formaggio, M. Crisma, C. Toniolo, Biopolymers 1995, 35 , 103–111. 82. L. Pauling, R. B. Corey, Proc. Natl. Acad. Sci. USA 1951, 37 , 729–740. 83. C. Toniolo, Macromolecules 1978, 11 , 437–438. 84. E. R. Blout, Poly-α-Amino Acids, Polypeptides and Proteins, M. A. Stahmann, ed. University of Wisconsin Press, Madison, WI, 1962, pp. 275–279. 85. C. Chothia, J . Mol. Biol . 1973, 75 , 295–302. 86. R. Nelson, D. Eisenberg, Curr. Opin. Struct. Biol . 2006, 16 , 260–265.

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121.

S. Brahms, J. Brahms, G. Spach, A. Brack, Proc. Natl. Acad. Sci. USA 1977, 74 , 3208–3212. N. J. Greenﬁeld, G. D. Fasman, Biochemistry 1969, 8 , 4108–4116. L. K. Li, A. Spector, J. Am. Chem. Soc. 1969, 91 , 220–222. J. S. Balcerski, E. S. Pysh, G. M. Bonora, C. Toniolo, J . Am. Chem. Soc. 1976, 98 , 3470–3473. M. M. Kelly, E. S. Pysh, G. M. Bonora, C. Toniolo, J . Am. Chem. Soc. 1977, 99 , 3264–3266. M. Palumbo, S. Da Rin, G. M. Bonora, C. Toniolo, Makromol. Chem. 1976, 177 , 1477–1492. C. Toniolo, M. Palumbo, Biopolymers 1977, 16 , 219–224. K.-C. Chou, M. Pottle, G. N´emethy, Y. Ueda, H. A. Scheraga, J. Mol. Biol . 1982, 162 , 89–112. K.-C. Chou, H. A. Scheraga, Proc. Natl. Acad. Sci. USA 1982, 79 , 7047–7051. M. D. Yoder, N. T. Keen, F. Jurnak, Science 1993, 260 , 1503–1507. D. E. Kamen, Y. Griko, R. W. Woody, Biochemistry 2000, 39 , 15932–15943. L. Tilstra, W. L. Mattice, Circular Dichroism and the Conformational Analysis of Biomolecules, G. D. Fasman, ed., Plenum Press, New York, 1996, pp. 261–283. M. Goodman, F. Naider, C. Toniolo, Biopolymers 1971, 10 , 1719–1730. C. Toniolo, G. M. Bonora, A. Fontana, Int. J. Pept. Protein Res. 1974, 6 , 283–285. M. Ramirez-Alvarado, T. Kortemme, F. J. Blanco, L. Serrano, Bioorg. Med. Chem. 1999, 7 , 93–103. F. J. Blanco, G. Rivas, L. Serrano, Nat. Struct. Biol . 1994, 1 , 584–590. A. R. Viguera, M. A. Jimenez, M. Rico, L. Serrano, J. Mol. Biol 1996, 255, 507–521. V. Sieber, G. R. Moe, Biochemistry 1996, 35 , 181–188. H. J. Dyson, M. Rance, R. A. Houghten, P. E. Wright, R. A. Lerner, J. Mol. Biol . 1988, 201 , 201–217. M. Ramirez-Alvarado, F. J. Blanco, H. Niemann, L. Serrano, J. Mol. Biol . 1997, 273 , 898–912. A. J. Maynard, G. J. Sharman, M. S. Searle, J. Am. Chem. Soc. 1998, 120 , 1996–2007. T. Kortemme, M. Ramirez-Alvarado, L. Serrano, Science 1998, 281 , 253–256. M. N. Boyden, S. A. Asher, Biochemistry 2001, 40 , 13723–13727. O. B. Ptitsyn, Protein Folding, T. E. Creighton, ed., Freeman, New York, 1992, pp. 243–300. M. L. de la Paz, E. Lacroix, M. Ramirez-Alvarado, L. Serrano, J. Mol. Biol . 2001, 312 , 229–246. S. V. Kuznetsov, J. Hilario, T. A. Keiderling, A. Ansari, Biochemistry 2003, 42 , 4321–4332. H. L. Schenck, S. H. Gellman, J. Am. Chem. Soc. 1998, 120 , 4869–4870. L. Zhong, W. C. Johnson, Jr., Proc. Natl. Acad. Sci. USA 1992, 89 , 4462–4465. C. Toniolo, G. M. Bonora, Makromol. Chem. 1974, 175 , 2203–2207. C. Toniolo, G. M. Bonora, M. Mutter, J. Am. Chem. Soc. 1979, 101 , 450–454. M. Mutter, S. Vuilleumier, Angew. Chem. Int. Edit. Engl . 1989, 28 , 535–554. M. Mutter, F. Maser, K.-H. Altmann, C. Toniolo, G. M. Bonora, Biopolymers 1985, 24 , 1057–1074. C. Toniolo, G. M. Bonora, M. Mutter, V. N. R. Pillai, Makromol. Chem. 1981, 182 , 2007–2014. V. Moretto, M. Crisma, G. M. Bonora, C. Toniolo, H. Balaram, P. Balaram, Macromolecules 1989, 22 , 2939–2944. F. Formaggio, A. Bettio, V. Moretto, M. Crisma, C. Toniolo, Q. B. Broxterman, J. Pept. Sci . 2003, 9 , 461–466.

539

540

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

122. C. Soto, M. S. Kindy, M. Baumann, B. Frangione, Biochem. Biophys. Res. Commun. 1996, 226 , 672–680. 123. E. Hughes, R. M. Burke, A. J. Doig, J. Biol. Chem. 2000, 275 , 25109–25115. 124. D. J. Gordon, K. L. Sciarretta, S. C. Meredith, Biochemistry 2001, 40 , 8237–8245. 125. D. T. S. Rijkers, J. W. M. H¨oppener, G. Posthuma, C. J. M. Lips, R. M. J. Liskamp, Chem. Eur. J . 2002, 8 , 4285–4291. 126. K.-H. Altmann, M. Mutter, Int. J. Pept. Protein Res. 1985, 26 , 373–380. 127. M. Goodman, C. Toniolo, P. Pallai, Forum Peptides, B. Castro, J. Martinez, eds., Dohr, Nancy, France, 1986, pp. 146–174. 128. M. Goodman, C. Toniolo, F. Naider, Peptides, Polypeptides and Proteins, E. R. Blout, F. A. Bovey, M. Goodman, N. Lotan, eds., Wiley, New York, 1974, pp. 308–319. 129. C. Toniolo, G. M. Bonora, Peptides: Chemistry, Structure and Biology, R. Walter, J. Meienhofer, eds., Ann Arbor Science, Ann Arbor, MI, 1975, pp. 145–150. 130. C. Toniolo, M. Palumbo, J. Polym. Sci., Polym. Chem. Edit. 1976, 14 , 515–518. 131. K.-H. Altmann, A. Fl¨orsheimer, M. Mutter, Int. J. Pept. Protein Res. 1986, 27 , 314–319. 132. C. M. Venkatachalam, Biopolymers 1968, 6 , 1425–1436. 133. A. J. Geddes, K. D. Parker, E. D. T. Atkins, E. Brighton, J.Mol. Biol . 1968, 32 , 343–358. 134. P. N. Lewis, F. A. Momany, H. A. Scheraga, Biochim. Biophys. Acta 1973, 303 , 211–229. 135. G. N´emethy, M. P. Printz, Macromolecules 1972, 5 , 755–758. 136. B. W. Matthews, Macromolecules 1972, 5 , 818–819. 137. A. S. Kolaskar, A. V. Lakshminarayanan, K. P. Sarathy, V. Sasisekharan, Biopolymers 1975, 14 , 1081–1094. 138. C. Toniolo, CRC Crit. Rev. Biochem. 1980, 9 , 1–44. 139. G. D. Rose, L. M. Gierasch, J. P. Smith, Adv. Protein Chem. 1985, 37 , 1–109. 140. B. L. Sibanda, J. M. Thornton, J. Mol. Biol . 1993, 229 , 428–447. 141. A. Perczel, M. Hollosi, Circular Dichroism and the Conformational Analysis of Biomolecules, G. D. Fasman, ed., Plenum Press, New York, 1996, pp. 285–380. 142. K. S. Rotondi, L. M. Gierasch, Biopolymers (Pept. Sci.) 2006, 84 , 13–22. 143. R. W. Woody, Peptides, Polypeptides, and Proteins, E. R. Blout, F. A. Bovey, M. Goodman, N. Lotan, eds., John Wiley & Sons, New York, 1974, pp. 338–350. 144. R. Deslauriers, D. J. Evans, S. J. Leach, Y. C. Meinwald, E. Minasian, G. N´emethy, I. D. Rae, H. A. Scheraga, R. L. Somorjai, E. R. Stimson, J. W. Van Nispen, R. W. Woody, Macromolecules 1981, 14 , 985–996. 145. C. A. Bush, S. K. Sarkar, K. D. Kopple, Biochemistry 1978, 17 , 4951–4954. 146. L. M. Gierasch, C. M. Deber, V. Madison, C. H. Niu, E. R. Blout, Biochemistry 1981, 20 , 4730–4738. 147. S. Laiken, M. Printz, L. C. Craig, J. Biol. Chem. 1969, 244 , 4454–4457. 148. A. C. Gibbs, L. H. Kondejewski, W. Gronwald, A. M. Nip, R. S. Hodges, B. D. Sykes, D. S. Wishart, Nat. Struct. Biol . 1998, 5 , 284–288. 149. S. Lee, H. Mizuno, H. Nakamura, Y. Kodera, T. Kato, N. Go, N. Izumiya, FEBS Lett. 1984, 174 , 310–313. 150. V. Ananthanarayan, N. Shyamasundar, Biochem. Biophys. Res. Commun. 1981, 102 , 295–301. 151. J. Bandekar, D. J. Evans, S. Krimm, S. J. Leach, S. Lee, J. R. McQuie, E. Minasian, G. N´emethy, M. S. Pottle, H. A. Scheraga, E. R. Stimson, R. W. Woody, Int. J. Pept. Protein Res. 1982, 19 , 187–205. 152. B. K. Sathyanarayanan, J. Applequist, Int. J. Pept. Protein Res. 1986, 27 , 86–94. 153. A. Perczel, M. Hollosi, G. Tusnady, G. D. Fasman, Protein Eng. 1991, 4 , 669–679.

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

154. A. Perczel, M. Hollosi, B. M. Foxman, G. D. Fasman, J. Am. Chem. Soc. 1991, 113 , 9772–9784. 155. A. Perczel, G. D. Fasman, Protein Sci . 1992, 1 , 378–395. 156. A. Perczel, M. Hollosi, P. Sandor, G. D. Fasman, Int. J. Pept. Protein Res. 1993, 41 , 223–236. 157. A. M. Tamburro, B. Bochicchio, A. Pepe, Biochemistry 2003, 42 , 13347–13362. 158. B. Bochicchio, N. Floquet, A. Pepe, A. J. P. Alix, A. Tamburro, Chem. Eur. J . 2004, 10 , 3166–3176. 159. A. M. Tamburro, Nanomedicine 2009, 4 , 469–487. 160. C. M. Venkatachalam, D. W. Urry, Macromolecules 1981, 14 , 1225–1229. 161. M. De Poli, A. Moretto, M. Crisma, C. Peggion, F. Formaggio, B. Kaptein, Q. B. Broxterman, C. Toniolo, Chem. Eur. J . 2009, 15 , 8015–8025. 162. V. Madison, K. D. Kopple, J. Am. Chem. Soc. 1980, 102 , 4855–4863. 163. Z. Liu, K. Chen, A. Ng, Z. S. Shi, R. W. Woody, N. R. Kallenbach, J. Am. Chem. Soc. 2004, 126 , 15141–15150. 164. R. W. Woody, The Peptides, Vol. 7, V. J. Hruby, ed., Academic Press, New York, 1985, pp. 15–114. 165. N. Sreerama, R. W. Woody, Circular Dichroism: Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, R. W. Woody, eds., Wiley-VCH, New York, 2000, pp. 601–619. 166. E. Vass, M. Hollosi, F. Besson, R. Buchet, Chem. Rev . 2003, 103 , 1917–1954. 167. E. Benedetti, A. Bavoso, B. Di Blasio, V. Pavone, C. Pedone, C. Toniolo, G. M. Bonora, Biopolymers 1983, 22 , 305–317. 168. P. M. Cowan, S. McGavin, Naure (London) 1955, 176 , 501–503. 169. W. Traub, U. Shmueli, Naure (London) 1963, 198 , 1165–1166. 170. M. W. MacArthur, J. M. Thornton, J. Mol. Biol . 1991, 218 , 397–412. 171. K. A. Williams, C. M. Deber, Biochemistry 1991, 30 , 8919–8923. 172. A. Yaron, F. Naider, Crit. Rev. Biochem. Mol. Biol . 1993, 28 , 31–81. 173. A. A. Adzhubei, M. J. E. Sternberg, J. Mol. Biol . 1993, 229 , 472–493. 174. M. P. Williamson, Biochem. J . 1994, 297 , 249–260. 175. R. Berisio, S. Loguercio, A. De Simone, A. Zagari, L. Vitagliano, Protein Pept. Lett. 2006, 13 , 847–854. 176. B. K. Ho, E. A. Coutsias, C. Seok, K. A. Dill, Protein Sci . 2005, 14 , 1011–1018. 177. R. Berisio, L. Vitagliano, L. Mazzarella, A. Zagari, Protein Sci . 2002, 11 , 262–270. 178. L. Stryer, R. P. Haugland, Proc. Natl. Acad. Sci. USA 1967, 58 , 719–726. 179. S. F. Sneddon, C. L. Brooks, III, J. Am. Chem. Soc. 1992, 114 , 8220–8225. 180. A. K. Mishra, R. Chandrasekar, M. Faraggi, M. H. Kappler, J. Am. Chem. Soc. 1994, 116 , 1414–1422. 181. H. Tamiaki, K. Nomura, K. Maruyama, Bull. Chem. Soc. Jpn 1994, 67 , 1863–1871. 182. S. Saxena, Y. H. Freed, Chem. Phys. Lett. 1996, 251 , 102–110. 183. W. S. Aldridge, III, B. J. Hornstein, S. Serrano, D. M. Dattelbaum, J. R. Schoonover, T. J. Meyer, J. Org. Chem. 2006, 71 , 5186–5190. 184. R. B. Best, K. A. Merchant, I. V. Gopich, B. Schuler, A. Bax, W. A. Eaton, Proc. Natl. Acad. Sci. USA 2007, 104 , 18964–18969. 185. S. Doose, H. Neuweiler, H. Barsch, M. Sauer, Proc. Natl. Acad. Sci. USA 2007, 104 , 17400–17405. 186. H. Sahoo, D. Roccatano, A. Hennig, W. M. Nau, J. Am. Chem. Soc. 2007, 129 , 9762–9772. 187. M. Cordes, A. K¨ottgen, C. Jasper, O. Jacques, H. Boudebous, B. Giese, Angew. Chem. Int. Edit. 2008, 47 , 3461–3463.

541

542

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

188. Z. Yang, D. Kise, S. Saxena, J. Phys. Chem. 2010, B 114, 6165–6174. 189. E. S. Price, M. S. DeVore, C. K. Johnson, J. Phys. Chem. 2010, B 114, 5895–5902. 190. B. Schuler, E. A. Lipman, P. J. Steinbach, M. Kumke, W. A. Eaton, Proc. Natl. Acad. Sci. USA 2005, 102 , 2754–2759. 191. G. N. Ramachandran, G. Kartha, Nature (London) 1954, 174 , 269–270. 192. A. Rich, F. H. C. Crick, Nature (London) 1955, 176 , 915–916. 193. A. V. Persikov, J. A. M. Ramshaw, B. Brodsky, Biopolymers (Pept. Sci.) 2000, 55 , 436–450. 194. L. Vitagliano, R. Berisio, L. Mazzarella, A. Zagari, Biopolymers 2001, 58 , 459–464. 195. K. Okuyama, C. Hongo, R. Fukushima, G. Wu, H. Narita, K. Noguchi, Y. Tanaka, N. Nishino, Biopolymers (Pept. Sci.) 2004, 76 , 367–377. 196. M. Goodman, M. Bhumralkar, E. A. Jefferson, J. Kwak, E. Locardi, Biopolymers (Pept. Sci.) 1998, 47 , 127–142. 197. S. N. Timasheff, H. Susi, R. Townend, L. Stevens, M. J. Gorbunoff, T. F. Kumosinski, Conformation of Biopolymers, G. N. Ramachandran, ed., Academic Press, New York, 1967, pp. 173–196. 198. M. L. Tiffany, S. Krimm, Biopolymers 1968, 6 , 1767–1770. 199. M. L. Tiffany, S. Krimm, Biopolymers 1969, 8 , 347–359. 200. R. W. Woody, J. Am. Chem. Soc. 2009, 131 , 8234–8245. 201. J. Applequist, Biopolymers 1981, 20 , 2311–2322. 202. V. Sasisekharan, Acta Crystallogr. 1959, 12 , 897–903. 203. H. Okabayshi, T. Isemura, S. Sakakibara, Biopolymers 1968, 6 , 323–330. 204. R. K. Dukor, T. A. Keiderling, Biopolymers 1991, 31 , 1747–1761. 205. M. Rothe, H. Rott, J. Mazanek, Peptides 1976 , A. Loffet, ed., Editions Universit´e de Bruxelles, Bruxelles, Belgium, 1976, pp. 309–318. 206. S. Kakinoki, Y. Hirano, M. Oka, Peptide Science 2004 , Y. Shimohigashi, ed., Japanese Peptide Society, Osaka, Japan, 2005, pp. 453–456. 207. E. R. Blout, Fundamental Aspects and Recent Developments in Optical Rotatory Dispersion and Circular Dichroism, F. Ciardelli, P. Salvadori, eds., Heyden, London, U.K., 1973, pp. 352–372. 208. R. W. Woody, Adv. Biophys. Chem. 1992, 2 , 37–79. 209. Z. Shi, R. W. Woody, N. R. Kallenbach, Adv. Protein Chem. 2002, 62 , 163–240. 210. R. W. Woody, Monatsh. Chem. 2005, 136 , 347–366. 211. Z. S. Shi, K. Chen, Z. Liu, N. R. Kallenbach, Chem. Rev . 2006, 106 , 1877–1897. 212. E. S. Pysh, J. Mol. Biol . 1967, 23 , 587–599. 213. E. S. Pysh, J. Chem. Phys. 1970, 52 , 4723–4733. 214. J. Applequist, K. A. Bode, J. Phys. Chem. 1999, B 103, 1767–1773. 215. F. A. Bovey, F. P. Hood, Biopolymers 1967, 5 , 325–326. 216. E. W. Ronish, S. Krimm, Biopolymers 1974, 13 , 1635–1651. 217. C. Moore, G. D. Fasman, Chemtracts-Biochem. Mol. Biol . 1993, 4 , 67–74. 218. S. K. Brahmachari, M. Bansal, V. S. Ananthanarayanan, V. Sasisekharan, Macromolecules 1979, 12 , 23–28. 219. N. Helbecque, M. H. Loucheux-Lefebvre, Int. J. Pept. Protein Res. 1982, 19 , 94–101. 220. K. A. Thomasson, J. Applequist, Biopolymers 1991, 31 , 529–535. 221. M. A. Young, E. S. Pysh, J. Am. Chem. Soc. 1975, 97 , 5100–5103. 222. N. Sreerama, R. W. Woody, Biochemistry 1994, 33 , 10022–10025. 223. M. K¨umin, L.-S. Sonntag, H. Wennemers, J. Am. Chem. Soc. 2007, 129 , 466–467.

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDES

224. M. K¨umin, S. Schweizer, C. Ochsenfeld, H. Wennemers, J. Am. Chem. Soc. 2009, 131 , 15474–15482. 225. D. G. McCafferty, C. A. Slate, B. M. Nakhle, H. D. Graham, Jr., T. L. Austell, R. W. Vachet, B. H. Mullis, B. W. Erickson, Tetrahedron 1995, 51 , 9859–9872. 226. L. Crespo, G. Sanclimens, B. Montaner, R. Perez-Tomas, M. Royo, M. Pons, F. Albericio, E. Giralt, J. Am. Chem. Soc. 2002, 124 , 8876–8883. 227. G. Sanclimens, L. Crespo, E. Giralt, F. Albericio, M. Royo, J. Org. Chem. 2005, 70 , 6274–6281. 228. M. A. Kelly, B. W. Chellgren, A. L. Rucker, J. M. Troutman, M. G. Fried, A.-F. Miller, T. P. Creamer, Biochemistry 2001, 40 , 14376–14383. 229. R. Schweitzer-Stenner, F. Eker, A. Perez, K. Griebenow, X. Cao, L. A. Naﬁe, Biopolymers (Pept. Sci.) 2003, 71 , 558–568. 230. R. Boni, A. S. Verdini, Macromolecules 1973, 6 , 517–520. 231. C. G. Overberger, M. J. Han, J. Polym. Sci., Polym. Chem. Edit. 1975, 13 , 2251–2258. 232. K.-D. Ahn, C. G. Overberger, J. Polym. Sci., Polym. Chem. Edit. 1983, 21 , 1699–1715. 233. W. W.-Y. Yang, C. G. Overberger, C. M. Venkatachalam, J. Polym. Sci., Polym. Chem. Edit. 1983, 21 , 1751–1761. 234. O. Tiba, C. G. Overberger, J. Polym. Sci., Polym. Chem. Edit. 1987, 25 , 2941–2952. 235. F. Rabanal, M. D. Ludevid, M. Pons, E. Giralt, Biopolymers 1993, 33 , 1019–1028. 236. E. C. Petrella, L. M. Machesky, D. A. Kaiser, T. D. Pollard, Biochemistry 1996, 35 , 16535–16543. 237. M. Mutter, T. W¨ohr, S. Gioria, M. Keller, Biopolymers (Pept. Sci.) 1999, 51 , 121–128. 238. A. Zhang, Y. Guo, Chem. Eur. J . 2008, 14 , 8939–8946. 239. L. Tooke, L. Duitch, T. J. Measey, R. Schweitzer-Stenner, Biopolymers 2010, 93 , 451–457. 240. E. Beausoleil, W. D. Lubell, Biopolymers 2000, 53 , 249–256. 241. P. Zhang, L. Polavarapu, J. Huang, T. Li, Chirality 2006, 19 , 99–106. 242. M. Wakahara, T. Nakamura, M. Oka, T. Hayashi, M. Hattori, Y. Hirano, Peptide Science 2001 , H. Aoyagi, ed., The Japanese Peptide Society, Osaka, Japan, 2002, pp. 289–292. 243. M. Yuge, S. Kakinoki, M. Kitamura, Y. Hirano, M. Oka, Peptide Science 2005 , T. Wakamiya, ed., The Japanese Peptide Society, Osaka, Japan, 2006, pp. 403–406. 244. M. Yuge, S. Kakinoki, M. Kitamura, Y. Hirano, M. Oka, Peptide Science 2006 , H. Ishida, H. Mihara, eds., The Japanese Peptide Society, Osaka, Japan, 2006, pp. 82–83. 245. M. Yuge, S. Kakinoki, Y. Hirano, M. Oka, Peptide Science 2006 , H. Ishida, H. Mihara, eds., The Japanese Peptide Society, Osaka, Japan, 2006, pp. 84–85. 246. L. Tterlikkis, F. M. Loxsom, W. Rhodes, Biopolymers 1973, 12 , 675–684. 247. W. C. Johnson, Jr., I. Tinoco, Jr., J. Am. Chem. Soc. 1972, 94 , 4389–4390. 248. D. D. Jenness, C. Sprecher, W. C. Johnson, Jr., Biopolymers 1976, 15 , 513–521. 249. F. R. Brown, III, J. Carver, E. R. Blout, J. Mol. Biol . 1969, 39 , 307–313. 250. G. M. Bonora, C. Toniolo, Biopolymers 1974, 13 , 1067–1078. 251. J. Kwak, A. De Capua, E. Locardi, M. Goodman, J. Am. Chem. Soc. 2002, 124 , 14085–14091. 252. G. A. Kinberger, J. P. Taulane, M. Goodman, Tetrahedron 2006, 62 , 5280–5286. 253. F. W. Kotch, I. A. Guzei, R. T. Raines, J. Am. Chem. Soc. 2008, 130 , 2952–2953. 254. W. C. Johnson, Jr., I. Tinoco, Jr., J. Am. Chem. Soc. 1972, 94 , 4389–4390. 255. M. L. Tiffany, S. Krimm, Biopolymers 1968, 6 , 1379–1382. 256. M. L. Tiffany, S. Krimm, Biopolymers 1972, 11 , 2309–2316. 257. A. F. Drake, G. Siligardi, W. A. Gibbons, Biophys. Chem. 1988, 31 , 143–146. 258. S. C. Yasui, T. A. Keiderling, J. Am. Chem. Soc. 1986, 108 , 5576–5581.

543

544

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

259. M. G. Paterlini, T. B. Freedman, L. A. Naﬁe, Biopolymers 1986, 25 , 1751–1765. 260. F. Eker, X. L. Cao, L. Naﬁe, R. Schweitzer-Stenner, J. Am. Chem. Soc. 2002, 124 , 14330–14341. 261. F. Eker, K. Griebenow, R. Schweitzer-Stenner, J. Am. Chem. Soc. 2003, 125 , 8178–8185. 262. F. Eker, K. Griebenow, X. L. Cao, L. A. Naﬁe, R. Schweitzer-Stenner, Proc. Natl. Acad. Sci. USA 2004, 101 , 10054–10059. 263. A. Hagarman, T. Measey, R. S. Doddasomayajula, I. Dragomir, F. Eker, K. Griebenow, R. Schweitzer-Stenner, J. Phys. Chem. 2006, B110 , 6979–6986. 264. R. Schweitzer-Stenner, T. J. Measey, Proc. Natl. Acad. Sci. USA 2007, 104 , 6649–6654. 265. R. Schweitzer-Stenner, J. Phys. Chem. 2009, B113 , 2922–2932. 266. Z. S. Shi, C. A. Olson, G. D. Rose, R. L. Baldwin, N. R. Kallenbach, Proc. Natl. Acad. Sci. USA 2002, 99 , 9190–9195. 267. J. Graf, P. H. Nguyen, G. Stock, H. Schwalbe, J. Am. Chem. Soc. 2007, 129 , 1179–1189. 268. I. Gokce, R. W. Woody, G. Anderluh, J. H. Lakey, J. Am. Chem. Soc. 2005, 127 , 9700–9701. 269. E. A. Bienkiewicz, A.-Y. M. Woody, R. W. Woody, J. Mol. Biol . 2000, 297 , 119–133. 270. J. M. Scholtz, H. Qiang, E. J. York, J. M. Stewart, R. L. Baldwin, Biopolymers 1991, 31 , 1463–1470. 271. R. S. Lord, D. J. Cox, Biopolymers 1973, 12 , 2359–2373. 272. C. Toniolo, G. M. Bonora, Can. J. Chem. 1976, 54 , 70–76. 273. S. Mammi, M. Goodman, J. Pept. Sci . 2005, 11 , 273–277. 274. M. Goodman, F. Naider, R. Rupp, Bioorg. Chem. 1971, 1 , 310–328. 275. C. Toniolo, G. M. Bonora, M. Mutter, Makromol. Chem. 1975, 176 , 2547–2558. 276. G. M. Bonora, C. Toniolo, M. Mutter, Polymer 1978, 19 , 1382–1386. 277. E. Scoffone, F. Marchiori, R. Rocchi. A. Scatturin, A. M. Tamburro, Peptides 1969 , E. Scoffone, ed., North-Holland, Amsterdam, The Netherlands, 1971, pp. 223–230. 278. J. W. Nelson, N. R. Kallenbach, Proteins: Struct., Funct., Genet. 1986, 1 , 211–217. 279. B. Zagrovic, J. Lipfert, E. J. Sorin, I. S. Millett, W. F. van Gunsteren, S. Doniach, V. S. Pande, Proc. Natl. Acad. Sci. USA 2005, 102 , 11698–11703. 280. J. Makowska, S. Rodziewicz-Motowidło, K. Bagi´nska, J. A. Vila, A. Liwo, L. Chmurzy´nski, H. A. Scheraga, Proc. Natl. Acad. Sci. USA 2006, 103 , 1744–1749. 281. J. Makowska, S. Rodziewicz-Motowidło, K. Bagi´nska, M. Makowski, J. A. Vila, A. Liwo, L. Chmurzy´nski, H. A. Scheraga, Biophys. J . 2007, 92 , 2904–2917. 282. K. Chen, Z. Liu, N. R. Kallenbach, Proc. Natl. Acad. Sci. USA 2004, 101 , 15352–15357. 283. K. Chen, Z. Liu, C. Zhou, Z. Shi, N. R. Kallenbach, J. Am. Chem. Soc. 2005, 127 , 10146–10147. 284. Z. Shi, K. Chen, Z. Liu, A. Ng, W. C. Bracken, N. R. Kallenbach, Proc. Natl. Acad. Sci. USA 2005, 102 , 17964–17968. 285. K. Bagi´nska, J. Makowska, W. Wiczk, F. Kasprzykowski, L. Chmurzy´nski, J. Pept. Sci . 2008, 14 , 283–289. 286. P. Salvador, A. Asensio, J. J. Dannenberg, J. Phys. Chem. 2007, B111 , 7462–7466. 287. B. W. Chellgren, T. P. Creamer, Biochemistry 2004, 43 , 5864–5869. 288. A. M. Firestine, V. M. Chellgren, S. J. Rucker, T. E. Lester, T. P. Creamer, Biochemistry 2008, 47 , 3216–3224. 289. V. Setnicka, J. Hlavacek, M. Urbanova, J. Pept. Sci . 2009, 15 , 533–539. 290. I. Jakli, A. Perczel, J. Pept. Sci . 2009, 15 , 738–752. 291. M. Narita, N. Ohkawa, S. Nagasawa, S. Isokawa, Int. J. Pept. Protein Res. 1984, 24 , 129–134. 292. A. Rath, A. R. Davidson, C. M. Deber, Biopolymers (Pept. Sci.) 2005, 80 , 179–185.

16 ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS Claudio Toniolo and Fernando Formaggio

16.1. INTRODUCTION Peptidomimetics, originally also termed pseudopeptides or peptide surrogates, are peptide analogues that can contain natural (but noncoded) or synthetic amino acids. They are generated by modifying existing peptides to alter their properties (e.g., enzymatic stability, lipophilicity, biological activity). As a result, they are relevant for the development of compounds with novel therapeutic proﬁles. In recent years they have found wide application in medicinal chemistry as biostable, bioavailable, and often potent mimetics of naturally occurring peptides. Initial synthetic efforts were centered on modiﬁcations of the peptide side chains, or involved amino acid additions, deletions, or substitutions only, but more recently the main interest of peptide chemists from academia and industrial laboratories as well have focused mainly on backbone modiﬁcations. Several review articles [1–14] have dealt with the molecular design of speciﬁc, receptor-selective peptidomimetic ligands and with their challenging synthetic issues and intriguing conformational preferences, but none of them has ever treated in detail the electronic circular dichroism (ECD) properties of the chromophores characterizing this class of peptide analogues. In this overview, for the ﬁrst time the results of the published ECD studies of peptidomimetics have been summarized and critically discussed. We have restricted our attention to the most relevant and popular subclass of peptidomimetics, namely those with modiﬁcations in the -NH–CH(R)–CO- backbone. For the ECD properties of unmodiﬁed peptides and poly-α-amino acids in their classical secondary structures, the reader is referred to Chapter 15 in this volume.

Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

545

546

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Oi

Hi

Ci′

Ni Ci

C′i−1

Oi−1

α

Ni+1

Figure 16.1. Intramolecular steric repulsions that H

induce warping of the i → i intramolecularly

R

Hi+1

H-bonded (C5 ) peptide conformation.

16.2. FULLY EXTENDED PEPTIDE CONFORMATION (2.05 -HELIX) The fully extended polypeptide conformation (φi = ψi = 180◦ ) (2.05 -helix) was proposed at an early stage in 3D-structural studies of ﬁbrous proteins. In this form, H-bonding takes place between the N–H groups of one chain and the C=O groups of the chains on either side, thus making a planar sheet held together by interchain H-bonds directed approximately perpendicular to the chain axis. Neighboring sheets are then held together by van der Waals forces. In 1951 Pauling and Corey [15] investigated the possibility of small contractions of the polypeptide chains and proposed precise conformations for parallel and antiparallel pleated -sheet β-forms which better satisfy stereochemical and H-bonding requirements and have chain-repeat lengths nearer to those found experimentally. These authors were also able to show that steric hindrance between adjacent chains prevents the onset of the planar sheet in case the side chain is anything but a H-atom, that is, this conformation could be formed only by poly(Gly). The repeating motif of the fully extended, ﬂat polypeptide conformation is the (i )N–H· · ·O=C (i ) intramolecularly H-bonded form (Figure 16.1). The relative disposition of the two dipoles, Ni –Hi and Ci =Oi , is such that there is obviously some interaction between them. These four atoms, together with the Cαi atom, are involved in a pentagonal cyclic structure. It is for this reason that this conformation is also called the C5 structure. The C5 form was considered in conformational energy calculations and its occurrence in apolar noninteracting solvents was proposed in solution studies using IR absorption and NMR measurements of model peptides [16–18]. Gly derivatives have the highest population of C5 structure if compared to the derivatives of residues carrying a side chain. The inﬂuence of the bulkiness of the side chain can easily be explained by considering the intramolecular nonbonded interactions between the group R and the atoms Hi +1 and Oi −1 , which induce a warping of these asymmetric molecules (Figure 16.1). Unequivocal veriﬁcation of the occurrence of the C5 form was obtained in the crystal state in a few peptides and proteins by X-ray diffraction analysis [18–21]. Among the coded amino acids, Gly was found to be involved in > 99% of the C5 structures, including an unusual stretch of four consecutive Gly residues which therefore forms a short twofold 2.05 -helix. Notably, a variant of the planar C5 form (with Y-conjugation) was reported in an X-ray diffraction investigation of homo-peptides from the natural, but noncoded, achiral Ala (α,β-didehydro Ala) residue [22]. The results of the theoretical and experimental (crystal-state and low-polarity solvents) studies from the Toniolo and Benedetti laboratories, summarized in references 23, and 24, strongly supported the view that the quaternary α-amino acids C α,α -diethylglycine (Deg, Figure 16.2), C α,α -di-n-propylglycine, C α,α -diphenylglycine, and C α,α -dibenzylglycine, all achiral (with Cα,α -symmetrically disubstituted side chains), overwhelmingly tend to adopt the C5 structure. This peptide conformation

547

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

H3C

H2C

CH2

CH3

H3C H2C

C

CH3

H3C H2C

C CO

NH

(CH2)3

Deg

CH3

C CO

NH

(CH2)4

(S)-Beg

NH

CO (S)-Epg

α

Figure 16.2. Chemical formulae for the C -ethylated achiral Deg, and chiral Beg and Epg α-amino acid residues.

Figure 16.3. Average geometrical parameters that characterize the fully-extended, intramolecularly H-bonded C5 structure from a statistical analysis of the known X-ray diffraction structures. (Adapted from reference 23.)

was characterized at atomic resolution (Figure 16.3). The narrowing of the tetrahedral ˚ N· · ·O distance for the N–Cα –C bond angle (τ ) to ∼103◦ and the very short (2.54 A) intramolecular H-bond are the most notable features. In the homo-peptides from these residues the multiple N–H and C=O groups involved in the intramolecularly H-bonded forms are not implicated in any intermolecular H-bonding, possibly because severe steric hindrance from the bulky side chains prevent adjacent peptide chains to approach sufﬁciently. In conclusion, it became evident that the multiple C5 structure (2.05 -helix) becomes stable enough, at least in homo-peptides, when both side-chain Cβ -atoms of the constituent residue are symmetrically substituted (but not interconnected in a cyclic system). In contrast, when both side-chain Cβ -atoms of the constituent residue are unsubstituted, as in α-aminoisobutyric acid (Aib), or when only one of them is substituted, as in chiral C α -methylated residues, then the 310 -(or α-)helix is the largely preferred conformation [24]. However, since 2000 this previously widely accepted dogma does not hold any more. Tanaka and his colleagues [25, 26] reported that the terminally protected, chiral homotetramer Tfa–[(S )-Beg]4 –OEt (Tfa, triﬂuoroacetyl; Beg, C α -n-butyl, C α -ethylglycine, Figure 16.2; OEt, ethoxy) adopts a fully planar 2.05 -helix in the crystal state. This surprising ﬁnding was subsequently supported by Toniolo et al. [27], who studied Epg (C α -ethyl, C α -n-pentylglycine, Figure 16.2) homo-oligomers in the crystal state. The X-ray diffraction structure of the double C5 conformation of the dipeptide Tfa–[(S )Epg]2 –OtBu (OtBu, tert-butoxy) is shown in Figure 16.4. At this point, the unavoidable conclusion is that an asymmetric Cα,α -disubstituted Gly residue, bearing two different

548

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 16.4. X-Ray diffraction structure of Tfa–[(S)-Epg]2 –OtBu. Intramolecular N–H· · ·O=C Hbonds are represented by dashed lines; F and O atoms are larger and black and gray, respectively, while N and H atoms are white. (Adapted from reference 27.)

side chains each with at least two carbon atoms, may represent an appropriate building block for the construction of a 2.05 -helix. These investigations have also shown that the integrity of this type of helical structure is preserved only in solvents of low polarity and provided that the C-terminal protection of the peptide sequence is devoid of any H-bonding donor group. It is evident that, based on the above discussed data, an analysis of the experimental ECD properties of chiral 2.05 -helices has become an accessible task. To this end, and toward the in-depth elucidation of the ECD spectrum by a theoretical approach, collaborative efforts are currently being conducted [28]. An important requirement for a successful experimental study in solution is that the solubility of the chiral oligopeptide would be sufﬁciently enhanced in UV-transparent, low-polarity solvents, as compared to that of the peptides prepared and investigated to date. Unfortunately, recent attempts to obtain reliable ECD spectra using the polar solvent 2,2,2-triﬂuoroethanol (TFE) did not furnish appreciable results [25, 29].

16.3. POLY-N(ALKYL)-α-AMINO ACIDS Poly(Pro) and poly-N (alkyl)-α-amino acids, abbreviated here as poly[N (R)AA], are characterized exclusively by tertiary peptide bonds. In this sense, the latter may be considered acyclic analogues of poly(Pro). The least congested polypeptides of this family, poly[N (alkyl)Gly] or peptoids, will be discussed separately (next section). Among the ECD properties of the poly[N(R)AA] from chiral α-amino acids, only those of polyN (methyl)-α-amino acids have been investigated (Figure 16.5a). Notably, N -methylation is an important modiﬁcation of the peptide bond [30–33]. It commonly occurs in natural peptides from plants, marine sources, and a variety of microorganisms. Several

549

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

(a)

R N

CH *

CO n

R = CH3

MeAla

R = (CH2)2-COOC2H5

MeGlu(OEt)

CH3 (b)

Figure 16.5. (a) General formulae of the N-methylated homo-poly-α-amino acids discussed here and (b) X-ray diffraction structure of H–(L-MeAla)6 –OH (O atoms larger and gray, N atoms white). (Adapted from reference 33.)

of these compounds possess interesting antibiotic, antitumor, and immunosuppressant activities. Synthetic, N -methylated analogues of bioactive peptides often exhibit high stability to enzymatic degradation. N -methylated peptides are also potent inhibitors of β-sheet structure and, as a result, of amyloid formation. In their pioneering work, Mark and Goodman [34, 35] demonstrated by energy calculations that the homo-polymer poly(l-MeAla), where MeAla is N -methylated Ala, is severely conformationally restricted and is preferentially folded in a right-handed, approximately threefold helix with all peptide bonds in the trans conformation. This unusual, semi -extended, nonintramolecularly H-bonded, secondary structure has backbone φ, ψ torsion angles of about −150◦ , 70◦ . The main factor favoring this conformation for poly(l-MeAla) is the severe steric repulsion between the C α -methyl and N -methyl groups that would occur in other, more canonical peptide conformations. These conclusions were essentially conﬁrmed by subsequent works from other laboratories [36, 37]. Madison and Schellman [38] calculated the ECD spectrum of the most stable, alltrans, conformation for the (l-MeAla)20 homo-oligomer. They found two bands of approximately the same intensity at 221 nm (negative), associated with the n → π * transition of the peptide chromophore, and at 193 nm (positive), associated with a complex pattern of π → π * transitions (Figure 16.6). The crossover point was predicted at 207 nm. Goodman and co-workers [39, 40] reported the experimental ECD spectrum of poly(l-MeAla) in the structure-supporting solvent TFE (Figure 16.7). It shows a broad negative Cotton effect centered at 223 nm followed by a slightly less intense, positive Cotton effect at 192 nm (the crossover point is seen at 204 nm). Thus, the theoretical and experimental ECD results ﬁt nicely. Moreover, their ﬁndings on poly(l-MeAla) differ somewhat from those obtained with poly(l-Pro) [38] in that no evidence for an all-cis form was detected upon changing solvent polarity. Furthermore, a few ECD studies were performed in triﬂuoroacetic acid (TFA), due to the generally poor solubility of poly(l-MeAla). However, the conclusions from these studies are doubtful, as TFA was many years later shown to easily hydrolyze tertiary peptide bonds [32, 41]. Arvidsson and co-workers [33] investigated the ECD properties of the monodisperse homo-oligomer H–(l-MeAla)6 –OH in methanol and aqueous solutions. The spectra parallel closely those described for the corresponding polymer in TFE [39, 40]. In addition, Peggion, Goodman, and their colleagues reported conformational energy calculations and ECD data for poly[l-MeGlu(OEt)] [where MeGlu(OEt) is N -methyl, γ -ethyl glutamate] [42]. Their ﬁndings agree well with those for poly(l-MeAla).

550

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

20

[θ] × 10–3 (deg × cm2 × dmol–1)

10

0

–10

–20

Figure 16.6. Computed ECD spectrum of the 200

220

240

Wavelength (nm)

260

(L-MeAla)20 homo-oligomer. (Adapted from reference 38.)

However, in contrast to this remarkably closely ﬁtting scenario, quite surprisingly, in their recent X-ray diffraction analysis of the two homo-oligomers H–(l-MeAla)5,6 –OH (Figure 16.5b) Arvidsson and co-workers [33] unambiguously showed that these peptides do not form a helix with all φ, ψ torsion angles of (−135◦ , 65◦ ), but this conformation strictly alternates with that characteristic of a distinct semi -extended conformation (−65◦ , 140◦ ). The former backbone conformation generates a torsion angle between the two MeAla methyl groups [(methyl)C–N–Cα –Cβ ] of about 70◦ , while the torsion angle in the latter conformation, which resembles that of poly(l-Pro) PPII [38], is reduced to approximately 20◦ . At this point, there is a clear need for a theoretical investigation of the ECD properties of this novel poly(l-MeAla), sequentially bis-semi -extended, 3D-structure.

16.4. POLY-PEPTOIDS Peptoids (Figure 16.8a) are polymers of Nα -substituted Gly residues [43]. Therefore, they are generally achiral, unless the nitrogen atom is substituted with a group (R) containing an asymmetric center (Figure 16.8b). Peptoids can be deﬁned as α-peptide mimics in which the side chain R is linked to the α-nitrogen instead of the α-carbon atom. They also represent a variant with respect to the chiral poly-N (alkyl)-α-amino acids, poly[N(R)AA],

551

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

40

[θ] × 10–3 (deg × cm2 × dmol–1)

20

0

–20

Figure 16.7. ECD spectrum of

–40 200

220

240

260

Wavelength (nm)

(a)

O

R

H

H

O

N

*

*

N

O

H O

H

R

H

R*

H

chiral peptide

* O

H O

(Adapted from reference 39.)

H

R N

H

H

N N O

R (b)

H3C

chiral peptoid

N *

H

H

*

R

Nsch C6H11

Figure 16.8. (a) Chemical formulae of chiral peptide and peptoid chains. Starred

O H3C

H

homo-poly[L-MeAla] in TFE solution.

carbon atoms or R groups indicate chiralities. (b) Chemical formulae of the two α-chiral, aliphatic, N-substituents investigated: Nsch

H Nssb

is N-(S)-(1-cyclohexylethyl)glycine and Nssb is N-(S)-(sec-butyl)glycine.

552

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

where AA is any α-amino acid other than Gly, discussed in the preceding section. These compounds are an attractive scaffold for biological applications (e.g., as antimicrobials) because they can be generated using a straightforward, modular synthesis that allows the incorporation of a wide variety of functionalities [44, 45]. Peptoids have been evaluated as tools to study biomolecular interactions, and they also hold signiﬁcant promise for therapeutic applications due to their enhanced proteolytic stabilities and increased cellular permeabilities relative to α-peptides. The unsubstituted α-carbon of peptoids should allow almost unhindered rotation about the φ, ψ torsion angles. In addition, both cis and trans conformers of the tertiary amide bonds should be accessible at room temperature. However, the substituent on the α-nitrogen is expected to confer 3D-structural properties on peptoids that might seriously limit their ﬂexibility. Here, we will discuss only peptoids with aliphatic N-substituents, in that aromatic chromophores can affect the ECD spectral shapes. Also, we will not treat hybrids of α-amino acids and Nα -substituted Gly residues. The lowest-energy conformation for the simplest model compound of the aliphatic peptoid subclass, Ac–Sar–NMe2 (Ac, acetyl; Sar, sarcosine or N -methylglycine; NMe2 , dimethylamino) corresponds to φ = ±90◦ , ψ = 180◦ for both the trans and cis Ac–Sar amide conformers [44]. In the crystal state, the homo-pentapeptoid H-(Nrch)5 -NH2 [Nrch, N -(R)-(1-cyclohexylethyl)glycine] adopts a left-handed helical conformation with repeating cis amide bonds (Figure 16.9) [46]. The periodicity of this helix is approximately three residues per turn. The C=O groups are aligned with the helix axis. The handedness of the helix is governed by the chirality of the Nα -substituent. The backbone φ, ψ torsion angles are similar to those observed for the PPI poly(Pro) helix, but with opposite signs (obviously, the signs depend on the Nch peptoid chirality). The bands in the ECD spectrum of the α-chiral aliphatic homo-oligopeptoid H–(Nsch)5 –NH2 [Nsch, N -(S )-(1-cyclohexylethyl)glycine] in CH3 CN are relatively weak, reﬂecting only a partial helical ordering in solution [46, 47]. However, those of the longer homo-oligomers investigated—for example, H–(Nsch)15 –NH2 (Figure 16.10)—show a distinct positive maximum at 210 nm, and two stronger negative maxima at 195 nm and 225 nm, respectively. The ECD bands are spectral characteristics that are typically associated with those of the PPI poly(Pro) helix [48]. Therefore, these

Figure 16.9. X-ray diffraction structure of the Nrch homo-pentapeptoid amide (left, perpendicular; right, parallel to the threefold helix axis). (Adapted from reference 46.)

553

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

4

[θ]R × 10–3 (deg × cm2 × dmol–1)

2

0

–2

–4

–6

–8

Figure 16.10. ECD spectrum of the Nsch

–10 180

200

220 Wavelength (nm)

240

homo-pentadecapeptoid in CH3 CN solution. (Adapted from reference 46.)

ECD ﬁndings are consistent with the results from the crystallographic analysis. The shapes of the ECD spectra and their steady increase in intensity with lengthening of the main chain are qualitatively similar for the Nssb [N -(S )-(sec-butyl)glycine] subclass of aliphatic sidechain containing peptoid homo-oligomers. This latter observation points to a more ordered helical fold for the longest oligo-peptoids. ECD also demonstrated that this helical structure is quite stable upon heating. This result is consistent with a 3D-structure predominantly stabilized by steric repulsion rather than by intrachain H-bonding. NMR data on these foldameric series provide strong evidence that the tertiary amide cis –trans isomeric ratio is main-chain length dependent and that at the n = 15 level the cis-amide family of dynamically interconverting conformers is overwhelmingly populated. Introduction of the achiral, water-solubilizing NLys [N -(4-aminobutyl)glycine] and NArg [N -(3-guanidinopropyl)glycine] guest residues into the host Nssb or Nsch homopeptoid chains allowed ECD measurements to be performed in aqueous buffer solutions [49–51]. Under these conditions, the ECD spectra are weak, reminiscent of that of a peptide unordered conformation. However, in methanol (MeOH) solution and in membranemimetic vesicles the curves are more intense and resemble those of PPI poly(l-Pro). Not surprisingly, the intensity of the ECD curves is proportional to the percentage of chiral Nα -substituted residues.

554

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

R peptide

CO

NH

CH

R CO NH

R depsipeptide

CO

NH CH

CH

CO NH

R′ CO

O

CH

R

R CH

CO NH

R CO NH

CH

CH

CO

R′ CO

O

CH CO

Figure 16.11. Chemical formulae of peptide and depsipeptide (the latter with strictly alternating amide-ester groups) chains.

16.5. POLY-DEPSIPEPTIDES In the classical deﬁnition, a poly-depsipeptide chain is generated by a strict alternation of an α-amino acid and an α-hydroxy acid in a linear sequence (Figure 16.11). Therefore, this kind of sequential polymer is characterized by a repeating dyad with an amide unit followed (or preceded) by an ester unit. Subsequently, this terminology was inappropriately extended to any backbone-modiﬁed peptide that incorporates one or more α-hydroxy acid building blocks. It is worth noting that increasing effort is currently being devoted to the conformational analysis of a variety of compounds characterized by the presence of α-(or β-)hydroxy acids. More speciﬁcally, these compounds include (a) oligo- and polyesters as biodegradable and biocompatible materials [52–55] and (b) depsipeptides and depsiproteins, in this broader sense, to mimic naturally occurring ion carriers [56, 57] or to check the inﬂuence of speciﬁc H-bonds on peptide or protein bioactivity and conformation [58–62]. In this section, we will discuss exclusively the chirospectroscopic properties of classical depsipeptides. As stated above, poly-depsipeptides represent a combination of polypeptides and poly-α-esters (for the 3D-structural and chirospectroscopic properties of polypeptides, see Chapter 15, this volume). The 3D-structure of the simplest chiral poly-α-ester, poly(lLac) where Lac is lactic acid, was investigated by energy computations and in the crystal state [63–68]. The results point to a regular or slightly distorted, right-handed, threefold helix. However, according to the optical rotatory dispersion data of Goodman and D’Alagni [69], this homo-polymer, the chemical structure of which rules out any possibility of intramolecular H-bonding, may not have a helical conformation in solution. The UV absorption and related ECD spectra of ester model compounds, oligo- and poly(lLac), and poly-β-esters exhibit a solvent-dependent band near 210 nm, arising from the n → π * transition of the –COOR chromophore [52, 53, 55, 69–73]. The X-ray diffraction structures of the l-Lac dimer and trimer were recently solved [74]. The only available X-ray diffraction structure of a classical depsipeptide, Z–(Aib–Hib)2 –Aib–OMe [Z, benzyloxycarbonyl; Hib, α-hydroxyisobutyric acid; OMe, methoxy] [59] (Figure 16.12), strongly supports the view that the α-hydroxy acid Hib guest unit is easily incorporated into the 310 -helical structure typical of the host α-amino acid Aib chain (in this depsipeptide both building blocks are characterized by two helicogenic gem-methyl groups at the Cα -atom). The resulting helical conformation, in which the scheme of consecutive intramolecular H-bonds is interrupted at alternate positions, is classiﬁed as a β-bend ribbon spiral. Two preferred depsipeptide secondary structures were identiﬁed from energy calculations on the poly-depsipeptide (l-Ala–l-Lac) [63, 64]. The most stable helix, termed R = 13 (Figure 16.13), has ϕ, ψ torsion angles (−65◦ , −35◦ for l-Ala; −63◦ , −47◦ for l-Lac) similar to those of the standard right-handed polypeptide α-helix. The amide and ester

555

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

Figure 16.12. X-ray diffraction structure of the oligo-pentadepsipeptide Z–(Aib–Hib)2 –OMe. Intramolecular C=O· · ·H–N H-bonds are represented by dashed lines, O atoms are larger and gray, and N and H atoms are white. (Adapted from reference 59.)

O O C

C

C

C

N

CH

C O

H

C C

O C

N

H O

C

C

H

O

C N

H

O H O

C

O

C C H

C

C N C

N C

C O

C

C

H C

C H

O

O

C C

H

O

H H

O

C

N

C

C

O

H

H

Figure 16.13. Left, the R = 10 helix, and right, the R = 13 helix, calculated for the polydepsipeptide (L-Ala– L-Lac). (Adapted from reference 64.)

C=O bonds are approximately parallel to the helix axis. Intramolecular H-bonding occurs between ester carbonyl oxygen atoms and amide NH atoms that are separated by three α-carbons. The pseudo-ring thus generated is characterized by 13 atoms. The other helix, termed R = 10 (Figure 16.13), is left-handed with backbone torsion angles 51◦ , −94◦ for l-Ala and −144◦ , 30◦ for l-Lac. The amide C=O and N–H bonds are

556

8 4 0 –4 –8 160 180 200 220 240

12

[θ] × 10–3 (deg × cm2 × dmol–1)

[θ] × 10–3 (deg × cm2 × dmol–1)

[θ] × 10–3 (deg × cm2 × dmol–1)

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

8 4 0 –4 –8 160 180 200 220 240

8

4

0

–4

–8 160 180 200 220 240

Wavelength (nm)

Wavelength (nm)

Wavelength (nm)

(a)

(b)

(c)

Figure 16.14. Theoretical ECD spectra of the poly-depsipeptide (L-Ala– L-Lac) R = 13 helix (a), R = 10 helix (b), and the unordered conformation (c, dashed line). The full line in (c) is the spectrum of a mixture of 50% unordered conformation and 50% R = 10 helix. (Adapted from reference 72.)

roughly parallel to the helix axis, whereas the ester C=O bonds are approximately perpendicular. Intramolecular H-bonds are observed between neighboring amide groups, which produce pseudo-ring structures formed by 10 atoms. Comparison of the theoretical (Figure 16.14) and experimental (Figure 16.15) ECD properties suggests that poly(l-Ala–l-Lac) is only about 50% ordered in solution (even in solvents of low polarity) [75–77]. The experimental ECD spectra of two other poly-depsipeptides, poly(l-Val–l-Lac) and poly(l-Ala–l-Hiv) where Hiv is α-hydroxyisovaleric acid, are similar. Not surprisingly, the calculated ECD spectrum for the R = 13 helix is close in shape to the ECD curves of polypeptides in the right-handed α-helix structure. However, it is clearly very different from the experimental poly-depsipeptide ECD spectra. Apparently, these polymers are not folded in the right-handed R = 13 helical conformation to any signiﬁcant extent in solution.

16.6. α,β-DIDEHYDRO-α-AMINO ACID-BASED POLY-PEPTIDES α,β-Didehydro(unsaturated)-α-amino acids (Figure 16.16), usually represented by the notation AA, have been frequently found in naturally occurring peptides of microbial origin and in a limited number of proteins [78–81]. They are also constituents of an important class of polycyclic peptide antibiotics (lantibiotics). The presence of AA in peptides alters lipophilicity and bioactivity as well as increases resistance to hydrolysis by proteolytic enzymes. AA residues have been incorporated by chemical synthesis in a variety of natural peptide sequences to obtain highly active analogues. This modiﬁcation has become a useful method to study structure–function relationships in bioactive peptides. The accumulation of three functionalities (the amide α-NH- and α-CO- groups, along with the C=C double bond) at position Cα of a AA residue has remarkable stereochemical and spectroscopic effects [82–86]. In particular, the presence of an sp 2 hybridized carbon (Cα ) atom in the backbone, the altered electronic distribution

557

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

9 2

6

[θ] × 10–3 (deg × cm2 × dmol–1)

1

3

0

3 –3

Figure 16.15. ECD spectra of the

–6

poly-depsipeptides (L-Ala– L-Lac) (1),

200

220

240

Wavelength (nm)

R1

R C

R=H

R = C6H5 R1 = H

C HN

R1 = H

CO

R=H

260

(L-Val– L-Lac) (2), and (L-Ala– L-Hiv) (3) in tetrahydrofuran solution at −50◦ C. (Adapted from reference 75.)

ΔAla ΔZPhe

R1 = C6H5 ΔEPhe

Figure 16.16. Chemical formulae of representative α,β-didehydro()α-amino acid residues, including the two Phe E- and Z- diastereoisomers.

(Y-conjugation) caused by the α –β π -system, and the change in the side-chain rotamer population all contribute signiﬁcantly to the preferred conformation of the peptide main chain. The 3D-structural ﬂexibility of the -peptide backbone as well as of the speciﬁc side chain of the AA residue is restricted on account of the double bond between the Cα and Cβ atoms. Although conjugation requires an extended conformation, the bulkiness of the side chain may play a signiﬁcant role in the overall conformation of the AA residue. In the -NH–C(=CRR1 )–CO- class (with the R and R1 groups of protein amino acids) all residues exhibit cis (Z )–trans (E ) isomerism around the C=C double bond, except Ala (R = R1 = H) (Figure 16.16) and Val (R = R1 = CH3 ). An example of this type of diastereoisomerism is shown in Figure 16.16 for Phe. In Z Phe, the C=O group is in the trans position with respect to the phenyl group, while in the E Phe it is in the cis position. Most of the conformational studies on AA-rich peptides were carried out on Phe peptides and particularly on peptides based on Z Phe because of the relative ease of

558

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

synthesis of this isomer and the fact that most preparation procedures lead to the Z Phe conﬁguration. The average bond lengths and bond angles of AA residues were calculated from a variety of crystal structures. The shortening of N–Cα and Cα –C single bond lengths and the elongation of the C =O double bond in Phe may be due to partial conjugation of the Cα =Cβ double bond with the adjacent peptide bonds. The intramolecular steric clash between Cδ H and NH of the Phe residue results in a remarkable opening up of the bond angles Cα –Cβ –Cγ and N–Cα –Cβ . From simple model building studies it was observed that the most favorable conformations for the Phe residue are (φ, ψ) ≈ (±60◦ , ±30◦ ) and (±60◦ , ∓150◦ ). Theoretical conformational studies conﬁrmed these ﬁndings and suggested that Phe strongly favors β-turns and 310 - (or α-) helical conformations. The near-UV absorption spectrum of peptides containing Phe is characterized by an intense, conformationally useful absorption band at about 280 nm, which has been assigned to a charge-transfer electronic transition from the electron donating styryl groups to the electron-accepting carbonyl group in the Phe moiety [87, 88]. The chromophoric system of the residue, therefore, is essentially the cinnamic moiety C6 H5 –C=C–C=O. The peptides containing two Phe residues show an analogous band, centered around the same wavelength. Not unexpectedly, its intensity is approximately twice as that of their mono-unsaturated counterparts. In addition, a rather strong band at about 220 nm dominates the far-UV absorption spectrum of Phe peptides. It is clear that this latter spectroscopic property, unfortunately, is potentially confusing in terms of a correct peptide conformational assignment. The reasons for the small differences, if any, between the absorptions of the Phe Z- and E-diastereoisomers are not well understood. Induced ECD proved to be an excellent tool to determine the solution conformation of peptides based on the achiral Phe residue, especially in detecting the screw sense of the helices formed. It is principally for this reason that the number of publications on this topic is huge [87,89–126]. The Z Phe-containing peptides show different ECD curves, depending on the main-chain length and on the position in the sequence and number of the Z Phe residues. ECD spectra of mono Z Phe compounds exhibit only low-intensity bands in the near-UV region (Figure 16.17). In general, the shape and sign of these bands are strongly affected by the nature of the nearby chiral α-amino acid. ECD spectra of Z Phe tri- and longer peptides exhibit a broad, relatively intense band at about

[θ] × 10–4 (deg × cm2 × dmol–1)

20 10 1 0 –10 2 –20

Figure 16.17. ECD spectra of the dipeptide

–30

Ac–Z Phe– L-Ala–OH (1) and the tripeptide Ac–(Z Phe)2 – L-Ala–OH (2) in MeOH solution.

240

280 Wavelength (nm)

320

The latter peptide is characterized by two consecutive Z Phe residues. (Adapted from reference 89.)

559

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

[θ] × 10–3 (deg × cm2 × dmol–1)

20

0

–20

Figure 16.18. ECD spectrum of the

240

280 Wavelength (nm)

320

tetrapeptide Ac–(Z Phe)3 – L-Ala–OH in MeOH solution. This peptide is characterized by three consecutive Z Phe residues. (Adapted from reference 90.)

280 nm, corresponding to the main absorption band of the cinnamoyl-like chromophore. The varying intensities of this band are indicative of the different propensities of the peptides to fold into a β-turn. This conclusion is conﬁrmed by the solvent dependence of the intensity of this band. Peptides containing two or more (up to eight, of which 2–4 are consecutive) Z Phe residues show a couplet of intense bands with opposite signs at 300 and 260 nm and a crossover point at ∼280 nm (Figures 16.17–16.22). This ECD pattern is typical of exciton splitting due to the dipole–dipole interactions between the Z Phe chromophores and is a strong indication that the two Z Phe residues are placed in a mutual, ﬁxed disposition within the structure of the molecule, generally a 310 -helix. The (−+) signs (in the direction of decreasing wavelengths) of the couplet correspond to a 310 -helix with a right-handed screw sense, while the (+−) couplet is assigned to the helix with the left-handed screw sense. Comparison of ellipticity of the peptides containing two and three Z Phe residues show similar values, in spite of the presence of an extra Z Phe in the latter case. This should actually be the case if the 310 -helical conformation of the peptide would terminate with a type I, instead of a type III, β-turn. It may also be a result of the change of the helix sense of the last residue or unwinding of the helix, a common observation at the helix C-terminus [127]. In the peptides where the chiral residue(s) are located in internal

560

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

30 1

[θ] × 10–3 (deg × cm2 × dmol–1)

2 15 3

0 4

–15

Figure 16.19. ECD spectra of the pentapeptide Boc– L-Ala–Z Phe–Gly–Z Phe– L-Ala–OMe, where Boc is tert-butyloxycarbonyl, in CH3 CN (1),

–30 240

320

280

solutions. (Adapted from reference 92.)

Wavelength (nm)

L-Leu

CH2 Cl2 (2), MeOH (3), and 1,1,1,3,3,3-hexaﬂuoroisopropanol (HFIP) (4)

L-Val

[θ] × 10–3 (deg × cm2 × dmol–1)

40

20

L-Pro L-Ala

0

–20

D-Pro no chiral additive

–40 240

Figure 16.20. ECD spectra of the achiral

D-Leu

280 Wavelength (nm)

320

nonapeptide H–Gly–(Z Phe–Aib)4 –OMe in the presence or absence of added Boc–Xxx–OH (each Xxx residue is indicated). (Adapted from reference 108.)

position(s) of the sequence, the helix is right-handed, while peptide esters (but not peptide amides) where the single chiral residue is at the C-terminus are found to adopt the lefthanded helical sense. If the single chiral residue is, however, at the N-terminus, the peptides tend to adopt 310 -helix conformations of both screw senses in the crystal state. In this case, the ECD results are consistent with the presence of right- and left-handed conformers also in solution, with a prevalence of the more stable right-handed helix. Interestingly, in a few peptides a reversible screw sense inversion of the 310 -helix was detected, depending on solvent (Figures 16.19 and 16.21) and temperature conditions. An ECD band was also observed at 320 nm in some of these compounds [90, 118] that also changes sign with changing solvent polarity (Figure 16.21). One reason for the

561

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

1 2 3 4 5 6 7 8 9

[θ] × 10–3 (deg × cm2 × dmol–1)

300 1 200

2 3

100

5

0

100:0 90:10 80:20 70:30 60:40 50:50 40:60 30:70 20:80

4

9 6

–100

7 8

Figure 16.21. ECD spectra of the decapeptide

–200 260

240

280

300

320

[θ] × 10–3 (deg × cm2 × dmol–1)

Boc– L-Ala–(Z Phe)4 – L-Ala–(Z Phe)3 –Gly–OMe in CHCl3 –MeOH solvent mixtures (curve 1: 100% CHCl3 ). (Adapted from reference 118.)

Wavelength (nm)

10

340

Z

E

0

–10

Figure 16.22. ECD spectra of the hexapeptides Boc–Gly–E Phe– L-Phe–Gly–E Phe– L-Phe–OH (E)

240

280 Wavelength (nm)

320

and Boc–Gly–Z Phe– L-Phe–Gly–Z Phe– L-Phe–OH (Z) in CHCl3 solution. (Adapted from reference 115.)

occurrence of this band may be the weak electronic transition polarized along the short axis of the benzene ring. This contribution to the ECD spectrum suggests that the benzene ring is not free to rotate, as expected, owing to the presence of the Cα =Cβ double bond. Its strong intensity also indicates that the phenyl ring is restricted to a preferred chiral disposition. The usefulness of the Z Phe chromophore for conformational analysis was further corroborated by a series of studies on interactions of achiral, Aib/Z Phe-based, helical peptides with external chiral molecules (“noncovalent chiral domino effect”) [97, 99–114] (Figure 16.20). The ECD properties of the H–L–Glu(l-Lys)–Z Phe–OH amphiphilic dipeptides self-assembled into nanovesicles were investigated [121]. Two hexapeptides, each characterized by two Phe residues (either in the E - or Z conﬁguration) at positions 2 and 5 in the sequence, were studied by ECD spectroscopy [115] (Figure 16.22). Interestingly, it was found that in solvents of low polarity, where their conformational freedom is restricted and a rigid helical structure is attained, the two ECD spectra are almost mirror images. ECD spectra were calculated in the far-

562

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

[θ] × 10–4 (deg × cm2 × dmol–1)

20

10

0

–10

Figure 16.23. Theoretical near-UV ECD spectrum of the dodecapeptide

–20 240

280 Wavelength (nm)

320

Boc–(L-Ala–Z Phe–Aib)4 –OMe in the right-handed 310 -helix with φ, ψ torsion angles −54◦ , −28◦ . (Adapted from reference 99.)

and near-UV regions for helical, Z Phe-containing, peptides [99, 100]. The simulations were performed for various conformers that differ in helix (310 or α) type, helix screw sense, and Z Phe side-chain orientation. An example is reported in Figure 16.23. The theoretical ECD-peptide conformation relationships provided useful guidelines for determination of the helix sense in the Z Phe peptides and for the estimation of their statistically averaged conformations in solution. Other α,β-didehydroaromatic residues, the ECD properties of which were studied both experimentally and theoretically, include α,β-didehydro-(1-pyrenyl)alanine and α,β-didehydro-(4,4 -biphenyl)alanine [104, 114]. The observed ECD curves were interpreted on the basis of the exciton chirality method. The ECD spectra of the terminally blocked dipeptides Ac–l-Pro–l-Xxx–NHMe (NHMe, methylamino), where Xxx is an aliphatic α,β-unsaturated residue, i .e. Z Leu, Val, Z Abu (Abu, α-aminobutyric acid), and E Abu were recorded in different solvents [125, 128]. The curves, although not assignable to any common spectral class, must be attributed to peptides in a preferred type II β-turn conformation as determined for these compounds by use of other spectroscopic techniques. The spectra exhibit a remarkable solvent dependence and suggest an unordered conformation in aqueous solution. In contrast to the aliphatic AA mentioned above, the simplest AA of this subclass, Ala, is known to overwhelmingly prefer the ﬂat, fully extended (C5 ), Y-conjugated conformation as established by solution and X-ray diffraction analyses of its homo-oligomers [22]. This is the reason why the Nα -protected Ala/chiral α-amino acid dipeptide amides investigated by ECD do not fold into a β-turn conformation [125, 127, 129]. It is worth mentioning that the UV absorption curves of the aliphatic AAcontaining peptides do not show separate bands for amide and C=C double bond electronic transitions. As a result, their ECD spectra do not exhibit any conformationally informative Cotton effect above 250 nm, in contrast to the properties of the aromatic AA-containing peptides.

16.7. POLY-β-PEPTIDES Among peptide foldamers, poly-β-peptides have special appeal because β-amino acids represent the next homologs of α-amino acids in the “backbone space” [130–135] (Figure 16.24). β-Amino acids are either achiral (β-alanine) or chiral. In the latter

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

(a)

NH

CH2

(b)

CH2

CO

(d)

Figure 16.24. Chemical formulae of (a) the acyclic, * HN

CH3 NH

(c)

CH2

CH *

CO

CH *

CH2

CO

unsubstituted, achiral residue β-alanine; (b, c) the acyclic, mono-substituted, chiral β 2 -residue β-aminoisobutyric acid (βAib) and β 3 -residue β-aminobutyric acid (βAbu), respectively; (d, e) the

(e) * HN

CH3 NH

* CO

* CO

cyclic, di-substituted, chiral β 2,3 -residues 2-aminocyclopentanecarboxylic acid (ACPC) and 2-aminocyclohexanecarboxylic acid (ACHC), respectively.

compounds, the chiral center can be found at the α-carbon (near the CO group; β 2 AA) or at the β-carbon (near the NH group; β 3 AA) or at both carbons (β 2,3 AA). The β 2,3 amino acids, which bear two chiral atoms, give rise to four diastereoisomers. Similarly to α-peptides, β-peptides are based on amide groups able to form stabilizing intramolecular C=O· · ·H–N H-bonds. A large body of work has laid a sufﬁciently solid conformational platform for spectroscopic investigations of β-peptides. Early studies of poly-β-peptides showed that these compounds are able to fold into helical structures, although details of the helix geometries were not elucidated [136–139]. Moreover, sheet structures were also proposed. An explosive breakthrough in this ﬁeld came from the papers originating from the laboratories of Gellman and DeGrado and their collaborators, initially in the late 1990s [131,140–150], followed by those from Seebach and co-workers [133, 134, 151–158]. These articles include a huge amount of experimental ECD work. Results of theoretical ECD analyses have also been reported [140, 152, 159–162]. The synthesis of monodisperse oligomers of deﬁned sequence enabled highresolution 3D-structural studies of β-peptides based on the conformationally constrained cyclic β-amino acids (S , S )- or (R, R)-trans-2-aminocyclohexanecarboxylic acid (ACHC) and (S , S )- or (R, R)-trans-2-aminocyclopentanecarboxylic acid (ACPC). The former compounds adopt the 14-helix conformation in the crystal state as well as in organic solvents. Also, series of β-peptides prepared from acyclic residues with a diverse collection of side chains fold into this type of helical structure. Depending on the chirality of the β-amino acids, either the left-handed or the right-handed 14-helix is generated. β-Peptides rich in β 3 -amino acids derived from naturally occurring (S )-amino acids adopt left-handed 14-helices. The 14-helix (Figure 16.25) is stabilized by H-bonding between an Ni –Hi group and a Ci +2 =Oi +2 group, forming a succession of 14-membered pseudo-rings. This 3D˚ versus 2.2 A ˚ radius) and longer (1.56 A ˚ versus 1.50 A ˚ structure is slightly wider (2.7 A rise per residue) than the α-peptide α-helix. While the α-helix has a 3.6 residue repeat, the 14-helix repeat is approximately every three residues (therefore, the alternative notation 314 -helix is often used), which positions the side chain of each residue directly atop one another along one face of the helix. The φ, θ , ψ backbone torsion angles are about −135◦ , 60◦ , −140◦ . The amide C=O and N–H groups project toward the N- and Ctermini, respectively, resulting in a net dipole opposite to that of the α-helix. The ECD spectra of several β-peptides that adopt the 14-helix, as determined by NMR or crystallography, show a band near 195 nm and a band of opposite sign near 215 nm [131, 134, 141, 146, 148–150, 155, 163, 170, 171, 179, 181, 184] (Figures 16.26 and 16.27). The magnitude of the ellipticity at 215 nm varies somewhat from peptide to peptide. However, it is possible that some or all of these peptides are not fully helical. NMR spectroscopy would not be sensitive to a small amount of nonhelical structure, so

563

564

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 16.25. Left, the idealized 14-helix of the (S, S)-trans-ACHC homo-deca-β-peptide; center, the idealized α-helix of (L-Ala)10 ; right, the idealized 12-helix of the (S, S)-trans-ACPC homo-deca-β-peptide, as viewed along (top) and perpendicular to (bottom) the helix axis. Intramolecular C=O· · ·H–N H-bonds are represented by dashed lines, O atoms are larger and gray, N and H atoms are white.

long as it is in rapid exchange with the helix conformation. Consistent with this suggestion, the ellipticity is greatest for those peptides with the most conformationally restricted amino acids, which lead to minimal fraying of the ends of the helices. The contribution of the amide π → π * transition to the ECD spectrum of the 14-helix was calculated. Because of excitonic coupling, this transition is split in two orthogonally polarized bands at 194 and 204 nm. The higher-energy band is in good agreement with experiment. The observed value of 214 nm for the lower-energy band probably reﬂects the presence of an overlapping amide n → π * transition, centered at a slightly longer wavelength. The intensity of the ECD spectrum of the α-helix is known to depend on main-chain length, becoming more intense as the helix is lengthened. Similar behavior was found for the 14-helix. The ECD spectra of many 10- to 15-residue-long peptides, which were designed to adopt a 14-helical conformation, are more intense than those of their shorter counterparts. For example, the ellipticities of a series of amphiphilic β-peptides were examined in the presence of micelles, which strongly stabilize the 14-helical conformation. Their mean residue ellipticities increase in a length-dependent manner. Systematic conformational searches and molecular dynamics calculations of the ACPC versus the ACHC β-amino acid revealed inherent preferences for different helical conformations [131–135, 140,142–145, 147, 166, 175, 176]. The cyclohexyl ring of ACHC stabilizes the θ torsional angle to a value near ±60◦ , which speciﬁcally stabilizes the 14-helical conformation. The smaller ring size of ACPC biases θ toward larger values, generating a different helical form, the 12-helix, as the most favorable conformer (Figure 16.25). The 3D-structure of the 12-helix is stabilized by a series of H-bonds between an amide C=O group at position i and an amide N–H group at position i +3 in the sequence. The helix has approximately 2.5 residues/turn and shows the same polarity as the α-helix, with the amide N–H groups exposed from the N-terminus of the helix.

565

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

8

[θ]R × 10–3 (deg × cm2 × dmol–1)

4

0

–4

–8

Figure 16.26. ECD spectra of the 14-helical –12 200

220 Wavelength (nm)

240

H–(ACHC)n –NH2 (n = 5, dashed line; n = 6, full line) homo-β-oligopeptides in MeOH solution. All ACHC residues have the (S, S)-trans conﬁguration. (Adapted from reference 163.)

The ability to switch between two completely different β-peptide helices by relatively modest alteration of residue structure calls attention to a signiﬁcant difference between α-amino acids and β-amino acids as building blocks. The chemist can exert much greater control over the intrinsic secondary structural propensity of β-amino acid residues than is possible with α-amino acid residues. The prediction that homo-β-peptides of ACPC should form the 12-helix was born out in experimental studies, in which relatively short oligomers were shown to adopt the 12-helix conformation, both in organic solution and in the crystal state. In organic solvents, the conformation is so stable that it is observed in peptides containing as few as six ACPC residues. However, β-peptides consisting of this amino acid were not soluble in water. To address this limitation, the pyrrolidinyl β-amino acid trans-3-aminopyrrolidine-4-carboxylic acid (APC) was prepared and incorporated into β-peptides along with ACPC residues. ECD (Figure 16.28) studies indicated that oligomers with as few as four residues show substantial populations of the 12-helix in water and that the helical content increases with main-chain length. The ECD curves exhibit a band at about 220 nm, followed by a more intense band of opposite sign near 200 nm. The zero-crossing is seen in the vicinity of 213 nm. Theoretical calculations indicated that the amide π → π * contribution to the ECD spectrum of the 12-helix should be similar in shape to that of the 14-helix but that the sign should be reversed

566

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a) O

H2N

N H

N H

O

O

O

O

N H

N H

N H

COOH

(b)

[θ] × 10–3 (deg × cm2 × dmol–1)

80

40

0

–40 200

220 240 Wavelength (nm)

Figure 16.27. Chemical formula (a) and ECD spectrum (b) of a 14-helical hexapeptide based on acyclic, mono-substituted, β 3 -amino acid building blocks in MeOH solution. (Adapted from reference 134.)

for a given helical handedness and the splitting between the parallel and perpendicular bands should be greater. The experimental spectra observed for a hexamer that forms the 12-helix are consistent with this analysis, showing a band near 205 nm followed by a band of opposite sign at approximately 190 nm. Additionally, a band is observed near 220 nm, which is probably associated with the amide n → π * transition. The presence of a band at 200–205 nm, together with another band near 220 nm, was not observed in other secondary structures of β-peptides and may be diagnostic of the 12-helix. Homo-oligomers from the (R, S )-cis-ACPC β-residue were shown to adopt preferentially a six-strand, extended, polar conformation [164]. This secondary structure is stabilized by intra-residue electrostatic interactions between the N–H and C=O groups, which form weak H-bonds in solution and lead to a six-membered H-bonded pseudoring structure (C6 ) (Figure 16.29). The ECD spectrum of the longest β-peptide of this series, the homo-heptamer, has a negative Cotton effect at 203 nm, the intensity of which decreases with shortening of the β-peptide backbone. β-Peptides with strictly alternating β 2 - and β 3 - (or β 3 - and β 2 -) mono-substituted residues tend to adopt the 10/12-helix conformation with 2.7 residues/turn [131, 134, 153, 154] (Figure 16.30). Depending on whether the sequence begins with a β 3 - or a β 2 -unit, the helix can start with a 10- or a 12-membered ring, respectively. This helix was studied by ECD spectroscopy. The spectrum shows a single, intense peak near 205

567

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

15

[θ]R × 10–3 (deg × cm2 × dmol–1)

10

5

0

Figure 16.28. ECD spectra of the 12-helical Ac–(APC–ACPC)2 – 4 –NH2 sequential β-peptide amides in H2 O solution. All ACPC β-amino acids residues have the (R, R)-trans conﬁguration,

–5 200

220 Wavelength (nm)

(b)

(a) [θ] × 10–3 (deg × cm2 × dmol–1)

while that of all APC β-amino acids is (R, S)-trans. (Adapted from reference 147.)

240

0.0

–0.4

–0.8

H

–1.2

N

C

H

O

NH2 7

–1.6 200

220 240 Wavelength (nm)

Figure 16.29. ECD spectrum (a) of the homo-β-heptapeptide amide (intramolecularly H-bonded into a series of C6 forms, b) based on the (R, S)-cis-ACPC building block. (Adapted from reference 164.)

568

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a) O N H

O N H

O N H

O N H

O N H

O N H

O N H

O N H

O N H

(b)

10

12

[θ] × 10–3 (deg × cm2 × dmol–1)

(c)

80

60

40

20

0 220 240 200 Wavelength (nm)

Figure 16.30. (a) The amino acid sequence of the sequential β3 /β2 -nonapeptide studied. (b) 3D-Structure of the 10/12-helix. The H-atoms, except those of the amides, are omitted for clarity. (c) The ECD spectrum of the nonapeptide in MeOH solution. (Adapted from reference 153.)

nm (Figure 16.30). In this helix, the amides surrounded by methylenes form H-bonds to one another (i , i +2), generating the 10-membered rings, while the 12-atom rings are produced between amides surrounded by side chains (i +1, i +3). In contrast to the uniform alignment of amide bonds with the helical axis for the 14- and 12-helices, there are two types of amide bond orientations in the 10/12-helix. The 10-atom-ring amides are approximately perpendicular to the helical axis, while the 12-atom-ring amides are nearly aligned with the helical axis. This arrangement results in a smaller overall helix dipole compared to that of the other helical conformations. Interestingly, the dramatic solvent-dependent change observed in the ECD spectrum of a β-peptide was suggested to arise from an environmentally induced switch between the 10/12-helix (in H2 O) and the 14-helix (in MeOH). A comparable change in ECD arising from end-group deprotection was similarly rationalized. We have discussed above the detailed geometries and chirospectroscopic properties of the most common helical structures formed by poly-β-peptides. For additional, remarkable scientiﬁc contributions on those topics, see references 163–186. Other related, rapidly expanding areas, not mentioned here, include oligomers based on γ - and δ-amino acids, and various types of “hybrid” (α/β-; α/γ -; β/γ -)peptides. Alternating heterochiral oligo-β-peptides were also investigated [165, 167]. β-Strand-like conformations of β-peptides attracted the attention of structural biochemists and ECD theoreticians [169]. It was also found that in a limited number of cases a given ECD pattern can be induced by spatially different 3D-structures [152]. To gain more insight into the relationship between

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

β-peptide conformation and ECD properties, more accurate methods to calculate the ECD spectra for β-peptides are required. It is clear that the ﬁeld of peptides based in part or fully on β-, γ -, and δ-amino acids is wide open and limited almost exclusively by scientists’ imagination.

ACKNOWLEDGMENTS The authors wish to thank Dr. R. W. Woody for critical reading of the manuscript.

REFERENCES 1. Houben-Weyl: Methods of Organic Chemistry, Vol. E 22c, Synthesis of Peptides and Peptidomimetics, M. Goodman, A. Felix, L. Moroder, and C. Toniolo, Eds., Thieme, Stuttgart, Germany, 2003. 2. P. S. Farmer, E. J. Ariens, Trends Pharmacol. Sci . 1982, 3 , 362–365. 3. A. F. Spatola, Chemistry and Biochemistry of Amino Acids, Peptides, and Proteins, Vol. 7, B. Weinstein, ed., Dekker, New York, 1983, pp. 267–357. 4. C. Toniolo, Int. J. Pept. Protein Res. 1990, 35 , 287–300. 5. J.-L. Fauch`ere, C. Thurieau, Adv. Drug Design 1992, 23 , 127–159. 6. A. Giannis, T. Kolter, Angew. Chem. Int. Edit. 1993, 32 , 1244–1267. 7. J. Gante, Angew. Chem. Int. Edit. 1994, 33 , 1699–1720. 8. R. M. J. Liskamp, Recl. Trav. Chim. Pays-Bas 1994, 113 , 1–19. 9. V. J. Hruby, G. Li, C. Haskell-Luevano, M. Shenderovich, Biopolymers 1997, 43 , 219–266. 10. Bioorg. Med. Chem. 1999, 7 , No. 1 (special issue), J. W. Kelly, ed. 11. Advances in Amino Acid Mimetics and Peptidomimetics, Vol. 2, A. Abell, ed., JAI Press, Stamford, CT, 1999. 12. Acc. Chem. Res. 2008, 41 , No. 10 (special issue), Y.-D. Wu and S. Gellman, eds. 13. A. Grauer, B. K¨onig, Eur. J. Org. Chem. 2009, 5099–5111. 14. G. H¨olzemann, Kontakte (Darmstadt) 1991, 55–63. 15. L. Pauling, R. B. Corey, Proc. Natl. Acad. Sci. USA 1951, 37 , 729–740. 16. S. Mizushima, T. Shimanouchi, M. Tsuboi, R. Souda, J. Am. Chem. Soc. 1952, 74 , 270–271. 17. M. T. Cung, M. Marraud, J. N´eel, Ann. Chim. (Paris) 1972, 7 , 183–209. 18. C. Toniolo, CRC Crit. Rev. Biochem. 1980, 9 , 1–44. 19. S. Hovm¨oller, T. Zhou, T. Ohlson, Acta Crystallogr. D 2002, 58 , 768–776. ˚ 20. Aberg, A. Yaremchuk, M. Tukalo, B. Rasmussen, S. Cusack, Biochemistry 1997, 36 , 3084–3094. 21. H. Birkedal, D. Schwarzenbach, Ph. Pattison, Chem. Commun. 2002, 2812–2813. 22. M. Crisma, F. Formaggio, C. Toniolo, T. Yoshikawa, T. Wakamiya, J. Am. Chem. Soc. 1999, 121 , 3272–3278. 23. C. Toniolo, E. Benedetti, Molecular Conformation and Biological Interactions, P. Balaram and S. Ramaseshan, eds., Indian Academy of Sciences, Bangalore, India, 1991, pp. 511–521. 24. M. Crisma, F. Formaggio, A. Moretto, C. Toniolo, Biopolymers (Pept. Sci.) 2006, 84 , 3–12. 25. N. Imawaka, M. Tanaka, H. Suemune, Helv. Chim. Acta 2000, 83 , 2823–2835. 26. M. Tanaka, Chem. Pharm. Bull. (Tokyo) 2007, 55 , 349–358. 27. C. Toniolo, M. Crisma, F. Formaggio, A. Moretto, C. Peggion, B. Kaptein, Q. B. Broxterman, in Peptides for Youth: The Proceedings of the 20th American Peptide Symposium, S. Del Valle, E. Escher, and W. D. Lubell, eds., Springer Science, New York, 2009, pp. 45–46.

569

570

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

28. 29. 30. 31.

32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43.

44.

45. 46. 47.

48. 49. 50. 51. 52. 53. 54. 55. 56.

57. 58. 59. 60.

C. Toniolo, F. Formaggio, R. W. Woody, 2012, work in progress. M. Oba, M. Tanaka, M. Kurihara, H. Suemune, Helv. Chim. Acta. 2002, 85 , 3197–3214. E. Schr¨oder, K. Lubke, The Peptides, Vol. II, Academic Press, New York, 1966, pp. 397–423. C. Gilon, M. A. Dechantsreiter, F. Burkhart, A. Friedler, H. Kessler, Houben-Weyl: Methods of Organic Chemistry. Synthesis of Peptides and Peptidomimetics, Vol. E 22c, M. Goodman, A. Felix, L. Moroder, and C. Toniolo, eds., Thieme, Stuttgart, Germany, 2003, pp. 215–271. A. Moretto, M. Crisma, B. Kaptein, Q. B. Broxterman, C. Toniolo, Biopolymers (Pept. Sci.) 2006, 84 , 553–565. S. Zhang, S. Prabpai, P. Kongsaeree, P. I. Arvidsson, Chem. Commun. 2006, 497–499. J. E. Mark, M. Goodman, J. Am. Chem. Soc. 1967, 89 , 1267–1268. J. E. Mark, M. Goodman, Biopolymers 1967, 5 , 809–814. A. M. Liquori, P. De Santis, Biopolymers 1967, 5 , 815–820. W. L. Mattice, Macromolecules 1973, 6 , 855–858. V. Madison, J. Schellman, Biopolymers 1972, 11 , 1041–1076. M. Goodman, M. Fried, J. Am. Chem. Soc. 1967, 89 , 1264–1267. M. Goodman, F. Chen, F. R. Prince, Biopolymers 1973, 12 , 2549–2561. J. R. Spencer, N. G. J. Delaet, A. Toy-Palmer, V. V. Antonenko, M. Goodman, J. Org. Chem. 1993, 58 , 1635–1638. A. Cosani, M. Terbojevich, M. Palumbo, E. Peggion, Macromolecules 1979, 12 , 875–878. R. J. Simon, R. S. Kania, R. N. Zuckermann, V. D. Huebner, D. A. Jewell, S. Banville, S. Ng, L. Wang, S. Rosenberg, C. K. Marlowe, D. C. Spellmeyer, R. Tan, A. D. Frankel, D. V. Santi, F. E. Cohen, P. A. Bartlett, Proc. Natl. Acad. Sci. USA 1992, 89 , 9367–9371. L. S. Richter, D. C. Spellmeyer, E. J. Martin, G. M. Figliozzi, R. N. Zuckermann, Combinatorial Peptide and Nonpeptide Libraries, G. Jung, ed., VCD, Weinheim, Germany, 1996, pp. 387–404. S. A. Fowler, H. E. Blackwell, Org. Biomol. Chem. 2009, 7 , 1508–1524. C. W. Wu, K. Kirshenbaum, T. J. Sanborn, J. A. Patch, K. Huang, K. A. Dill, R. N. Zuckermann, A. E. Barron, J. Am. Chem. Soc. 2003, 125 , 13525–13530. K. Kirshenbaum, A. E. Barron, R. A. Goldsmith, Ph. Armand, E. K. Bradley, K. T. V. Truong, K. A. Dill, F. E. Cohen, R. N. Zuckermann, Proc. Natl. Acad. Sci. USA 1998, 95 , 4303–4308. F. A. Bovey, F. P. Hood, Biopolymers 1967, 5 , 325–326. J. A. Patch, A. E. Barron, J. Am. Chem. Soc. 2003, 125 , 12092–12093. S. L. Seurynck, J. A. Patch, A. E. Barron, Chem. Biol . 2005, 12 , 77–88. S. L. Seurynck-Servoss, M. T. Dohm, A. E. Barron, Biochemistry 2006, 45 , 11809–11818. J. Delsarte, G. Weill, Macromolecules 1974, 7 , 450–453. D. Grenier, R. E. Prud’homme, A. Leborgne, N. Spassky, J. Polym. Sci., Polym. Chem. Ed . 1981, 19 , 1781–1793. C. Aleman, B. Lotz, J. Puiggali, Macromolecules 2001, 34 , 4795–4801. M. Rueping, A. Dietrich, V. Buschmann, M. G. Fritz, M. Sauer, D. Seebach, Macromolecules 2001, 34 , 7042–7048. V. T. Ivanov, I. I. Mikhaleva, Houben-Weyl: Methods of Organic Chemistry, Vol. E 22c, M. Goodman, A. Felix, L. Moroder, C. Toniolo, eds., Thieme, Stuttgart, Germany, 2003, pp. 274–291. M. Dobler, Ionophores and Their Structures, John Wiley & Sons, New York, 1981. H. Oku, K. Yamada, R. Katakai, Biopolymers 2008, 89 , 270–283. M. Crisma, G. Valle, G. M. Bonora, C. Toniolo, G. Cavicchioni, Int. J. Pept. Protein Res. 1993, 41 , 553–560. G. Boussard, M. Marraud, J. N´eel, B. Maigret, A. Aubry, Biopolymers 1977, 16 , 1033–1052.

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78.

79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95.

C. N. Pace, Nat. Struct. Mol. Biol . 2009, 16 , 681–682. J. Gao, D. A. Bosco, E. T. Powers, J. W. Kelly, Nat. Struct. Mol. Biol . 2009, 16 , 684–690. M. Goodman, C. Gilon, M. Palumbo, R. T. Ingwall, Israel J. Chem. 1974, 12 , 67–77. R. T. Ingwall, M. Goodman, Macromolecules 1974, 7 , 598–605. P. De Santis, A. J. Kovacs, Biopolymers 1968, 6 , 299–306. D. A. Brant, A. E. Tonelli, P. J. Flory, Macromolecules 1969, 2 , 228–235. A. E. Tonelli, Macromolecules 1992, 25 , 3581–3584. X. Yang, S. Kang, S. L. Hsu, H. D. Stidham, P. B. Smith, A. Leugers, Macromolecules 2001, 34 , 5037–5041. M. Goodman, M. D’Alagni, J. Polym. Chem. B, Polym. Lett. 1967, 5 , 515–521. C. Toniolo, V. Perciaccante, J. Falcetta, R. Rupp, M. Goodman, J. Org. Chem. 1970, 35 , 6–10. L. R. Caswell, M. F. Howard, T. M. Onisto, J. Org. Chem. 1976, 41 , 3312–3316. T. Kimura, T. Fukuda, S. Shimada, H. Matsuda, Chem. Lett. 2004, 33 , 608–609. A. B. Sugiharto, C. M. Johnson, I. E. Dunlop, S. Roke, J. Phys. Chem. 2008, C112 , 7531–7534. M. Lutz, J. van Krieken, Acta Crystallogr. 2010, C66 , o401–o405. R. T. Ingwall, C. Gilon, M. Goodman, Macromolecules 1976, 9 , 802–808. M. Goodman, W. Becktel, R. Katakai, G. Wouters, Makromol. Chem. 1981, Suppl. 4 , 100–115. W. J. Becktel, G. Wouters, D. M. Simmons, M. Goodman, Macromolecules 1985, 18 , 630–634. ¨ U. Schmidt, J. H¨ausler, E. Ohler, H. Poisel, Progress in the Chemistry of Organic Natural Products, Vol. 37, W. Herz, H. Grisebach, and G. W. Kirby, eds., Springer, Berlin, 1979, pp. 251–327. C. H. Stammer, Chemistry and Biochemistry of Amino Acids, Peptides, and Proteins, Vol. 6, B. Weinstein, ed., Dekker, New York, 1982, pp. 33–74. K. Noda, Y. Shimohigashi, N. Izumiya, The Peptides: Analysis, Synthesis, Biology, Vol. 5, E. Gross and J. Meienhofer, eds., Academic Press, New York, 1983, pp. 285–339. U. Schmidt, A. Lieberknecht, J. Wild, Synthesis 1988, 159–172. K. Uma, P. Balaram, Indian J. Chem. 1989, 28B , 705–710. T. P. Singh, P. Narula, H. C. Patel, Acta Crystallogr. 1990, B46 , 539–545. T. P. Singh, P. Jaur, Prog. Biophys. Mol. Biol . 1996, 66 , 141–165. R. Jain, V. S. Chauhan, Biopolymers (Pept. Sci.) 1996, 40 , 105–119. P. Mathur, S. Ramakumar, V. S. Chauhan, Biopolymers (Pept. Sci.) 2004, 76 , 150–161. O. Pieroni, A. Fissi, G. Montagnoli, Biopolymers 1973, 12 , 1445–1449. Y. Shimohigashi, J. W. Dunning, Jr., M. D. Grim, C. H. Stammer, J. Chem. Soc., Perkin Trans. II 1981, 1171–1175. O. Pieroni, G. Montagnoli, A. Fissi, S. Merlino, F. Ciardelli, J. Am. Chem. Soc. 1975, 97 , 6820–6826. O. Pieroni, A. Fissi, S. Merlino, F. Ciardelli, Israel J. Chem.. 1976/77, 15 , 22–28. O. Pieroni, A. Fissi, G. Montagnoli, F. Ciardelli, Biopolymers 1977, 16 , 1677–1686. O. Pieroni, A. Fissi, C. Pratesi, P. A. Temussi, F. Ciardelli, J. Am. Chem. Soc. 1991, 113 , 6338–6340. O. Pieroni, A. Fissi, C. Pratesi, P. A. Temussi, F. Ciardelli, Biopolymers 1993, 33 , 1–10. A. Tuzi, M. R. Ciajolo, G. Guarino, P. A. Temussi, A. Fissi, O. Pieroni, Biopolymers 1993, 33 , 1111–1121. O. Pieroni, A. Fissi, R. M. Jain, V. S. Chauhan, Biopolymers 1996, 38 , 97–108.

571

572

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

96. A. Tuzi, M. R. Ciajolo, D. Picone, O. Crescenzi, P. A. Temussi, A. Fissi, O. Pieroni, J. Pept. Sci . 1996, 2 , 47–58. 97. H. Miyake, H. Kamon, I. Miyahara, H. Sugimoto, H. Tsukube, J. Am. Chem. Soc. 2008, 130 , 792–793. 98. Y. Demizu, N. Yamagata, Y. Sato, M. Doi, M. Tanaka, H. Okuda, M. Kurihara, J. Pept. Sci . 2010, 16 , 153–158. 99. Y. Inai, T. Ito, T. Hirabayashi, K. Yokota, Biopolymers 1993, 33 , 1173–1184. 100. Y. Inai, Y. Kurokawa, T. Hirabayashi, Macromolecules 1999, 32 , 4575–4581. 101. Y. Inai, Y. Kurokawa, T. Hirabayashi, Biopolymers 1999, 49 , 551–564. 102. Y. Inai, K. Tagawa, A. Takasu, T. Hirabayashi, T. Oshikawa, M. Yamashita, J. Am. Chem. Soc. 2000, 122 , 11731–11732. 103. Y. Inai, T. Oshikawa, M. Yamashita, I. Kato, T. Hirabayashi, Chem. Lett. 2001, 472–473. 104. Y. Inai, T. Oshikawa, M. Yamashita, T. Hirabayashi, Y. Kurokawa, Bull. Chem. Soc. Jpn. 2001, 74 , 959–966. 105. Y. Inai, Y. Kurokawa, N. Kojima, J. Chem. Soc., Perkin Trans. 2 , 2002, 1850–1857. 106. Y. Inai, Y. Ishida, K. Tagawa, A. Takasu, T. Hirabayashi, J. Am. Chem. Soc. 2002, 124 , 2466–2473. 107. Y. Inai, H. Komori, Biomacromolecules 2004, 5 , 1231–1240. 108. N. Ousaka, Y. Inai, T. Okabe, Biopolymers 2006, 83 , 337–351. 109. Y. Inai, N. Ousaka, Biopolymers 2006, 82 , 471–481. 110. H. Komori, Y. Inai, J. Phys. Chem. 2006, A110 , 9099–9107. 111. N. Ousaka, Y. Inai, J. Am. Chem. Soc. 2006, 128 , 14736–14737. 112. H. Komori, Y. Inai, J. Org. Chem. 2007, 72 , 4012–4022. 113. N. Ousaka, Y. Inai, R. Kuroda, J. Am. Chem. Soc. 2008, 130 , 12266–12267. 114. N. Ousaka, Y. Inai, J. Org. Chem. 2009, 74 , 1429–1439. 115. R. Latajka, M. Jewginski, M. Makowski, A. Krezel, S. Paluch, Biopolymers 2008, 89 , 691–699. 116. R. Jain, M. Singh, V. S. Chauhan, Tetrahedron 1994, 50 , 907–920. 117. R. M. Jain, K. R. Rajashankar, S. Ramakumar, V. S. Chauhan, J. Am. Chem. Soc. 1997, 119 , 3205–3211. 118. U. A. Ramagopal, S. Ramakumar, P. Mathur, R. Joshi, V. S. Chauhan, Protein Eng. 2002, 15 , 331–335. 119. Rudresh, S. Ramakumar, U. A. Ramagopal, Y. Inai, S. Goel, D. Sahal, V. S. Chauhan, Structure 2004, 12 , 389–396. 120. P. Mathur, U. A. Ramagopal, S. Ramakumar, N. R. Jagannathan, V. S. Chauhan, Biopolymers (Pept. Sci.) 2006, 84 , 298–309. 121. A. Mishra, J. J. Panda, A. Basu, V. S. Chauhan, Langmuir 2008, 24 , 4571–4576. 122. A. K. Padyana, S. Ramakumar, P. Mathur, N. R. Jagannathan, V. S. Chauhan, J. Pept. Sci . 2009, 9 , 54–63. 123. D. H. Rich, P. K. Bhatnagar, J. Am. Chem. Soc. 1978, 100 , 2212–2218. 124. C. Shin, A. Kisuno. Y. Yonezawa, Chem. Lett. 1988, 1469–1472. 125. M. Lisowski, G. Pietrzynski, B. Rzeszotarska, Int. J. Pept. Protein Res. 1993, 42 , 466–474. 126. A. Anantharaman, D. Sahal, J. Med. Chem. 2010, 53 , 6079–6088. 127. C. Toniolo, E. Benedetti, Trends Biochem. Sci . 1991, 16 , 350–353. 128. G. Pietrzynski, B. Rzeszotarska, E. Ciszak, M. Lisowski, Z. Kubica, G. Boussard, Int. J. Pept. Protein Res. 1996, 48 , 347–356. 129. A. Gupta, V. S. Chauhan, Biopolymers 1990, 30 , 395–403. 130. S. H. Gellman, Acc. Chem. Res. 1998, 31 , 173–180.

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDOMIMETICS

131. R. P. Cheng, S. H. Gellman, W. F. DeGrado, Chem. Rev . 2001, 101 , 3219–3232. 132. G. Lelais, D. Seebach, Biopolymers (Pept. Sci.) 2004, 76 , 206–243. ˇ 133. D. Seebach, T. Kimmerlin, R. Sebesta, M. A. Campo, A. K. Beck, Tetrahedron 2004, 60 , 7455–7506. 134. D. Seebach, A. K. Beck, D. J. Bierbaum, Chem. Biodivers. 2004, 1 , 1111–1239. 135. D. Seebach, D. F. Hook, A. Gl¨attli, Biopolymers (Pept. Sci.) 2006, 84 , 23–37. 136. J. M. Fernandez-Santin, J. Aymami, A. Rodriguez-Galan, S. Mu˜noz-Guerra, J. A. Subirana, Nature 1984, 311 , 53–54. 137. J. M. Fernandez-Santin, S. Mu˜noz-Guerra, A. Rodriguez-Galan, J. Aymami, J. Lloveras, J. A. Subirana, E. Giralt, M. Ptak, Macromolecules 1987, 20 , 62–68. 138. J. Bella, C. Aleman, J. M. Fernandez-Santin, C. Alegre, J. A. Subirana, Macromolecules 1992, 25 , 5225–5230. 139. F. Chen, G. Lepore, M. Goodman, Macromolecules 1974, 7 , 779–783. 140. J. Applequist, K. A. Bode, D. H. Appella, L. A. Christianson, S. H. Gellman, J. Am. Chem. Soc. 1998, 120 , 4891–4892. 141. D. H. Appella, L. A. Christianson, I. L. Karle, D. R. Powell, S. H. Gellman, J. Am. Chem. Soc. 1999, 121 , 6206–6212. 142. D. H. Appella, L. A. Christianson, D. A. Klein, M. R. Richards, D. R. Powell, S. H. Gellman, J. Am. Chem. Soc. 1999, 121 , 7574–7581. 143. Y. Hamuro, J. P. Schneider, W. F. DeGrado, J. Am. Chem. Soc. 1999, 121 , 12200–11201. 144. R. P. Cheng, W. F. DeGrado, J. Am. Chem. Soc. 2001, 123 , 5162–5163. 145. D. Liu, W. F. DeGrado, J. Am. Chem. Soc. 2001, 123 , 7553–7559. 146. R. F. Epand, T. L. Raguse, S. H. Gellman, R. M. Epand, Biochemistry 2004, 43 , 9527–9535. 147. X. Wang, J. F. Espinosa, S. H. Gellman, J. Am. Chem. Soc. 2000, 122 , 4821–4822. 148. M. Lee, T. L. Raguse, M. Schinneri, W. C. Pomerantz, X. Wang, P. Wipf, S. H. Gellman, Org. Lett. 2007, 9 , 1801–1804. 149. W. C. Pomerantz, T. L. R. Grygiel, J. R. Lai, S. H. Gellman, Org. Lett. 2008, 10 , 1799–1802. 150. E. Vat, W. C. Pomerantz, M. Geyer, S. H. Gellman, L. Brunsveld, ChemBioChem 2008, 9 , 2254–2259. 151. D. Seebach, J. V. Schreiber, P. I. Arvidsson, J. Frackenpohl, Helv. Chim. Acta 2001, 84 , 271–279. 152. A. Gl¨attli, X. Daura, D. Seebach, W. F. van Gunsteren, J. Am. Chem. Soc. 2002, 124 , 12972–12978. 153. M. Rueping, J. V. Schreiber, G. Lelais, B. Jaun, D. Seebach, Helv. Chim. Acta 2002, 85 , 2577–2593. 154. G. Lelais, D. Seebach, B. Jaun, R. I. Mathad, O. Fl¨ogel, F. Rossi, M. Campo, A. Wortmann, Helv. Chim. Acta 2006, 89 , 361–403. 155. R. I. Mathad, B. Jaun, O. Fl¨ogel, J. Gardiner, M. L¨oweneck, J. D. C. Cod´ee, P. H. Seeberger, D. Seebach, M. K. Edmonds, F. H. M. Graichen, A. D. Abell, Helv. Chim. Acta 2007, 90 , 2251–2273. 156. D. Seebach, S. Abele, K. Gademann, G. Guichard, T. Hintermann, B. Jaun, J. L. Mathews, J. V. Schreiber, L. Oberer, U. Hommel, H. Widmer, Helv. Chim. Acta 1998, 81 , 932–982. 157. M.-O. Ebert, J. Gardiner, S. Ballet, A. D. Abell, D. Seebach, Helv. Chim. Acta 2009, 92 , 2643–2658. 158. J. V. Schreiber, D. Seebach, Helv. Chim. Acta 2000, 83 , 3139–3152. 159. K. A. Bode, J. Applequist, Macromolecules 1997, 30 , 2144–2150. 160. J. Applequist, K. A. Bode, J. Phys. Chem. 2000, A104 , 7129–7132. 161. J. Applequist, J. Phys. Chem. 2000, A104 , 7133–7139.

573

574

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

162. X. Daura, D. Bakowies, D. Seebach, J. Fleischhauer, W. F. van Gunsteren, P. Kr¨uger, Eur. Biophys. J . 2003, 32 , 661–670. 163. A. Hetenyi, I. M. Mandity, T. A. Martinek, G. K. Toth, F. F¨ulop, J. Am. Chem. Soc. 2005, 127 , 547–553. 164. T. A. Martinek, G. K. Toth, E. Vass, M. Hollosi, F. F¨ulop, Angew. Chem. Int. Edit. 2002, 41 , 1718–1721. 165. T. A. Martinek, I. M. Mandity, L. F¨ulop, G. K. Toth, E. Vass, M. Hollosi, E. Forro, F. F¨ulop, J. Am. Chem. Soc. 2006, 128 , 13539–13544. 166. A. Hetenyi, Z. Szakonyi, I. M. Mandity, E. Szolnoki, G. K. Toth, T. A. Martinek, F. F¨ulop, Chem. Commun. 2009, 177–179. 167. G. V. M. Sharma, K. R. Reddy, P. R. Krishna, A. R. Sankar, K. Narsimulu, S. K. Kumar, P. Jayaprakash, B. Jagannadh, A. C. Kunwar, J. Am. Chem. Soc. 2003, 125 , 13670–13671. 168. G. V. M. Sharma, K. R. Reddy, P. R. Krishna, A. R. Sankar, P. Jayaprakash, B. Jagannadh, A. C. Kunwar, Angew. Chem. Int. Edit. 2004, 43 , 3961–3965. 169. S. Chandrasekhar, B. N. Babu, A. Prabhakar, A. Sudhakar, M. S. Reddy, M. U. Kiran, B. Jagadeesh, Chem. Commun. 2006, 1548–1550. 170. A. S. Norgren, P. I. Arvidsson, Org. Biomol. Chem. 2005, 3 , 1359–1361. 171. A. S. Norgren, P. I. Arvidsson, J. Org. Chem. 2008, 73 , 5272–5278. 172. A. Martinez de Ilarduya, C. Alem´an, M. Garc´ıa-Alvarez, F. Lopez-Carrasquero, S. Mu˜nozGuerra, Macromolecules 1999, 32 , 3257–3263. 173. M. G. Garc´ıa-Martin, M. Garc´ıa-Alvarez, A. Martinez de Ilarduya, J. A. Galbis, S. Mu˜nozGuerra, Biopolymers 2005, 77 , 121–127. 174. E. Torres, E. Gorrea, K. K. Burasco, E. Da Silva, P. Nolis, F. Rua, S. Boussert, I. Diez-Perez, S. Dannenberg, S. Izquierdo, E. Giralt, C. Jaime, V. Branchadell, R. M. Ortu˜no, Org. Biomol. Chem. 2010, 8 , 564–575. 175. Y. Imamura, N. Watanabe, N. Umezawa, T. Iwatsubo, N. Kato, T. Tomita, T. Iguchi, J. Am. Chem. Soc. 2009, 131 , 7353–7359. 176. G. L. Simpson, A. H. Gordon, D. M. Lindsay, N. Pramsawar, M. P. Crump, K. Mulholland, B. R. Hayter, T. Gallagher, J. Am. Chem. Soc. 2006, 128 , 10638–10639. 177. E. Vaz, L. Brunsveld, Org. Lett. 2006, 8 , 4199–4202. 178. J. Cheng, T. J. Deming, Macromolecules 2001, 34 , 5169–5174. 179. S. A. Hart, A. B. F. Bahadoor, E. E. Matthews, X. J. Qiu, A. Schepartz, J. Am. Chem. Soc. 2003, 125 , 4022–4023. 180. J. A. Kritzer, J. D. Lear, M. E. Hodsdon, A. Schepartz, J. Am. Chem. Soc. 2004, 126 , 9468–9469. 181. J. A. Kritzer, J. Tirado-Rives, S. A. Hart, J. D. Lear, W. L. Jorgensen, A. Schepartz, J. Am. Chem. Soc. 2005, 127 , 167–168. 182. J. A. Kritzer, N. W. Luedtke, E. A. Harker, A. Schepartz, J. Am. Chem. Soc. 2005, 127 , 14584–14585. 183. J. A. Kritzer, O. M. Stephens, D. A. Guarracino, S. K. Reznik, A. Schepartz, Bioorg. Med. Chem. 2005, 13 , 11–16. 184. D. A. Guarracino, H. R. Chiang, T. N. Banks, J. D. Lear, M. E. Hodsdon, A. Schepartz, Org. Lett. 2006, 8 , 807–810. 185. E. Abraham, C. W. Bailey, T. D. W. Claridge, S. G. Davies, K. B. Ling, B. Odell, T. L. Rees, P. M. Roberts, A. J. Russell, A. D. Smith, L. J. Smith, H. R. Storr, M. J. Sweet, A. L. Thompson, J. E. Thomson, G. E. Tranter, D. J. Watkin, Tetrahedron: Asymmetry 2010, 21 , 1797–1815. 186. C. Fernandes, S. Faure, E. Pereira, V. Th´ery, V. Declerck, R. Guillot, D. J. Aitken, Org. Lett. 2010, 12 , 3606–3609.

17 CIRCULAR DICHROISM SPECTROSCOPY OF NUCLEIC ACIDS ´ Klara ´ Bednaˇ ´ rova, ´ and Michaela Vorl´ıcˇ kova´ Jaroslav Kypr, Iva Kejnovska,

17.1. INTRODUCTION Circular dichroism (CD) arises from differential absorption of right-handed and left-handed circularly polarized light by chiral molecules. This phenomenon has been described in detail in previous books (e.g., reference 1) and reviews [2–6; see Chapter 18, this volume]. In nucleic acids there are three sources of chirality. First is the asymmetric sugar (especially position C1 ); this chirality causes monomeric nucleosides to exhibit CD. The second source is the helicity of the secondary structures adopted by nucleic acids. The third source of CD results from long-range tertiary ordering of DNA in some environments. CD of monomeric constituents of nucleic acids and short single-stranded fragments were described previously [2]. The theory of CD is well-developed [1] and complex. Nevertheless, the use of CD spectroscopy to elucidate nucleic acid secondary structure is mainly based on empirical grounds. Conventional CD spectroscopy operates within the spectral range of about 200 nm to 320 nm. For these measurements, conventional spectrometers are used. CD spectroscopy is even more sensitive and informative in the far UV region below 200 nm, but these measurements are difﬁcult to perform and the specialized instruments required are expensive [2]. CD spectra of nucleic acids can also be measured in the infrared region (Chapters 18, 22, and 23, this volume), but here the method is much less sensitive. In this chapter we will focus on CD results obtained in the 200- to 320-nm range, the range mostly used to study secondary structures of nucleic acids.

Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

575

576

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Nucleic acids (NA) are of two types, DNA and RNA. They mainly differ by the type of constituent sugar, which is deoxyribose in DNA and ribose in RNA. DNA generally contains adenine, guanine, thymine, and cytosine. RNA contains uracil instead of thymine and other noncanonical bases. In vivo, RNA strands are much shorter than DNA and the two molecules also differ in secondary structure. DNA is mostly double-stranded. In RNAs, helices, bulges, loops, and mismatches contribute to complex tertiary structures. CD spectroscopy is more useful in studies of DNA than RNA. Below we will review problems where CD spectroscopy contributed to the understanding of the conformational polymorphism of DNA.

17.2. DENATURED DNA, B-FORM AND THE HELIX-COIL TRANSITION CD spectra of denatured DNAs are weak because the strands are mostly disordered and exhibit little chirality. The B-form helix of natural DNAs (i.e., the basic ordered double-stranded conformation) provides a rather weak CD spectrum containing a positive band around 275 nm and a negative band at around 245 nm (Figure 17.1, left). The bands have about the same intensity and the integral of the spectral curve is close to zero. We call such a spectrum conservative. The base pairs are more or less perpendicular to the long axis of the double helix, which introduces relatively weak chirality into the structure. Furthermore, contributions to the CD spectrum from different regions of the compositionally heterogeneous parts of the molecules mutually compensate. CD spectra of helices formed by synthetic DNA molecules are stronger and do have characteristics dependent on nucleotide sequence [6]. This is caused not only by different chromophores but also by rather different B-family conformations. CD spectra of native and denatured synthetic oligo- and polynucleotides differ from one another, depending on the variants of the B-type conformation adopted. CD spectroscopy thus allows one to follow the helix–coil transition of DNAs (Figure 17.1, right). The temperature dependence monitored at a selected wavelength can be used to characterize the stability of the DNA structure. The plot of CD intensity versus temperature results in a characteristic S-shaped curve reﬂecting cooperativity of the helix–coil transition. The presence of isoelliptic points in the spectra is indicative of the two-state nature of the transition. The course of the melting enables determination of the melting point (Tm ) and thermodynamic parameters of the studied structure.

17.3. THE B-FORM AND THE HAIRPIN A hairpin arises from folding back of a DNA single strand on itself to form a helix capped at one end with a loop of single-stranded residues. For hairpin formation to occur, the DNA sequence must contain at least approximately a dyad symmetry. The B-form doublestranded structure-to-hairpin transition is also reﬂected by CD spectroscopy. We illustrate this with a (G + C)-rich DNA fragment (Figure 17.2). For this sequence, electrophoresis conﬁrmed that the duplex-to-hairpin transition occurred [6]. It is noteworthy that the transition is often irreversible, indicating a large kinetic barrier between the two-stranded form and the hairpin.

CIRCULAR DICHROISM SPECTROSCOPY OF NUCLEIC ACIDS

Figure 17.1. B-form DNA spectrum and the helix–coil transition. Left: CD spectrum of B-form DNA from calf thymus (42% G + C). The spectrum was measured in 10 mM sodium acetate, pH 7. Spectra of the natural DNAs were measured on a Roussel–Jouan dichrograph, Model CD 185. Spectra of the polynucleotides in this and in the following ﬁgures were measured on the Jobin–Yvon dichrograph Mark VI and, unless stated otherwise, in 1-cm cells (absorption ∼ 0.7) at room temperature. CD in all ﬁgures is expressed as ε (in M−1 cm−1 ), molarity being related to nucleoside residues in the DNA samples. Sketches of the particular DNA structures were taken from the NDB database. Right: CD spectra of poly(dA-dT) in 10 mM sodium acetate, pH 7 at 19◦ C (dash). The spectra reﬂecting a helix-coil transition were taken at 21.9◦ C (dash–dot), 24.3◦ C (long dash) and 26.6◦ C (solid line). Insert: Temperature-induced changes in poly(dA-dT) monitored at (circles) 220 nm and (squares) 262 nm.

17.4. THE A-FORM AND THE B–A TRANSITION The A-form of DNA provides a much stronger CD spectrum than the B-form (Figure 17.3). The A-form spectrum is dominated by a strong positive band at 260 nm and a strong negative band at 210 nm. These spectral features probably originate from base-pair tilting that is characteristic of the A-form. A low-ionic strength DNA solution (∼1–5 mM) can be transformed from the B-form to the A-form by addition of ethanol [7] or other agents. Higher salt concentrations result in precipitation of DNA or cause its psi condensation (see below). The B–A transition is highly cooperative, is reversible, and has fast kinetics, and the CD spectra recorded during the transition intersect in isoelliptic points. RNA provides a CD spectrum similar to the A-form DNA of the

577

578

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 17.2. Temperature-induced irreversible duplex-hairpin transition. The spectrum of d(C6 –G6 ) was measured in 1mM sodium phosphate, 0.3 mM EDTA, pH 7. Left: CD spectra measured at 0◦ C corresponding to duplex before denaturation (dashed line) and hairpin after denaturation (solid line). Right: Spectral changes induced by increasing (open symbols) and decreasing (ﬁlled symbols) temperatures monitored at 260 nm (squares) and 280 nm (triangles). CD spectra were measured in 0.1 cm cells. (Redrawn from reference 6.)

corresponding nucleotide sequence (Figure 17.3). CD spectroscopy has proven to be a useful tool to study the B–A transition in various DNAs.

17.5. THE B/A CONFORMATION OF (dC)n · (dG)n DNA SEQUENCES The (dC)n · (dG)n sequences of DNA adopt an unusual B-conformation. A CD spectrum of the self-complementary sequence d(C6 –G6 ) is shown in Figure 17.2. It has a distinct positive CD band at 260 nm, similar to the A-form, even in the absence of alcohol. Combined CD spectroscopy, molecular dynamics, and NMR studies showed that this sequence adopts an A-like structure with B-like sugar puckering [8]. This and similar sequences undergo a cooperative and two-state transition to A-form induced by alcohols. Even the reverse self-complementary sequences d(Gn –Cn ) have A-like features. The (dG)n sequence in the 5 half of the molecule is A-like, whereas the 3 (dC)n half is B-like [9]. These ﬁndings extended our understanding of the conformational polymorphism of DNA.

CIRCULAR DICHROISM SPECTROSCOPY OF NUCLEIC ACIDS

Figure 17.3. B–A transition of DNA. Left panel: CD spectra reﬂecting triﬂuorethanol-induced B–A transition of d(GCGGCGACTGGTGAGTACGC) duplex with its complementary strand: 0% TFE (dashed), 80% TFE (solid) lines. Insert: The transition monitored at 266 nm. Right panel: CD spectra of RNA of the same sequence (U instead of T) duplexed with a complementary DNA strand: 0% TFE (dashed), 80% TFE (solid) lines. (Redrawn from reference 6.)

17.6. THE Z-FORMS AND THE B–Z TRANSITION Like the B–A transition, the B–Z transition was ﬁrst detected in solution by CD spectroscopy [10]. The Z-form is a left-handed helix. The transition is highly cooperative, and it has a slow kinetics caused by the need for base-pair ﬂipping. It is speciﬁc for alternating pyrimidine–purine, namely dC–dG, sequences, and it is facilitated by the presence of methyl5 dC [11]. The CD spectrum of the Z-form is more or less a mirror image of that of the B-form (Figure 17.4), although variants of the Z-form (called Z ) have rather different spectra [6]. The B–Z transition is also reﬂected by UV absorption spectroscopy (Figure 17.4, upper insert).

17.7. GUANINE QUADRUPLEXES Guanine quadruplexes are interesting alternative arrangements to the Watson–Crick double helix. These structures are based on guanine tetrads held together by Hoogsteen

579

580

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 17.4. CD spectra reﬂecting triﬂuorethanol-induced B–Z transition of poly(dG–dC) duplex. Spectra were measured in 59% TFE (dashed) and 67% TFE (solid) lines. TFE was added to the DNA dissolved in 1 mM sodium phosphate, 0.3 mM EDTA, pH 7, and CD spectra were measured at 0◦ C. Inserts: The transition monitored at 291 nm. UV absorption spectra measured at the same conditions as the CD measurements. (Redrawn from reference 5.)

hydrogen bonds. The tetrads are stacked on top of one another, and a cation is inserted in the cavity between the neighboring tetrads. Thus, the stability of guanine quadruplexes is sensitive to the cation type. There are several types of guanine quadruplexes. The quadruplexes can be built by one, two, or four DNA molecules and the strand orientation can be parallel, antiparallel, or hybrid parallel–antiparallel. CD spectroscopy can distinguish among particular guanine quadruplex topologies (Figure 17.5). The spectrum of parallel quadruplexes contains a dominating positive CD band at 260 nm; antiparallel quadruplexes are characteristic by a positive CD band at 295 nm and a negative one around 260 nm. The bands at 260 and 295 nm probably reﬂect populations of anti and syn glycosidic angles of the dG residues in the particular quadruplex arrangements. The 260-nm band is similar in shape to that of the A-form, which indicates similar base stacking [12]. The quadruplex, however, has a positive band at 210 nm where the A-form has a negative band (compare Figures 17.3 and 17.5). Quadruplex formation is also reﬂected by UV absorption spectroscopy (Figure 17.5, insert). Study of quadruplex structures is important as the human genome contains thousands of sequences prone to formation of guanine quadruplexes under favorable conditions. CD spectroscopy, especially

CIRCULAR DICHROISM SPECTROSCOPY OF NUCLEIC ACIDS

Figure 17.5. CD spectra of guanine quadruplexes. Left: Time-dependent formation of a parallelstranded quadruplex of d(G4 ) stabilized by 16 mM K+ : spectrum immediately after K+ addition (dashed) and after 24 h (solid line). Right: Na+ -induced formation of an antiparallel bimolecular quadruplex of d(G4 T4 G4 ): 1 mM Na+ (dashed), 500 mM NaCl (solid line). Oligonucleotides were dissolved in 1 mM sodium phosphate, 0.3 mM EDTA, pH 7, thermally denatured (5 min at 90◦ C), and slowly cooled before measurements. Insert: UV absorption spectra of d(G4 T4 G4 ) measured at the same conditions as the CD measurements. (Redrawn from reference 6.)

in combination with gel electrophoresis, has already contributed signiﬁcantly to our knowledge of various quadruplex topologies adopted by, for example, the sequences in human telomeres [13].

17.8. CYTOSINE QUADRUPLEXES DNA strands rich in cytosine also form quadruplexes. These consist of two parallel homoduplexes connected through hemi-protonated C · C+ pairs [14]. The two homoduplexes are mutually intercalated in an antiparallel orientation (Figure 17.6). The cytosine quadruplexes provide a characteristic CD spectrum dominated by a positive band at 290 nm (Figure 17.6). Formation of cytosine quadruplexes requires acidic pH, which is needed to protonate cytosine. Interestingly, some cytosine quadruplexes are stable even at pH 7.

581

582

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 17.6. Acid-induced transition of a C-rich oligonucleotide into a cytosine quadruplex. CD spectra and UV absorption spectra (insert) reﬂecting the acid-induced transition of a human telomere DNA fragment d[(C3 TAA)3 C3 ] into an intercalated cytosine quadruplex (CD dependence measured at 287 nm). The oligonucleotide was dissolved in 10 mM potassium phosphate, 0.1 M KCl, pH 7.1 (dashed line); the pH value was changed to pH 5 by addition of dilute HCl and the CD spectrum was measured (solid line). pH values were determined using a Sentron Red-Line electrode and a Sentron Titan pH meter.

17.9. DNA STRANDS RICH IN GUANINE AND ADENINE DNA strands rich in guanine and adenine can form various ordered structures that melt cooperatively. The conformers may be either (a) antiparallel duplexes, which provide CD spectra similar to B-form DNA [15], (b) parallel duplexes formed under physiological salt conditions and neutral pH, or (c) ordered single strands induced by acid pH [16], ethanol [17], or even dimethylsulfoxide [18]. The parallel duplex provides a CD spectrum with a dominating positive CD band at 260 nm, like the A-form of DNA or parallel guanine quadruplex, suggesting similar guanine–guanine stacking in various G-rich DNAs [19]. The CD spectrum of the ordered single strand is very similar to the spectrum of the homoduplex, indicating that the homoduplex arises though a dimerization of the ordered single strands [17].

CIRCULAR DICHROISM SPECTROSCOPY OF NUCLEIC ACIDS

17.10. EXTENSIVE CHANGES IN THE CIRCULAR DICHROISM OF POLY(dA-dT) Poly(dA-dT) provides a more or less B-like CD spectrum under physiological conditions, but the CD spectrum changes drastically at molar concentrations of CsF or alcohol in the presence of cesium ions as shown in Figure 17.7 [20]. We call this conformation the X form [21]. The X form is characterized by a very deep negative band in the longwavelength part of the CD spectrum. Remarkably, the changes are speciﬁc for CsF. For example, NaCl does not induce these changes. Ethanol induces a transition of poly(dAdT) into the A-form if sodium is the counterion, but the X-form is induced if cesium is the counterion [21]. The CsF-induced changes are accompanied by large changes in the phosphorus NMR spectrum of poly(dA-dT) (Figure 17.7, insert) [20]. Two widely separated resonances appear, as they do with the Z-form, but the X-form is not Z-form because assignment of the two resonances is opposite. A structure of (dA-dT)3 observed by X-ray that contains Hoogsteen base pairing [22] may be the crystal counterpart of the X-form.

17.11. POLY(AMINO2 dA-dT) AND POLY(dG-METHYL5 dC) Poly(amino2 dA-dT) and poly(dG-methyl5 dC) are much more similar DNAs than their designations might indicate (Figure 17.8, inserts). Both contain alternating purine–pyrimidine sequences, and both have amino groups in the double helix minor groove and methyl groups in the major groove. Poly(dG-methyl5 dC) undergoes the B–Z transition even under physiological conditions [11]. Poly(amino2 dA–dT), unlike poly(dA-dT), also undergoes a salt-induced transition under physiological conditions, and the resulting structure has features of both the A-form and X-form [23]. It is clear from these studies that both the amino group in the minor groove and the methyl group in the major groove strongly inﬂuence conformational behavior of the alternating purine–pyrimidine DNAs.

17.12. PSI DNA Under some conditions (PEG, ethanol with moderate salt concentrations, polylysine, etc.), DNA condenses into so-called psi DNA, a form that can be seen in the electron microscope. In the condensates, a long-range structure introduces chirality. In the CD spectra, psi DNA is characterized by huge positive or negative amplitudes and signal beyond 300 nm, where DNA bases no longer absorb light. The CD signal originates from differential scattering of light by helical arrangement of the condensates [24]. Psi DNA is an ordered structure, and it is considered to be a simple approximation of the DNA organization in chromosomes.

17.13. CONCLUSIONS This review summarized the signiﬁcant role played by CD spectroscopy in the history of nucleic acid studies. We focused on problems where relevant CD studies have essentially been ﬁnished (B-form, A-form, Z-form) as well as on those that remain open to further

583

584

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

2

–3.6 –4.0 –4.4 –4.8 NMR chemical shift (ppm)

31P

Δe [M–1 m–1]

0

–2

–2 Δe275 nm –4

–4 –6

–6

–8

–10 –8 50 240

260

280

55 60 65 Methanol [%]

300

Wavelength [nm]

Figure 17.7. Methanol-induced transition of poly(dA–dT) into an X-form. CD spectra of poly(dA–dT) measured in 0.5 mM sodium phosphate, 3 mM CsCl, and 0% (dot), 57% (dash dot dot), 60% (short dash), 62% (dash dot), and 64% (solid line) methanol (v/v). Temperature was 0◦ C. Insert: The transition monitored by ε at 275 nm. The conditions are as above, but 0.002 mM CaCl2 was present in the methanol solution. Top: 31 P NMR spectrum of the X-form of poly(dA-dT) induced by 6.3 M CsF. (Redrawn from reference 21.)

studies like those of B/A-DNA, quadruplexes, X-form, structures of poly(amino2 dA-dT) or (G + A)-rich DNA sequences. CD spectroscopy does not provide molecular structures at atomic resolution, but it has many advantages: •

It is a simple, fast, and relatively cheap method. It is extremely sensitive and therefore requires small amounts of material. This assures solubility of the sample even under extreme conditions. • It works over a wide range of DNA concentrations, from one order of magnitude higher than those used for NMR measurements to very dilute solutions. • Long as well as short DNA molecules can be studied. •

585

CIRCULAR DICHROISM SPECTROSCOPY OF NUCLEIC ACIDS

(a)

Figure 17.8.

(b)

MgCl2 -induced B–Z and B–X transitions of poly(dG–methyl5 dC) and

2

poly(amino dA–dT). Polynucleotides were dissolved in 0.6 mM potassium phosphate, 0.03 mM EDTA, pH 6.8. Left: CD spectra taken over the course of the B–Z transition of poly(dG–methyl5 dC) measured 1, 4, 17, and 88 min (dashed to solid lines, respectively) after addition of MgCl2 to 0.05 mM concentration. Right: CD spectra reﬂecting the B–X transition of poly(amino2 dA–dT) in 0, 0.028, 0.056, 0.070, and 0.190 mM MgCl2 (from thin to solid lines as concentration is increased). Top: Sketches of G·methyl5 C and amino2 A · T base pairs. (Redrawn from reference 6.)

•

The method is used empirically, and dozens of molecules can be compared in a single study. • CD can be measured under various conditions of temperature, pH, organic solvents, salt types, and concentrations, which enables researchers to map the whole conformational space of the studied molecule. • Gradual changes within a single DNA conformation and cooperative changes between discrete conformational states can be distinguished. The two types of changes have distinct physical properties and biological relevance. Not a single band, but instead the whole spectrum, should be followed and taken into account in the interpretation of the CD spectra. In combination with simple UV absorption spectroscopy and gel electrophoresis, CD spectroscopy is a powerful complementary method to X-ray diffraction and NMR spectroscopy in studies of DNA.

586

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

ACKNOWLEDGMENTS The authors thank Professor B. Woody for his valuable advice and comments. This work was supported by grant IAA500040903 and IAA100040701 kindly provided by the Grant Agency of the Academy of Sciences of the Czech Republic.

REFERENCES 1. N. Berova, K. Nakanishi, R. Woody, Circular Dichroism, Principles and Applications, 2nd ed., Wiley-VCH, New York 2000. 2. W. C. Johnson, CD of nucleic acids, in Circular Dichroism, Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. Woody, eds., Wiley-VCH, New York, 2000, pp. 703–718. 3. J. C. Maurizot, Circular dichroism of nucleic acids: Nonclassical conformations and modiﬁed oligonucleotides, in Circular Dichroism, Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. Woody, eds., Wiley-VCH, New York, 2000, pp. 719–739. 4. D. M. Gray, R. L. Ratliff, M. R. Vaughan, Meth. Enzymol . 1992, 211 , 389–406. 5. M. Vorlickova, J. Kypr, V. Sklenar, Nucleic acids: Spectroscopic methods, in Encyclopedia of Analytical Science, 2nd ed., P. J. Worsfold, A. Townshend, C. F. Poole, eds., Elsevier, Oxford, 2005, 6 , pp. 391–399. 6. J. Kypr, I. Kejnovska, D. Renciuk, M. Vorlickova, Nucl. Acids Res. 2009, 37 , 1713–1725. 7. V. I. Ivanov, L. E. Minchenkova, E. E. Minyat, M. D. Frank Kamenetskii, A. K. Schyolkina, J. Mol. Biol . 1974, 87 , 817–833. 8. L. Trantirek, R. Steﬂ, M. Vorlickova, J. Koca, V. Sklenar, J. Kypr, J. Mol. Biol . 2000, 297 , 907–922. 9. R. Steﬂ, L. Trantirek, M. Vorlickova, J. Koca, V. Sklenar, J. Kypr, J. Mol. Biol . 2001, 307 , 513–524. 10. F. M. Pohl, T. M. Jovin, J. Mol. Biol . 1972, 67 , 375–396. 11. M. Behe, G. Felsenfeld, Proc. Natl. Acad. Sci. USA 1981, 78 , 1619–1623. 12. J. Kypr, M. Fialova, J. Chladkova, M. Tumova, M. Vorlickova, Eur. Biophys. J . 2001, 30 , 555–558. 13. M. Vorlickova, J. Chladkova, I. Kejnovska, M. Fialova, J. Kypr, Nucl. Acids Res. 2005, 33 , 5851–5860. 14. M. Gueron, J. L. Leroy, Curr. Opin. Struct. Biology 2000, 10 , 326–331. 15. M. Ortiz-Lombardia, R. Eritja, F. Azor´ın, J. Kypr, I. Tejralova, M. Vorl´ıcˇ kova, Biochemistry 1995, 34 , 14408–14415. 16. N. G. Dolinnaya, J. R. Fresco, Progr. Nucl. Acids Res. 2003, 75 , 321–347. 17. M. Vorlickova, I. Kejnovska, J. Kovanda, J. Kypr, Nucl. Acids Res. 1999, 27 , 581–586. 18. J. Kypr, M. Vorlickova, Biopolymers 2001, 62 , 81–84. 19. J. Kypr, M. Vorlickova, Biopolymers 2002, 67 , 275–277. 20. M. Vorlickova, J. Kypr, V. Sklenar, J. Mol. Biol . 1983, 166 , 85–92. 21. M. Vorlickova, J. Kypr, J. Biomol. Struct. Dynam. 1985, 3 , 67–83. 22. N. G. A. Abrescia, A. Thompson, T. Huynh-Dinh, J. A. Subirana, Proc. Natl. Acad. Sci. USA 2002, 99 , 2806–2811. 23. M. Vorlickova, J. Sagi, A. Szabolcs, A. Szemzo, L. Otvos, J. Kypr, J. Biomol. Struct. Dynam. 1988, 6 , 503–510. 24. C. Bustamante, M.F., M. F. Maestre, I. Tinoco, Jr., J. Chem. Phys. 1980, 73 , 4273–4281.

18 ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES Roberto Corradini, Tullia Tedeschi, Stefano Sforza, and Rosangela Marchelli

18.1. INTRODUCTION Peptide nucleic acids (PNAs) are a class of oligonucleotide analogues with a polyamide backbone, depicted in Figure 18.1, ﬁrst described by Nielsen et al. in 1991 [1, 2]. These molecules have been largely utilized in DNA and RNA recognition in a variety of applications, ranging from modulation of gene expression [3] to diagnostic methodologies [4], and have been proposed as robust materials for micro- and nanofabrication [5]. The interesting features of this class of compounds are: (a) the ability to interact with complementary sequences of DNA and RNA forming very stable duplexes, with thermal stability higher than the corresponding duplexes formed by DNA oligonucleotides [6], and even more stable PNA–PNA duplexes [7]; (b) the high sequence selectivity of the duplex formation, which again is superior to DNA oligonucleotides in the recognition of even a single mispairing of the bases [8, 9]; (c) the ability of poly-pyrimidine PNA to form PNA–DNA–PNA triplexes of remarkable stability, which can produce strandinvasion of duplex DNA [10]; and (d) the high stability in biological ﬂuids, due to their unnatural skeleton, which prevents degradation by nucleases and proteases [11]. In their simplest version, PNAs have an achiral structure; yet it has been demonstrated that even in the achiral PNA–PNA duplexes they form helical (hence chiral) structures [12]. If a chiral bias is introduced, a preference for both right- or left-handed structures can be obtained. Thus these molecules are also a very interesting case of simpliﬁed nucleic acid structures in which the helicity can be modulated by an appropriate design. In this chapter we summarize the results on the electronic CD properties of these very interesting artiﬁcial biopolymers. Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

587

588

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

B

B O

O *

O O

O

P O

*

*

N H

n

O

N

DNA

n

*

Figure 18.1. Structure of peptide nucleic acid (PNA) outlining their similarity to DNA.

PNA

The formation of PNA-containing structures with nucleic acids is governed by the Watson–Crick base pairing for double helix, and by both Watson–Crick and Hoogsteen hydrogen bonding in the case of triplex formation, as illustrated in Figure 18.2a. A peculiar feature of PNA is that both antiparallel and parallel complexes with DNA and RNA can be formed, conventionally described as depicted in Figure 18.2b. Although the H-bonding interactions, as well as the stacking interactions between adjacent base pairs, are the same occurring in DNA and RNA structures, PNA have a preferred helical structure (named P-form) which differs from that of DNA signiﬁcantly, with 18 base pairs per turn and a smaller twist angle, as inferred from PNA–PNA solidstate studies (Figure 18.2c) [12]. The available PNA:DNA crystal structures [13, 14] show similar characteristics, with a pitch of 15.5 base pairs and a twist angle (23◦ ) signiﬁcantly smaller than both B- and A-DNA (36◦ and 32.7◦ , respectively). Therefore, conformational analysis based on CD spectra by analogy with known DNA structures (A, B, Z, etc.) should be avoided or considered only tentative. A series of other techniques, such as NMR or induced circular dichroism, should be used in order to produce experimental evidence for the conformation proposed.

(a)

H N T

O H

N

O Hoogsteen H-bonds

H

N

N A N

N

H N

N C+ + N H O

O H

N

T

N

H N G

N

H

O

Hoogsteen H-bonds N

O Watson-Crick H-bonds

H

N

N

N

N H

(b)

N

H H

C

N

O Watson-Crick H-bonds

(c) Parallel

N-Term

C-Term PNA

5′

N-Term

3′ Antiparallel

3′

PNA

DNA

C-Term PNA 5′

DNA PNA

Figure 18.2. (a) Type of interactions formed by PNA: Watson–Crick for duplex formation and Hoogsteen in PNA–DNA–PNA triplexes. (b) Parallel and antiparallel orientation of PNA–DNA duplexes. (c) PNA:PNA duplex in the solid state, forming an helical P-helix. (Data from Protein Data Bank, Code 1PUP, from reference 12.)

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

Several different problems have been addressed using CD spectroscopy: (a) preorganization of PNA single strand; (b) formation of PNA–PNA duplexes and their preferred helical features; (c) formation of duplexes or triplexes with naturally occurring nucleic acids by conformational changes detected by CD; (d) kinetics of duplex and triplex formation; (e) thermal stability of the duplexes and triplexes and conformational changes occurring as a function of temperatures; (f) ability of achiral dyes to bind to PNA-containing duplexes. In this chapter we will show the techniques and the most signiﬁcant results in this ﬁeld, with some speciﬁc examples illustrating the approaches mostly used for the study of PNA properties. Related subjects on the study of biopolymers by Electronic Circular Dichroism are reported in a previous book [15, Chapter 26: DNA–Drug Interactions] and in Chapters 14 (Electronic Circular Dichroism of Proteins), 15 (Electronic Circular Dichroism of Peptides), 17 (Circular Dichroism Spectroscopy of Nucleic Acids), 19 (Circular Dichroism of Protein–Nucleic Acid Interactions), and 20 (Drug and Natural Product Binding to Nucleic Acids Analyzed by Electronic Circular Dichroism) from the current volume.

18.2. CIRCULAR DICHROISM AS A TOOL FOR STUDYING PREORGANIZATION: SINGLE STRAND PNAS AND PNA–PNA DUPLEXES Electronic circular dichroism (ECD) has been used since early studies on PNA properties mostly by Norden and co-workers, who showed how the formation of PNA–PNA and PNA–DNA duplexes and of PNA–DNA–PNA triplexes could be followed by variation of the CD signal [7, 16]. Since then, other authors have used CD for studying the conformation of PNA analogues containing chiral moieties or the complexation of both chiral and achiral PNA with nucleic acids. Achiral standard single-strand PNAs and PNA–PNA duplexes do not show circular dichroism. However, since early studies, PNAs have been usually synthesized with a llysine moiety linked at the C-terminus, inserted for increasing their solubility in aqueous media; thus they were actually chiral [1]. Furthermore, following the successful use of PNA in many applications, a variety of molecules appeared in the literature which were designed in taking the polyamidic PNA chain as a model, and introducing either rings or substituents in their backbone, as illustrated in Figure 18.3 [17]. Many modiﬁcations contain chiral moieties and can be studied by electronic CD. Though the main interest for PNA resides in the possibility to bind DNA and RNA, the study of single-stranded PNA and of PNA–PNA double helices has provided precious insights into the “preorganization” of PNA—that is, their preference for one helix handedness—which, in turn, affect their ability to bind natural nucleic acids. Of particular relevance has been the introduction of chiral monomers derived from amino acids in the polyamide backbone (model F in Figure 18.3), for which a systematic study based on CD spectroscopy has allowed us to obtain a clear-cut rationale, which is now commonly used for the design of PNA with special properties [18]. The introduction of amino acids at the end of an achiral PNA strand does not induce circular dichroism signal in the nucleobase region, due to the ﬂexibility of these molecules (see P1 in Figure 18.4b). Furthermore, the presence of both E and Z isomers of the tertiary amide connecting the nucleobase to the backbone, as demonstrated by NMR spectra of single-stranded PNA in solution, generates a series of 2n (n = number of PNA monomers) different stereoisomers [19], which also contributes to a random arrangement of nucleobases in this simple version of chiral PNAs.

589

590

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a)

(b)

Base

Ring-closure

O

O O *

N

*

n*

N H

O

O *

N

N H

Base

(c)

Base

n*

N

N H

n* R2

R2

Substituents Base Base

(d)

(e)

O *

N H

Base O

*

N

H N

(f) O

N n*

n* n

O

R1 *

N H 5 (γ)

N

2(α α)

O n*

R2

Figure 18.3. Examples of modiﬁcations of the PNA backbone described in the literature. (a) Principal strategies for inducing preorganization of PNA. (b) Example of a six-member ring structure involving the aminoethyl group. (c) Example of a ﬁve-member ring structure involving the aminoethyl group. (d) Example of a ﬁve-member ring built on the glycine part. (e) Substitution of the aminoethyl moiety with cyclic 1,2-diamines. (f) Introduction of substituents on C2, C5, or both in the aminoethylglycine backbone.

Conjugation of a peptide is a common strategy to modulate the properties of PNA in biomedical applications, especially cellular uptake. PNA–peptide chimeras can be synthesized easily using the same strategies for both the PNA and the peptide segments. Also, these compounds as single strands do not show signiﬁcant circular dichroism bands in the nucleobase region, though showing typical CD of the peptide moiety [20]. Introduction of functional groups in the backbone induces a certain degree of preorganization in the ssPNA, and these products have CD signals. The PNAs bearing a C2-modiﬁed lysine monomer at the C-terminus showed no CD signals in the single strands, whereas PNAs containing C2-modiﬁed lysine monomers, either d- or l-, in the middle of a strand of achiral monomers showed a spectrum with alternate maxima in the 270–280 nm, 250–260 nm, 240–245 nm, and 220 nm regions (Figure 18.4a), with opposite signs for the two enantiomers and molar ellipticity (calculated with respect to bases) on the order of ±2000 deg/M cm for the 250–260 nm band. The modiﬁed PNAs that contain monomers derived from other amino acids have been found to have similar features, with the exception of the 270–280 nm band, which was not always present. Since these data were obtained on PNA bearing portions of self-complementary sequences, it cannot be excluded that such features are related to partially formed self-pairing PNA complexes, as observed in the solid state [21]. A much higher degree of preorganization was observed for the ssPNA containing C5-modiﬁed monomers. Ly and co-workers reported the strong preference of l-Ser- and l-Ala-derived C5 (γ )-chiral PNA monomers for a right-handedness, using NMR and molecular modeling [22]. PNA strands containing increasing numbers of l-Ser monomers showed a CD signature (Figure 18.4b) similar to that obtained for C2-modiﬁed chiral PNAs derived from d-amino acids, with similar intensity ([θ ]260 = 2000–4000 deg M−1 cm−1 per base). This suggests a similarity in the preorganization of the PNA. The synthesis of PNA containing cyclic structures was reported by several authors as a tool to induce a preorganization of the PNA strand, thus minimizing the entropy loss

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

(b)

(a) Molar Ellipticity (Deg M−1 cm−1)

7000 Base

6000

O

5000 *

4000 3000 2000

N H

O

N n*

N3N+

1000 0 –1000 –2000 –3000 220

240

260

280

300

wl (nm)

(c) (i)

(ii)

Figure 18.4. Examples of CD spectra of single-stranded PNA. (a) Chiral PNA containing one C2-modiﬁed monomer based on D-lysine in the middle of the strand (molecular ellipticity calculated for nucleobase residue, data from reference 33). (b) PNA containing a C5-modiﬁed L-Ser residue (indicated as γ T) obtained changing the position and the number of chiral residues, measured in sodium phosphate buffer at 2 μM strand concentration. (Reprinted with permission from reference 22, copyright 2006 American Chemical Society.) (c) Example of PNA containing monomers derived from cyclic structures: PNA containing one T monomer based on D- (Dt.T*) or L-trans-4-aminoproline (Lt.T*) at the N-terminus; (i) (Lt.T*)-AT2 AT2 AT2 , (ii) (Dt.T*)-AT2 AT2 AT2 . (Reprinted from reference 24, copyright 1999, with permission from Elsevier.)

during PNA–DNA or PNA–RNA duplex formation [23]. This resulted in a much more structured CD signal of the corresponding PNA, although common patterns could not be inferred. As an example, we report in Figure 18.4c the CD spectra of two enantiomers of PNA containing one chiral monomer derived from d- or l-trans-hydroxyproline at the N-terminus, which showed an intense, yet peculiar, CD signal. As a general conclusion, every new structure implies a different type of preorganization and a preferred conformation and therefore has its own CD signature [24]. Although the presence of one terminal amino acid is not enough to give rise to signiﬁcant CD effects for single-strand PNAs, two complementary antiparallel PNA strands with a C-terminal lysine give rise to well-deﬁned CD spectra upon mixing, very similar to those obtained in the case of DNA–DNA duplexes, which suggest the formation of a duplex with an helical arrangement [7]. The intensity of the CD spectrum reaches a maximum when the stoichiometry of the two PNA strands is at 1:1 ratio. The effect of different amino acids linked to the Cterminus, as well as the effect of different nucleobases placed at the C-terminus close to

591

592

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

the aminoacidic moiety, has been thoroughly investigated [25]. With a l-lysine at the Cterminus, the different nucleobases gave always CD spectra with the same sign. However, the intensity of the CD signal was found to be dependent upon the nucleobase closest to the C-terminus. In particular, either a guanine or a cytosine gave a well-deﬁned circular dichroism spectrum, and in the case of cytosine the subsequent base should preferably be a purine base. The higher thermal stability of a G–C base pair, as compared to an A–T base pair, is thought to be associated to the structural stability required for the amino acid to induce a preferred handedness. A typical CD spectrum of an antiparallel PNA–PNA duplex with a terminal l-amino acid is characterized by alternate maxima, in the 270–280-nm, 250–260 nm, 240–245 nm, and 220–230 nm regions; the molar ellipticity for the 250–260 nm band depends both on the length of the PNA and on the type of amino acid used for induction, as shown in Figure 18.5, with differences [θ ]275 − [θ ]255 per base in the range 2000 to 10,000 deg M−1 cm−1 . The development of the CD spectrum upon PNA mixing is quite slow (few minutes) as compared to the appearance of the hypochromicity effect (seconds), suggesting that the base-pairing process occurs quite fast, followed by a slower reorganization with the emergence of a preferred helical handedness [25]. Replacement of the l-lysine by other l-amino acids gives essentially the same pattern, but of different intensity. In general, all the tested amino acids linked at the C-terminus with the same conﬁguration (Lys, Leu, Phe, Ala, Glu) induced the same CD pattern, irrespective of the side chain, with the only exception of l-Glu at pH 5, whose CD spectrum appeared to be the mirror image of l-Glu (and of all the other l-amino acids) at pH 7. The presence of a positively charged side chain was found to give rise to stronger CD signals, as inferred from a systematic study involving PNA with His, Lys, Arg, and N ε -acetylated-Lys [26]. On the base of antiparallel PNA–PNA duplexes formed by PNAs of different length (4-, 6-, 8-, 10-, 12-mer) a leveling of helicity was proposed to occur from 10-mer to 12-mer [25]. A subsequent systematic study completed this series with oligomers up to 19-mer, showing that propagation of the helicity can increase the CD signal beyond this limit. Furthermore, the process of generation and propagation of the helical structure was found to have a complex behavior, due to multiple-state conformations, depending on

(b) 6

2

4

0 220 –2

240

260

280 300 Wavelength (nm)

–4 –6 –8

10-Arg 10-Lys 10-His

2 CD (mdeg)

CD (mdeg)

(a) 4

0 220 –2 –4 –6

–10

–8

–12

–10

240

260

280 300 Wavelength (nm) 17-Arg 17-Lys 17-His

Figure 18.5.

Example of CD spectra of duplexes formed by an achiral PNA and its complementary PNA strand bearing an amino acid residue at C-terminus. (a) 10-mer (HAGTGATCTAC-aa-NH2 /H-GTAGATCACT-NH2 ) and (b) 17-mer (H-CTGTGACAGTGATCTAC-aa-NH2 / H-GTAGATCACTGTCACAG-NH2 ). The CD intensity is a function of both the amino acid used

(indicated in the graph) and the length of the duplex. Spectra were recorded in water at pH 7.0, with a PNA strand concentration of 5 μM. (Reprinted with permission from reference 26, copyright 2010, American Chemical Society.)

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

the amino acid, the length of the duplex (Figure 18.5), and the solvent used [26]. This is in line with the observation that, in the solid state, PNA–PNA duplexes bearing a chiral amino acid at the C-term were found to form equal amounts of right- and left-handed helices, whereas in solution different external constraints give rise to a prevalence of one helix handedness [27]. The experimental CD spectrum of the antiparallel PNA–PNA duplex with a l-lysine at the C-terminus was found to be quite similar to the theoretical CD spectrum calculated for a canonical B-form DNA of the same sequence, supporting a right-handed structure for the antiparallel PNA–PNA duplex [8]. The CD spectrum predicted for a helical stack of base pairs was calculated using the Schellman matrix quantum mechanical method [25, 28, 29]. However, it was later shown by Rasmussen et al. [12] that the PNA–PNA structure is signiﬁcantly different from canonical forms such as B-DNA. The antiparallel PNA–PNA helices were found to be very wide, largely pitched, and with the base pairs perpendicular to the helix axis. Thus, PNA molecules were found to have a unique structure very different from standard B- or A-helix, suitably called P-helix. The actual helical sense was later demonstrated by several experimental evidences to be left-handed (vide infra). Interestingly, despite the analogies in CD spectra and the presence of similar chromophores, different molecules do not necessarily have the same type of structure and not even the same type of helicity. The correct assignment of the handedness to the antiparallel PNA–PNA duplexes took a more deﬁnitive turn with the appearance of PNAs incorporating chiral amino acidderived monomers [30], which can be easily synthesized with amino acidic side chain in position 2 or 5 or both (Figure 18.3f) [31, 32]. PNA–PNA duplexes of PNA with an aminoacidic side chain in position 2 show CD spectra analogous to those observed for PNAs bearing a lysine residue at the C-terminus. PNAs incorporating monomers based on l-amino acids give spectra analogous to those with a l-lysine at the C-terminus whereas PNAs incorporating monomers based on d-amino acids give spectra analogous to those with a d-lysine at the C-terminus [33]. Chiral monomers inserted in the middle of the strand induce a stronger CD signal, as compared to PNA bearing a chiral monomer at the N- or C-terminus (Figure 18.6a). Thus a stronger preference in the helix handedness of an antiparallel PNA–PNA duplex is linked to a restricted conformational mobility of the chiral residue. In all cases, as in the case of amino acids linked to the C-terminus, the CD spectra of the PNA single strand are much weaker or nearly absent [33]. Since PNAs containing substituted chiral monomers with C2-modiﬁcation, obtained from d-amino acids (2d), were found to bind complementary DNA with higher afﬁnity than their enantiomers, as determined by melting temperature measurements [30, 31], it was assumed that they are in a more favorable preorganization to bind right-handed DNA and thus it was proposed that these tend to form right-handed, rather than lefthanded, helices. This assignment was independently conﬁrmed by a method in which the handedness of antiparallel PNA–PNA duplexes can be inferred by CD of aggregates templated by the PNA–PNA duplex (see Section 18.5) [34]. The preferential handedness of PNA–PNA duplexes, both antiparallel and parallel, were further investigated by using the so-called “chiral box” PNAs—that is, PNAs bearing three consecutive chiral monomers with amino acidic side chain (derived from lysine) in position 2 in the middle of the sequence [35, 36]. Quite interestingly, in the antiparallel PNA–PNA duplexes, l-Lys PNAs form left-handed helices, while d-Lys PNAs form right-handed helices, as previously shown. The helicity assignments can be inferred directly observing the signal of the CD spectra around 260 nm according to the model previously proposed; and the induction, as deduced from the intensity of

593

594

Mol.Ellp. (Deg/M cm) 30000 (a) 25000 20000 15000 10000 5000 0 –5000 –10000 220 230 240 250 260 270 280 290 300

θ (mdeg)

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

40 30 20 10 0 –10 –20 –30 –40 220

par. L-PNA-PNA

par. D-PNA-PNA

240

260

280

300

320

30

(c)

(d) 20 10 CD (mdeg)

θ (mdeg)

antipar. L-PNA-PNA

Wavelength (nm)

Wavelength (nm) 50 40 30 20 10 0 –10 –20 –30 –40 –50

(b)

antipar. D-PNA-PNA

0 15

–10 225

250

275

300

Wavelength (nm)

325

5

350

0 120 240 360 480 600 –5

–20

–15

–30 200

220

240

260

280

300

320

Wavelength (nm)

Figure 18.6. (a) CD spectra (molar ellipticity calculated as a function of base concentration) of PNA–PNA duplexes formed by an achiral PNA (H-AGTGATCTAC-NH2 ) and complementary chiral PNAs incorporating D-Lys monomers in different positions: H-GTAGAT(2D−Lys) CACT-NH2 (dotted line, one chiral monomer in the middle), H-G T(2D−Lys) AGA T(2D−Lys) CAC T(2D−Lys) -NH2 (thin line, three scattered chiral monomers), H-GTAG A(2D−Lys) T(2D−Lys) C(2D−Lys) ACT-NH2 (thick line, chiral box PNA). (From reference 35 copyright 2000, Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission.) (b) Comparison between the CD spectra of antiparallel and parallel duplexes obtained with the chiral-box PNA containing either D-or L-Lys-derived C2-modiﬁed monomers H-GTAG A(2D−Lys) T(2D−Lys) C(2D−Lys) ACT-NH2 and H-GTAG A(2L−Lys) T(2L−Lys) C(2L−Lys) ACT-NH2 . (From reference 36, copyright 2005, John Wiley & Sons, reproduced with permission.) (c) CD spectra of PNA–PNA duplexes formed by PNA containing a single chiral monomer with substitution both at C2 and at C5 (H-GTAGAT(2,5−Lys) CACT-NH2 ) and complementary achiral antiparallel PNA with chiral conﬂict. Solid line: one monomer containing 2L,5L lysine-derived side chains. Broken line: one monomer containing 2D,5D lysine-derived side chains (data from reference 18). All measurements were carried out at 25◦ C in phosphate buffer at pH 7.0, with 5 μM concentration of each strand. (d) CD spectra of (PNA-TTTTTSS TTTTT-L-LysNH2 )2 /PNA-A10 -GlyNH2 (dotted line) and (PNATTTTTRR TTTTT-L-LysNH2 )2 /PNA-A10 -GlyNH2 (solid line) triplexes, containing a chiral monomer with either a (S, S)- or (R, R)-diaminocyclohexane in place of the aminoethyl moiety. The insert shows the CD at 255 nm as a function of time (in seconds) upon mixing the PNAs in a 2:1 base ratio (20◦ C). (Reprinted with permission from reference 38, copyright 1997, American Chemical Society.)

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

the CD signal, is higher than that observed with the same type and number of chiral monomers scattered (Figure 18.6a,b). The CD signals of parallel PNA–PNA duplexes were found to be very weak and with a maximum in the 260- to 270-nm region with a reversed preferential helicity (l-Lys, right-handed; d-Lys, left-handed, Figure 18.6b) [36]. Accordingly, the preferred mode of binding and the best mismatch recognition of the d-Lys containing PNA with (right-handed) DNA is in the antiparallel orientation, while that of l-Lys PNA is in the parallel mode, conﬁrming the empirical assumption that chiral PNAs give a better performance in DNA recognition when having an intrinsic preference for right-handed helicity [36]. Finally, the helical preference of PNAs including monomers bearing chiral amino acid side chain in position 5, or position 5 and 2 simultaneously, has also been studied by using CD spectra in order to infer the helical preference of antiparallel PNA–PNA duplexes [18, 32]. In contrast with what has been observed for PNAs with chiral monomers substituted at C2, monomers with d-amino acids at C5 induce a preference for left-handed helices, and monomers with l-amino acids show a preference for right-handed helices. When both substitution are present simultaneously (2 and 5), a “chiral accord” (induction of the same handedness from both stereogenic centers) or a “chiral conﬂict” (induction of opposite helices) can arise. In the case of chiral accord, CD spectra conﬁrm that the helical preference of the antiparallel PNA–PNA duplexes is exactly what can be expected (right-handed in the case of 2d,5l and left-handed in the case of 2l,5d). In the case of “chiral conﬂict”, CD spectra clearly indicate that the induction exerted by the stereogenic center in position 5 is prevalent (right-handed in the case of 2l,5l and left-handed in the case of 2d,5d, Figure 18.6c). Again, also in these cases the PNA–DNA duplex stability is related to the strength of the preference for the right-handed helical conformation [18]. Two stereogenic centers were also present in the modiﬁed PNAs synthesized by Lagriffoule et al. [37], including monomers obtained by replacing the aminoethyl portion of the backbone by a 1,2-diaminocyclohexyl moiety, either in the (S , S ) or the (R, R) conﬁguration, thus introducing stereogenic centers at C5 and C4 (PNA structure as in Figure 18.3e). CD spectra of PNA single strands and of PNA–PNA antiparallel duplexes were very similar to those already seen for the previous chiral PNAs based on standard amino acids, both in terms of wavelength maxima and minima and in terms of CD intensities. Quite interestingly, the (S , S )-monomer, whose conﬁguration at C5 corresponded to that obtained by the insertion of an l-amino acid side chain, also induced a right-handed helix in the PNA–PNA duplexes, further conﬁrming that only the conﬁguration of the stereogenic centers determine the PNA helix handedness, irrespectively of the groups inserted. Consistently (R, R)-monomers gave, as C5 d-amino acid-based monomers, left-handed PNA–PNA helices. This type of PNA were also reported to form PNA–PNA–PNA triplexes; the corresponding spectrum is reported in Figure 18.6d [38], which has unique features, showing a series of bands of the same sign in the 280- to 290-nm, 250- to 260-nm, and 220- to 230–nm regions.

18.3. DUPLEXES AND TRIPLEXES OF PNA WITH DNA AND RNA Normally, in ECD measurements of PNA with DNA and RNA the concentration of both components is calculated from the absorbance using the Lambert–Beer law and with the approximation of considering the single base extinction coefﬁcients (ε) as additive. For PNA, the following values of ε are used, according to Nielsen’s work [2]: 13,700 for

595

596

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

A, 6600 for C, 11,700 for G, and 8600 for T, whereas for DNA and RNA the molar absorbtivity is either provided by the supplier or can be calculated with available free software [39]. For some PNAs, especially those without terminal or internal modiﬁcations, aggregation or adsorption on the walls of the cuvette is possible. Thus, the stock solution concentration is best measured at higher temperature (50–60◦ C). After mixing the various components in the measuring buffer, incubation at high temperature (normally 90◦ C for 5 min) and slow cooling at the desired temperature is performed, in order to disrupt eventual secondary structures and to favor hybridization; this step is necessary when the PNA is hybridized to long DNA or RNA fragments, or in the presence of strong self-aggregation of PNA.

18.3.1. PNA–DNA and PNA–RNA Duplexes of Achiral PNA The CD spectrum of achiral PNA–DNA antiparallel duplex was systematically described by Nord´en and co-workers in an early work on PNA-based structures [40]. The PNA–DNA antiparallel duplex formed by an achiral PNA shows a complex spectrum, containing several CD bands (Figure 18.7). Common features of PNA–DNA spectra are an intense Cotton effect in the range 260–265 nm, with typical [θ ] per base in the range of 10,000–20,000 deg M−1 cm−1 , and a second Cotton effect at 240–245 nm, which is much more variable both in sign and intensity; both bands are slightly shifted toward shorter wavelengths if compared with DNA–DNA duplexes of the same sequence. Since the PNA–DNA duplex is different from the ssDNA, in particular at 260–265 nm, where the ssDNA CD intensity is nearly zero, the process of formation of a PNA–DNA duplex can be followed by (a) titration of the DNA with increasing amounts of PNA and (b) monitoring the formation of the duplex at 260–265 nm. Job-plot can provide information on the stoichiometry of the complex, which is very useful when both PNA–DNA duplexes and PNA–DNA–PNA triplexes are likely to be formed, in particular with pyrimidine-rich PNAs. Sometimes, a longer-wavelength transition with lower intensity is present in the range 280–290 nm. The intensity and sign of this band is strongly dependent on the sequence and on the length of the duplex. Sugimoto et al. [41] reported a systematic study on the formation of PNA–DNA duplexes as a function of the length and sequence at 5◦ C, with the aim of providing evidences of the validity of the nearest-neighbor model for structure and stability. Some of their results are reported in Figure 18.7. Several general features are evident from these data: (a) The longer-wavelength band is either positive or negative, depending mainly on the sequence and not on the length; (b) the CD spectrum is very similar when the same nearest neighbors are present, but only for short oligomers; (c) for longer oligomers, the CD spectrum is entirely depending on the sequence used, regardless of the nearest-neighbor components; this effect is evident from the 10-mer duplexes on. Unfortunately, no theoretical prediction of electronic CD of these duplexes has so far succeeded to explain this rather complex behavior. The shorter-wavelength (200–230 nm) spectrum is even more complicated due to superposition of nucleobase and peptide chromophores; thus common features of different duplexes cannot be found. However, since the PNA–DNA and DNA–DNA duplexes are often similar in the 230–300 nm region, formation of a PNA–DNA duplex from hairpin DNA could be followed by Armitage, Nielsen, and co-workers using the 225 nm CD signal [42]. The parallel DNA–PNA duplex (with the N-terminus of PNA facing the 5 terminus of DNA) is formed with slow kinetics, and its spectrum is completely different from that

1.0 0 –1.0 –2.0 –3.0 200

10–5 [θ] (deg cm2 dmol–1)

10–5 [θ] (deg cm2 dmol–1)

(a) 2.0

(c) 2.0 1.0 0 –1.0 –2.0 –3.0 200

240 280 320 Wavelength (nm)

(b) 2.0 1.0 0 –1.0 –2.0 –3.0 200

240 280 320 Wavelength (nm) 10–5 [θ] (deg cm2 dmol–1)

10–5 [θ] (deg cm2 dmol–1)

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

240 280 320 Wavelength (nm)

(d) 2.0 1.0 0 –1.0 –2.0 –3.0 200

240 280 320 Wavelength (nm)

Figure 18.7. CD spectra of antiparallel PNA:DNA duplexes of various sequences and length: (a) 6-mers PNA(CCGACG)/d(CGTCGG) (thick line) and PNA(CGACCG)/d(CGGTCG) (thin line); (b) 8-mers PNA(CTCACGGC)/d(GCCGTGAG) (thick line) and PNA(CACGGCTC)/d(GAGCCGTG) (thin line); (c) 10mers PNA(GCTAACAGCG)/d(CGCTGTTAGC) (thick line) and PNA(GCGCTACAAG)/d(CTTGTAGCGC) (thin line); (d) pna(ATAAATTGGATACAAA)/d(TTTGTATCCAATTTAT) (thick line) and pna(CAAATGGATTAAATAC)/d(GTATTTAATCCATTTG) (thin line). Sample concentration was 70 μM, and measurements were done phosphate buffer (pH 7.0) containing 1 M NaCl, at

5.0◦ C. (Reprinted with permission from reference 41, copyright 2001, American Chemical Society.)

of the antiparallel one (Figure 18.8b), with a negative band at 285 nm and a positive one at 260 nm [40]. The increase in ionic strength causes a decrease in the stability of the parallel duplex (Figure 18.8b), unlike the antiparallel (Figure 18.8a, where the spectra are superimposable). This suggests that the conformation of the DNA is highly distorted by this complexation process since a severe rearrangement of the conformation is required, in line with the lower stability and the slow kinetics of formation of these duplexes. The structural features of the parallel duplexes are at present unclear, although some molecular modeling calculations have been performed to predict their features, which were found to resemble more to the canonical B-form of DNA [43]. Although the spectra of both antiparallel and parallel PNA–RNA duplex were reported in early studies, the CD spectra of PNA–RNA have been less extensively studied. However, RNA complexation is one of the most interesting processes in the prospective of the use of PNA as antisense drugs (i.e., targeting mRNA and blocking translation) and even more in recent years for the use of PNA for blocking microRNA (miR) expression. Measurements involving RNA strands require special care in order to avoid contamination by ubiquitous RNases; therefore, RNase free water should be

597

598

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(b) 120 100 80 60 40 20 0 –20 –40 –60 –80 –100 –120 200 320 Δe (M–1 cm–1)

Δe (M–1 cm–1)

(a) 120 100 80 60 40 20 0 –20 –40 –60 –80 –100 –120 200

220

240 260 280 Wavelength (nm)

300

220

240 260 280 Wavelength (nm)

300

320

Figure 18.8. Circular dichroism spectra of (a) antiparallel PNA–DNA duplex (sequence: PNA N-GTAGATCACT-C/DNA 5 -AGTGATCTAC-3 ) at low (0 M NaCl, solid line) and high (500 mM NaCl, broken line) ionic strength; (b) parallel PNA:DNA duplex (sequence: PNA N-GTAGATCACT-C/DNA 5 -CATCTAGTGA-3 ) at low (0M NaCl, solid line) and high (500 mM NaCl, broken line); in both cases a 10 mM phosphate buffer at pH 7.0 was used with 5 mM strand concentration (Reprinted with permission from ref. [40] Copyright 1996 American Chemical Society.)

used in the preparation of samples and stock solutions of RNA should be frozen and checked for degradation before being reused after the ﬁrst preparation. Two examples are reported in Figure 18.9 (unpublished data from our laboratory) in which PNA–DNA and PNA–RNA duplexes of the same sequence and length are compared. The overall shape is conserved, but slightly shifted for the PNA–RNA duplex. Both the 260–265 nm maximum and the 240–245 nm minimum are shifted to shorter wavelength, whereas the 280–290 nm band has the same sign and similar wavelength in the two complexes. Since the RNA–PNA complexes have higher melting temperatures than the PNA–DNA duplexes, it is reasonable to infer that the conformation corresponding to the PNA–RNA spectrum is more compatible with the PNA structure.

(a) 20

(b) 50 40 30

10 5 0 220 230 240 250 260 270 280 290 300 l (nm) –5

θ mdeg

θ (mdeg)

15

20 10 0 –10

220 230 240 250 260 270 280 290 300 λ (nm)

–20

Figure 18.9.

Comparison between PNA–DNA (thin lines) and PNA–RNA (thick lines) duplexes of the same sequence and length: (a) PNA 10-mer: H-GTAGATCACT-NH2 ; DNA (or RNA) 5 -AGTATCTAC-3 ; (b) PNA 18-mer: H-CCGCTGTCACACGCACAG-NH2 DNA (RNA) 5 CTGTGCGTGTGACAGCGG-3 . Measurements were done in a 10 mM phosphate buffer at pH 7 containing 0.1 M NaCl (PBS-buffer), concentration of each strand was 5 μM.

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

PNA bearing methyl groups on amide nitrogens were described; upon introduction of 30% (three out of 10), N -methyl units in the PNA strand(s) no major changes were detected in PNA–DNA and PNA–RNA CD spectra, although full methylation gave distinctly altered CD spectra. However, no major changes were detected by CD spectroscopy in the formation of PNA–PNA duplexes for the fully methylated PNA [44].

18.3.2. PNA–DNA and PNA–RNA Complexes of Modiﬁed Chiral PNA As described in Section 18.2, the use of modiﬁed chiral PNA can induce pre-organization of the single-stranded PNA. This, in turn, affects the ability of the PNA to bind to DNA with the proper conformation. A duplex formed by an achiral PNA and DNA of the same length and sequence can be used as a reference for a not-distorted conformation. An example of this approach comes from the comparison of the spectra obtained for the series of PNA containing a single chiral (either C2- or C5-modiﬁed) monomer (depicted in Figure 18.3f), with complementary antiparallel DNA (Figure 18.10). Since the sequence presented is the same as in Figure 18.9a, direct comparison can be done with the achiral PNA spectrum. It turns out that the C2-modiﬁed PNA synthesized from d-Lys (2d) and C5-modiﬁed PNA derived from l-Lys (5l) have a spectrum with a (b) 20

(a) 20 PNA (2D)/apDNA

10 5

10 5 0

0 –5

PNA (5L)/apDNA

15 θ (mdeg)

θ (mdeg)

15

225

PNA (2L)/apDNA 250 275 300 325 Wavelength (nm)

–5

PNA (5D)/apDNA 225

250

275

300

325

350

Wavelength (nm)

θ (mdeg)

(c) 10 8 6 4 2 0 –2 –4 –6 –8 –10 220 230

350

240 250 260 270 280 290 300 Wavelength (nm)

Figure 18.10. CD spectra in solution of PNA containing one C2- (a) or C5-modiﬁed chiral monomer (b) derived from D- or L-Lys (H-GTAGATLys CACT-NH2 ) with antiparallel DNA (apDNA, 5 -AGTGATCTAC-3 ) in phosphate buffer at pH 7.0 (concentration: 5 μM of each strand) (from reference 18, copyright 2007, Wiley-VCH Verlag GmbH & Co. KGaA, reproduced with permission.) (c) CD spectra of PNA:DNA duplexes formed by achiral PNA (H-GTAGATCACT-NH2, thin lines with squares) or PNA (H-GTAGAD(L)-Lys TD(L)-Lys CD(L)-Lys ACT-NH2 ) containing a ‘‘chiral box’’ of C2-modiﬁed monomers derived from D- (thick solid line) or L-Lys (thin, broken line) with parallel DNA (5 -CATCTAGTGT-3 ) in phosphate buffer at pH 7.0 (concentration: 5 μM of each strand) (from reference 36, copyright 2005, John Wiley & Sons, reproduced with permission.)

599

600

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

pattern similar to the achiral PNA–DNA duplex, though with more intense bands, whereas the corresponding enantiomers (2l or 5d) have a different spectrum with a less intense band in the 260–265 nm region. This behavior parallels the preference of the corresponding PNA to form right-handed PNA–PNA helices (see above), which is an index of the preference for right-handed helical structures; this preference is also reﬂected in the melting temperature of these PNA–DNA duplexes, which were found to be higher for the 5l and 2d PNA (56◦ C and 52◦ C, respectively) than that of achiral PNA (50◦ C) and those of 5d and 5l (32◦ C and 47◦ C, respectively). A similar effect, but with a reversed stereoselectivity, was observed for the parallel PNA–DNA duplexes formed by PNA containing a stretch of three C2-modiﬁed chiral residues in the middle (“chiral box”); as shown in Figure 18.10, the l-Lys-containing PNA gave a spectrum of higher intensity than the achiral one, whereas the corresponding d-Lys PNA gave a different spectrum that was the superposition to that of the single strands. Thus the 2D “chiral box” PNA was demonstrated to bind DNA only in an antiparallel orientation, giving rise to a complete direction control [35]. Chiral cyclic PNA monomers have been largely used as alternative structures to increase PNA performances; several reviews are available which describe the synthesis and DNA binding properties of these derivatives [22, 45]. However, chiroptical properties of the PNA and of their complexes have not been described in all cases. The PNA–DNA spectra for chiral cyclic PNA derivatives can be very different from those of acyclic PNA–DNA and PNA–RNA duplexes with modiﬁed backbones. As an example, CD spectra of antiparallel PNA–DNA duplexes formed by a PNA containing a d- or l-trans-4-aminoprolyl monomer (tT*) at the N-terminus ((tT*)-AT2 AT2 AT2 ) were described by Ganesh and co-workers (Figure 18.11b,c). In this case, both the intensity and the maxima observed were different from those of unmodiﬁed PNA–DNA and strongly dependent on stereochemistry of the monomer. In spite of the drastic changes in the overall conformation, both PNAs were reported to bind to DNA better than unmodiﬁed aminoethylglycine-based PNA (aeg-PNA). CD spectra were also utilized in order to probe complexation in the case of ﬂuorinated oleﬁnic PNA (FOPA) [46], (1S,2R/1R,2S )-ciscyclopentyl-[47], aminoethylprolyl-[48], piperidine-[49], or pyrrolidinyl-PNA [50]. Since prediction of CD spectra from modeling studies is still not available, CD is sometimes used for assessing a possible similarity of the conformation of the duplexes formed with those of natural oligonucleotides or with those formed by unmodiﬁed PNA. A 1:1 complex was detected by CD for pyrrolidinyl PNA which showed preferential binding for DNA over RNA [50]. Similarly, Vilaivan and co-workers used the CD spectrum to demonstrate the occurrence of 1:1 complex of a T10 PNA based on a (S , S )-prolyl-2-aminocyclopentanecarboxylic acid (ACPC) backbone with A10 DNA with a conformation resembling that of dA10 :dT10 , but with higher stability [51]. Ganesh and co-workers used the CD intensity of PNA–RNA and PNA–DNA duplexes to give a rationale for the preference of cyclohexanyl-PNA for RNA over DNA, since the former gave a well-structured spectrum similar to that of unmodiﬁed PNA, whereas the latter showed a highly distorted pattern [52, 53]. Kumar and Ganesh described the properties of aminoethylpipecolyl (aepip)-aegPNA chimerae using CD. The insertion of one (2S , 5R)-monomer into an unmodiﬁed PNA sequence gave a spectrum closely resembling the PNA–DNA duplex, with slightly higher intensity (Figure 18.11d). This was in line with the observed higher stability of the antiparallel duplex formed by the modiﬁed PNA [54].

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

Similarity between the CD spectra of polyA RNA complexes with the backboneextended pyrrolidine-PNA (bepPNA) and that of aegPNA was used by Kumar and coworkers to explain the preference of these molecules for binding RNA over DNA [55]. Finally, many PNA-related molecules such as pyrrolidine-based oxy-peptide nucleic acids [56] and amino acid-modiﬁed oligonucleotides [57, 58] have been studied using approaches similar to those described above.

18.3.3. PNA–DNA–PNA Triplexes PNA–DNA–PNA triplexes are formed by Watson–Crick and Hoogsteen base pairing in TAT and CGC+ (C+ is protonated cytosine) triplets and are the only complex formed when the PNA is entirely composed of pyrimidine and the DNA entirely of purine nucleobases [1, 2]. A typical example of the CD spectrum of a PNA–DNA–PNA triplex in the case of a T8 PNA–LysNH2 with poly dA in a 2:1 ratio (on a nucleobase ratio) is Base

(a)

(b) 2 O

Aminoprolyl (ap)

D-trans-ap

1

N

CD m deg

H N O Base

0

240

–1

260 l nm

280

300

–2 –3

Aepip

–4

N

(c) +4

–5

O

L-trans-ap

CD m deg

+3 +2 +1 0 –1

240

260 l nm

–2

280

300

(d) Ellipticity (mdeg)

N H

3 2

(ii) (i)

1 0 –1 220

240 260 280 Wavelength (nm)

300

320

Figure 18.11. Examples of CD spectra of duplexes of modiﬁed PNA. (a) Structure of trans4-aminoprolyl-PNA (ap-PNA) and of aminoethylpipecolyl (aepip-PNA) monomer, (b) D-trans-4aminoprolyl-PNA, and (c) L-trans-4-aminoprolyl-containing PNA with antiparallel DNA (sequence: PNA H-(Lt.T* or Dt-T*)-ATTATTATT-CONHCH2 CH2 COOH, DNA: 5 dAAT AAT AAT A 3 ) in buffer, 10 mM phosphate at pH = 7.3, c = 6 μM of each strand (b,c: reprinted from reference 24, copyright 1999, with permission from Elsevier); (d) aepip-PNA monomer inserted in an achiral PNA strand (sequence: A T G T* T C T C T T T-(b-Ala)-OH, where T∗ = aminoethylpipecolyl monomer aepip) hybridized with antiparallel DNA (i) and compared to the unmodiﬁed PNA (ii), thus showing little distortion from PNA–DNA structure. Buffer: 10 mM sodium phosphate, pH 7.30, concentration = 2 μM of each strand. (Reprinted from reference 24, copyright 1999, with permission from Elsevier.)

601

602

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

25 20 15 10 5 0

(b) 8 35°C 0 0.5 1 1.5 2 2.5 [PNA]/[poly(dA)]

15

0

CD at 255 nm

CD (m deg)

(a) 30

6 21°C 4 12°C 2 0

–15 230

250 270 290 Wavelength (nm)

310

0

200

400 600 800 1000 1200 Time (seconds)

Figure 18.12. (a) Evolution of the CD signal of DNA poly(dA) (25 μM in bases) with increasing concentration of PNA T8 from bottom to top at 260 nm: 0.00, 0.33,0.67,1.00, 1.33,1.67,1.83,2.00, and 2.33. Inset: signal at 260 nm as a function of PNA–DNA base ratio. (Reprinted with permission from reference 14, copyright 1993, American Chemical Society.) (b) Kinetics followed by CD at 255 nm upon mixing 80 μM bases of PNA-T8 with 20 μM base pairs of poly(dA)–poly(dT) performed at 12◦ C, 21◦ C, and 31◦ C (in a 5 mM phosphate buffer, pH 7.0, containing 50 mM NaCl). (Reprinted with permission from reference 59, copyright 1996, American Chemical Society.)

reported in Figure 18.12a. This combination has been shown to give rise to a very stable triplex structure, which is formed also in long dsDNA segments by a strand displacement process [59]. The triplex has a typical spectrum with maxima at 285 nm, 275 nm, and 255 nm and a minimum at 240–245 nm. The 255 nm signal was used by Nielsen, Nord´en, and co-workers to establish the stoichiometry of the complex; it was possible, following the CD signal at this wavelength, to measure the kinetics of formation of triplex structures, which was found to be in the timescale of minutes (Figure 18.12b). The spectrum of a transient species could also be inferred from kinetic data at different wavelengths [59]. A case of PNA–DNA2 triplex formed by cytosine-rich homopyrimidine PNA was also reported, and it was documented following the variation of the CD signal as a function of PNA concentration [60]. Similarly, oligopyrimidine PNA containing one or two modiﬁed aepip monomers were shown to form PNA2 –DNA triplexes with polypurine DNA, and this was accompanied by a CD signal typical of PNA2 –DNA triplexes (Figure 18.13a) [54]. One interesting case is that of ornithine-based PNA, depicted in Figure 18.13b [61]. This system was shown to be not very efﬁcient in binding DNA and to give slightly more stable triplexes with polyA RNA. A careful analysis of its optical purity revealed that severe epimerization had taken place during the solid-phase synthesis of these compounds. The optically pure enantiomers (TD-Orn )10 and (TL-orn )10 were synthesized using special reaction procedures [61, 62], and were found to bind to complementary RNA forming triplexes with low stability and very peculiar spectral features. In fact, upon complexation with both PNAs, a reversal of the CD spectrum was observed, indicating that the RNA conformation was completely modiﬁed by triplex formation (Figure 18.13). The spectrum was also very different from that of a typical PNA–DNA–PNA triplex. Analysis at different temperatures revealed two major transitions: One led to a CD spectrum similar to that of RNA and was more visible in the CD spectrum, and the second was more evident in the UV melting and was attributed to dissociation of base pairs [61]. Analogous results were obtained for systems based on Lys and 2-aminobutyric acids as backbones [63].

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

(a)

(b) 30 (i)

1

(ii) (iii)

(iv)

0

θ (mdeg)

Ellipticity (mdeg)

2

Base

−1 N

−2

N H

220

240

260

(a)

10

(b)

0

(c)

−10

(d)

300

320

−30 250

Wavelength (nm)

O

Base H N

−20 O

280

20

(e) 270

HN O

290

310

λ (nm)

Figure 18.13. (a) Triplex formation by homopyrimidine PNA containing aepip-PNA monomer (T*, inset) with DNA 5 -GCAAAAAAAACG-3 showing similarities of the triplex formed; PNA: (i) H-TTTTTTTT*-(β-Ala)-OH, (ii) H-T T T T*TTTT*-(β-Ala)-OH, (iii) H-TTTTTTTT-(β-Ala)-OH, (iv) HT*TTTTTTT-(β-Ala)-OH; measurements were made in 10 mM sodium phosphate, pH 7.30, strand concentration: 1.5 μM. (Reprinted from reference 54, copyright 2004, with permission from Elsevier.) (b) Circular dichroism of: (a) RNA polyA (50 μM in bases); (b) D-ornithine T10 (structure shown in inset, 100 μM in bases); (c) L-ornithine T10 (100 μM in bases); (d) RNA polyA (50 μM in bases) L-ornithine T10 (100 μM in bases); (e) RNA polyA (50 μM in bases) D-ornithine T10 (100 μM in bases). Measurents were made in PBS buffer containing 2.5% of DMF. (From reference 61,2002, copyright John Wiley & Sons, reproduced with permission.)

Similarly, Lowe and co-workers reported the formation of triplexes of PNA poly-T made entirely of N -aminoethyl-d-proline monomers; the low thermal stability of these triplexes were accompanied by a drastic change in the CD spectrum, with negative maxima observed in the 260–280 nm region [64].

18.3.4. PNA in Quadruplexes and i-Motifs Several different noncanonical DNA structures have found application in the formation of nanostructures and nanomotors [65]. G-quadruplex and i-motifs are two main examples of this class. The G-quadruplex motif is formed by several inter- or intrastrand interactions based on stacked guanine tetrads formed by guanine-rich DNA or RNA strands. Guanine-rich sequences with quadruplex forming potential are located in the promoter regions of regulatory genes [66]. I-motifs are structures occurring with C-rich oligonucleotides, consisting of two parallel duplexes, formed by interaction of cytosine and protonated cytosine at appropriate pH, that are intercalated in an antiparallel orientation [67]. It has been shown that carefully designed G-rich PNA oligomers were capable of forming PNA4 quadruplexes [68], hybrid PNA2 –DNA2 [69], PNA2 –RNA [70, 71], or (PNA–DNA chimeras)4 [72] tetramolecular and trimolecular [73]. C5-modiﬁed chiral PNA could be used in order to induce selectivity in quadruplex versus duplex formation [74]. CD spectroscopy is highly informative in these type of studies. For example, quadruplex-forming oligonucleotides with one PNA at either 3 - or 5 -end were analyzed by CD measurements, showing that the overall conformation was conserved, as inferred from the persistence of a positive band at 264 nm and a negative band at 243 nm, typical of parallel DNA quadruplexes [75]. Thermal denaturation of these structures can be easily followed by CD measurements at different temperatures; the kinetics of quadruplex

603

604

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

formation was studied by monitoring the ellipticity at 264 nm as a function of time after heating. A G3 PNA oligomer bearing an acrydone unit at the N-terminus was mixed with a three-repeat fragment of human telomeric DNA d(GGGTTAGGGTTAGGG). The measured UV melting temperature was only one degree higher than that of the DNA quadruplex. However, the CD spectrum changed signiﬁcantly, showing different intensity of the 265 nm and 295 nm bands, characteristic of parallel and antiparallel quadruplex structures, respectively [76]. The interaction of guanine-rich PNA with a portion of c-Myc RNA was studied, comparing the formation of PNA–DNA duplex via Watson–Crick base pairing and mixed quadruplex complexes based on potassium-promoted guanine tetrad formation. A 2:1 PNA–RNA complex with structural features resembling those of parallel quadruplex was demonstrated using CD signals [74]. DNA4 i-motifs show a characteristic CD proﬁle with a positive maximum near 285 nm, which is followed by a negative trough with a minimum centered near 265 nm [67]. PNA were shown to be able to form tetrameric PNA4 [77, 78] and mixed PNA2 –DNA2 [79] and PNA2 –RNA2 [80] i-motifs. Only for the mixed PNA2 –DNA2 structure, CD data are available: A 1:1 mixture of DNA and PNA showed a CD signal that is ampliﬁed nearly twofold compared to the DNA i-motif, with the same shape, suggesting that the chirality of the DNA duplex is transferred to the nucleobases forming the PNA duplex [80].

18.4. THERMAL DENATURATION STUDIES The thermal stability of duplexes and other structures formed by PNA are normally carried out by UV measurements at variable temperatures and are based on the hyperchromic effect that is observed as a consequence of strand separation [81]. Similarly, since the CD spectra of either PNA–DNA or PNA–RNA are different from the sum of the single components (especially in the case of achiral PNA or PNA conjugated with amino acids and peptides which show no CD absorption), measurement of CD as a function of temperature can allow to detect the melting transition and, by ﬁtting into two-state models, can give the same thermodynamic information as the UV melting. Melting of PNA–DNA duplexes is reversible, whereas the occurrence of hysteresis in the cooling curve compared to the heating can be considered as signiﬁcant evidence for triplex formation, since the annealing process is slow. There are several advantages in using CD spectroscopy in these type of measurements. First of all, the intensity of a CD band chosen at an appropriate wavelength at which the two single strands give weak or no signal allows to follow the melting in a more speciﬁc way. The study of the interaction of an antitumor PNA 18-mer conjugated with a cationic peptide, able to act as anti-gene inhibitor of the MYCN oncogene in cells [82], with its target DNA provided an interesting example. A very unusual melting curve was observed in the UV, which showed an increase in absorbance near the temperature of 90◦ C (Figure 18.14a). Since a typical sigmoidal curve could not be detected in this case, in order to rule out that this increase was due to an aspeciﬁc effect, the CD melting curve at 264 nm, where both ssDNA and ssPNA CD was zero, was studied. Since the PNA–DNA mixture showed a maximum, it was possible to clearly prove the occurrence of a melting transition with mid-point near 90◦ C (Figure 18.14b), thus conﬁrming the exceptional stability of this adduct. Careful analysis of ternary mixtures of dsDNA and PNA at variable temperature allowed us to prove a duplex invasion

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

(a)

(b)

40

1.8 35

1.75 1.7

30 θ (mdeg)

25

1.6 A260

i)

i)

1.65

1.55 1.5 1.45

20 15 10

1.4 5

1.35 15

35

55

75

15

t (°C)

35

55

75

t (°C)

(c) 1.05

(d) 1.00

1

0.80

Tm (UV) = 84°C Norm CD

Norm. Abs

ii)

0

1.3

0.95

iii)

0.9 0.85

0.60 0.40 0.20

T1/2(CD) = 75°C

0.00

0.8 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95

20 25 30 35 40 45 50 55 60 65 70 75 80 85

Temperature (°C)

Temperature (°C)

Figure 18.14. Comparison between UV and CD melting curves. (a) UV melting curve of a high-melting PNA–DNA duplex (PNA: H-ATG CCG GGC ATG ATC T—PKKKRKV-NH2 , DNA 5 -A GAT CAT GCC CGG CAT-3 ); (b) CD melting curves of PNA and DNA as in (a) showing speciﬁcity of the 260-nm signal for PNA–DNA (i) and lack of interference by PNA alone (ii) or DNA alone (iii). Measurements in (a) and (b) were made phosphate buffer (10 mM phosphate, 0.1 M NaCl, 0.1 mM EDTA); C = 5 μM of each strand; scan speed 18◦ C/min (From reference 83, 2008, copyright John Wiley & Sons, reproduced with permission). (c) UV melting curve of a PNA–PNA 15-mer duplex (H–GTGACAGTGATCTAC–Lys–NH2 and H–GTAGATCACTGTCAC–NH2 ) in water at pH 6.8, showing a clear inﬂection point at 84◦ C (c = 5 μM of each strand); (d) normalized CD at 260 nm of the same solution as in (c), showing a mid-point for the transition at lower temperature [84].

process—that is, formation of a PNA–DNA duplex from dsDNA by displacement of one strand [83]. A second advantage in using CD is represented by the possibility to detect conformational transitions that are not or scarcely detected by UV. The conformational changes observed for the ornithine-PNA mentioned above are examples of this effect. However, this implies that the transition observed in the CD melting curve not always coincides with the UV transition, since the former is sensitive to conformational changes, whereas the latter is dictated by unpairing of the nucleobases. An example of how the two effect are correlated but different comes from our own experience on the helicity of PNA–PNA duplexes. As shown in Figure 18.14c,d, by measuring the melting transition of a PNA–PNA duplex in which one of the strands has l-Lys at its C-term, a curve with mid-transition at 84◦ C was observed in the UV [84]. By measuring the CD of the duplex

605

606

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

at 260 nm, the loss of helicity accompanying the melting process could be measured, with a ﬁnal zero CD intensity as expected. However, the mid-point of the transition was in this case at 75◦ C, signiﬁcantly lower than that observed in the UV measurement. This implies that a transition to a different conformation with lower or no CD signal precedes the actual base dissociation, thus suggesting a higher degree of complexity of the melting process. It is evident that the CD “melting temperature” should not be considered as the Tm of the duplex in this case. It is a common experience that CD and UV melting temperatures can be different, an effect that is probably due to some degree of conformational rearrangement that precedes the actual base dissociation. However, in most cases the same results are abtained with both techniques.

18.5. INTERACTIONS OF DYES WITH PNA–XNA DUPLEXES The binding of dyes on PNA:DNA duplexes is very useful for structural diagnostics. Though DNA intercalators do not interact with the PNA–DNA duplex, due to structural difference, several groove binders were found to interact with PNA–DNA duplexes. CD spectroscopy was used to assess the binding of several molecules with PNA containing duplexes and triplexes. The 4 ,6-diamidino-2-phenylindole (DAPI), a DNA minor groove binder, was found to exhibit a circular dichroism with a positive sign and amplitude consistent with minor groove binding. Similarly, a PNA–DNA duplex containing a central AATA motif, a typical minor groove binding site for distamycin A and DAPI, showed the interaction for both drugs, though with strongly reduced afﬁnity compared to that of dsDNA [85]. A very important class of molecules able to interact with PNA–DNA and PNA–PNA duplexes is represented by cyanine dyes, intensely colored compounds that have found a widespread application in numerous ﬁelds [86]. The typical dye structure consists of two heteroaromatic fragments linked by a polymethine chain. The extensive conjugation in the cyanines leads to long-wavelength absorption maxima and large molar absorptivities. For example, 3,3 -diethylthiacarbocyanine (DiSC2 (5)) dye (Figure 18.15a) is constituted by two benzothiazole units connected by a pentamethine linker. The extensive conjugation yields an absorption maximum at 651 nm and an extinction coefﬁcient of 260,000 M−1 cm−1 in methanol. In many cases, cyanine dyes exist not as isolated monomers but rather as aggregates of multiple chromophores. The photophysical and photochemical properties of these aggregates have been studied in great detail and are often quite distinct from those of the monomeric dye. Armitage and co-workers [87] demonstrated the spontaneous assembly of helical cyanine dye aggregates using double-helical DNA as a nanotemplate to precisely control the spatial dimensions of the aggregate (Figure 18.15b). The binding of the dye into the minor groove of DNA, followed by additional insertion of dimers to adjacent sites, is highly cooperative. Due to the lack of hydrogen bond donor and acceptor groups on the dye, there is minimal hydrogen-bonding interaction between the cyanine dye and the DNA. Instead, the binding appears to be driven primarily by hydrophobic effect and van der Waals interactions. Since the PNA backbone is less hydrophilic than DNA, the possibility that DNA-binding cyanine dyes would also associate with PNA-containing duplexes has been also considered [34]. Indeed the dye did not bind to these structures as an isolated monomer or dimers, but rather spontaneously assembled into a helical aggregate using the PNA as a template. Upon binding to PNA–DNA, the dyes show signiﬁcant changes in optical properties such as UV–vis absorbance, circular dichroism,

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

(a)

S

(d)

S

+

30

N

N

20 10

(b)

0 θ (mdeg)

Cyanine Dye Aggregate Exciton Chirality

–10 –20 –30 –40

(c)

S

–50

S

+

–60

N

N

C2H5

θ (mdeg)

(e)

250 200 dye bound to antiparallel 150 L-PNA - PNA 100 50 0 –50 –100 –150 –200 –250 400 450

450

CH3

CH3

500

550

600

650

C2H5

(f) 250 dye bound to antiparallel D-PNA - PNA

500

700

Wavelength (nm)

550

600

Wavelength (nm)

650

700

200 dye bound to parallel 150 D-PNA - PNA 100 50 0 –50 –100 –150 –200 –250 400 450

dye bound to parallel L-PNA - PNA

500

550

600

650

700

Wavelength (nm)

Figure 18.15. (a) Structure of the 3,3 -diethylthiacarbocyanine dye (DiSC2 (5)) and design of aggregate interaction with PNA–DNA and PNA–PNA duplexes. (b) Scheme of the interaction of DiSC2 (5) with PNA–PNA or PNA–DNA duplexes. (c) Structure of a DiSC2 (5) analogue bearing chiral pendant arms. (d) CD titration of a PNA–DNA duplex with DiSC2 (5). [PNA–DNA] = 5.0 μM; dye is added in 0.5 μM aliquots. Spectra were recorded at 15◦ C. (Reprinted with permission from reference 34, copyright 1999, American Chemical Society.) (e) CD spectra of the cyanine dye DiSC2 (5) bound to antiparallel PNA–PNA duplexes: (solid line) D-Lys chiral box PNA–achiral antiparallel PNA; (broken line) L-Lys chiral box PNA–achiral antiparallel PNA. The PNA duplex concentrations are 5 μM, and the dye concentration is 15 μM; spectra were recorded at 20◦ C. (f) CD spectra of the cyanine dye DiSC2 (5) bound to parallel PNA–PNA duplexes: (solid line) D-Lys chiral box PNA–achiral antiparallel PNA; (broken line) L-Lys chiral box PNA–achiral antiparallel PNA, same experimental conditions as in e. (e, f: from reference 36, copyright 2005, John Wiley & Sons, reproduced with permission.)

and ﬂuorescence. For example, the dye DiSC2 (5) exhibits upon binding an ∼120 nm shift to shorter wavelength of the main absorption band in the UV–vis. This shift is correlated to the formation of H-type aggregates, with decrease of the ground-state energy and an increase of excited-state energy [88]. This results in an instantaneous color change from blue to purple, providing a simple colorimetric indicator for PNA hybridization. This is useful for colorimetric test development, since sometimes the color change can be perceived by the naked eye [89] and easily measured by UV–vis spectroscopy [90]. Nevertheless, the use of circular dichroism is much more convenient, since only the bound dye-duplex aggregates are detected, whereas the eventual excess of free dye is CD-silent. Furthermore, the CD spectra of these aggregates can be used to establish the helix handedness of the PNA–XNA duplexes.

607

608

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

18.5.1. Cyanine Dye Aggregation Templated by PNA–DNA Duplexes Multimeric binding of DiSC2 (5) to antiparallel PNA–DNA duplex can be inferred from CD spectra, as shown in Figure 18.15c [34]. Solutions for these CD analyses are usually prepared by dissolving the PNA and DNA strands in an aqueous sodium phosphate buffer (10 mM, pH 7.0; NaCl). Stock solutions of the dye are prepared by solving the dye in methanol and measuring the concentrations spectrophotometrically using the manufacturer’s extinction coefﬁcient (ε651 = 260,000 M−1 cm−1 in methanol). Very little amount of dye is required for most assays (c = 5–30 μM). The concentration of the target can be varied, although it is possible to detect 1–5 μM strand concentrations, with the usual dye:PNA ratio being 2:1 or 3:1. The buffer should also include at least 10% methanol in order to prevent adsorption of the dye to the walls of the sample container and aspeciﬁc aggregation. A detergent can also be used for this purpose, or other additives, such as succinyl-β-cyclodextrin [91]. DiSC2 (5) (Figure 18.15a), being a symmetrical achiral molecule, has no CD spectrum. Upon interaction with either the right-handed PNA–DNA duplex or the DNA–DNA duplex, the dye adopts the chirality of the target duplex and exhibits an induced circular exciton split CD spectrum, with crossover at 534 nm. Increasing dye concentration results in a positive CD signal at 558 nm and a negative signal at 529 nm. An isoelliptic point is observed at 544 nm, except for the ﬁrst two spectra with lower dye concentration, where the crossover is around 534 nm (Figure 18.15d). As mentioned above, the bisignate bands observed are attributed to exciton coupling among multiple DiSC2 (5) molecules [15]. Exciton coupling is favored when the individual chromophores have large extinction coefﬁcients and/or are positioned close to one another [15, 92]. Thus, the observed exciton coupling in this experiment indicates that multiple dye molecules are simultaneously bound to the PNA–DNA duplex. In addition, the positive splitting provides evidence of a right-handed helical relationship for the dye chromophores, consistent with the template of the aggregate by the PNA–DNA duplex, which is itself a right-handed helix. Since the ratio of DiSC2 (5):PNA–DNA is <1.0 throughout the titration, cyanine dye binding to the duplex must be cooperative. This is consistent with the UV–vis data that indicated binding of the dye as a higher aggregate. It is also noteworthy that the shape of the induced CD spectrum is invariant once the ratio of dye:duplex exceeds 0.2, indicating that there is a single CD-active species bound to the PNA–DNA duplex above this ratio [34]. In the same work [34], temperature-dependent CD experiments showed that the exciton feature appeared at 30◦ C and became more intense with decreasing temperature. A CD experiment at variable temperatures that monitored the CD intensity at 529 nm yielded a transition at 28◦ C, which shows both the binding of the dye and the induction of chirality. Similar CD studies were also reported for modiﬁed PNA strands containing 2dor 2l-leucine monomeric units [93]. In these cases the binding of DiSC2 (5) to the PNA–DNA duplex induced a strong positive CD signal at 552 nm and a negative signal at 521 nm. The intensity of the exciton signal (−25 mdeg at 521 nm and +10 mdeg at 552 nm, using [PNA–DNA] 5.0 μM and [DiSC2 (5)] 10 μM at 15◦ C) demonstrates clearly that incorporation of C2 l-leucine monomers into a PNA–DNA duplex signiﬁcantly hinders cyanine dye aggregation, whereas the d-leucine analogue is less inhibitory (−100 mdeg at 521 nm and +50 mdeg at 552 nm under the same conditions as reported above). These results are in agreement with the studies performed in our group, showing that 2d-PNAs give a preferential right-handed helicity, while 2l-PNAs give a preferential left-handed helicity [33]. Therefore, since the double helix has a more organized structure

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

in the case of 2d-PNA–DNA duplexes, the dye aggregates more easily on this complex. However, the CD signal of the unmodiﬁed PNA–DNA duplex remains the most intense, on account of the lack of steric hindrance caused by the amino acid residues. The speciﬁcity of the CD PNA/cyanine dye probe was applied for the identiﬁcation and quantiﬁcation of DNA upon extraction from peanuts and peanut-containing foods and ampliﬁcation by PCR [94]. The appearance of the CD signal in the 500–600 nm region with a right-handed exciton coupling conﬁrms the presence and the identity of the ampliﬁed DNA at a very low level (few picomoles). Photobleaching of the cyanine dye 3,3 -diethylthiacarbocyanine iodide (DiSC2 (3) − I ) was found to be catalyzed by its interaction with PNA–DNA duplexes, thus allowing to reveal these complexes by destaining on acrylamide gels. The formation of the multimolecular complexes responsible for the catalytic activity was studied by circular dichroism and was found to be due to similar helical aggregates [95].

18.5.2. Cyanine Dyes in PNA–PNA Duplexes Complementary strands of PNA hybridize to form stable double-helical complexes. CD spectra of antiparallel PNA–PNA duplexes were ﬁrst measured by Armitage and coworkers [34] on PNA–PNA duplexes with C-terminal l-lysine residues. A mirror-image relationship between the CD spectra for DiSC2 (5) bound to PNA–DNA duplexes versus this PNA–PNA was found. This effect likely arises from opposite twists for the two helices: right-handed for the ﬁrst one and left-handed for the latter. This can be considered to be a clear-cut experimental evidence of the left-handedness of the PNA–PNA duplex containing a terminal l-amino acid residue, as discussed in Section 18.2. A new dicarbocyanine dye bearing branched, chiral N -alkyl substituents (Figure 18.15c) was synthesized, and its ability to form helical aggregates on PNA double-helical templates was studied [96]. The dye aggregates less effectively than an analogous dye bearing linear, achiral substituents, presumably due to steric problems with packing the branched substituents compared with the linear ones. CD experiments clearly show that when the PNA duplex has a left-handed helicity, addition of the achiral dye leads to formation of a left-handed dye aggregate. However, when the chiral dye aggregates in the presence of this duplex, a right-handed structure is formed, suggesting that the dye alters the helicity of the underlying template. When a racemic PNA duplex (i.e., equal amounts of right- and left-handed helices) is used, no chirality is observed for the dye aggregate formed by the achiral dye, but a right-handed helical aggregate is once again formed by the chiral dye. It is interesting to note that these results indicate that chirality is transferred from the dye to the PNA. According to the systematic work of Armitage and co-workers [88], the cyanine dye DiSC2 (5) was established as a reliable probe of duplex helicity, and it was largely applied in order to characterize the preferential induced helicity of the modiﬁed chiral PNAs. In particular, the strong chiral constraint of two enantiomeric chiral PNAs bearing three adjacent d- or l-lysine based residues in the middle of the strand (“chiral box” PNAs) was studied also by CD in the presence of DiSC2 (5) [36]. The exciton coupling effect observed in the CD spectra in the 400–700 nm region conﬁrms that the d-Lys-PNA–achiral PNA antiparallel duplex forms a right-handed helix, whereas the l-Lys-PNA–achiral PNA antiparallel duplex forms a left-handed helix (Figure 18.15e) [36]. In the case of CD spectra in the 400-700 nm region of the parallel duplexes after the addition of the DiSC2 (5) cyanine, the CD spectra are reversed, compared to the

609

610

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

previous ones: the right-handed helix is formed by the l-Lys-PNA–achiral PNA duplex, whereas the left-handed helix is formed by the d-Lys-PNA–achiral PNA (Figure 18.15f). These results demonstrate how, in the case of PNAs bearing a stereocenter at the position C2, the handedness of the PNA duplex is not linked to the conﬁguration of the stereogenic center in an absolute way, but it is strictly dependent on the directionality of the binding: If the PNA–PNA duplex is antiparallel, the helix is righthanded by using d-lysine-derived PNA residues and left-handed by using l-lysine-derived PNA residues, whereas if the PNA–PNA duplex is parallel, the opposite correlation is observed. Moreover, a systematic study on the combined effects of two lysine-derived stereogenic centers at C2 and C5 of a PNA monomer inserted within an achiral PNA strand was performed by circular dichroism, by taking into account all four diastereomers. Also in this case antiparallel PNA–PNA duplex-preferred handedness was conﬁrmed by using DiSC2 (5) [18].

18.5.3. Forced Intercalation Interaction of asymmetric cyanine dyes with PNA–DNA duplexes was also studied. For example, orange thiazole (TO) dye derivative was covalently linked to the PNA backbone in place of a nucleobase residue. These new PNA–dye conjugates, which show an increase of ﬂuorescence emission upon binding to DNA, were called “FIT” probes (Force Intercalation probes). The possible modes of TO–nucleic acid interactions was also described by circular dichroism spectroscopy [97]. In particular, the interaction of free TO with a DNA–DNA and a PNA–DNA duplex was compared with the case of FIT probe. When the unbound thiazole orange is added to a solution containing DNA–DNA or PNA–DNA duplexes, at high dye/base-pair ratios, two bands of different sign appear in the region corresponding to the UV absorption. Such a CD proﬁle can originate from a superimposed exciton CD induced by dye–dye interactions in the minor groove. Instead, the TO chromophore in the FIT–PNA · DNA complex shows a negative band with a maximum at 514 nm, which coincides with the absorption band, which were assigned to the intercalation mode.

18.6. CONCLUSION Peptide nucleic acids and their analogues are unnatural structures created by synthesis which have been described in the last two decades. Due to their structural variability, a large number of combinations are possible, including homo- (PNA–PNA) and heteroduplexes (PNA–DNA, PNA–RNA), triplex and quadruplexes, i-motifs, and parallel and antiparallel orientations, using unmodiﬁed or modiﬁed PNA. This prevents us from making an exhaustive description of all the spectral features of these molecules. However, the present chapter describes the main features of the most relevant structures containing PNA and the use of electronic circular dichroism as a tool to study their essential properties, from structure to binding properties and kinetics. Not all the possible structures have been explored so far, and circular dichroism is certainly one of the techniques which will play a role in further studies. New structural information from NMR and X-ray diffraction, together with the development of reliable computing models for interpreting circular dichroism, will allow us to use CD data in a more rational and predictable way.

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

ACKNOWLEDGMENTS This work was partially supported by grants from the Ministero dell’Universit`a e della Ricerca (MIUR) (PRIN2007, Grant No. 2007F9TWKE).

REFERENCES 1. P. E. Nielsen, M. Egholm, R. H. Berg, O. Buchardt, Science 1991, 254 , 1497–1500. 2. P. E. Nielsen, ed., Peptide Nucleic Acids: Protocols and Applications. Horizon Bioscience, Norwich, UK, 2004. 3. P. E. Nielsen, Chem. Biodivers. 2010, 7 , 786–804. 4. P. E. Nielsen, Curr. Opin. Biotech. 2001, 12 , 16–20. 5. A. Gourishankar, S. Shukla, R. Pasricha, M. Sastry, K. N. Ganesh, Curr. Appl. Phys. 2005, 5102–5107. 6. M. Egholm, C. Behrens, L. Christensen, R. H. Berg, P. E. Nielsen, O. Buchardt, J. Chem. Soc. Chem. Commun. 1993, 800–801. 7. P. Wittung, P. E. Nielsen, O. Buchardt, M. Egholm, B. Nord´en, Nature 1994, 368 , 561–563. 8. M. Egholm, O. Buchardt, L. Christenses, C. Behrens, S. M. Freier, D. A. Driver, R. H. Berg, S. K. Kim, B. Nord´en, P. E. Nielsen, Nature 1993, 365 , 566–568. 9. H. Orum, P. E. Nielsen, M. Egholm, R. H. Berg, O. Buchardt, C. Stanley, Nucleic Acids Res. 1993, 21 , 5332–5336. 10. D. Y. Cherny, B. P. Belotserkovskii, M. D. Frank-Kamenetskii, M. Egholm, O. Buchardt, R. H. Berg, P. E. Nielsen, Proc. Natl. Acad. Sci. USA 1993, 90 , 1667–1670. 11. V. V. Demidov, V. N. Potaman, M. D. Frank-Kamenetskii, M. Egholm, O. Buchard, S. H. Sonnichsen, P. E. Nielsen, Biochem. Pharmacol . 1994, 48 , 1310–1313. 12. H. Rasmussen, J. S. Kastrup, J. N. Nielsen, J. M. Nielsen, P. E. Nielsen, Nat. Struct. Biol . 1997, 4 , 98–101. 13. V. Menchise, G. De Simone, T. Tedeschi, R. Corradini, S. Sforza, R. Marchelli, D. Capasso, M. Saviano, C. Pedone, Proc. Natl. Acad. Sci. USA 2003, 100 , 12021–12026. 14. S. K. Kim, P. E. Nielsen, M. Egholm, O. Buchardt, R. H. Berg, B. Nord`en, J. Am. Chem. Soc. 1993, 115 , 6477–6481. 15. K. Nakanishi, N. Berova, R. W. Woody, Circular Dichroism: Principles and Applications, VCH, New York, 1994. 16. R. Corradini, S. Sforza, T. Tedeschi, F. Totsingan, R. Marchelli, Curr. Top. Med. Chem. 2007, 7 , 681–694. 17. J. I. Yeh, B. Shivachev, S. Rapireddy, M. J. Crawford, R. R. Gil, S. Du, M. Madrid, D. H. Ly, J. Am. Chem. Soc. 2010, 132 , 10717–10727. 18. S. Sforza, T. Tedeschi, R. Corradini, R. Marchelli, Eur. J. Org. Chem., 2007, 5879–5885. 19. M. Leijon, A. Graslund, P. E. Nielsen, O. Buchardt, B. Nord´en, S. M. Kristensen, M. Eriksson, Biochemistry 1994, 33 , 9820–9825. 20. T. Koch, M. Naesby, P. Wittung, M. Jorgensen, C. Larsson, O. Buchardt, C. J. Stanley, B. Nord´en, P. E. Nielsen, H. Orum, Tetrahedron Lett. 1995, 36 , 6933–6936. 21. B. Petersson, B. B. Nielsen, H. Rasmussen, I. K. Larsen, M. Gajhede, P. E. Nielsen, J. S. Kastrup. J. Am. Chem. Soc. 2005, 127 , 1424–1430. 22. A. Dragulescu-Andrasi, S. Rapireddy, B. M. Frezza, C. Gayathri, R. R. Gil, D. H. Ly, J. Am. Chem. Soc. 2006, 128 , 10258–10267. 23. V. A. Kumar, Eur. J. Org. Chem. 2002, 2021–2032. 24. B. P. Gangamani, V. A. Kumar, K. N. Ganesh, Tetrahedron 1999, 55 , 177–192.

611

612

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

25. P. Wittung, M. Eriksson, R. Lyng, P. E. Nielsen, B. Norden, J. Am. Chem. Soc. 1995, 117 , 10167–10173. 26. F. Totsingan, V. Jain, W. C. Bracken, A. Faccini, T. Tedeschi, R. Marchelli, R. Corradini, N. R. Kallenbach, M. M. Green, Macromolecules, 2010, 43 , 2692–2703. 27. H. Rasmussen, T. Liljefors, B. Petersson, P. E. Nielsen, J. S. Kastrup, J. Biomol. Struct. Dyn. 2004, 21 , 495–502. 28. E. Rizzo, J. Schellman, Biopolymers 1984, 23 , 435–470. 29. R. Lyng, A. Rodger, B. Norden, Biopolymers 1991, 31 , 1709–1720. 30. G. Haaima, A. Lohse, O. Buchardt, P. E. Nielsen, Angew. Chem. Int. Ed . 1996, 35 , 1939–1942. 31. A. Puschl, S. Sforza, G. Haaima, O. Dahl, P. E. Nielsen, Tetrahedron Lett. 1998, 39 , 4707–4710. 32. T. Tedeschi, S. Sforza, F. Maffei, R. Corradini, R. Marchelli, Tetrahedron Lett. 2008, 49 , 4958–4961. 33. S. Sforza, G. Haaima, R. Marchelli, P. E. Nielsen, Eur. J. Org. Chem. 1999, 197–204. 34. J. O. Smith, D. A. Olson, B. A. Armitage, J. Am. Chem. Soc. 1999, 121 , 2686–2695. 35. S. Sforza, R. Corradini, S. Ghirardi, A. Dossena, R. Marchelli, Eur. J. Org. Chem. 2000, 2905–2913. 36. T. Tedeschi, S. Sforza, A. Dossena, R. Corradini, R. Marchelli, Chirality 2005, 17 , S196–S204. 37. P. Lagriffoule, P. Wittung, M. Eriksson, K. K. Jensen, B. Norden, O. Buchardt, P. E. Nielsen, Chem. Eur. J . 1997, 3 , 912–919. 38. P. Wittung, P. Nielsen, B. Nord´en, J. Am. Chem. Soc. 1997, 119 , 3189–3190. 39. As examples see http://www.basic.northwestern.edu/biotools/oligocalc.html or http://www. ambion.com/techlib/misc/oligo_calculator.html 40. S. Tomac, M Sarkar, T Ratilainen, P. Wittung, P. E. Nielsen, B. Nord´en, A. Gra1slund, J. Am. Chem. Soc. 1996, 118 , 5544–5552. 41. N. Sugimoto, N. Satoh, K. Yasuda, S. Nakano, Biochemistry 2001, 40 , 8444–8451. 42. S. A. Kushon, J. P. Jordan, J. L. Seifert, H. Nielsen, P. E. Nielsen, B. A. Armitage, J. Am. Chem. Soc. 2001, 123 , 10805–10813. 43. S. Sen, L. Nilsson, J. Am. Chem. Soc. 1998, 120 , 619–631. 44. G. Haaima, H. Rasmussen, G. Schmidt, D. K. Jensen, J. S. Kastrup, P. Wittung Stafshede, B. Nord´en, O. Buchardt, P. E. Nielsen, New J. Chem. 1999, 23 , 833–840. 45. V. A. Kumar, K. N. Ganesh, Acc. Chem. Res. 2005, 38 , 404–412. 46. M. Hollenstein, C. J. Leumann, J. Org. Chem. 2005, 70 , 3205–3217. 47. T. Govindaraju, V. A. Kumar, K. N. Ganesh, J. Org. Chem. 2004, 69 , 5725–5734. 48. M. D’Costa, V. A. Kumar, K. N. Ganesh, Organic Lett. 2001, 3 , 1281–1284. 49. P. S. Lonkar, V. A. Kumar, Bioorg. Med. Chem. Lett. 2004, 14 , 2147–2149. 50. T. Vilaivan, G. Lowe, J. Am. Chem. Soc. 2002, 124 , 9326–9327. 51. C. Suparpprom, C. Srisuwannaket, P. Sangvanich, T. Vilaivan, Tetrahedron Lett. 2005, 46 , 2833–2837. 52. T. Govindaraju, V. Madhuri, V. A. Kumar, K. N. Ganesh, J. Org. Chem. 2006, 71 , 14–21. 53. T. Govindaraju, V. A. Kumar, K. N. Ganesh, J. Am. Chem. Soc. 2005, 127 , 4144–4145. 54. P. S. Shirude, V. A. Kumar, K. N. Ganesh, Tetrahedron 2004, 60 , 9485–9491. 55. T. Govindaraju, V. A. Kumar, Tetrahedron 2006, 62 , 2321–2330. 56. M. Kitamatsu, M. Shigeyasu, M. Saitoh, M. Sisido, Biopolymers 2006, 84 , 267–273. 57. K. Gogoi, V. A. Kumar, Chem. Commun. 2008, 706–708. 58. S. Bagmare, M. D’Costa, V. A. Kumar, Chem. Commun. 2009, 6646–6648.

ELECTRONIC CIRCULAR DICHROISM OF PEPTIDE NUCLEIC ACIDS AND THEIR ANALOGUES

59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90.

P. Wittung, P. E. Nielsen, B. Nord´en, J. Am. Chem. Soc. 1996, 118 , 7049–7054. P. Wittung, P. E. Nielsen, B. Nord´en, Biochemistry 1997, 36 , 7973–7979. S. Sforza, G. Galaverna, A. Dossena, R. Corradini, R. Marchelli, Chirality 2002, 14 , 591–598. R. Corradini, S. Sforza, A. Dossena, G. Palla, R. Rocchi, F. Filira, F. Nastri, R. Marchelli, J. Chem. Soc. Perkin Trans 1 2001, 2690–2696. G. N. Roviello, D. Musumeci, C. Pedone, E. M. Bucci Amino Acids 2010, 38 , 103–111. T. Vilaivan, C. Khongdeesameor, P. Harnyuttanakorn, M. S. Westwell, G. Lowe, Bioorg. Med. Chem. Lett. 2000, 10 , 2541–2545. N. C. Seeman, Mol. Biotechnol . 2007, 37 , 246–257. J. L. Huppert, S. Balasubramanian, Nucleic Acids Res. 2007, 35 , 406–413. E. L. Edwards, M. H. Patrick, R. L. Ratliff, D. M. Gray, Biochemistry 1990, 29 , 828–836. B. Datta, M. E. Bier, S. Roy, B. A. Armitage, J. Am. Chem. Soc. 2005, 127 , 4199–4207. B. Datta, C. Schmitt, B. A. Armitage, J. Am. Chem. Soc. 2003, 125 , 4111–4118. V. L. Marin, B. A. Armitage, J. Am. Chem. Soc. 2005, 127 , 8032–8033. S. Roy, F. A. Tanious, W. D. Wilson, D. H. Ly, B. A. Armitage, Biochemistry 2007, 46 , 10433–10443. V. Esposito, A. Randazzo, A. Messere, A. Galeone, L. Petraccone, C. Giancola, G. Piccialli, L. Mayol, Eur. J. Org. Chem. 2003, 17 , 3364–3371. Y. Krishnan-Ghosh, E. Stephens, S. Balasubramanian, J. Am. Chem. Soc. 2004, 126 , 5944–5945. S. Lusvarghi, C. T. Murphy, S. Roy, F. A. Tanious, I. Sacui, W. D. Wilson, D. H. Ly, B. A. Armitage, J. Am. Chem. Soc. 2009, 131 , 18415–18424. L. Petraccone, B. Pagano, V. Esposito, A. Randazzo, G. Piccialli, G. Barone, C. A. Mattia, C. Giancola, J. Am. Chem. Soc. 2005, 127 , 16215–16223. A. Paul, P. Sengupta, Y. Krishnan, S. Ladame, Chem. Eur. J . 2008, 14 , 8682–8689. Y. Krishnan-Ghosh, E. Stephens, S. Balasubramanian, Chem. Commun. 2005, 5278–5280. N. K. Sharma, K. N. Ganesh, Chem. Commun. 2005, 4330–4332. S. Modi, A. H. Wani, Y. Krishnan, Nucleic Acids Res. 2006, 34 , 164354–4363. S. Chakraborty, S. Modi, Y. Krishnan, Chem. Commun. 2008, 70–72. C. R. Cantor, P. R. Schimmel, Biophysical Chemistry: Part II: Techniques for the Study of Biological Structure and Function, W. H. Freeman, San Francisco, 1980, Chapter 7. R. Tonelli, R. Fronza, S. Purgato, C. Camerin, F. Bologna, S. Alboresi, M. Franzoni, R. Corradini, S. Sforza, A. Faccini, R. Marchelli, A. Pession, Mol. Cancer Ther. 2005, 4 , 779–786. A. Faccini, A. Tortori, T. Tedeschi, S. Sforza, R. Tonelli, A. Pession, R. Corradini, R. Marchelli, Chirality 2008, 20 , 494–500. F. Totsingan, Ph.D. thesis, University of Parma, 2008. P. Wittung, S. K. Kim, O. Buchardt, P. Nielsen, B. Norden, Nucleic Acids Res. 1994, 22 , 5371–5377. A. Mishra, R. K. Behera, P. K. Behera, B. K. Mishra, G. B. Behera, Chem. Rev . 2000, 100 , 1973–2011. J. L. Seifert, R. E. Connor, S. A. Kushon, M. Wang, B. A. Armitage, J. Am. Chem. Soc. 1999, 121 , 2987–2995. K. C. Hannah, B. A. Armitage, Acc. Chem. Res. 2004, 37 , 845–853. M. Komiyama, S. Ye, X. Liang, Y. Yamamoto, T Tomita, J. M. Zhou, H. Aburatani, J. Am. Chem. Soc. 2003, 125 , 3758–3762. L. M. Wilhelmsson, B Nord´en, k. Mukherjee, M. T. Dulay, R. N. Zare, Nucleic Acids Res. 2002, 30 , e3.

613

614

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

91. T. Tedeschi, S. Sforza, S. Ye, R. Corradini, A. Dossena, M. Komiyama, R. Marchelli, J. Biochem. Biophys. Methods 2007, 70 , 735–741. 92. N. Berova, D. Gargiulo, F. Derguini, K. Nakanishi, N. Harada, J. Am. Chem. Soc. 1993, 115 , 4769–4775. 93. I. Dilek, M. Madrid, R. Singh, C. P. Urrea, B. A. Armitage, J. Am. Chem. Soc. 2005, 127 , 3339–3345. 94. S. Sforza, E. Scaravelli, R. Corradini, R. Marchelli, Chirality 2005, 17 , 515–521. 95. M. Wang, R. Holmes-Davis, Z. Raﬁnski, B. Jedrzejewska, K. Y. Choi, M. Zwick, C. Bupp, A. Izmailov, J. Paczkowski, B. Warner, H. Koshinsky, Anal. Chem. 2009, 81 , 2043–2052. 96. B. R. Renikuntla, B. A. Armitage, Langmuir 2005, 21 , 5362–5366. 97. D. V. Jarikote, N. Krebs, S. Tannert, B. Roder, O. Seitz, Chem. Eur. J . 2007, 13 , 300–310.

19 CIRCULAR DICHROISM OF PROTEIN–NUCLEIC ACID INTERACTIONS Donald M. Gray

19.1. INTRODUCTION This chapter describes the use of electronic CD measurements to study the interactions of proteins with nucleic acids. Many of the practical aspects of monitoring the CD spectra of biopolymers during titrations with ligands, denaturants, or osmolytes have been covered in recent summaries by Greenﬁeld [1] and Cary and Kneale [2]. These practical aspects, as well as examples from the literature previously reviewed [3, 4], including examples of the CD spectra of viruses [3], will not be repeated here. Moreover, readers of this chapter will ﬁnd a wealth of related information on methodology and techniques in other chapters of these volumes on chiroptical methods. Of particular relevance to the study of biopolymers by electronic circular dichroism are Chapters 14. (The Electronic Circular Dichroism of Proteins), 15 (Electronic Circular Dichroism of Peptides), 17 (Circular Dichroism Spectroscopy of Nucleic Acids), 18 (Electronic Circular Dichroism of Peptide Nucleic Acids and Their Analogues), and 20 (Drug and Natural Product Binding to Nucleic Acids Analyzed by Electronic Circular Dichroism) in this volume. Electronic CD spectroscopy is a choice technique for the study of conformational changes of interacting biomolecules in solution, especially when samples are limited in amount and a relatively rapid survey of conformational changes is desired. CD spectroscopy is regularly applied together with other techniques that may include ﬂuorescence, electrophoretic mobility shift assays, calorimetry, or NMR, among many others, in the study of a particular DNA- or RNA-binding protein. This chapter focuses on the contributions of CD measurements in such studies. The applications of CD measurements are so numerous and varied that insights into possible strategies are perhaps best appreciated by reviewing examples from the literature. Thus, the following section illustrates Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

615

616

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

the variety of approaches to the use of CD techniques and interpretations of results in published studies of protein–nucleic acid interactions. It is not a comprehensive review of the literature. It will be noted in these examples that the CD spectra of protein–DNA complexes are frequently monitored in one of two main wavelength ranges, a far-UV range from about 190 to 250 nm and a near-UV range from about 250 to 320 nm. The use of these two ranges is dictated by the fact that the absorption of peptide bonds dominates in the far-UV range, with the result that CD spectra of samples containing proteins and nucleic acids are often measured at lower concentrations, or with shorter cell pathlengths, in the far UV than in the near UV. Nevertheless, the wavelength range that can be monitored by a CD spectral scan in a speciﬁc situation depends on the total absorbance and relative magnitudes of the protein and nucleic acid CD bands and their changes during binding.

19.2. STRUCTURAL CHANGES OF NUCLEIC ACIDS AND PROTEINS 19.2.1. Transitions Between DNA Secondary Structures The use of CD has been diagnostic in the identiﬁcation and study of protein domains that are able to convert alternating d(G–C)n sequences from the right-handed B conformation to the left-handed Z-DNA conformation. A classic example is the Zα domain of the human RNA editing enzyme ADAR1 (double-stranded RNA deaminase I). This enzyme can cause a change in mRNA coding by the deamination of adenosine to inosine, which acts like guanosine. Two ADAR1 Zα domains (of 63 residues) can bind to a d(CGCGCG):d(CGCGCG) duplex. A combined CD and NMR study indicated that binding of one Zα domain actively converts B- to Z-DNA, resulting in an equilibrium between the two conformations. The Z form is then stabilized by the binding of a second Zα domain [5]. Zα domains have been found in many proteins, with CD spectroscopy providing a key means of identiﬁcation. Poxviruses such as vaccinia virus have an E3L protein that is an important determinant of the host range of the virus. The protein contains a Zα domain as well as a double-stranded RNA (dsRNA) binding domain. CD measurements were important in showing that the Zα domains (63–67 residues) of the E3L protein of vaccinia and four other poxviruses differed from the Zα domain of the human ADAR1 enzyme in their capabilities of converting poly[d(G–C)] from the B to the Z form [6]. Figure 19.1 includes a CD spectrum of B-DNA, which has a positive CD band at 295 nm and a negative band at about 250 nm. The signs of these bands are largely reversed for the Z form of this polymer, resulting in a spectrum with a characteristic long-wavelength negative band at 290–295 nm. The extents and kinetics of the B- to Z-DNA conversion differed greatly in the presence of the various domains, as seen in the panels of Figure 19.1. The human ADAR1 and yab (yaba-like disease virus) Zα domains brought about the fastest and the most complete conversions. The vaccinia vZα domain did not shift the equilibrium toward the Z-DNA form under the same conditions. This effect is related to variability in the β-wing region residues of the Zα domains, which have a winged helix–turn–helix fold [6]. While the function of the Zα domain of poxvirus E3L protein remains unknown, it is interesting that this protein also contains a dsRNA binding domain. A Zα domain from the human Z-DNA binding protein 1 (ZBP1) can form cytoplasmic aggregates with mRNA [7], and a dsRNA-dependent protein kinase isolated from zebraﬁsh has two Zα domains instead of two dsRNA binding domains that are generally found in analogous

CIRCULAR DICHROISM OF PROTEIN–NUCLEIC ACID INTERACTIONS

617

10 8 6 4 Ellipticity (mdeg)

2 0 –2 –4 B-DNA hZαADAR1 yabZαE3L IsZαE3L orfZαE3L spZαE3L vZαE3L

–6 –8 –10 –12 –14 –16 230

240

250

260

270 280 Wavelength (nm)

290

300

310

320

2800

3200

3600

(a) 6

Ellipticity (mdeg) at 255 nm

4 2 0 hZαADAR1 yabZαE3L IsZαE3L orfZαE3L spZαE3L vZαE3L

–2 –4 –6 –8 –10 0

400

800

1200

1600

2000

2400

Time (sec) (b)

Figure 19.1. (a) CD spectra of poly[d(G–C)] in the B form and in the presence of Zα domains from human ADAR1 editing enzyme (hZα), yaba-like disease virus (yabZα), lumpy skin disease virus (lsZα), orf virus (orfZα), swinepox virus (spZα), and vaccinia virus (vZα), listed in decreasing order of their abilities to convert B-DNA to Z-DNA. CD spectra of the added proteins contributed the negative signals at wavelengths shorter than about 250 nm. (b) Kinetics of the B to Z conversion in the presence of the same domains. Proteins were added to poly[d(G–C)] at a protein to base-pair ratio of 0.4 (with the ﬁnal protein concentration being 90 μM), in a buffer of 10 mM HEPES, pH 7.4, 10 mM NaCl, and 0.1 mM EDTA, except for yabZα where the buffer included 100 mM NaCl. Spectra were taken at 25◦ C using a 2-mm-pathlength cell. CD values are in mdeg ellipticity. (Reproduced from Quyen et al. [6] by permission of Oxford University Press, copyright 2007.) (See insert for color representation of the ﬁgure.)

618

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

dsRNA-activated mammalian enzymes [8]. That is, Zα domains may play important roles in binding RNA, although they are typically characterized by an ability to bring about a B to Z conformational transition in DNA, as conﬁrmed by CD spectroscopy. In addition to Zα domains, peptides with alternating Lys residues, such as Lys–Gly–Lys–Gly–Lys–Gly–Lys, are able to convert poly[d(G-m5 C)] to the Z form [9]. A novel Z-DNA binding protein has been constructed with a Lys–Gly–Lys–Gly– Lys–Gly–Lys peptide linked to a leucine zipper dimerization sequence [10]. CD measurements indicated that the engineered protein, KGZIP, can dimerize into an α-helical leucine zipper. CD spectra also showed that the construct can convert poly[d(G-m5 C)] to the Z form, and other measurements demonstrated that it can bind to a Z-DNA substrate with an afﬁnity comparable to that of a Zα domain [10]. There is CD spectral evidence for protein-induced transitions between the B and A, or A-like, conformations of DNA [11, 12]. Chen et al. [11] found that the CD spectrum of a 23-bp DNA target sequence was signiﬁcantly altered from that of the B form by additions of saturating amounts of proteins with three, four, or six Cys2 –His2 zinc ﬁnger domains of the human metallothionein response element transcription factor 1 (MTF-1). Magnitudes of a positive band at 280 nm and a negative band at about 210 nm were increased in CD spectra of the protein-bound DNA, with the increases being greater as the number of zinc ﬁnger domains in the added peptides was increased from three to six [4, 11]. It was assumed that the spectra of the proteins, which have signiﬁcant bands at wavelengths shorter than 250 nm, did not change upon binding. With this assumption, spectral features of the complexed DNA were found to be similar to those of the A form of DNA except that the positive band was not shifted to shorter wavelengths (see Chapters 17 and 20, this volume). However, NMR and crystallographic structures of DNAs in complexes with Cys2 –His2 zinc ﬁngers from other proteins show that the DNA remains in a B conformation, but with a widened major groove [13, 14]. Therefore, the CD spectral features of the DNA bound by MTF-1 Cys2 –His2 zinc ﬁngers should perhaps be designated as those of an “enlarged-groove” B-DNA conformation with A-like features, as suggested by Chen et al. [11]. In a different study, the B-form CD spectrum of calf thymus DNA was converted to an approximate A form spectrum, with a positive band at 271 nm and a negative band at 208–210 nm, by the addition of bovine pancreatic DNase I under conditions that prevented DNA digestion [12].

19.2.2. Alterations of DNA Quadruplexes An especially important use of CD spectroscopy has been in the characterization of DNA G-quadruplex structures and how they are perturbed by cellular proteins, since G-quadruplexes have distinctive spectral signatures depending on the type of quadruplex fold; see Chapter 17, this volume. Many genomic DNA sequences have the potential to form G-quadruplex structures, including G-rich sequences in telomeres, promoters, and DNA that is reverse transcribed from the human immunodeﬁciency virus type 1 (HIV-1) RNA genome. Studies of cellular proteins that may recognize and unfold G-quadruplexes rely heavily on CD measurements [15–20]. As an example, Kankia et al. [15] used CD spectra to investigate the ability of the HIV-1 nucleocapsid (NC) protein to destabilize an intramolecular DNA G-quadruplex formed by the 15-mer thrombin binding aptamer, TBA, which has the sequence d(GGTTGGTGTGGTTGG) [21]. The small (55 residue) HIV-1 NC protein has two zinc ﬁngers of the Cys2 –His–Cys type that are involved in binding to single-stranded

CIRCULAR DICHROISM OF PROTEIN–NUCLEIC ACID INTERACTIONS

DNA and RNA, and the protein has been termed a nucleic acid chaperone since it is involved in nucleic acid helix destabilizing and annealing functions [22]. The quadruplex structure of TBA in 10 mM KCl (plus 10 mM cesium–HEPES, pH 7.5) has a major positive CD band at 293 nm and a negative band with a maximum near 268 nm as seen in Figure 19.2a (dashed curve) [15]. These features are indicative of a quadruplex fold that has stacked G-quartets of opposite H-binding polarity [23]. Spectra of unfolded TBA at the same concentration without added KCl have small magnitudes of about 2 and 1 mdeg, respectively, at these wavelengths [15]. The addition of the NC protein at a 2:1 NC:TBA molar ratio decreases the magnitudes of all of the CD bands of TBA in the presence of 10 mM KCl (Figure 19.2a, solid curve), indicating that the protein under these conditions can partially unfold the TBA quadruplexes in the sample. The unfolding appears to be a two-state transition, given the presence of isodichroic points in the family of spectra obtained by subsequent dissociation of the NC–TBA complexes with increasing concentrations of KCl. That is, at 10 mM KCl the CD spectrum may be that of a mixture of TBA in its quadruplex folded form and in an unfolded form bound by NC. Figure 19.2b gives the CD magnitudes at 293 nm of the NC plus TBA mixture as a function of KCl concentration.

19.2.3. Bending of DNA Changes in the secondary structure of DNA, such as from right-handed B-DNA to lefthanded Z-DNA, or unwinding due to melting, have possible functional importance in their effects on tertiary superhelical structures of constrained segments of double-helical DNA [5]. Bending and kinking of the DNA double helix also have functional importance in their inﬂuence on DNA tertiary structure in packaging and gene regulation [24]. As in the study of transitions that are between more deﬁned secondary structures, CD spectroscopy provides a sensitive means of investigating local bending distortions of the DNA helix. An illustration of the use of CD spectra in monitoring bending is in studies of the DNA binding of the small chromatin protein Sac7d (66 residues) from the hyperthermophile Sulfolobus acidocaldarius [24]. The Sac7d protein has a binding site size of 3.3 ± 0.3 base pairs and kinks duplex DNA by an angle of about 66◦ with intercalation of Val26 and Met29 side chains between two adjacent base pairs. Figures 19.3a and 19.3b show that the long-wavelength 285-nm CD band of poly[d(G-C)] increases in magnitude by over threefold when the DNA is saturated with a Val26 to Ala26 (V26A) mutant protein, neglecting possible minor contributions by the protein. This increase in the CD magnitude at 285 nm reﬂects the effect of DNA bending and unwinding, and the isodichroic point at 272 nm implies that the transition is just between two states. The enhancement of the band magnitude upon binding of the V26A mutant protein was about 70% of the increase that occurred upon binding of the wild-type protein under the same conditions, although the afﬁnity of the mutant protein was reduced by ﬁvefold. Similar changes in the CD spectra of DNA afﬁrmed that the same type of structural change took place during binding of the wild-type protein and mutants of the V26 and M29 residues, and that mutational effects on the thermodynamic properties of binding could be directly compared [24]. Kneale and collaborators [2, 25] have also been able to correlate an increase in the positive long wavelength CD band of DNA with distortion and bending of the duplex. These workers focused on C.AhdI, a 74-residue transcriptional regulator of the restriction endonuclease of the restriction modiﬁcation system of Aeromonas hydrophila. C.AhdI has a helix–turn–helix DNA-binding motif and binds in dimeric and tetrameric forms

619

620

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

12

230 mM KCI + 2 NC

10 mM KCl no NC

10

CD (mdeg)

8 6 4

10 mM KCI + 2 NC

2 0 –2 –4 240

250

260

270 280 290 Wavelength (nm)

300

310

320

(a)

CD at 293 nm

11 10 9 8 7 6

0

50

100 150 [KCI] (mM)

200

250

(b)

Figure 19.2. (a) The KCl concentration dependence of CD spectra of the G-quadruplex thrombinbinding aptamer (TBA) sequence in the presence of two equivalents of the nucleocapsid protein (NC) of human immunodeﬁciency virus type 1. As the KCl concentration was increased from 10 to 230 mM, the NC:TBA complex was dissociated and bands of the TBA spectrum reached maximal values indicative of free TBA. A spectrum of TBA at 10 mM KCl in the absence of NC is shown for reference. (b) A plot of the CD values at 293 nm as a function of KCl concentration from spectra in panel A. The dashed line is the value for TBA in 10 mM KCl without added NC protein. The NC alone does not signiﬁcantly contribute to the CD signal in this wavelength region. The concentration of TBA was about 2 μM. Spectra were taken at 20◦ C using a 10-mm-pathlength cell. The buffer was 10 mM cesium–HEPES, pH 7.5. CD values are in mdeg ellipticity. (Reproduced from Kankia et al. [15] by permission of Oxford University Press, copyright 2007.)

to a 35-bp sequence containing two operator sites. CD spectroscopy demonstrated that there is an 80% increase in the 278-nm CD band when tetramers of C.AhdI bind to saturate the two operator sites, but there is only a 40% increase in the CD band when dimers bind to a mutated sequence containing only one operator site. Thus, the structural changes measured by CD spectroscopy are additive. However, electrophoretic bending assays using 277-bp DNA constructs that contained the wild type and mutated 35-bp sequences did not show greater bending when the tetrameric DNA–protein complex was formed. Analysis of the mobility shift assays indicated that DNA with two dimers of

CIRCULAR DICHROISM OF PROTEIN–NUCLEIC ACID INTERACTIONS

621

6 4

εL – εR

2 0 –2

free DNA

–4 –6 –8 240

250

260

270 280 290 300 Wavelength (nm) (a)

310

320

7

εL – εR (285 nm)

6 5 4 3 2 1 0

0

0.1

0.2 0.3 Protein/DNA (b)

0.4

0.5

Figure 19.3. (a) Changes in the CD spectrum during titration of poly[d(G–C)] with mutant V26A of Sac7d protein as the protein/DNA bp ratio increased from 0 (solid curve) to 0.02, 0.04, 0.067, 0.1, 0.125, 0.166, 0.25, and 0.5 (curve with largest magnitude at 285 nm). (b) A plot of the CD values at 285 nm as a function of the protein/DNA nucleotide ratio showed that the DNA became saturated. Spectra were taken at 25◦ C. The buffer was 0.01 M KH2 PO4 and 0.025 M KCl, pH 7. CD values are in M−1 cm−1 , per mole of nucleotide. (Reprinted with modiﬁcations and permission from Peters et al. [24], copyright 2005, American Chemical Society.)

C.AhdI bound was bent by about 38◦ , which was actually less than the bending of about 47◦ when there was only one dimer bound to the DNA. A possible interpretation is that, while there are identical bends at the sites of protein binding, there is twisting of the intervening base pairs when a tetrameric complex is formed so that the overall bend angle is not the sum of the individual bend angles [25].

19.2.4. Distortions of Single-Stranded Nucleic Acids Numerous proteins involved in DNA metabolism are capable of binding to and altering the CD spectra of single-stranded DNAs (ssDNAs) and single-stranded RNAs (ssRNAs),

622

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

many of which bind the nucleic acid by means of a characteristic β-strand OB-fold (oligonucleotide/oligosaccharide binding fold) domain [4, 26, 27]. The small 87-residue gene 5 protein (g5p) encoded by Ff ﬁlamentous phage is involved in both replication of the viral genome and control of mRNA translation, and it is typical of many, but not all, OB-fold proteins with respect to the effect it has on the CD spectra of bound nucleic acids. The g5p is a stable dimer that binds antiparallel segments of DNA devoid of any base pairing between the strands [28]. The major binding mode has four nucleotides sequestered per protein monomer. Figures 19.4a and 4b show CD spectra of 36-nucleotide oligomers of dA36 and rA36 as they are titrated with the g5p up to a protein monomer/nucleotide ratio of 0.25, a ratio at which the oligomers are effectively saturated under the conditions of 2 mM Na+ (phosphate buffer), pH 7.0, and 20◦ C [27, 29, 30]. Upon binding of g5p, the positive long-wavelength CD bands of dA36 decrease to negative values, and the crossover wavelength undergoes a dramatic red shift, but there is minimal change in the magnitude of the CD band near 250 nm (Figure 19.4a). These changes are typical of those seen for binding of ssDNAs by a number of other ssDNA binding proteins [3, 27]. Similar CD effects are also observed for duplex DNAs wrapped into nucleosomes and packaged into some phage heads [3, 4], as well as for B-DNA DNA unwound from 10.4 bp/turn to 10.2 bp/turn by solvents such as methanol [31]; also see Chapter 17, this volume. There was no evidence of an anomalous contribution from differential light scattering, extending to wavelengths outside of the absorption band, in titrations of dA36 with the g5p. The CD spectrum of g5p itself has small bands above 250 nm that contribute less than 0.2 ε units on the scale shown in Figure 19.4 at a protein/nucleotide ratio of 0.25. A simple interpretation of the DNA structural change(s) that are correlated with the long-wavelength CD changes of dA36 upon binding g5p remains elusive. The CD effects caused by g5p binding are not simply the result of nucleotide unstacking that might be mimicked by heating, since the CD spectrum of dA36 undergoes much smaller changes in shape and in band magnitudes at 250 and 280 nm upon heating to 86◦ C (Figure 19.5a). Indeed, the temperature-induced CD changes of free dA36 are not obviously explained by just the loss of stacking and uncoupling of base transition moments, although the

(a)

(b)

Figure 19.4. CD titrations of (a) dA36 and (b) rA36 with the Ff gene 5 protein (g5p). Spectra are shown for the free oligomers in the absence of protein (solid curves) and at protein/nucleotide ratios of 0.05–0.07 (solid circles), 0.14 (dashed curve for dA36 + g5p), 0.20–0.21 (crosses), and 0.25 (dash–dotted curve). Final nucleotide concentrations were typically 40–60 μM in the titrated

samples. Spectra were taken at 20◦ C using a 10-mm-pathlength cell. The buffer was 2 mM Na+ (phosphate buffer), pH 7.0. CD values are in M−1 cm−1 , per mole of nucleotide. Measurement errors from repeat experiments were smaller than the differences between the spectra shown. (Data taken from references 27 and 29.)

CIRCULAR DICHROISM OF PROTEIN–NUCLEIC ACID INTERACTIONS

absorbance of dA36 at 260 nm is signiﬁcantly hyperchromic (by 22%) upon heating to 90◦ C [30, 32]. It is interesting that CD titrations of poly[d(A)] with Escherichia coli RNA polymerase core enzyme [32] and with the single-stranded DNA binding protein from Pf3 phage [4] result in perturbations of the polymer that are more like those seen upon an increase in temperature than are the effects seen upon titration with the Ff g5p. CD spectral changes of ssDNAs in protein–ssDNA complexes apparently reﬂect a composite of alterations in nucleotide conformations as well as in base–base interactions. On the other hand, the titration of rA36 with g5p results in a decrease in the magnitudes of both positive and negative CD bands above 240 nm (Figure 19.4b). In this case, the band magnitudes are decreased almost exactly as when the oligomer itself is heated to 86◦ C in the absence of protein (Figure 19.5b). Thus, when the backbone conformation is restricted as in RNA, the binding of the protein appears to result in just the unstacking of bases and a monotonic decrease in the optical activity of the nucleic acid. G5p binds to rA36 with only slightly higher afﬁnity than to dA36 at 0.2 M NaCl (1.3 × 10−5 M versus 0.7 × 10−5 M, respectively) [30], and complexes formed between g5p and poly[d(A)] or poly[r(A)] are similar as visualized by electron microscopy [28].

19.2.5. Interactions at the Nucleotide Level Perturbations of the CD spectrum of a nucleic acid that occur upon binding of a protein are generally attributed to changes in short-range interactions between the bases, providing evidence for transitions between secondary structures or alterations in base pairing, base stacking, or bending of a DNA duplex. Support for attributing such changes to short-range interactions is provided in a study by Repges et al. [33] that combined experimental measurements and theoretical calculations of the spectral effects of the binding of a DNA methyltransferase from Thermus aquaticus, M.TaqI, on a 14-base pair (bp) DNA containing the M.TaqI 4-bp target sequence. For this study, 6-thioguanine base substitutions were made at six different positions in the target sequence. The 6-thioguanine base absorbs at 342 nm where unmodiﬁed DNA and the enzyme are transparent. Calculations of spectral changes of this base upon protein binding were made using a matrix method in a dipole–dipole approximation. (The method used by Repges et al. [33] was also applied to calculate the CD spectra of different conformations of G-quadruplex DNA structures [23]. See Chapter 20 in volume 1 for a treatment of independent systems theory.) Calculated spectra for the six 6-thioguanine-substituted DNAs were in reasonable

(a)

(b)

Figure 19.5. CD spectra of (a) dA36 and (b) rA36 as a function of increasing temperature. Spectra are shown at 20◦ C (solid curves) and at temperatures of 38◦ C (solid circles), 57◦ C (dashed curve), and 86◦ C (crosses). The buffer was 2 mM Na+ (phosphate buffer), pH 7.0. CD values are in M−1 cm−1 , per mole of nucleotide. (Data taken from reference 27.)

623

624

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

agreement with the measured spectra, although the calculated spectra were blue-shifted by about 30 nm. Nevertheless, the study indicated that binding of the protein affected the CD spectrum of the 6-thioguanine mainly via changes in the coupling of the lowestenergy π → π ∗ transitions of the substituted base with the lowest-energy transitions of its stacked nearest-neighbors, without direct optical effects due to the enzyme. Von Hippel and colleagues [34, 35] used 2-aminopurine substitutions to investigate conformational changes of DNA and RNA at speciﬁc nucleotide sites in protein–nucleic acid complexes. The 2-aminopurine base, an isomer of adenine, has an absorbance band in the 300- to 350-nm region that is at longer wavelengths than the absorbance wavelengths of the common bases. This band exhibits a CD spectrum that is dependent on the local nucleotide interaction. For example, if there is stacking of two 2-aminopurine bases in a B-like conformation, the CD spectrum above 300 nm is split by exciton interaction into two overlapping bands that result in a sigmoidal shape. The combination of CD and ﬂuorescence studies with site-speciﬁc 2-aminopurine labels has been used to conﬁrm structural aspects of complexes of the λ phage boxB RNA hairpin with the N protein, as required for λ phage transcription antitermination [34]. Another important application of this approach has allowed the mapping of nucleic acid conformations throughout a T7 polymerase elongation complex, using long-wavelength CD measurements of 2-aminopurine labels in promoter-initiated complexes and at various positions of a preassembled RNA-primed DNA bubble in complexes with T7 RNA polymerase [35].

19.2.6. Alterations in Protein Secondary Structures One of the most important applications of CD spectroscopy has been in the analysis of protein secondary structures and detection of changes in secondary structures upon binding of nucleic acids and other ligands (see Chapters 14 and 20, this volume). Transcription factors can undergo signiﬁcant changes in their α helical contents as they bind to nucleic acids, and they have been among the proteins most studied by CD measurements at wavelengths <250 nm. 19.2.6.1. Basic/Helix–Loop–Helix/Leucine Zipper Transcription Factors. Members of the b/HLH/Zip family of eukaryotic transcription factors have a basic (b) recognition domain that binds to speciﬁc DNA sequences, while a helix–loop–helix (HLH) and/or a leucine zipper (Zip) region allows dimerization. An example is the CREB (cAMP response element-binding) transcription factor. In work by Moll et al. [36], the binding of the CREB b/Zip region to a 32-bp DNA duplex containing a consensus CRE binding site resulted in ≥ 20% increase in the magnitudes of CD bands at 208 and 222 nm, presumably due to the basic region of the protein becoming more α-helical upon DNA binding. The CD signal at 222 nm was further used to obtain thermal melting proﬁles of the protein bound to various DNA sequences and to demonstrate that the presence of 10 mM MgCl2 enhanced the sequence speciﬁcity of binding. The yeast transcription factor GCN4, another b/Zip protein, was studied by Chan et al. [37]. A 57-residue b/Zip protein was chemically synthesized to contain the 27residue basic recognition region of GCN4 hybridized to a leucine zipper domain from C/EBP (CCAAT/enhancer binding protein) [37, 38]. The hybrid protein was found to have an enhancement of α helicity from 49% to 60–75% (or increases of 22–53%) when bound to duplex DNAs of 24 bp, due to the basic region becoming more helical.

CIRCULAR DICHROISM OF PROTEIN–NUCLEIC ACID INTERACTIONS

The increase occurred when the protein was added at a 1:1 ratio to either speciﬁc or nonspeciﬁc 24-mer duplex DNAs. Since there was no relationship between the induction of α helicity and binding afﬁnities ascertained from EMSA (electrophoretic mobility shift assay) measurements, the results signiﬁed that the basic region of the GCN4 protein can apparently be preorganized by binding to nonspeciﬁc sites on the genome [37]. In addition to the importance of CD in monitoring the binding of the basic region of b/Zip proteins to DNA, CD measurements can also be useful in detecting the formation of leucine zipper dimers, which consist of two coiled α-helices. Morgan et al. [39] engineered a chimeric protein in which the GCN4 bZip region of 58 residues was fused to the N-terminus of photoactive yellow protein, PYP (125 residue). The linkage was engineered so that the zipper residues would be likely to pack as a monomer against folded PYP in a dark adapted state, but would be released and be able to dimerize into an α-helix following irradiation at 446 nm and partial unfolding of the PYP. CD spectra at 200–250 nm of the engineered chimera in a dark state were consistent with the sequestration of the bZip region, and, even in the dark state, dimerization could be selectively induced by the addition of a 20-bp DNA duplex containing a speciﬁc target binding site of 7 bp. See Figure 19.6. Afﬁnity for the speciﬁc target DNA was further enhanced after irradiation, as determined by EMSA experiments, suggesting that such chimeras could eventually allow photocontrol of transcription processes regulated by b/Zip proteins. CD measurements have also been important in investigating mutants that permit dimerization of the Zip region of the Mad1 b/HLH/Zip transcription factor [40]. Under physiological conditions, Mad1 does not form a homodimer and bind to DNA, but it does act as a heterodimer with Max, another b/HLH/Zip protein, in regulating cellular processes such as growth and apoptosis in a complex pathway. Formation of Mad1 homodimers is thought to be unfavorable because of the opposition of charged side chains, such as a critical Asp112, that prevents the formation of leucine zipper α-helix dimers. Montagne et al. [40] used CD to show that the substitution of Asp112 resulted in an increased α-helical content of Mad1 when bound to DNA, due to the enhanced presence of a leucine zipper, especially at a physiological temperature (37◦ C). The spectrum of the protein without the Asp112 mutation in the presence of the DNA at 20◦ C has a 222-nm band magnitude indicative of about 50% α-helicity, which could be accounted for by the contributions of helices of the basic region bound to DNA plus the HLH region (open circles in Figure 19.7a). The increased relative α helicity of the mutant–DNA complex at temperatures above 20◦ C (closed circles in Figures 19.7a and 19.7b) is persuasive evidence for the addition of substantial leucine zipper content and that Asp112 is a critical residue in normally repressing Mad1 homodimerization. 19.2.6.2. Zinc Finger Transcription Factors. Zinc ﬁnger motifs are important because they are found in one of the largest families of transcription factors. CD spectroscopy has been valuable in investigating possible structural changes of both the DNA and protein components of zinc ﬁnger–DNA complexes. (See Sections 19.2.1 and 19.2.2 above for examples of the changes in DNA spectra.) The Zic and Gli subfamilies of zinc ﬁnger transcription factors have ﬁve tandem Cys2 –His2 ﬁngers that are involved in regulating aspects of vertebrate development [41]. Sakai-Koto et al. [41] inferred from increases in CD bands at 208 and 222 nm that the α-helical contents of zinc ﬁngers of Gli1, Gli2, and Gli3 transcription factors increase from a range of 16.5–21.1% to a range of 24.9–37.3% upon binding a 35-bp DNA duplex that contained a consensus 9-bp Gli-binding site. Gli zinc ﬁngers showed larger increases in binding afﬁnities and

625

626

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Mean residue ellipticity (deg. cm2. dmol–1)

0

–5

–10

–15

–20×103 200

210

220

230

240

250

260

Wavelength (nm)

Figure 19.6. CD spectrum of a dark-adapted chimeric protein consisting of a GCN4 b/Zip region fused to a photoactive yellow protein, PYP (solid curve). The protein was at a concentration of 15 μM. Spectra of the protein are also shown after addition of an equimolar amount of nonspeciﬁc DNA (short dashed curve) or speciﬁc target DNA (medium dashed curve) and subtraction of the CD spectra of the DNAs. Spectra were taken at 20◦ C using a 1-mm-pathlength cell. The buffer was 1× TAE buffer (Tris–acetate–ethylenediaminetetraacetic acid buffer), pH 7.5, plus 100 mM NaCl. CD values are in mean residue molar ellipticity units, deg cm2 dmol−1 , after subtracting any DNA contribution. (Reprinted from J. Mol. Biol., Vol. 339, S.-A. Morgan, S. Al-Abdul-Wahid, G. A. Woolley, Structure-based design of a photocontrolled DNA binding protein, p. 104, copyright 2010, with permission from Elsevier.)

in apparent α-helical contents than did Zic zinc ﬁngers when binding the same DNA, suggesting that the α-helical content of the bound protein is related to its binding afﬁnity [41]. Spectral effects attributed to the proteins upon forming the zinc ﬁnger–DNA complexes were obtained by subtracting the spectrum of the free DNA, it being assumed that an altered DNA structure did not contribute to changes in the short-wavelength CD spectra of the complexes. Yan et al. [42] were able to partially separate the CD effects of binding into DNA and protein components in a study involving three artiﬁcial peptides containing six Cys2 –His2 zinc ﬁngers (from the Sp1 transcription factor) that bind to GC-rich sequences (GC boxes). In three different peptide constructs, two groups of three zinc ﬁngers each were separated from each other by linkers of either (Gly–Gly–Gly–Gly–Ser)4 , (Glu–Ala–Ala–Ala–Arg)4 , or (Lys–Ala–Ala–Glu–Ala)4 . Each group of three zinc ﬁngers binds to a GC box of 9 bp. When each of the three peptide constructs was added to a DNA containing two GC boxes separated by a spacer of 10 AT bp, named 2GC(10), the positive CD band of the DNA at long wavelengths (245–300 nm) was increased to the

CIRCULAR DICHROISM OF PROTEIN–NUCLEIC ACID INTERACTIONS

627

[θ] (deg⋅cm2⋅dmol–1)

0

–5000

–10000

–15000

–20000 200

210

220 230 Wavelength (nm)

240

250

(a)

[θ]222 (deg⋅cm2⋅dmol–1)

0

–5000

–10000

–15000

–20000

20

40 60 Temperature (°C) (b)

80

Figure 19.7. (a) CD spectra of the b/HLH/Zip region of 80 residues of transcription factor Mad1 in the presence of a 21-bp DNA containing a 6-bp E-box-speciﬁc target sequence. The spectrum of the Mad1 protein (with Cys replaced by Ser at two positions) in the presence of the DNA is shown by open circles. The spectrum of the same protein with a critical Asp112 to Asn112 mutation had an enhanced α helical content when bound to the E-box-containing sequence, indicated by the spectrum with closed circles. The contribution of the DNA has been subtracted. (b) Melting proﬁles of the complexes in panel a. Spectra in panel a were taken at 20◦ C. The protein dimer and DNA duplex concentrations in the mixtures were 15 μM and 20 μM, respectively, in a cell of 1-mm pathlength. The buffer was 50 mM phosphate, 50 mM KCl, and 5 mM DTT (dithiothreitol), pH 6.8. CD values are in mean residue molar ellipticity units, deg cm2 dmol−1 . (Reprinted with modiﬁcations and permission from Montagne et al. [40], © 2005, American Chemical Society.)

same extent, indicating the same extent of binding to the DNA and perturbation of the DNA structure; see Figure 19.8. At short wavelengths below 245 nm there were differences in the CD bands that could be attributed to the differential induction of α-helicity in the linker regions of the peptides, overlapping any contribution at short wavelengths due to the altered DNA structure. That is, it was reasonable to assume that the DNA

628

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

θ (mdeg)

5

0

–5

Sp1ZF6(ER)4 + [2GC(10)] Sp1ZF6(KE)4 + [2GC(10)] Sp1ZF6(G4S)4 + [2GC(10)] Free [2GC(10)]

–10 200

220

240

260

280

300

320

Wavelength (nm)

Figure 19.8. CD spectra of a DNA containing two GC-box sequences separated by a 10-bp spacer, 2GC(10), complexed with each of three peptides containing six zinc ﬁngers but with different linkers between zinc ﬁngers 3 and 4: Sp1ZF6(ER)4 with linker (Glu–Ala–Ala–Ala–Arg)4 , Sp1ZF6(KE)4 with linker (Lys–Ala–Ala–Glu–Ala)4 , and Sp1ZF6(G4 S)4 with linker (Gly–Gly–Gly–Gly–Ser)4 . Spectra were taken at 20◦ C using a 1-mm-pathlength cell. Samples contained 4.5 μM peptide–DNA complex in 10 mM Tris–HCl (pH 8.0), 50 mM NaCl, 0.005% Nonidet P-40, and 1 mM dithiothreitol. CD values are mdeg ellipticity. (Reprinted with permission from Yan et al. [42], ©2005, American Chemical Society.) (See insert for color representation of the ﬁgure.)

spectral contribution at short wavelengths was the same for all three mixtures, given the similarity of the spectra at long wavelengths. Differences in the CD bands at short wavelengths could thus be ranked and assigned mainly to the different peptide linkers, plus possibly some contribution from an indirect effect on the structures of the zinc ﬁngers themselves. A determination of the absolute magnitudes of the spectra of the complexed peptides would require knowing and subtracting the actual perturbed spectrum of the DNA at short wavelengths. 19.2.6.3. Homeodomain Transcription Factors. Homeodomains are a third type of transcription factor that have been investigated by CD measurements. Homeodomains contain helix–turn–helix (HTH) motifs, where the C-terminal recognition helix increases in length upon binding DNA. For example, Vecchio et al. [43] estimated from short-wavelength CD spectra that a 68-residue segment of the rat thyroid transcription factor 1 homeodomain (TTF-1HD), with about 40% α-helix, increases in α-helical content to about 50% when binding a 14-bp duplex containing the 4-bp TTF-1HD binding site. This was consistent with about six residues of the recognition helix becoming helical upon binding DNA. It was assumed that the CD contribution of the DNA did not change during binding. 19.2.6.4. Recombination Activating Protein. The RAG1 (recombination activating gene 1) protein is involved in immunoglobulin gene V(D)J recombination. Ciubotaru et al. [44] used CD to investigate the binding of the core domains of the murine RAG1 protein to a 40-bp oligomer containing 7-mer and 9-mer recombination signal

CIRCULAR DICHROISM OF PROTEIN–NUCLEIC ACID INTERACTIONS

sequences. They found that the magnitudes of CD bands due to the protein’s α-helicity decrease dramatically in the bound state, to less than 25% of their values in the spectrum of the free protein, indicative of a signiﬁcant change in the protein’s secondary structure. This is in contrast to the transcription factors discussed above, which undergo an enhancement of α-helical structure upon DNA binding. Mixtures of the RAG1 protein with a nonspeciﬁc DNA exhibited much less of a decrease in these CD bands. 19.2.6.5. Translin Protein. Human translin protein is an unusual octameric protein, with subunits of 228 residues, that binds to both DNA and RNA and that has relatively high afﬁnity for (GT)n microsatellite repeats and d(TTAGGG)n telomeric repeats [45, 46]. From CD measurements at short wavelengths <250 nm, Kaluzhny et al. [46] determined that the α-helical content of translin decreases from about 67% to 50% upon binding d(TG)12 , consistent with the loss of helical structure by 38–39 residues. They also investigated the spectra of complexes with d(TG)12 and d(TTAGGG)5 at wavelengths >250 nm in titrations with translin and obtained evidence for possible unstacking of d(TG)12 and unfolding of a d(TTAGGG)5 quadruplex structure by translin. It was assumed that the long-wavelength CD contribution of the translin protein was not perturbed by binding the DNAs.

19.2.7. Effects on Protein Aromatic Residues In addition to the widespread use of CD spectra to identify differences in protein secondary structures when free and when bound to nucleic acids, as illustrated by the above examples, CD measurements have been used to detect the effects of nucleic acid substrate binding on aromatic residues at both long and short wavelengths. In a study of the formation of a ternary complex between human polynucleotide kinase plus ATP and a 20-mer DNA oligonucleotide, Mani et al. [47] found that the CD bands of Trp, Tyr, and Phe residues at wavelengths above 250 nm were differentially perturbed by the binding of ATP and the oligonucleotide. Their data indicated that the substrates each induced different conformations of the protein. Aromatic residues can also be signiﬁcant contributors to CD spectra at wavelengths shorter than 250 nm, with perturbations induced by binding to nucleic acids. As referred to in Section 19.2.4, CD has been important in investigating protein–DNA interactions of proteins such as the Ff g5p that bind ssDNA; see reference 4 for additional examples. The spectrum of the g5p protein is unusual in the 210- to 250-nm short-wavelength region in that it is dominated by a positive band at 229 nm due to the combined La bands of ﬁve Tyr [48]. (The g5p has aromatic residues consisting of ﬁve Tyr, three Phe, and no Trp). The protein also mainly has β secondary structure, with only 7% of distorted α-helix [48], and there is no evidence for large CD signal changes caused by alterations in α-helical structure upon DNA binding as found in studies with other proteins reviewed above. The short-wavelength spectra of single-stranded nucleic acids such as dA36 titrated with the g5p exhibit a peak near 229 nm that is clearly inﬂuenced by the protein, as seen in Figure 19.9 (dash–dotted curve). Moreover, as may be seen by inspection, the magnitude of the CD at 229 nm in the titrated sample (dash–dotted curve) is less than would be predicted by summing the individual spectra of the free g5p and free dA36 (circles plus solid curve). Since the spectra of single-stranded nucleic acids at wavelengths above 250 nm are perturbed by the protein (Figure 19.4), where the protein spectrum itself is negligible on the same scale, it is likely that the spectrum of the nucleic acid in the short-wavelength region is also perturbed by protein binding. Therefore, the

629

630

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

30

εL– εR (M–1cm–1)

20 10 0 –10 –20

220

240 260 Wavelength (nm)

280

300

Figure 19.9. CD spectra of free dA36 (solid curve), the free Ff g5p (solid circles), and a mixture at a protein monomer/nucleotide ratio of 0.25 (dash–dotted curve). Spectra were taken at 20◦ C. The buffer was 2 mM Na+ (phosphate buffer), pH 7.0. CD values of the free dA36 and the mixture are in M−1 cm−1 , per mole of nucleotide. The spectrum of the free g5p is shown at one-fourth the magnitude per mole of protein monomer so that the spectrum of the free protein is comparable with that of the protein component of the mixture. The range of values at 229 nm for repeat CD spectra of the protein was ±1ε on this scale. (Data taken from references 27, and 29. Also see legend to Figure 19.4.)

typical challenge arises as to how to separate the short-wavelength spectral effects of g5p binding into those due to the protein versus the DNA. Fortunately, the spectra of some ssDNA sequences like poly[d(T)] have crossover wavelengths close to 229-nm, and a decrease in the magnitude of the 229-nm band of the protein in complexes with such sequences can be more conﬁdently quantitated [49]. Another approach was to study the short-wavelength spectra of complexes with poly[r(A)], since the effect of protein binding on the spectrum of the polymer at long wavelengths is similar to that of heating (see Figure 19.5). In this case, it was reasonable to assume that g5p binding proportionally decreased the magnitudes of spectral bands of poly[r(A)] at short wavelengths, which could then be subtracted from the spectrum of the complex [50]. A third approach to reveal the origin of a binding effect on Tyr residues was to study mutants of g5p with each of the ﬁve Tyr individually substituted by Phe or His [48, 51, 52]. It was possible by these and other approaches to evaluate the contribution of each Tyr to the 229-nm band and to show that perturbation of a speciﬁc Tyr (Tyr 34) at the interface of cooperatively bound proteins (and not, for example, a Tyr in the DNA-binding site) is responsible for most of the decrease of the 229-nm CD band during formation of g5p-nucleic acid complexes.

19.3. SUMMARY First, this chapter has emphasized the structural information available from CD measurements of interacting proteins and nucleic acids, since the identiﬁcation and monitoring of structural changes has been a primary goal of researchers in the ﬁeld. Unless there are effects of aggregates or liquid crystalline states of the materials, solution CD spectra of protein–nucleic acid complexes can reasonably be interpreted as the sum of the separate

CIRCULAR DICHROISM OF PROTEIN–NUCLEIC ACID INTERACTIONS

contributions of the protein and nucleic acid secondary structures. The main challenge in using CD spectra to study structural effects of protein–nucleic acid binding has been in separating overlapping CD effects into those of the protein and nucleic acid components. As illustrated by the examples above, differences in the near-UV (250–320 nm) CD spectra that report on nucleic acid conformational changes can be less ambiguous than are changes in the far UV (190–250 nm), where protein and nucleic acid CD bands overlap and may well be similar in magnitude, although substantial variations in α-helical contents of proteins can be apparent in the far-UV region. Site-speciﬁc substituents (e.g., 6-thioguanine or 2-aminopurine in nucleic acids) that absorb at wavelengths longer than 300 nm have an advantage in that monitoring and interpreting optical activity at such long wavelengths is not complicated by the presence of overlapping CD bands from the normal nucleic acid or protein chromophores. Second, empirical interpretations of CD spectra in terms of solution structural changes are most convincing when the interpretations are consistent with evidence from crystallography, NMR, and other techniques. Comparisons of CD experiments using different protein mutants or constructs, as well as different nucleic acid sequences, can help distinguish between protein and nucleic acid structural perturbations. A ﬁnal point is that knowing the exact contributions to the CD spectrum of a protein–nucleic acid complex is not a requirement for using CD measurements to determine parameters such as the stoichiometry or afﬁnity of binding, or to monitor the effects of varying temperature, time, or solvent conditions. Methodologies for using CD spectroscopy in such experiments have been addressed by others [1, 2].

ACKNOWLEDGMENTS The author gratefully acknowledges the colleagues who responded to requests for information and permissions. High-resolution ﬁgures were generously provided by Dr. Y.-G. Kim (Sungkyunkwan University, Suwon, Korea) for use in Figure 19.1 and by Dr. S. P. Edmondson (University of Alabama, Huntsville) for use in Figure 19.3. Special thanks are expressed to editor Dr. Robert W. Woody for his support and cogent comments and to Dr. Carla W. Gray for a critical reading of the manuscript. Work at the University of Texas at Dallas was supported by Grant AT-503 from the Robert A. Welch Foundation.

REFERENCES 1. N. J. Greenﬁeld, Nat. Protoc. 2006, 1 , 2733–2741. 2. P. D. Cary, G. G. Kneale, Circular dichroism for the analysis of protein–DNA interactions, in Methods in Molecular Biology, DNA-Protein Interactions, Vol. 543, T Moss and B. Leblanc, eds., Humana Press, New York, 2009, pp. 613–624. 3. D. M. Gray, Circular dichroism of protein-nucleic acid interactions, in Circular Dichroism and the Conformational Analysis of Biomolecules, G. D. Fasman, ed., Plenum Press, New York, 1996, pp. 469–500. 4. D. M. Gray, CD of protein–nucleic acid interactions, in Circular Dichroism: Principles and Applications, 2nd edition, N. Berova, K. Nakanishi, and R. W. Woody, eds., John Wiley & Sons, New York, 2000, pp. 769–796. 5. Y.-M. Kang, J. Bang, E.-H. Lee, H.-C. Ahn, Y.-J. Seo, K. K. Kim, Y.-G. Kim, B.-S. Choi, J.-H. Lee, J. Am. Chem. Soc. 2009, 131 , 11485–11491.

631

632

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

6. D. V. Quyen, S. C. Ha, K. Lowenhaupt, A. Rich, K. K. Kim, Y.-G. Kim, Nucl. Acids Res. 2007, 35 , 7714–7720. 7. N. Deigendesch, F. Koch-Nolte, S. Rothenburg, Nucl. Acids Res. 2006, 34 , 5007–5020. 8. S. Rothenburg, N. Deigendesch, K. Dittmar, F. Koch-Nolte, F. Haag, K. Lowenhaupt, A. Rich, Proc. Natl. Acad. Sci. USA 2005, 102 , 1602–1607. 9. H. Takeuchi, N. Hanamura, I. Harada, J. Mol. Biol . 1994, 236 , 610–617. 10. Y.-G. Kim, H.-J. Park, K. K. Kim, K. Lowenhaupt, A. Rich, Nucl. Acids Res. 2006, 34 , 4937–4942. 11. X. Chen, M. Chu, D. P. Giedroc, Biochemistry 1999, 38 , 12915–12925. 12. C. N. N’soukpo´e-Kossi, S. Diamantoglou, H. A. Tajmir-Riahi, Biochem. Cell Biol . 2008, 86 , 244–250. 13. R. Stoll, B. M. Lee, E. W. Debler, J. H. Laity, I. A. Wilson, H. J. Dyson, P. E. Wright, J. Mol. Biol . 2007, 372 , 1227–1245. 14. S. Lee, K. Doddapaneni, A. Hogue, L. McGhee, S. Meyers, Z. Wu, J. Mol. Biol . 2010, 397 , 1055–1066. 15. B. I. Kankia, G. Barany, K. Musier-Forsyth, Nucl. Acids Res. 2005, 33 , 4395–4403. 16. T. R. Salas, I. Petruseva, O. Lavrik, A. Bourdoncle, J. L. Mergny, A. Favre, C. Saintom´e, Nucl. Acids Res. 2006, 34 , 4857–4865. 17. Y. Qin Y, L. H. Hurley, Biochimie 2008, 90 , 1149–1171. 18. Y. Wu, N. Rawtani, A. K. Thazhathveetil, M. K. Kenny, M. M. Seidman, R. M. Brosh Jr., Biochemistry 2008, 47 , 5068–5077. 19. J.-H. Fan, E. Bochkareva, A. Bochkarev, D. M. Gray, Biochemistry 2009, 48 , 1099–1111. 20. M. Paramasivam, A. Membrino, S. Cogoi, H. Fukuda, H. Nakagama, L. E. Xodo, Nucl. Acids Res. 2009, 37 , 2841–2853. 21. J. A. Kelly, J. Feigon, T. O. Yeates, J. Mol. Biol . 1996, 256 , 417–422. 22. M. Cruceanu, M. A. Urbaneja, C. V. Hixson, D. G. Johnson, S. A. Datta, M. J. Fivash, A. G. Stephen, R. J. Fisher, R. J. Gorelick, J. R. Casas-Finet, A. Rein, I. Rouzina, M. C. Williams, Nucl. Acids Res. 2006, 34 , 593–605. 23. D. M. Gray, J.-D. Wen, C. W. Gray, R. Repges, C. Repges, G. Raabe, J. Fleischhauer, Chirality 2008, 20 , 431–440. 24. W. B. Peters, S. P. Edmondson, J. W. Shriver, Biochemistry 2005, 44 , 4794–4804. 25. I. Papapanagiotou, S. D. Streeter, P. D. Cary, G. G. Kneale, Nucl. Acids Res. 2007, 35 , 2643–2650. 26. A. G. Murzin, EMBO J . 1993, 12 , 861–867. 27. T.-C. Mou, M. Shen, S. Abdalla, D. DelaMora, E. Bochkareva, A. Bochkarev, D. M. Gray, Chirality 2006, 18 , 370–382. 28. G. A. Olah, D. M. Gray, Carla W. Gray, D. L. Kergil, T. R. Sosnick, B. L. Mark, M. R. Vaughan, J. Trewhella, J. Mol. Biol . 1995, 249 , 576–594. 29. T.-C. Mou, C. W. Gray, T. C. Terwilliger, D. M. Gray, Biochemistry 2001, 40 , 2267–2275. 30. T. C. Mou, D. M. Gray, Nucl. Acids Res. 2002, 30 , 749–758. 31. W. C. Johnson, Jr., CD of nucleic acids, in Circular Dichroism: Principles and Applications, 2nd edition, N. Berova, K. Nakanishi, and R. W. Woody, eds., John Wiley & Sons, New York, 2000, pp. 703–718. 32. R. R. Reisbig, A-Y. M. Woody, R. W. Woody, J. Biol. Chem. 1979, 254 , 11208–11217. 33. R. Repges, C. Beuck, Y. E. Weinhold, G. Raabe, J. Fleischhauer, Chirality 2008, 20 , 978–984. 34. N. P. Johnson, W. A. Baase, P. H. von Hippel, J. Biol. Chem. 2005, 280 , 32177–32183. 35. K. Datta, N. P. Johnson, P. H. von Hippel, J. Mol. Biol . 2006, 360 , 800–813. 36. J. R. Moll, A. Acharya, J. Gal, A. A. Mir, C. Vinson, Nucl. Acids Res. 2002, 30 , 1240–1246.

CIRCULAR DICHROISM OF PROTEIN–NUCLEIC ACID INTERACTIONS

37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52.

I.-S. Chan, A. V. Fedorova, J. A. Shin, Biochemistry 2007, 46 , 1663–1671. A. R. Lajmi, T. R. Wallace, J. A. Shin, Protein Expression Puriﬁcation 2000, 18 , 394–403. S.-A. Morgan, S. Al-Abdul-Wahid, G. A. Woolley, J. Mol. Biol . 2010, 399 , 94–112. M. Montagne, J.-F. Naud, F.-O. McDuff, P. Lavigne, Biochemistry 2005, 44 , 12860–12869. K. Sakai-Kato, A. Ishiguro, K. Mikoshiba, J. Aruga, N. Utsunomiya-Tate, Biochim. Biophys. Acta 2008, 1784 , 1011–1019. W. Yan, M. Imanishi, S. Futaki, Y. Sugiura, Biochemistry 2007, 46 , 8517–8524. P. D. Vecchio, P. Carullo, G. Barone, B. Pagano, G. Graziano, A. Iannetti, R. Acquaviva, A. Leonardi, S. Formisano, Proteins 2008, 70 , 748–760. M. Ciubotaru, L. M. Ptaszek, G. A. Baker, S. N. Baker, F. V. Bright, D. G. Schatz, J. Biol. Chem. 2003, 278 , 5584–5596. E. Jacob, L. Pucshansky, E. Zeruya, N. Baran, H. Manor, J. Mol. Biol . 2004, 344 , 939–950. D. Kaluzhny, O. Laufman, E. Timofeev, O. Borisova, H. Manor, A. Shchyolkina, J. Biomol. Struct. Dyn. 2005, 23 , 257–265. R. S. Mani, F. Karimi-Busheri, M. Fanta, C. E. Cass, M. Weinfeld, Biochemistry 2003, 42 , 12077–12084. T.-C. Mou, N. Sreerama, T. C. Terwilliger, R. W. Woody, D. M. Gray, Protein Sci . 2002, 11 , 601–613. D. M. Gray, C. W. Gray, T.-C. Mou, J.-D. Wen, Enantiomer 2002, 7 , 49–58. B.-C. Sang, D. M. Gray, Biochemistry 1989, 28 , 9502–9507. B. L. Mark, T. C. Terwilliger, M. R. Vaughan, D. M. Gray, Biochemistry 1995, 34 , 12854–12865. T. M. Thompson, B. L. Mark, C. W. Gray, T. C. Terwilliger, N. Sreerama, R. W. Woody, D. M. Gray, Biochemistry 1998, 37 , 7463–7477.

633

20 DRUG AND NATURAL PRODUCT BINDING TO NUCLEIC ACIDS ANALYZED BY ELECTRONIC CIRCULAR DICHROISM George A. Ellestad

20.1. INTRODUCTION Electronic circular dichroism (CD) provides signature spectra for nucleic acid polymorphic secondary structures as a function of sequence and environmental conditions [1–4].1 Because of this exquisite sensitivity, due primarily to base-pair stacking interactions, CD is a useful technique for studying drug–DNA interactions in which small-molecule drugs may induce a conformational change in the DNA secondary structure and thereby inﬂuence biological function by interfering with the binding of DNA regulatory proteins or by introducing frameshift mutations. There have been a number of excellent reviews over the last 20 years describing the use of CD to monitor drug–DNA interactions covering both natural and synthetic drug binding. A 1992 review by Zimmer and Luck [5] on the CD of drug–DNA binding interactions gave examples of a wide range of DNA-binding ligands including proteins and some of the natural products discussed below. In 2000 Nord´en and co-workers [6] gave an overview of DNA–drug interactions covering the history and theoretical basis of induced circular dichroism (ICD) due to DNA–achiral drug interactions with examples of small synthetic dye molecules, metallated porphyrins, and other aromatic compounds including peptide–nucleic acids. In 2001 Eriksson and Nord´en [7] published a guide covering the use of linear dichroism in combination with CD in the study of drug–DNA complexes and Allenmark [8] reviewed ICD in 2003. In 2007 Chaires and co-workers [9] gave an overview on ligand-induced CD for determining the binding mode and afﬁnity of ligand–DNA interactions, while in 2010 Rodger [10] reported on the use of circular and linear dichroism for drug–DNA binding. All of 1

J. Kypr discusses the CD of DNAs in Chapter 16 of this volume.

Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

635

636

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

these methods and review papers provide theoretical and practical background into this technique with helpful information on experimental design, instrument parameters, and a list of many important references for more detailed discussions. The purpose of this chapter is to provide an introduction in the use of CD spectroscopy for monitoring drug–DNA interactions. We ﬁrst present some general background information concerning secondary DNA structures that are encountered along with their characteristic CD spectra, and then we discuss how CD measurements of the drug–DNA complex can be used to gain immediate insight into the type of binding: groove, intercalation, or binding outside the double helix. This is followed by some recent examples from the literature which illustrate the prominent role that CD continues to play in the characterization of the DNA binding mode of a number of biologically interesting natural occurring antitumor agents and porphyrins. Most of the CD studies described here, conﬁned primarily to the last 10 years, were carried out in conjunction with other biophysical methods such as absorption, ﬂuorescence and 1 HNMR spectroscopy, dialysis, and, in some cases, X-ray crystallography. In contrast to NMR and X-ray crystallography, CD offers advantages in that it requires less material (as low as 25 μg) and is measured in solution using DNA concentrations as low as 25 μg/L [4]. There are three principal DNA secondary structural forms that are important to keep in mind for drug binding studies: the right-handed A and B forms and the left-handed Z form, and these are depicted in Figure 20.1. Guanine-quadruplex DNA is sometimes the target of DNA-binding drugs and will be described later. B-form DNA is called the “canonical” form and is considered to be the native conformation. In certain cases the B form can adopt the A, Z, or other forms depending on the sequence, solvent, hydration, salt, pH, counterion concentrations, temperature, methylation of the bases, and, in some cases, drug or ligand binding. The B-form DNA family of structures has a deep and narrow minor groove while the major groove is less deep and wider (Figure 20.1). In contrast, the A-form DNA has a wider and shallower minor groove than that of B-form DNA; the base pairs are also inclined at a large angle from the helix axis as opposed to the B and Z forms where they are essentially perpendicular. In contrast to the A and B forms, the thinner and more rigid double helix of the higher-energy Z-form DNA describes a left-handed screw arrangement with a zig-zag sugar phosphate backbone pattern. In Z-form DNA the bases are observed in an alternating anti–syn conformation relative to the ribose sugars in which the imidazole ring of guanine is located on the outside of the helix, as opposed to the B form where the bases are observed in the anti conformation and the imidazole ring is more shielded. This is an especially important feature for DNA-binding metallated porphyrins that target the Z form as it provides the possibility for coordination with the nucleophilic N7 of guanine. Each of these principal secondary structures has a characteristic CD spectrum, and these are described below. The recent 2009 review by Kypr’ et al. [4] is especially useful, as is their chapter (Chapter 16) in this volume, in providing the reader with background for correlating the various structural types of DNA with their characteristic CD spectra. CD signals of nucleic acids are observed in the UV region between 200 and 300 nm. In general, the B form has a positive signal between 260 and 280 nm and an equally intense negative signal at ∼245. The A form tends to exhibit a positive and more intense (relative to the positive B form peak) signal at ∼260 nm, with a minimum at ∼210 nm. In contrast, the CD of the left-handed double helix of Z-form DNA exhibits almost the reverse pattern to that of the B form with a negative long-wavelength signal at ∼290 nm and an equally intense

D R U G A N D N AT U R A L P R O D U C T B I N D I N G T O N U C L E I C A C I D S A N A LY Z E D

A-DNA

B-DNA

Z-DNA

Binding modes minor groove binding

major groove minor groove

major groove

intercalation

minor groove

Figure 20.1. Representation of the three principal secondary structures of DNA. The right¨ handed A and B form are obtained from standard parameters within the Schrodinger–Maestro graphical interface. The thinner and more elongated Z form is obtained from X-ray parameters of a hexamer as imported from the protein data bank (PDB). In this representation, three units of the hexamer are stacked in order to display the overall left-handed zig-zag helicity. The structure on the far right depicts drug–DNA double-helix interactions with the drug colored black: minor groove binding (top) and intercalation between base pairs (bottom). (See insert for color representation of the ﬁgure.)

positive signal at ∼260 nm [1]. Figure 20.2 depicts the CD spectra of the right-handed A and B forms and the left-handed Z form of poly(dGdC)·poly(dGdC). These spectra are speciﬁc to this sequence and are somewhat different from those of more mixed base-pair sequences. Yet they show the characteristic features in the 230 to 300 nm region of the spectrum that are most useful for studying DNA conformational changes arising from drug binding. As mentioned above, the B form may be converted to A- and Z-like forms, the ease of which depends on the DNA sequence, experimental conditions, and, in certain cases, drug binding [1–4; also see footnote 1 on page 635.]

20.2. DIAGNOSTIC ELECTRONIC CD SPECTRAL CHANGES ON DRUG–DNA BINDING CD spectroscopy of achiral or chiral molecules bound to DNA may give rise to dramatic new or altered Cotton effects, the proper interpretation of which provides information regarding the mode of binding [6–10]. The CD response may reﬂect an altered conformational state of the DNA, the guest ligand, or both and provide a characteristic spectroscopic “footprint” of the binding event. We have outlined three different situations where a new or altered Cotton effect may be observed upon binding of a drug or natural product to a chiral and chromophoric DNA host: (a) When an achiral and chromophoric drug binds to the DNA host, an induced circular dichroism (ICD) may appear within the absorption band of the asymmetrically perturbed chromophore; (b) DNA binding of a chiral, chromophoric drug may result in a CD change within the

637

638

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

6

A-DNA

Circular dichroism

4 2 0 –2 –4

B-DNA Z-DNA

–6

Figure 20.2. Overlapped CD curves of A-, B-, and 225

250

275 λ (nm)

300

325

Z-form poly(dGdC).poly(dGdC) DNA. (Courtesy of Milan Balaz, University of Wyoming.)

absorption band of the drug as well as a CD change in the absorption region of the DNA due to perturbations of structures with already optically active transition moments; (c) DNA binding of a chiral drug which is CD silent because it contains weak UV–vis chromophores may give rise to CD changes within the absorption bands of the DNA only. The observation of an ICD immediately reveals that the ligand actually binds to DNA, while its sign provides qualitative information about the binding mode: whether it is an intercalator or a groove binder, or binds on the outside of the polyphosphate backbone of the nucleic acid (see below) [5–10]. DNA-targeting natural products, especially those that are relatively ﬂat and with elongated conformations, generally bind in the minor groove of B-form DNA because this deep and narrow groove provides surface dimensions that are more complementary to those of these relatively small molecular weight ligands than the wide and shallower major groove (see Figure 20.1 for labeling of the grooves and for depictions of ligand minor groove and intercalative binding). DNA regulatory proteins may target both the minor and major grooves. However, if there is both minor groove and intercalative binding, interpretation of CD changes can be complicated and CD alone will not be sufﬁcient for a reliable answer. In general, a strong, positive ICD is typical of a minor groove binder where the ligand, such as the crescent shaped, achiral and chromophoric pyrrole-amidine antibiotics, netropsin (1), distamycin (2) [11], and related synthetic compounds, bind tightly and deep within the minor groove with transition moments of the ligand chromophore parallel to the groove. Intercalators usually exhibit negative and weaker ICDs if their long-axis transition moment lies perpendicular to the DNA helical axis. The ICD is generally positive if the intercalator’s long axis transition moment lies parallel with the groove of the DNA [5–10]. In practice, however, application of these generalities may be problematic, especially with structurally complex DNA-binding natural products as these observations apply only to intercalators whose transition moments are oriented along the long axis of the bound ligand. With structurally complex intercalators in particular, it is imperative to assign the transition moments of their UV absorption bands before attempting to interpret the observed CD Cotton effects upon binding. Daunorubicin (3) and actinomycin D (4)

D R U G A N D N AT U R A L P R O D U C T B I N D I N G T O N U C L E I C A C I D S A N A LY Z E D

H2N +

NH2

H2N +

NH2

O O

HN

HN

NCH3

NCH3 O

HN HN

O

NCH3

NCH3

H2N

H + N

HN O

HN O

NCH3

HN

NH2

H O

Netropsin (1)

Distamycin (2)

O

OH

O CH3 OH

OCH3 O

OH

O

O

H3C OH

NH3+

(+) Daunorubicin (3)

CH3

CH3

O

O

H2N

N

NH CH3Val

O

O

NH Thr

Thr Val-D

Sar

ValCH3

D-Val

Sar Pro

Pro

Actinomycin D (4)

639

640

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

are two examples of intercalators whose aromatic chromophores intercalate between the DNA base pairs while the corresponding appended chiral sugar or peptide moieties lie in the minor groove. Some recent applications of CD in the study of their DNA-binding properties are described later in this chapter.

20.3. ILLUSTRATIVE EXAMPLES OF MINOR GROOVE AND INTERCALATIVE NATURAL PRODUCT–DNA BINDING The following examples were chosen to illustrate the use of CD in monitoring the minor groove and intercalative binding of some antitumor natural products that target B-form DNA, enantiomeric intercalators that discriminate between B and Z forms of DNA, porphyrins that speciﬁcally target B- or Z-form DNA, and porphyrins and alkaloids that induce and stabilize quadruplex DNA. The structures of several of these DNA–ligand complexes have also been characterized by X-ray crystallography, which provides a solid molecular basis for the CD interpretations.

20.3.1. DNA Minor Groove Binding Natural Products 20.3.1.1. Netropsin and Distamycin. As mentioned in the previous section, netropsin (1) and distamycin (2) are classic examples of minor groove-speciﬁc binders with crescent shaped and elongated conformations that match well the complementary shape of the DNA minor groove. These achiral, minor groove-binding antibiotics have been under investigation for many years using a variety of biophysical methods, and Zimmer and Luck [5] have summarized the original CD studies with these antibiotics. An X-ray structure published in 1985 of an AT-rich 12-mer complexed with the dicationic netropsin provided a strong molecular basis for the interaction and showed a 1:1 molar ratio of drug to DNA oligomer [11]. A few years later, Pelton and Wemmer [12] published a surprising NMR study demonstrating that in contrast to the dicationic netropsin, the monocationic distamycin could bind in a side-by-side manner to certain duplex oligomers with a relatively wide minor groove in a 2:1 molar ratio. Then, Rentzeperis et al. [13] carried out an ICD titration study of netropsin and distamycin with an AAATT-containing DNA 11-mer as part of a thermodynamic binding study of these agents. These authors took advantage of the dramatic ICDs that were observed on binding, in conjunction with titration calorimetry and optical melting, to correlate binding thermodynamics with results obtained from NMR titrations. Netropsin and distamycin are optically inactive and thus CD silent. Yet as seen in Figure 20.3, the addition of netropsin (Figure 20.3a) and distamycin (Figure 20.3b) to the DNA oligomer results in signiﬁcant ICDs in the CD spectra. The wavelengths of these ICDs are clearly beyond the DNA region, and changes in ellipticity are easily monitored. The strong ICDs of the DNA-bound drugs, so typical of this type of minor groove binder, have their origin in the intermolecular coupling of the electric transition moments of the DNA base pairs and those of the extended π system of the drug when it lies deep within the chiral environment of the minor groove. CC-1065 (5, Scheme 20.1) is another example of a minor groove binder and is discussed below. In order to interpret these ICDs properly, however, an understanding of the magnitude and direction of the transition moments is critical [7]. The binding stoichiometry for each antibiotic was obtained by plotting the intensities of the ICD bands at 310 nm for netropsin and 333 nm for distamycin against the ratio of

D R U G A N D N AT U R A L P R O D U C T B I N D I N G T O N U C L E I C A C I D S A N A LY Z E D

25

641

(a) 296

11 –3

240

262

–17

θ, mdeg

–31 –45

(b)

25

319

0 243.6

290

Figure 20.3. CD spectra with isoelliptic points of the –25

–50 220

ligand–A3 T2 complexes at several ligand/duplex ratios in 10 mM sodium phosphate buffer, 10 mM NaCl, and 0.1 mM Na2 EDTA at pH 7.0 and 20◦ C with (a) netropsin

260

300 Wavelength, nm

340

and (b) distamycin. (Reproduced with permission from ACS, reference 13, Figure 3.)

380

25

θ, mdeg

20 15 10

Figure 20.4. CD titration curves monitored at 310 for netropsin and 333 nm for distamycin. Solutions were 10 mM in 10 mM sodium phosphate, 10 mM

5 0

0

[Ligand]/[A3T2]

NaCl buffer, and 0.1 mM Na2 EDTA at pH 7; 2.7 mL of duplex solution in 1-cm quartz cell at a concentration of 5.8 μM duplex titrated with 10-μL

5′-CGCAAATTGGC-3′ 3′GCGTTTAACCG-5′ A 3T 2

aliquots of 0.31 mM netropsin (open circles) or 0.19 mM distamycin (solid circles). (Reproduced with permission from ACS, reference 13, Figure 4.)

1

2

3

ligand/DNA oligomer concentrations (Figure 20.4). These binding isotherms clearly show from the titration break points that binding saturation for each antibiotic is reached at a ratio of 1:1 for netropsin (open circles) and 2:1 for distamycin (solid circles). Furthermore, the slope for the linear binding curve of the second mole of distamycin is less steep than for the ﬁrst, suggesting a lower afﬁnity. This was conﬁrmed by titration calorimetry measurements which showed about a 10-fold lower afﬁnity for the second molecule of

642

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

drug (Kb ∼106 M−1 ) compared to that of the ﬁrst (∼107 M−1 ). Thus the results obtained by these CD measurements correspond well with the NMR and titration calorimetry ﬁndings. 20.3.1.2. Cellular Uptake and Target of Anticancer CC-1065 Prodrug Analogues in Live Tumor Cells. In a new and innovative application of CD, Tietze et al. [14] measured the ICD of live tumor cells to monitor the cellular uptake and DNA targeting of some CC-1065 prodrug analogues. This circumvented the need for radiolabeled drug or ﬂuorescent dyes to aid in the evaluation of the cytotoxicity of these new agents (Scheme 20.1a–c). CC-1065 (5, Scheme 20.1a) is a chiral and potent DNA alkylating agent isolated by Upjohn scientists [15] from fermentation broths of Stretomyces zelensis that targets the minor groove of DNA. Since its discovery, it has been the object of a great deal of structure–activity and DNA-binding studies by not only the Upjohn Company, but also by several academic groups [16]. Because the natural curvature is complementary to that of the DNA minor groove, it binds deep within the groove in a sequence speciﬁc manner with a strong preference for AT-rich regions, similar to netropsin and distamycin. Upon binding, alkylation takes place via the N3 of adenine by nucleophilic attack on the methylene carbon of the conjugated cyclopropyl group (Scheme 20.1b), resulting in a long-wavelength positive ICD at 370 nm, which is shifted from 390 nm for the noncovalently bound drug [5]. Even though CC-1065 is optically

N HCl

O

Cl H

(a)

(c)

NH2

N

N

2 O

N O

OH N

N H

O N H

OCH3

prodrug

(+) CC-1065 (5)

O

OH

OR R= HO

OH O OH

β-D-galactosidase

seco-drug R = H

(b)

N HCl

NH2

NH2

N

N

N

N

N H

N

N

R

H N

O O

O

N H

N

R

N O

O O

O

N H

drug

Scheme 20.1. Mechanism of CC-1065 cyclopropyl cleavage and prodrug activation.

D R U G A N D N AT U R A L P R O D U C T B I N D I N G T O N U C L E I C A C I D S A N A LY Z E D

active (α cyclopropyl ring), it exhibits only a weak CD in buffer. Yet when bound to the DNA minor groove, it exhibits the intense ICD, similar to the previously mentioned examples [17]. A number of relatively nontoxic glycosylated prodrugs were developed by Tietze’s group for antibody-directed enzyme prodrug therapy, the intent being that on incubation of the prodrugs with live cells the sugars would be removed by cellular glycohydrolases to yield the seco-drugs (Scheme 20.1c), which would then cyclize in situ to form the cytotoxic drugs. To monitor the drugs interaction with cellular DNA, the CD of the secodrugs was ﬁrst examined free in solution, then as a mixture with synthetic oligomers, and ﬁnally with live cells. Alone, the seco-drugs initially exhibit a weakly negative Cotton effect at 250 nm and weakly positive ones at 275, 320, and 390 nm. The 390 nm band is diagnostic for the seco-structure. However, over time the CD spectrum changes with the disappearance of the 390 nm band due to intramolecular spirocyclization forming a structure similar to CC-1065 itself. On mixing the seco-drugs with synthetic doublestranded DNA oligomers containing an AT-rich region, a negative ICD band appears at 305 nm along with a positive one at 335 nm, Cotton effects that are not observed for the free drug and DNA oligomers. These new ICDs are attributed to the covalently linked drug–oligomer complex and are stronger than those of the drugs without DNA, possibly due to restricted movement of the drug when bound to speciﬁc sequences. Incubation of the seco-drugs with serum-free, live tumor cell suspensions followed by measuring the CD of the drug-free cell suspension resulted in the appearance of a negative ICD at 305 nm and a positive one at 335 nm, essentially identical to those observed upon drug treatment of synthetic DNA oligomers. The ICDs are well-resolved from the overlapped signals due to cellular DNA and proteins. These ﬁndings indicate that the cellular DNA had reacted with the cyclized drugs formed in situ from the secodrugs. In addition, DNA-binding kinetics and rates of conversion of the seco-analogues to the active cyclopropyl compounds were obtained, thus aiding in the prioritization of the various prodrugs for further biological evaluation. This is the ﬁrst time that CD has been used to actually prove that cellular DNA is a target for supposed DNA-damaging drugs. Clearly, this methodology has great potential for other DNA–drug complexes that give well resolved ICDs. 20.3.1.3. Calicheamcin and Esperamicin. Calicheamicin γ1 I (6) [18] and esperamicin A1 (7) [19] are two members of a class of unusual natural products, enediynes, that exert their potent cytotoxicity by minor groove binding and cleaving of duplex DNA. The double-stranded cleavage chemistry is mediated by a thiol-induced Bergman cycloaromatization of the enediyne to give a biradical intermediate that abstracts proximal hydrogen atoms from the DNA deoxyribose sugars, thereby leading to strand cleavage [20]. A molecular basis for the minor groove binding of calicheamicin has been obtained by extensive NMR studies [21–24]. CD spectroscopic studies gave direct insight into the DNA conformational changes on binding of both calicheamicn and esperamcin and conﬁrmed suggestions for an “induced ﬁt” in the minor groove to account for the sequence speciﬁcity observed with DNA cleavage experiments [20]. These two enediyne containing antitumor agents are quite different structurally from the more classical minor groove binders such as netropsin and distamycin. With calicheamicin and esperamicin, binding to the minor groove is directed primarily by the relatively hydrophobic carbohydrate tail portion (deoxy and methoxylated sugars), the curvature of which is isohelical

643

644

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

O NHCOOCH3 CH3

HO

O

C H3

I

S

O

OCH3 OCH3

O H3C HO H3CO OH

H3CSSS

O

O H N HO

OH

H3C

O O

H

O

H CH3CH2 N

O

CH3O Calicheamicin γ1I (6)

OCH3 OH

H3CO

CH3 O

O

CH2

O NH

O O H HO

H3CO O H3C H3CS

O

H3CSSS O H N HO

OH

H3C

NHCOOCH3

O O

H

O H (CH3)2CHN

O

H3CO Esperamicin A1 (7)

with the minor groove of DNA [21, 22]. Because the CD signal attributed to a negative dienone–enediyne exciton couplet (negative band at 313 and a positive band at 272 nm) [25] overlaps with that of the DNA signal, the aromatic derivatives calicheamicin ε (8) [26] and esperamicin Z (9) [27] were used for DNA-binding studies because they are essentially CD silent in the DNA region. The glycosylated iodobenzoate in calicheamicin ε and the anthranilate–pyruvate amide in esperamicin Z, along with the aromatized enediynes common to both molecules, exihibit only weak CD Cotton effects between 210 and 220 nm [26–28]. DNA binding of both molecules results in a reorganization of the DNA secondary structure as evidenced by a signiﬁcant reduction in the intensity of the positive band at ∼280 nm and an increase in the intensity of the negative band at ∼230 nm in the CD spectrum. Figure 20.5 depicts the dehydrating effect of alcohols and esperamicin Z on a B-form DNA 13-mer. Calicheamicin ε binding with a DNA 12-mer resulted in a similar change in the CD spectrum. Such a change has been attributed to an increase in the winding angle and a decrease in the twist angle. This results in a slight decrease in the number of base pairs per helical turn [26, 27].

D R U G A N D N AT U R A L P R O D U C T B I N D I N G T O N U C L E I C A C I D S A N A LY Z E D

O NHCOOCH3 HO CH3O

I

OCH3 OH OCH3

O H3 C HO H3CO

CH3 O O

S

O

H H3C N HO O

H O

O

CH3 CH2 NH

OH

S O

H3CO

Calicheamicin ε (8)

OCH3

OH

H3 CO CH2 H3CO

CH3 O

O NH

O

O

O

O

NHCOOCH3

H HO

H3C

O

H3CS OH

O H N HO

H3C

S O O

H

O

O

(CH3)2CHN H3 CO

Esperamicin Z (9)

20.3.1.4. Chromomycin. Chromomycin A3 (10, panel A, Figure 20.6) is a wellstudied antitumor antibiotic of the aureolic acid class isolated from fermentations of Streptomyces griseus and Streptomyces plicatus and whose biological target is DNA [29]. Much information about the chromomycin binding to DNA and the role of Mg+2 has been obtained by a number of biophysical methods including CD [5, 30, 31] and particularly NMR [32]. NMR studies showed that chromomycin targets the minor groove of GC-rich regions of DNA as a 2:1 chromomycin–Mg+2 complex at pH 7.4. This is another example of a minor groove-binding agent that does not contain a cationic charge. But as seen in the structure the molecule has an elongated shape in common with more typical minor groove binders with the tri- and disaccharides attached at opposite ends of the planar tricyclic chromophore. In 2004, Wang’s group reported the ﬁrst crystal structure of a chromomycin–Mg+2 –DNA complex in which the antibiotic is found as a [(Chro)2 Mg+2 ] complex imbedded in the minor groove of a DNA oligomer, d(TTGGCCAA)2 [33]. These authors also used CD spectroscopy, as had Dasgupta’s group earlier [30, 31], to monitor the stoichiometry of complexation of chromomycin with Mg+2 with and without DNA, as well as showing interesting ICD spectral changes on [(chro)2 Mg+2 ] binding to the DNA oligomer. We start by describing the CD of chromomycin without Mg+2 , then chromomycin plus Mg+2 , and ﬁnally chromomycin plus Mg+2 in the presence of duplex DNA. These

645

646

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

CD (mdeg)

5.00

10% MeOH 50% EtOH

Figure 20.5. The effect of 10% methanol,

80% glycol

50% ethanol, and 80% ethylene glycol and esperamicin Z on the CD spectrum of 5 -CAGGACGCGTCCT (2.9 μM) with 8 μM

8 μM EPMZ –4.00 220.0

320.0 Wavelength (nm)

esperamicin Z (20 mM Tris-NaCl, pH 7.5, and 0.1 M NaCl). (Reproduced with permission from ACS, reference 27, Figure 7B.)

CD studies are summarized in Figures 20.6b, 20.6c, and 20.d. The CD spectrum of chromomycin alone (40 μM) without Mg+2 at pH 7.3 in aqueous buffer shows a pronounced negative couplet centered at ∼285 nm with a ﬁrst negative Cotton effect at 300 nm followed by a positive one at ∼275 nm (solid line in Figure 20.6b). This negative couplet appears to be the result of intermolecular coupling of the 1 Bb long-axis polarized transitions of naphthaleneone chromophores in a dimer or larger aggregate in which the chromophores are in a left-handed relationship [33, 34]. As seen in the titration spectra in Figure 20.6b, addition of up to 8 mM Mg+2 results in a signiﬁcant decrease in the intensity of the Cotton effects at 413 and 400 nm in the visible region (1 La short-axis polarized transition) along with the disappearance of the above-mentioned negative couplet (negative band at 300 nm, positive band at 275 nm). Instead, only a weak negative Cotton effect is now apparent at ∼275 nm, which reﬂects a signiﬁcant change in the chromophoric relationship in [(Chro)2 Mg+2 ] compared to the left-handed aggregated form without Mg+2 . Upon addition of Mg+2 to a chromomycin solution in the presence of d(TTGGCCAA)2 , binding of the drug–Mg+2 complex to the oligomer takes place as evidenced by a strong new positive CD band in the DNA region at 287 nm and a strong negative one at 275 nm, indicative of the conversion of the DNA oligomer to a modiﬁed B form. A very broad and weak positive Cotton effect between 400 and 475 nm is also observed. Figure 20.6c shows the spectra of the free drug (40 μM, solid line) followed by the addition of up to 8 mM Mg+2 in the presence of 20 μM of d(TTGGCCAA)2 at pH 7.3. Figure 20.6d illustrates the CD difference spectrum between 200 and 350 nm of free d(TTGGCCAA)2 (solid line) with that of [(Chro)2 Mg+2 ]DNA minus that of [(Chro)2 Mg+2 ] (dashed line). The spectrum of [(Chro)2 Mg+2 ] was subtracted from

D R U G A N D N AT U R A L P R O D U C T B I N D I N G T O N U C L E I C A C I D S A N A LY Z E D

(a)

(c)

(b)

(d)

Figure 20.6. (a) Structure of chromomycin A3. (b) CD spectrum chromomycin (40 μM as monomer) showing the effect of Mg 2+ in the presence of 0–8 mM Mg2+ . (c) The effect of Mg2+ on binding of chromomycin to d(TTGGCCAA)2 ; Mg2+ concentration varied from 0 to 8 mM and the drug concentration as monomer was 40 μM. (d) CD comparison of d(TTGGCCAA)2 (solid Line) with the CD of [Chro)2-Mg2+ ] - d(TTGGCCAA)2 minus the signal due to [Chro)2-Mg2+ ] (dashed line) in Tris—HCl (pH 7.3) with 100 mM NaCl at room temperature. (Reproduced with permission from Oxford University Press, reference 33, Figure 1.)

that of the ChroDNA complex in order to minimize any CD interference from [(Chro)2 Mg+2 ]. This is valid because the relationship between the two chromophores in the [(Chro)2 Mg+2 ] complex is not signiﬁcantly changed from that when bound to the DNA [33]. The appearance of the strong, drug-induced positive CD band at 287 along with the strong negative band at 275 nm, compared with the CD spectrum of the free B-form d(TTGGCCAA)2 , indicates that DNA binding of [(Chro)2 Mg+2 ] has altered the DNA conformation from that of the free oligomer. This is conﬁrmed by the X-ray crystal structure of the drug–DNA complex. Although the drug-bound DNA maintains a right-handed helical B-form structure, it is kinked by ∼33◦ with a signiﬁcantly widened minor groove at the targeted GGCC sequence. The strongly positive and conservative exciton couplet at ∼260 nm (positive Cotton effect at 287 nm and a negative one at 275 nm), arising from formation of the [(Chro)2 Mg+2 ]DNA complex, can be used as a reliable sensor for determining binding afﬁnities between the drug and various DNA oligomers. 20.3.1.5. Actinomycin D. The DNA binding of the potent antitumor antibiotic actinomycin D (4) is an instructive example of both intercalation and minor groove binding, and CD played a role in providing evidence for a drug-induced DNA conformational

647

648

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

change. Initial modeling and X-ray crystal structures of the drug–DNA complex indicated that the ﬂat phenoxazone ring intercalates at GpC base-pair steps and the two cyclic pentapeptide moieties bind in the minor groove [35, 36]. Free actinomycin D exhibits strong UV absorption at 240 and 445 nm in ethanol [37], and the CD spectrum [38, 39] (data not shown) exhibits negative Cotton effects at ∼265 and 372 nm and a weaker negative effect at ∼445 nm when recorded in aqueous buffer at a concentration of 20 μM. In addition, there is a positive extremum at ∼245 nm and a weaker one at 296 nm. At a higher concentration of 140 μM, the spectrum becomes more complex with increased intensities of the original peaks but with some new positive and negative Cotton effects in the visible region [39]. This change in the CD spectrum at high concentration of drug is attributed to self-aggregation and is a phenomenon that needs to be considered when monitoring drug-DNA interactions using CD. The long-wavelength absorption was used in the early CD DNA binding studies because it is well removed from the DNA CD region. A binding study with ctDNA [38] showed an increase in intensity of the two intrinsic long-wavelength negative CD bands at ∼375 nm and ∼450 nm (the antibiotic is chiral due to the two peptide lactone appendages) upon addition of actinomycin to DNA. But these ﬁndings by themselves did not give much information as to the nature of the binding other than to show that there is binding. This and other subsequent actinomycin–DNA binding studies have been reviewed by Zimmer and Luck [5]. Because the phenoxazone chromophore of actinomycin intercalates between GpC base pairs and the cyclic pentapeptide lactone moieties bind in the DNA minor groove, interpretation of the observed CD changes is difﬁcult without an X-ray crystal structure of the drug–DNA complex or an NMR analysis to provide a solid structural basis for the binding. Large numbers of CTG triplet repeats have been observed in genes associated with neurological diseases, and actinomycin has been shown to stabilize such repeats by tight binding and thus prevent their expansion. A recent CD and X-ray analysis of a actinomycin–ATGCTGCAT complex, which contains a CTG triplet sequence, showed that the drug induces an A-like conformation (Figure 20.1) on the oligomer [40]. The same group had previously shown by NMR that a T:T mismatch adjacent to a GpC step is an excellent binding site for actinomycin [41]. CD studies with this mismatched oligomer, TT1, along with two other canonical sequences, AT0 and AT1, shown in Figure 20.7, showed changes in the DNA region as a result of drug binding. The top panel shows the CD spectra for the oligomers alone and the bottom panel shows the spectra of the drug–DNA complexes. Compared with the CD of the oligomers alone, the CD spectra of the TT1–and the canonical AT0–and AT1–actinomycin complexes show a signiﬁcant red shift of the positive CD peak from 275 to 290 nm as well as sharpening and increased intensities of both positive and negative DNA peaks (bottom panel, Figure 20.7). The CD spectrum of the free actinomycin was subtracted from that of the drug–DNA complex in order to remove any contribution of unbound actinomycin. Thus the suggested A-like conformation (Figure 20.1) of the drug-induced DNA oligomers based on the CD study is consistent with the crystal structure of the TT1–drug complex, which showed that actinomycin caused a kink in the helix, a widened minor groove, and unwinding of the base pairs. The crystal structure showed that intercalation of the phenoxazone chromophore occurs between both GpC base-pair steps of the TT1 oligomer, even with a ﬂanking T:T base-pair mismatch [40]. Furthermore, it showed that the cyclic peptides of the drug are embedded in the DNA minor groove, consistent with the earlier crystal structures with other canonical sequences.

649

D R U G A N D N AT U R A L P R O D U C T B I N D I N G T O N U C L E I C A C I D S A N A LY Z E D

1 TT1 AT0 AT1 0 [θ]

DNA

–1

CTG CAT TT1: ATG TAC GTC GAT

–2 200

220

240

260

280

300

320

Wavelength (nm)

CG CAT AT0: ATG TAC GC GTA

2

CATG CAT AT1: ATG TAC GTAC GTA

1

[θ]

0 DNA + ActD

–1 –2 –3 –4 200

220

240

260 280 Wavelength (nm)

300

320

Figure 20.7. (Left) DNA duplexes used in the CD studies. The rectangles inserted in the duplexes represent binding sites of the phenoxazone ring of the drug. (top right) CD spectra of TT1, AT0, and AT1 duplexes (4 μM) in standard buffer: alone (top right) and with 10 μM drug (bottom right). The CD spectra of the complexes were obtained by subtracting the spectrum of free actinomycin D. (Reproduced with permission from Oxford University Press, reference 40, Figure 1.)

20.3.1.6. Enantio-Selective Daunorubicin Binding to DNA. For some time there has been much interest in identifying molecules that speciﬁcally target a speciﬁc DNA structure to inhibit the biological function in which these structures play an important role. Most of these efforts have involved enantiomers of chiral metal complexes and have met with mixed success [3, 42]. A striking example of targeting right- and left-handed DNA with relatively high speciﬁcity has now been reported by Chaires and co-workers [43] for the (+) and (−) enantiomers of the clinically useful anthracycline antitumor agent daunorubicin (3) (daunomycin). Initial studies of the CD changes on (+) duanorubicin (natural occurring enantiomer) binding to DNA have been reviewed previously [5]. A crystal structure showed that the positively charged amino sugar, daunosamine, binds in the minor groove and the anthraquinone chromophore intercalates between GC base pairs [44]. Selectivity studies for the B and Z forms using the

650

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

daunomycin (+) and (−) enantiomers were carried out by examining the CD of the dialysate for enrichment of either enantiomer after dialysis of a racemic mixture against either B or Z form [poly(dGdC)]2 . CD spectra of the (+) enantiomer (curve a), a 1:1 mixture of the two enantiomers (curve b) and the (−) enantiomer (curve c), are shown in Figure 20.8a. CD spectral analyses indicated the enrichment of the (−) enantiomer (curve a in Figure 20.8b) in the dialysate when a racemic mixture of the enantiomers was dialyzed against B form [poly(dGdC)2 ] in 0.2 M NaCl. This enrichment of the (−) enantiomer in the dialysate indicates the preferential binding of the (+)enantiomer to the right-handed B form of this sequence. The (+) enantiomer was clearly enriched in the dialysate from binding of the 1:1 enantiomeric mixture to Z-form DNA in 3.0 M

(a) 2 × 104 a 1×

104 b 0

– 1 × 104

Molar ellipticity

c – 2 × 104

2 × 104

(b)

1 × 104

b

0

– 1 × 104

a

– 2 × 104 300

400 500 Wavelength (nm)

600

Figure 20.8. (a) CD spectra of (+)daunorubicin (curve a) and its (−)enantiomer (curve c); curve b, a 1:1 molar ratio mixture of the two enantiomers. (b) Curve a, CD of the dialysate after dialysis of the racemic mixture against 100 μM B-form [poly(dGdC)]2 in 0.2 M NaCl. Curve b, CD of the dialysis of the mixture against 100 μM Z-form [poly(dGdC)]2 in 3.0 M NaCl. (Reproduced with permission from PNAS, reference 43, Figure 2.)

D R U G A N D N AT U R A L P R O D U C T B I N D I N G T O N U C L E I C A C I D S A N A LY Z E D

651

NaCl (curve b in Figure 20.8b), indicating a preference of the (−) enantiomer for the left-handed Z form. In another qualitative experiment, Figure 20.9 provides CD evidence for the ability of both enantiomers to allosterically bind to their preferred DNA structures upon their addition to a 50:50 mixed population of B and Z form [poly(dGdC)]2 (100 μM in base pairs) using a salt concentration of 2.25 M. Curve a depicts the CD of the DNA alone and is characteristic of the mixture of B and Z forms resulting from the high concentration of salt. Curve b is the CD after the addition of the (+) enantiomer at a ﬁnal concentration of 2 μM showing conversion of the B- and Z-form mixture to a preponderance of the B form. And curve c is the CD after the addition of 3 μM of the (−) enantiomcer indicating conversion to the Z form. Quantitative binding isotherms obtained from ﬂuorescence titration experiments showed that binding of (−) daunorubicin allosterically converts the right-handed B conformation to the Z form in a positively cooperative manner. In an earlier ﬂuorescence titration experiment, the opposite had been observed with the (+) enantiomer when it was titrated into Z-form DNA [45]. These interconversions are driven by the different ratio of binding constants for the B and Z forms of DNA: 44 for (+) daunorubicin and 5.2 for the (−) enantiomer. Noting that the binding ratios for the two enantiomers are signiﬁcantly different, the authors point out that the B and Z forms of DNA are not mirror images (see Figure 20.1) and that a high salt concentration was necessary to maintain the [poly(dGdC)]2 duplex in the Z conformation. The high salt no doubt inﬂuences in a negative way the binding of the positively charged daunorubicin. In any event, the CD spectral changes clearly reﬂect these conformational changes on binding of the two enantiomers.

1.4 × 104

7.0 × 103

Molar ellipticity

b 0.0

a

Figure 20.9. CD spectra of [poly(dGdC)2 ] at

– 7.0 × 103

100 μM (in base pairs) in buffered 2.25 M NaCl. Curve a is the spectrum of a 50:50 mixed population of B- and Z-form DNA. Curve b is

c

the CD of the DNA with (+)daunorubicin added to a ﬁnal concentration of 2 μM. The molar ratio of the drug:bp is 0.02. Curve c is the

– 1.4 × 104

CD of the DNA with added (−) daunorubicin to a ﬁnal concentration of 3 μM. with a drug:bp

– 2.1 × 104 240

260

280

300

Wavelength (nm)

320

340

molar ratio of 0.03. (Reprinted with (Reproduced with permission from, PNAS, reference 43 Figure 3.)

652

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

20.4. ILLUSTRATIVE EXAMPLES OF THE SELECTIVE BINDING OF PORPHYRINS TO B- AND Z-FORM DNA Circular dichroism has been used for many years to characterize noncovalent, achiral porphyrin and metalloporphyrin binding to DNA. This topic has been studied extensively by Pasternak and co-workers, and the reader is referred to a 2003 minireview by Pasternak for the general spectroscopic features that give direct insight into the binding mode of these ligands [46]. The reader is also referred to three recent porphyrin DNA binding studies that made use of CD and will be helpful in interpreting CD spectral changes. Lee et al. [47] used CD to study the binding of a tetracationicporphyrin to triplex DNA. Nitta and Kuroda [48] carried out a quantitative analysis of the multiple modes of binding, including major and minor groove and outside binding, of a Mn(III) tetracationic porphyrin with several synthetic and natural DNAs. And Shelton et al. [49] described DNA binding studies of unmetallated, copper(II), and zinc(II) dicationic porphyrins using CD along with other spectroscopies. A brief overview will be given here before going into some new applications that include: (a) porphyrins as spectroscopic sensors for DNA conformational transitions, and (b) porphyrin binding to DNA quadruplexes. Based on their own ﬁndings and those of others, Pasternak and co-workers proposed that the sign and shape of an ICD signal in the Soret region provides a useful sensor for determining the binding mode of achiral porphyrins to DNA. In general, for natural DNAs at low drug-to-DNA ratios of less than 0.1 (calculated in base pairs), intercalation results in a negative ICD in the Soret region with a ε of about −25 M−1 cm−1 . In contrast, a positive ICD with a slightly larger ε of about +35 M−1 cm−1 is suggestive of groove-binding porphyrins. Outside stacking yields bi- or multisignate ICDs [46–50]. The ICD signal proﬁle was found to vary mainly as a function of the magnitude and spatial relationship of the transition dipoles of the bound porphyrin relative to those of the DNA bases. Porphyrins that contain axially liganded metals such as Zn(II), Mn(III), and Fe(III) do not intercalate because of steric hindrance and bind either to one of the DNA grooves, leading to a positive ICD, or do not bind at all resulting in no ICD in the Soret region.

20.4.1. A Cationic Zinc Porphyrin as a Probe for Z-DNA A zinc derivative of meso-tetrakis(4-N -methylpyridyl)porphyrin (11) has been found to be an extremely sensitive chiroptical probe for the left handed Z form of DNA [50]. Although the biological relevance of Z-form DNA is still somewhat uncertain, a recent X-ray structure has provided molecular details about its structure and the molecular basis for its binding to some DNA regulatory proteins [51]. Figure 20.10 shows diagnostic ICD signals in the Soret region of porphyrin B- and Z-DNA poly(dGdC)2 complexes formed as a function of added porphyrin. Only a very small ICD signal is apparent for the B-form DNA (Figure 20.10, curve d) complex in the Soret region in the presence of added Zn-porphyrin, whereas an intense, negative, bisignate proﬁle in the Soret region with Z-DNA begins to develop at concentrations as low as 1 μM of added Zn-porphyrin (Figure 20.10, curves a, b, and c with 1, 4, and 8 μM of porphyrin, respectively), indicative of outside binding to the left-handed duplex of Z DNA. An intercalative binding mode is precluded by the fact that Zn-porphyrin is pentacoordinate, in which the zinc is axially liganded by a water molecule. The sharp and negative bisignate proﬁle of the ICD suggests through-space, exciton coupling between the bound achiral porphyrins in which the bound porphyrins form a left-handed screw-like relationship on the chiral DNA.

D R U G A N D N AT U R A L P R O D U C T B I N D I N G T O N U C L E I C A C I D S A N A LY Z E D

653

10 d CD (mdeg)

0 a

c b

b

a –10

c

Figure 20.10. CD titration spectra of Z-DNA –20

(50 μM) in the presence of 1 μM (a), 4 μM (b), and 8 μM (c) of zinc porphyrin (11), and B-DNA (50 μM) in the presence of 1 μM of zinc porphyrin

250

300

350 400 λ (nm)

450

500

(11) (d). (Reproduced with permission from Wiley-Verlag, reference 50, Figure 2.)

CH3 N

N

N

NCH3

Zn

H3CN N

N

N CH3 Meso-tetrakis(4-N-methylpyridyl)porphyrin (11)

As seen in the ﬁgure, the ICD changes in the Soret region are well removed from signals in the DNA region between 240 and 280 nm, which allows for a straightforwar interpretation. In the DNA region the preference for binding of the porphyrin to the left-handed Z form is obvious as the signal intensity attributed to this DNA secondary structure increases with increasing concentrations of the ligand. The binding was also characterized using absorption, ﬂuorescence and resonance light scattering (RLS) spectroscopies. Both the ICD and RLS results suggest that the bound porphyrins are dispersed randomly on the B form DNA and thus not involved in interporphyrin coupling. In contrast, this metallated porphyrin apparently binds to the outside of Z-form DNA in an ordered distribution in which the N7 of guanine is more accessible than in B-form DNA. And it is this ordered distribution that permits interporphyrin coupling to give rise to the sharp bisignate CD pattern with Z form DNA.

654

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The neighboring porphyrins, however, are not in such close π –π contact that would cause quenching of the ﬂuorescence and broadening of the ICD signals.

20.4.2. Recognition of a Spermine-Induced Z-form DNA by a Tetraanionic Nickel Porphyrin A 2009 publication describes the recognition of a spermine-induced Z-form DNA conformation by a tetraanionic nickel (II) meso-tetrakis(4-sulfonatophenyl)porphyrin, NiTPPS (12), as deduced by CD (Figure 20.11) [52]. Spermine is predominantly tetracationic at physiological pH and is a micromolar inducer of Z-form DNA in poly(dGdC)2 . It is present in millimolar concentrations in eukaryotic cells. No ICDs in the Soret region of the CD spectrum were observed when the porphyrin was added to a right-handed poly(dGdC)2 oligomer in the absence of spermine indicating no binding (Figure 20.11). In contrast, in the presence of spermine and a pH of 6.8 an intense negative exciton couplet developed in the Soret region with an intensity increase as a function of added porphyrin reﬂecting binding to left-handed Z-DNA. This was accompanied by the inversion of the right-handed B form secondary structure to that typical of the left-handed duplex form in the DNA region of the spectrum. Similar to the above cationic Zn-porphyrin, this result indicates selectivity for outside binding to the left-handed Z-form, no doubt due to the accessibility of the nucleophilic N7 of guanine for coordination to the nickel-metallated porphyrin. Again an intercalative binding mode is precluded by the axially liganded nickel. A CD titration carried out at pH 6.8 indicated that saturation of the ICD signal occurred at ∼9 μM of added porphyrin per base pair, indicating the presence of no more than two molecules of porphyrin for each DNA helical turn in this construct. The neighboring porphyrin–porphyrin electronic interaction accounts for the observed ICD. In addition, it was observed that raising the pH to 8.5 (triprotonated spermine) (Figure 20.12, compare curves 1 and 2) resulted in disappearance of the Soret ICD band. Lowering the pH back to 6.8 restored the Soret signal (Figure 20.12, curve 3, cycle 1). No major change in the CD signal in the DNA region was observed between pH 6.8 and 8.5. This indicates that the Z-form conformation was maintained throughout these pH adjustments. Raising the pH to 9.5 from 6.8, however, caused not only the loss of the Soret ICD signal but also reversion of the high-energy Z form back to the B form (Figure 20.12, curve 4). Acidiﬁcation again to pH 6.8 (cycle 2) resulted in the restoration (less than an hour) of the Z-form DNA as indicated by the negative bisignate Soret ICD 20

–

15

12 + Z DNA

10 CD (mdeg)

SO3–

O3S N N Ni N

12 + B DNA

N

5

12

–

O3S

SO3–

0 –5

Figure 20.11. CD spectrum of 4 μM NiTPPS (12) in the presence of a left-handed or Z-conformation of 50 μM of poly(dGdC)2 and a

12 + Z DNA

right-handed or B conformation of 50 μM of poly(dGdC)2 . Conditions: spermine (14 μM); NaCl

–10 –15 250

300

350

400

450

Wavelength (nm)

500

550

(10 mM); cacodylate buffer (1 mM); pH 6.8; room temperature. (Reproduced with permission from ACS, reference 52, Figure 1.)

655

D R U G A N D N AT U R A L P R O D U C T B I N D I N G T O N U C L E I C A C I D S A N A LY Z E D

15 10

(5)

5

(3)

(4)

0

(2) (3) (5) (1)

–10 –15

6.8

250

300

400 350 Wavelength (nm)

B-DNA

NITPPS (411 nm) DNA (292 nm)

9 6 3 0 –3 –6 –9 –12

Z-DNA

–5 CD (mdeg)

CD (mdeg)

(1) ph = 6.8 (2) ph = 8.5 (3) ph = 6.8 (cycle 1) (4) ph = 9.5 (5) ph = 6.8 (cycle 2)

(1)

450

9.5

6.8

9.5

500

6.8 pH

550

Figure 20.12. CD titration spectrum of 4 μM of NiTPPS 1 in the presence of 50 μM poly(dGdC)2 at pH 6.8 and after sequential pH variations. Conditions: spermine (14 μM); NaCl (10 mM); 1 mM cacodylate buffer; room temperature. Inset: Reversible CD signal changes at 48◦ C of the porphyrin–poly(dGdC)2 –spermine complex at different pHs recorded at 411 nm (closed circles, Soret region) and 292 nm (open circles, DNA region). (Reproduced with permission from ACS, reference 52, Figure 3.)

Figure 20.13. Summary of modulation of porphyrin–poly(dGdC)2 –spermine complex as a function of pH. (Reproduced with permission from ACS, reference 52, Figure 2.)

signal shown by curve 5 in Figure 20.12. The proposed binding relationships between the DNA, spermine and porphyrin as a function of pH are summarized in Figure 20.13. This study shows once again the ability of CD to clearly monitor the discrimination of an anionic sulfonated nickel (II) porphyrin between left- and right-handed DNA secondary structures.

656

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

20.5. ILLUSTRATIVE EXAMPLES OF THE SELECTIVE BINDING OF PORPHYRINS AND ALKALOIDS TO G-QUADRUPLEX DNA 20.5.1. CD Analyses of the Binding Selectivity of Cationic Porphyrins to a DNA Quadruplex CD spectroscopy has been used to monitor porphyrin binding to G-quadruplexes [53], that is, DNA structures formed from telomeres that provide stability to the end of chromosomes by inhibiting telomerase-catalyzed chain extension. The 3 -single-stranded, guanine-rich overhang of telomeres can self-assemble to form a G-quadruplex or tetrad consisting of a planar arrangement of four guanine bases assembled through hydrogen bonds (middle structure in Figure 20.14). These G-quadruplexes can then assemble into four-stranded helical structures by stacking on top of one another to form helical assemblies that are stabilized by cations, predominantly K+ , and π –π hydrophobic stacking interactions between the guanines [54, 55]. Intra- and intermolecular G-quadruplexes can be formed with either parallel or antiparallel sugar-phosphate backbones and are illustrated in Figure 20.14 [4]. Their structures have been characterized by NMR [55], X-ray [56, 57] and CD [3, 4, 58–60]. In Figures 20.14a and 20.14b the guanine bases are depicted as triangles and point in the

parallel

anti-parallel H

20

N

N

T H

N

H O

N H H

H N

O

N

Δε [M–1 cm–1]

N

G-tetrad H

H N

H

O

N

N N

T

N

O

H

10

T

T

N

N

N

15

N

N

H

N

N

T

N

T

T

H

T

5

0

–5 (a) –10 200

(b) 0′

10′ 240

3.5h 280

24h

0.001

320 220 Wavelength [nm]

0.01 260

0.1

0.5 M 300

Figure 20.14. Structures and CD spectra of parallel and antiparallel G-quadruplexes. A G-tetrad is shown in the middle. (a) Time-dependent formation of a parallel-stranded quadruplex of d(G4 ) stabilized by 16 mM K+ . (b) Na+ -induced anti-parallel bi-molecular quadruplex of d(G4 T4 G4 ). Both oligonucleotides were dissolved in 1 mM Na phosphate, 0.3 mM EDTA, pH 7 and thermally denatured (5 min at 90◦ C) and slowly cooled before starting spectral measurements. (Reproduced with permission from Nucleic Acids Research, reference 4, Figure 4.)

D R U G A N D N AT U R A L P R O D U C T B I N D I N G T O N U C L E I C A C I D S A N A LY Z E D

5 –3 direction of the appended ribose sugars. In the parallel quadruplex (Figure 20.14a) all the guanines are in an anti conformation with respect to the glycosyl bonds. In the antiparallel quadruplex (Figure 20.14b) the guanines exist in alternating anti and syn relationships with respect to the glycosyl bonds in each strand. The sugar-phosphate moieties connecting the guanines at each corner of the quadruplexes are shown as lines for clarity and the appended sugars are not shown, also for clarity [4]. CD in particular, has been found to be a sensitive diagnostic tool to characterize various G-quadruplex structures. Belmont et al. [3], Kypr et. al. [4], and Gray et al. [60] give experimental and theoretical evidence that the CD spectral differences between these two structures are related to either parallel or antiparallel polarities (direction of the hydrogen bonding donors and acceptors) of the adjacent quartets. Parallel G-quadruplexes form a righthanded helix in which all the connecting sugar-phosphate moieties between the stacked quartets have the same 5 to 3 orientation (Figure 20.14a). In general, but not always, these structures exhibit a positive CD signal between 260 and 265 nm and a negative signal between 240 and 245 nm. Antiparallel quadruplexes formed by two folded-back sugar-phosphate strands connecting the stacked quartets also form a right-handed helix and exhibit a characteristic positive signal at ∼295 nm and a negative signal between 260 and 265 nm (Figure 20.14b) along with a small positive Cotton effect at ∼245 nm. Planar small molecules that can bind to the external quartet of a quadruplex and thereby stabilize these structures and inhibit telomere extension have potential for the treatment of cancer. Sanders and co-workers [53] used CD spectroscopy to determine if the cationic porphyrins A2 trans, A3 , and A2 cis were selective binders for parallel- or antiparallelG-quadruplex structures. For substrates they used a human telomeric DNA sequence, 5 -GGATTGGGATTGGGATTGGGATTGGG-3 (Htelo), that can exist as a mixture of intramolecular parallel and antiparallel conformations after annealing in KCl buffer at pH 7. Figure 20.15 shows the CD spectra of a 10 mM solution of the quadruplex upon addition of the three porphyrins. The CD of the mixture of parallel and antiparallel Gquadrulexes is depicted by the solid line (Htelo) in Figure 20.15a which shows two, partially overlapped Cotton effects: one at ∼270 nm for the parallel structure and one at ∼290 nm for the antiparallel structure. Based on the appearance of the characteristic positive signal at ∼290 nm for the antiparallel G-quadruplex structure and the absence of the strong positive Cotton effect at ∼270 nm (parallel G-quadruplex), it was concluded that A2 trans (short dash–dotted line in Figure 20.15) and A3 (dotted line) porphyrins preferred an antiparallel conformation. The A2 cis isomer (dashed line), however, did not show a speciﬁc preference for either of the two G-quadruplex conformations, which was suggested by the lack of a signiﬁcant change in the quadruplex CD spectra. The authors also used CD to see if the above two mentioned porphyrins could induce G-quadruplex formation from nonannealed Htelo DNA in a K+ free buffer (solid line in Figure 20.16). Only the A2 trans porphyrin induced this transition (dashed and dotted lines in Figure 20.16) as indicated by the suppression of the 260 nm positive signal of the nonannealed Htelo DNA in favor of the positive extremum at 290 nm. Furthermore, the presence of a clear isoelliptic point around 270 nm during the titration gives a strong indication of the transition between the two G-quadruplex conformations upon addition of the A2 trans porphyrin. Thus CD, along with surface plasmon resonance, provided compelling evidence that two meso 4-N -methylpyridinium substituents in a trans relationship on the poprphyrin macrocycle are capable of binding to an antiparallel, intramolecular, human telomeric quadruplex.

657

658

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

CD (mdeg)

240

260

280

300

320

3,5

3,5

3,0

3,0

2,5

Htelo

2,5

2,0

A2 cis

2,0

1,5

A3

1,5

1,0

A2 trans

1,0

Figure 20.15. CD spectra of 10 mM of Htelo

0,5

0,5

0,0

0,0

–0,5

–0,5

–1,0

quadruplex (solid line) in the presence of: A2 trans (short dash–dotted line); A3 (dotted line); and A2 cis (dashed line). Measurements at 20◦ C

–1,0 240

260

280

300

in 50 mM Tris–HCl, pH 7.4, 150 mM KCl. (Reproduced with permission from The Royal Society of Chemistry, reference 53, Figure 3.)

320

λ (nm)

CH3

CH3 N

NH

N

NH

N N CH3

N

N

HN

N HN

A2 cis (13)

N CH3 CH3 N

NH

N

N

HN

A2 trans (14)

N CH3

N CH3 A3 (15)

20.5.2. CD Analyses of Berberine Binding to a DNA Quadruplex Berberine (16) is a ﬂat and positively charged alkaloid originating from Chinese herbal medicine with antibiotic activity against a number of organisms. Berberine has recently been shown to exhibit antitumor activity by the inhibition of telomerase and binding to

D R U G A N D N AT U R A L P R O D U C T B I N D I N G T O N U C L E I C A C I D S A N A LY Z E D

220

240

260

280

3,0

CD (mdeg)

2,5

No

300

320 3,0

12 eq. A2 trans

K+

2,5

2,0

2,0

1,5

1,5

1,0

2.2 eq. A2 trans

0,5

1,0 0,5

0,0

0,0

–0,5

–0,5 Isoelliptic point

–1,0

–1,0

–1,5 –2,0 220

–1,5 –2,0 240

260

280 λ (nm)

300

659

320

Figure 20.16. CD spectra of 10 μM of nonannealed Htelo DNA in a 50 mM Tris–HCl, pH 7.4 buffer with no K+ (solid line); in the presence of A2 trans: 2.2 equiv. (dashed line) and 12 equiv. (dotted line) (Reprinted with permission from The Royal Society of Chemistry, reference 53, Figure 4.)

G-quadruplex DNA [61]. A recent study by Zhang et al. [62] used CD in addition to ﬂuorescence spectroscopy, polymerase chain reaction inhibition, and competition dialysis, to determine the binding selectivity of berberine and some derivatives to a G-rich human telomeric sequence, telo21 (5 -GGGTTAGGGTTAGGGTTAGGG-3 ). As mentioned in Section 20.5.1, formation of G-quadruplexes inhibits telomerase, the enzyme responsible for elongation of telomeres in tumor cells. Oligomer telo21 can form different types of Gquadruplex structures, depending on incubation conditions, and thus is a useful substrate for selectivity studies with different small molecules for the identiﬁcation of antitumor drugs. The CD spectra shown in Figure 20.17 reﬂect the conformational changes in human telomeric telo21 upon addition of Na+ , K+ , berberine, and a 9-O-substituted piperidino derivative. The CD curve (Figure 20.17, open triangles) of 5 μM telo21 in a buffered pH 7.4 solution without cations or drugs reveals a positive peak at ∼258 nm, a weak negative peak at ∼272 nm, and a shallow positive one at 298 nm. In the presence of 0.1 M Na+ , however, a spectrum characteristic of an antiparallel quadruplex (solid triangles) is observed with a new strong positive band at 295 nm along with a negative peak at ∼265 nm. And in the presence of 0.1 M K+ , a spectrum reﬂecting a mixture of parallel and antiparallel G-quadruplexes is observed with a strong positive band at 291 nm and a positive shoulder near 270 nm along with a negative band at 235 nm (dots). The addition of 10 μM each of berberine (upside down solid triangles) and a 9-O-alkylpiperidino derivative (17) (open circles) resulted in a major positive peak at ∼295 nm and a negative one at ∼262 nm, characteristic CD features of an antiparallel G-quadruplex structure similar to that induced by Na+ . Thus CD provided a simple and sensitive spectroscopic method for showing that berberine and a 9-substituted piperidino derivative selectively induce and bind to an antiparallel G-quadruplex structure using the telo21 oligomer as a substrate. In a related study, Ma et al. [63] used CD, along with other biological and biophysical techniques, to examine the binding of a 9-N -substituted alkylamino berberine (18) (see structures) to a another biologically important G-rich sequence, in this case a truncated sequence, Pu18 (AGGGTGGGGAGGGTGGGG), of the hypersensitive element III1 sequence (NHE III1 ) contained in the c-myc gene. This new study is another illustration of the sensitivity of CD for monitoring the selective induction and binding

660

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

4

5

O O

N 1

Cl– 7 OCH3

13 9 12

OCH3

Berberine (16) O O

N

Cl– O

N

OCH3

9-O-Alkylpiperidino derivative (17) O O

N

Cl– H N

NH2

OCH3

9-N-Alkylamino derivative (18)

to G-quadruplexes, in this case, to an intramolecular parallel G-quadruplex structure. Figure 20.18 depicts the conformational change induced by the 9-N -substituted alkylamino derivative (18) on Pu18 as monitored by CD. A strong positive peak at ∼275 nm is observed for 10 mM of Pu18 alone in a pH 7.2 buffered solution (solid line). The addition of 25 μM of the 9-N -substituted alkylamino derivative, however, caused a signiﬁcant downﬁeld shift of the positive band at 275 nm to ∼260 nm along with the appearance of a weak negative band at 240 nm (open circles). This new CD spectrum is similar to the CD spectra of Pu18 in the presence of K+ (data not shown). An equivalent concentration of berberine caused no such change in the CD spectrum (dots). In contrast to the above-mentioned 9-O-substituted piperidino derivative which selectively induced and bound to an antiparallel G-quadruplex structure, this new 9-substituted alkyamino berberine induced the formation of a parallel G-quadruplex structure and is in agreement with polyacrylamide gel-shift assays. Thus these examples show that depending on the DNA substrate, 9-substituted berberines are selective for either parallel or antiparallel Gquadruplexes and are good candidates for lead optimization for inhibitors of telomerase and the treatment of cancer.

20.6. CONCLUDING COMMENTS As seen from the above examples, electronic CD spectroscopy has played a central role in the study of DNA-binding properties of agents whose biological activity is a result of

661

D R U G A N D N AT U R A L P R O D U C T B I N D I N G T O N U C L E I C A C I D S A N A LY Z E D

0.1 M K+

16 14 12

10 mM piperidino berberine

10 mM berberine

10 CD (mdeg)

8 6 4

0.1 M Na+

2 0 –2 telo21 without cations or drug

–4 –6 –8 220

240

260 280 Wavelength (nm)

300

320

Figure 20.17. CD spectra of 5 μM of the DNA oligomer telo21 in 10 mM Tris–HCl, pH 7.4 without cations or drug (); 0.1 M Na+ (); 0.1 M K+ (•); 10 mM piperidino derivative (); 10 mM berberine (). (Reproduced with permission from Elsevier Limited, reference 62, Figure 1.)

14

Pu18 plus 25 μM alkylamino derivative

12 10

Pu18 Pu18 plus 25 μM berberine

CD (mdeg)

8 6 4 2

Figure 20.18. CD spectra of 10 mM Pu18 in pH

0

7.2 Tris–HCl buffer in the absence of cations. Solid line, Pu18 without drug; •, Pu18 plus 25 μM of berberine: , Pu18 plus 25 μM of

–2 –4 220

240

260

280

300

Wavelength (nm)

320

340

9-N-substitued alklyamino derivative. (Reproduced with permission from Elsevier Ltd, reference 63, Figure 4A.)

their interaction with DNA. We have chosen a number of natural product and porphyrin DNA-binding agents that illustrate the role of CD in characterizing DNA binding modes such as groove binding, intercalation, and outside binding. In many of these examples, X-ray crystal analyses were also carried out and conﬁrmed the DNA structural transitions suggested by the CD measurements. The DNA-binding studies with the naturally occurring antitumor drugs are important because several of these compounds are of clinical importance as are certain porphyrins for use in photodynamic therapy. The targeting and stabilization of G-quadruplexes for cancer therapy is particularly signiﬁcant as the telomerase-catalyzed extension of telomeres occurs in most cancers, and thus inhibition of this enzyme by inducing and stabilizing G-quadruplexes with small-molecular-weight

662

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

drugs may lead to the discovery of new and speciﬁc agents for cancer chemotherapy. CD spectroscopic studies of drug–DNA binding, in conjunction with other biophysical techniques, will no doubt continue to provide useful insight into these interactions with the aim of obtaining a better understanding of how DNA-binding drugs express their biological activity.

ACKNOWLEDGMENTS The author thanks Dr. Ana Petrovic for the computer-derived DNA duplex structures and is grateful to Ghislaine Vantomme and Edward Chen for help in the editing of several of the ﬁgures.

REFERENCES 1. V. A. Bloomﬁeld, D. M. Carothers, I. Tinoco, Jr., Electronic and vibrational spectroscopy, in Nucleic Acids, Structures, Properties, and Functions, University Science Books, Sausilito, CA, 2000, pp. 185–196. 2. W. C. Johnson, CD of nucleic acids, in Circular Dichroism, Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, pp. 703–718. 3. P. Belmont, J-F Constant, M. Demeunynck, Chem. Soc. Rev . 2001, 30 , 70–81. 4. J. Kypr, I. Kejnovsk´a, D. Renˇciuk, and M. Vorl´ıcˇ kov´a, Nucleic Acids Res. 2009, 37 , 1713–1725. 5. C. Zimmer, G. Luck, Use of circular dichroism to probe DNA structure and drug binding to DNA, in Advances in DNA Sequence Speciﬁc Agents, Vol. 1, L. H. Hurley, ed., JAI Press, Greenwich, Ct, 1992, pp. 51–88. 6. M. Ardhammar, B. Nord´en, T. Kurucsev, Drug DNA Interactions, in Circular Dichroism, Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, eds., WileyVCH, New York, 2000, pp. 741–768. 7. M. Eriksson, B. Nord´en, Linear and circular dichroism of Drug–DNA complexes, in Methods in Enzymology, Vol. 340, Academic Press, New York, 2001, pp. 68–98. 8. S. Allenmark, Chirality 2003, 15 , 409–422. 9. N. C. Garbett, P. A. Ragazzon, J. B. Chaires, Nature Protocols 2007, 2 , 3166–3172. 10. A. Rodger, Circular and linear dichroism of drug–DNA systems, in Methods in Mol. Biol . 2010, 613 , 37–54. 11. M. L. Kopka, C. Yoon, D. Goodsell, P. Pjjura, R. Dickerson, Proc. Natl. Acad. Sci. USA 1985, 82 , 1376–1380. 12. J. G. Pelton, D. E. Wemmer, Proc. Natl. Acad. Sci. USA 1989, 86 , 5723–5727. 13. D. Rentzeperis, L. A. Marky, T. J. Dwyer, B. H. Geierstanger, J. G. Pelton, D. E. Wemmer, Biochemistry 1995, 34 , 2937–2945. 14. L. F. Tietze, B. Krewer, F. Major, I. Schuberth, J. Am. Chem. Soc. 2009, 131 , 13031–13036. 15. L. J. Hanka, A. Dietz, S. A. Gerpheide, S. L. Kuentzel, D. G. Martin, J. Antibiot. 1978, 31 , 1211–1217. 16. K. S. MacMillan, D. L. Bogar, J. Med. Chem. 2009, 52 , 5771–5780. 17. W. C. Krueger, M. D. Prairie, Chem.–Biol. Interactions 1991, 79 , 137–149. 18. M. D. Lee, T. S. Dunne, C. C. Chang, M. M. Siegel, G. O. Morton, G. A. Ellestad, W. J. McGahren, D. B. Borders, J. Am. Chem. Soc. 1992, 114 , 985–987.

D R U G A N D N AT U R A L P R O D U C T B I N D I N G T O N U C L E I C A C I D S A N A LY Z E D

19. (a) J. Golik, J. Clardy, G. Dubay, G. Groenewold, H. Kawaguchi, M. Konishi, B. Krishnan, Ohkuma, K Saitoh, T. W. Doyle, J. Am.. Chem. Soc. 1987, 109 , 3461–4462. (b) J. Golik, G. Dubay, G. Groenewold, H. Kawaguchi, M. Konishi, B. Krishnan, Ohkuma, K Saitoh, T. W. Doyle, J. Am. Chem. Soc. 1987, 109 , 3462–3464. 20. N. Zein, A. M. Sinha, W. J. McGahren, G. A. Ellestad, Science 1988, 240 , 1198–1201. 21. S. Walker, D. Yang, D. Kahne, D. Gange, J. Am. Chem. Soc. 1991, 113 , 4716–4717. 22. A. Kalben, S. Pal, A. M. H. Andreotti, S. Walker, D. Gange, K. Biswas, D. Kahne, J. Am. Chem. Soc. 2000, 122 , 8403–8412. 23. L. G. Paloma, J. A. Smith, W. J. Chazin, K. C. Nicolaou, J. Am. Chem. Soc. 1994, 116 , 3697–3708. 24. N. Ikemoto, R. A. Kumar, T. Ling, G. A. Ellestad, S. K. Danishefsky D. J. Patel, Proc. Natl. Acad. Sci . 1995, 92 , 10506–10510. 25. E. Giorgio, K. Tanaka, W-d. Ding, G. Krishnamurthy, K. Pitts, G. A. Ellestad, C. Rosini, N. Berova, Bioorg. Med. Chem. 2005, 13 , 5072–5079. 26. G. Krishnamurthy, W-D. Ding, L. O’Brian, G. A. Ellestad, Tetrahedron 1994, 50 , 1341–1349. 27. M. Uesugi, Y. Sugiura, Biochemistry 1993, 32 , 4622–4627. 28. J. Golik, Structure determination of the esperamicins, in Enediyne Antibiotics as Antitumor Agents, D. B. Borders and T. W. Doyle, eds., Marcel Decker, New York, 1995, pp. 187–215. 29. M. Slavik, S. K. Carter, Adv. Pharmacol. Chemother. 1975, 12 , 1–30. 30. P. Aich, R. Sen, D. Dasgupta, Biochemistry 1992, 31 , 2988–2997. 31. S. Chakrabarti, P. Aich, D. Sarker, D. Bhattacharyya, D. Dasgupta, J. Biomol. Struct. Dyn. 2000, 18 , 209–218. 32. X. Gao, D. Patel, Biochemistry 1990, 29 , 10940–10956. 33. M-H. Hou, H. Robinson, Y.-G. Gao, A. H-J. Wang, Nucleic Acids Res. 2004, 32 , 2214–2222. 34. G. Cai, N. Bezhkova, J. Odingo, N. Berova, K. Nakanishi, J. Am. Chem. Soc. 1993, 115 , 7192–7198. 35. H. M. Sobell, S. C. Jain, J. Mol. Biol . 1972, 68 , 21–34. 36. S. Kamatori, F. Takusagawa, J. Mol. Biol . 1992, 225 , 445–456. 37. E. Bullock, A. W. Johnson, J. Chem. Soc. 1957, 3280–3285. 38. L. Blau, R. Bittmann, Mol. Pharmacol . 1975, 11 , 716–721. 39. H. E. Auer, B. E. Pawlowski-Konopnicki, T. R. Krugh, FEBS Lett. 1977, 73 , 167–170. 40. M-H. Hou, H. Robinson, Y.-G. Gao, A. H-J. Wang, Nucleic Acids Res. 2002, 30 , 4910–4917. 41. C. Lian, H. Robinson, A. Wang, J Am. Chem Soc. 1996, 118 , 8791–8801. 42. B. Nord´en, P. Lincoln, B. Akerman, E. Tuite, Met. Ions Biol. Syst. 1996, 33 , 177–252. 43. X. Qu, J. O. Trent, I. Fokt, W. Priebe, J. B. Chaires, Proc. Natl. Acad. Sci. USA 2000, 97 , 12032–12037. 44. A. H. Wang, G. Ughetto, G. J. Quigley, A. Rich, Biochemistry 1987, 26 , 1152–1163. 45. J. B. Chaires, J. Biol. Chem. 1986, 261 , 8899–8907. 46. R. F. Pasternack, Chirality 2003, 15 , 329–332. 47. Y-A Lee, J-O Kim, T-S Cho, R. Song, S. K. Kim, J. Am. Chem Soc. 2003, 125 , 8106–8107. 48. Y. Nitta, R. Kuroda, Biopolymers 2006, 81 , 376–391. 49. A. H. Shelton, A. Rodger, D. R. McMillin, Biochemistry 2007, 46 , 9143–9154. 50. M. Balaz, M. De Napoli, A. E, Holmes, A. Mammana, K. Nakanishi, N. Berova, R. Purello, Angew. Chem. Int. Ed . 2005, 44 , 4006–4009. 51. S. C. Ha, K. Lowenhaupt, A. Rich, Y-G Kim, K. K. Kim, Nature 2005, 437 , 1183–1186. 52. A. D’Urso, A. Mamanna, M. Balaz, A. E. Holmes, N. Berova, R. Lauceri, R. Purrello, J. Am. Chem. Soc. 2009, 131 , 2046–2047.

663

664

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

53. D. P. N. Goncalves, S. Ladme, S. Balasubramanian, J. K. M. Sanders, Org. Biomol. Chem. 2006, 4 , 3337–3342. 54. J. T. Davis, Angew. Chem. Int. Ed . 2004, 43 , 668–698. 55. D. J. Patel, A. T. Phan, V. Kuryavyi, Nucleic Acids Res. 2007, 35 , 7429–7455. 56. S. Burge, G. N. Parkinson, P. Hazel, A. K. Todd, S. Neidle, Nucleic Acids Res. 2006, 34 , 5402–5415. 57. G. N. Parkinson, R. Ghosh, S. Neidle, Biochemistry 2007, 46 , 2390–2397. 58. P. Balagurumoorthy, S. K. Brahmachari. D. Mohanty, M. Bansal, V. Sasisekharan, Nucleic Acids Res. 1992, 20 , 4061–4067. 59. M. Lu, Q. Guo, N. R. Kallenbach, Biochemistry 1992, 31 , 2455–2459. 60. D. M. Gray, J-D Wen, C. W. Gray, R. Repges, C. Repges, G. Raabe, J. Fleischhauer, Chirality 2008, 20 , 431–440. 61. M. Franceschin, L. Rossetti, A. D’Ambrosio, S. Schirripa, A. Bianco, G. Ortaggi, M. Savino, C. Schultes, S. Neidle, J. Bioorg. Med. Chem. Lett. 2006, 16 , 1707–1711. 62. W.-J. Zhang, T.-M. Ou, Y.-J. Lu, Y.-Y. Huang, W.-B. Wu, Z.-S. Huang, J.-L. Zhou, K.-Y. Wong, L.-Q. Gu, Bioorg. Med. Chem. 2007, 15 , 5493–5501. 63. Y. Ma, T.-M. Ou, J.-Q. Hou, Y-J. Lu, J-H. Tan, L-Q. Gu, Z.-S. Huang, Bioorg. Med. Chem. 2008, 16 , 7582–7591.

21 PROBING HSA AND AGP DRUG-BINDING SITES BY ELECTRONIC CIRCULAR DICHROISM ´ Simonyi Miklos

21.1. INTRODUCTION Subsequent to recent reviews [1, 2], this chapter covers newer ﬁndings and treats the subject by pointing out pharmacological relevance of the cited examples.

21.1.1. Proteins Binding Drugs in Blood Human blood contains a large variety of proteins serving different functions, such as blood clotting, circulatory transport of molecules (including enzymes, hormones, food constituents), defense against bacteria and viruses, and ﬁnally drug binding that attenuates the effect of dose and performs drug transport to different tissues. Once a drug is taken either orally or otherwise, it appears in blood and immediately interacts with blood proteins. This interaction is an adsorption (called binding) to protein surfaces. Drug binding results in the partitioning of drug molecules into free and bound phases, and its pharmacological signiﬁcance stems from the fact that both therapeutic and toxic effects are associated with the free rather than total concentration of substances. In most cases, drug molecules are oriented to speciﬁc spots, or cavities of the protein to which they could be bound by substantial force. These spots are the binding sites classically characterized by equilibrium constants and displacement experiments brought about by standard competitors. For drug-binding ability, out of the multiplicity of blood proteins, only two have utmost importance: human serum albumin (HSA) and human α1 acid glycoprotein (AGP). This chapter deals with examples casting light on the binding mechanisms of these proteins from circular dichroism spectroscopy. Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

665

666

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

21.1.2. Detection of Drug Binding by Circular Dichroism (CD) Discrimination of free and bound drugs can be achieved by measuring a chiroptical spectral peak induced by protein binding. A simple case is enantioselective binding of racemates, provided that the pool of drug molecules remaining in the free state does not gain optical activity owing to the loss of the bound enantiomer. However, many drugs are nonchiral, but produce measurable signal in the bound state as a consequence of conformational torsion brought about by accommodation to the binding site. Other molecules defy spectral detection even if bound, and their study requires chiroptical probes. These probes are well-known ligands of the proteins, and their binding is affected by drugs simultaneously present. The modiﬁed binding of probes indirectly reﬂects the binding of those molecules that escape direct chiroptical identiﬁcation.

21.2. HUMAN SERUM ALBUMIN Albumin is the most abundant protein in blood plasma with a concentration range of 580–670 μM, depending on age. It is responsible for binding almost all kinds of drugs. For detailed information outside our scope, the reader is directed to a comprehensive monograph [3]. Early studies established a classiﬁcation of the many different drugs bound by albumin to two distinct groups classiﬁed as Site I and Site II types [4] and represented by the binding of rac-warfarin and diazepam, respectively [5]. After the establishment of the complete genetic base sequence [6], the three-dimensional structure of human serum albumin has been determined by X-ray crystallography [7]. HSA is composed of 585 amino acids of a single chain organized by disulﬁde bridges into eight and a half double loops forming three domains. Of these, domains II and III form two principal binding sites able to bind a very wide range of compounds. The structure is brought about by 35 cysteine units, of which 32 are located pairwise for the double loops, two form a single disulﬁde bond, and one, Cys-34, is uncoupled. The stereoview of HSA structure is shown in Figure 21.1. Since fatty acids are physiological ligands of HSA, the next step was the crystallographic determination of fatty acid-binding sites on the protein [8]. The most detailed

Figure 21.1. Binding sites are indicated by speciﬁc ligands in white, warfarin (Site I, right-hand side) and diazepam (Site II, left-hand side). (Reprinted with permission from reference 3, copyright 1996, Elsevier.) (See insert for color representation of the ﬁgure.)

667

PROBING HSA AND AGP DRUG-BINDING SITES

picture emerged from a recent analysis on a wide variety of drug–HSA complexes revealing the precise architecture of fatty acid sites, the two primary drug-binding sites, and the ability of fatty acids to modify Sites I and II. In addition, numerous secondary binding sites for drugs were found that distribute across the protein [9] (Figure 21.2). Although binding site location is much more complicated than the existence of two separate and independent cavities, this chapter follows the early classiﬁcation, since the characterization of drugs to Site I and Site II types is generally accepted. While reviewing numerous examples, exceptions and deviations from the classical ideas will be noted.

21.2.1. Examples for Site I Binding 21.2.1.1. Bilirubin. Bilirubin is formed in the metabolism of heme through physiological oxidation by the enzyme, heme oxygenase, followed by a reduction brought about by biliverdin reductase [10] (Figure 21.3). As a result, the pigment loses iron and the ﬂat porphyrin ring opens up to form two planar conjugated chromophores connected by a methylene group around which the compound gains free rotation. Intensive interest has been focused on bilirubin owing to its toxicity leading to jaundice, and the defense against it provided by HSA through highafﬁnity binding [11]. Bilirubin was suggested to form intramolecular hydrogen bonds [12], an idea supported by infrared investigations [13]. The earliest realization that Cotton effects can be characteristic of protein binding is related to bilirubin: Its binding to bovine

Cleft Thyroxine 5 2°: lodipamide

IIIB FA5 Thyroxine 2,3 2° : Oxyphenbutazone 2° : Propofol

IIIA: Drug Site 2 FA 3,4 Thyroxine 4 Diflunisal Diazepam Halothane Ibuprofen Indoxyl sulphate Propofol 2° : CMPF

IB FA1 Hemin 2° : Azapropazone 2° : Indomethacin 2° : TIB

FA 2 IIA: Drug Site 1 FA7 Thyroxine 1 Azapropazone CMPF DIS Indomethacin Iodipamide Oxyphenbutazone IIA-IIB Phenylbutazone FA6 2° : Diflunisal TIB 2° : Halothane Warfarin 2° : Ibuprofen 2° Indoxyl sulphate 3° Diflunisal

Figure 21.2. Ligand-binding capacity of HSA deﬁned by crystallographic studies. (Reprinted with permission from reference 9, copyright 2005, Elsevier.) (See insert for color representation of the ﬁgure.)

668

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 21.3. Formation of bilirubin in the human body (atoms enlarged affected by transformation).

serum albumin was detected by ORD [14] followed by a CD study on HSA binding [15]. The CD spectra were spectacular: They were bisignate with relatively high intensity, showed inversion upon pH change (Figure 21.4), and were interpreted as arising from coupling between electric transition dipole moments of bound bilirubin leading to exciton splitting [15, 16]. Lacking a chiral center and producing exciton-coupled spectra, bilirubin was suggested to exist as enantiomerically related conformers from which the binding site, depending on the pH, preferred one or another of the mirror imaged forms. As later proved [17], a conformational change of HSA within pH 6–9 called neutral-to-base (N–B) transition involves histidine residues of domain I. A model was suggested in which the two bilirubin chromophores are bound to different half-domains of albumin, and one of these half-domains may rotate relative to the other [18, 19]. A comparison with related bichromophoric tetrapyrrole pigments [20] led to the establishment of intramolecularly hydrogen-bonded enantiomeric conformations for bilirubin existing in dynamic equilibrium as a racemic mixture of interconverting structures (Figure 21.5). In terms of the exciton coupling theory [21], HSA bound the right-handed P conformer [20] at physiological pH of 7.3 while the selection changes in acidic conditions and depends on the species of serum albumin: BSA shows reversed selectivity, as does complexation with cyclodextrins [22] or with quinine [23]. CD spectropolarimetry was suggested [24] as an analytical tool for the determination of bilirubin concentration in blood plasma. However, recent site-directed mutagenesis of HSA revealed dramatic changes in the CD spectra of different HSA mutants [25]. Multiple binding of bilirubin has been demonstrated [26] and the stoichiometry of 1:1 was conﬁned to the high-afﬁnity site [11]. Early competition experiments indicated several drugs to be effective displacers of bilirubin from albumin binding [27]. Drug displacement has been of special clinical interest in the treatment of newborn babies whose metabolizing enzymes are insufﬁcient in the ﬁrst two weeks of life, and their albumin can be saturated with bilirubin [28]. Upon displacement, deposition of bilirubin in tissues takes place, leading to bilirubin encephalopathy [29]. Since bilirubin binds to multiple sites [30], concern was expressed about the application of the Site II drug, ibuprofen, to

669

PROBING HSA AND AGP DRUG-BINDING SITES

350

400

450

500

550

Obs ellipticity × 103

+20 pH 7.3

+10

0

−10

pH 4.8

−20 1.2

1.0

Absorbance

0.8

pH 7.3 pH 4.8

0.6

0.4

0.2

0.0

Figure 21.4. Exciton-coupled CD spectra of HSA-bound 400

350

450

500

Wavelength (nm)

550

bilirubin at two pH values. (Reprinted with permission from reference 15, copyright 1970, Elsevier.)

O H O

O N

H

Z 15

C O

H O

N H

Z 15

O C

H

H N

N 8

8 Z

N H

4

10

12

N H O

Z

N H

12

4 H N

O

C

C

O

H

O

O

H

O

Figure 21.5. Interconverting enantiomeric conformers of intramolecularly hydrogen-bonded bilirubin. (Reprinted with permission from reference 23, copyright 1987, ACS.)

670

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

preterm infants [31]. However, a recent paper proved that ibuprofen will only replace bilirubin from HSA at high dose, but causes no risk at therapeutic dose regimen [32]. Paclitaxel, an agent used in multidrug chemotherapy, was recently demonstrated to have a high-afﬁnity binding site shared by warfarin and bilirubin and located in HSA Site I, subdomain IIA [33]. Consulting Figure 21.2, the above experimental results allow us to identify subdomain IIA Site I as the primary, high-afﬁnity binding site of bilirubin. This cavity is large enough to accommodate the large molecule (cf. Figure 21.2), close to IIA–IIB where ibuprofen is bound and the conformation of bound bilirubin can be affected by neighboring IA responsible for the N–B transition. 21.2.1.2. Warfarin. In the 1920s a hemorrhagic disorder afﬂicted cattle throughout the northern plains of North America. The responsible toxic principle present in spoiled sweet clover hay was identiﬁed as 3,3 -methylene-bis-4-hydroxycoumarin (dicumarol, Figure 21.6a). Its therapeutic usefulness as an orally effective aniticoagulant failed because of poor absorption. Among its synthetic analogues, warfarin emerged as the best studied and most popular oral anticoagulant [34] (Figure 21.6b). Induced CD by warfarin binding to HSA was ﬁrst found in 1972 [35]. Location of the binding site was unequivocally determined by comparing the warfarin-binding strengths of two fragments of HSA [36]. A tryptic fragment of HSA (T45, residues 198–585) comprising domains II and III displayed pH-independent warfarin-induced ellipticity, and its binding afﬁnity was considerably lower than that of HSA. By contrast, a peptic fragment of HSA (P46, residues 1–387) comprising domains I and II gave pH-dependent ellipticity induced by warfarin with binding afﬁnity practically the same as that of HSA (Figure 21.7a). In addition, the CD spectra in the 300- to 350-nm range showed similar relation between the full protein and the fragments (Figure 21.7b). Thus, the warfarin-binding site was located in domain II (Site I) with indication that domain I had an important role in the neutral-to-base (N–B) transition of albumin. Resolution of warfarin led to the establishment of S (−)-warfarin to be about 5 times more potent than R(+)-warfarin in humans [34]. Warfarin binding to HSA shows slight stereoselectivity; however, related coumarin anticoagulants, phenprocoumon and acenocoumarol (Figure 21.8), display higher enantiomer selectivity. Interestingly, for phenprocoumon the S enantiomer binds more strongly [37] (KS /KR = 2), while this ratio is 0.5 for acenocoumarol [38]. Warfarin in solution has been shown to exist as three interconverting tautomeric structures [39]: two cyclic diastereomeric hemiketals and a third minor component, the open-chain intermediate form (Figure 21.6b). Furthermore, NMR studies indicated the preferred solution conformation of phenprocoumon to maintain identical spatial relationship to the cyclic

(a)

(b)

Figure 21.6. (a) Dicoumarol. (b) Warfarin with asterisk at the chiral center.

671

+0.40

0.75 0.50

P46 HSA

0.25

T45

0.00

+0.30 +0.20 +0.10 0.00

−0.25 −0.50

θ obs (degree)

[θ]

× 10−4 (degr.cm2 dmol−1)

PROBING HSA AND AGP DRUG-BINDING SITES

−0.10 6

7

8

300 310 320 330 340 350 360 wavelength (nm)

9

pH (a)

(b)

Figure 21.7. (a) pH dependence of molar ellipticity at 310 nm, warfarin–HSA (), warfarin-P46 ( ), and warfarin-T45 (O). (b) Difference CD spectra at pH 9.0 of warfarin–albumin minus albumin (— —); warfarin-P46 minus P46 (· · ·); warfarin-T45 minus T45 (– – –). (Reprinted with permission from reference 36, copyright 1988, Elsevier.)

Figure 21.8. (a) Acenocoumarol. (a)

(b)

(b) Phenprocoumon.

Figure 21.9. Preferred conformations of R-warfarin and S-phenprocoumon in solution.

hemiketal form of 4-hydroxy-coumarins [40]. In addition, CD studies revealed that the corresponding enantiomers display mirror image spectra in the 240- to 340-nm range [41]. The similar spatial orientations of coumarin and phenyl rings in the opposite enantiomers [34] can be seen in Figure 21.9. Warfarin enantiomers differ in competition experiments. Salicylate completely abolished the Cotton effect of R-warfarin, while it only decreased the intensity of that of S -warfarin [42]. Recent crystal structure determination showed that the two enantiomers of warfarin adopt very similar conformations when bound to HSA-myristate (i.e., to the protein having bound fatty acid) and make many of the same speciﬁc contacts with amino acid side chains at the binding site, thus accounting for the relative lack of stereospeciﬁcity of the HSA–warfarin interaction [8]. The latter study also revealed that warfarin is bound to HSA in the open-chain conformation [40] (Figure 21.10).

672

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 21.10. R-(+)-Warfarin binding in subdomain IIA, Site I. (Reprinted with permission from reference 8, copyright 2001, American Society for Biochemistry and Molecular Biology.)

The warfarin-binding site was located primarily on domain II, while on domain I a secondary binding site was found [43]. Clearly, it should be realized that apart from the high-afﬁnity primary binding, most drugs ﬁnd secondary binding sites on HSA, and this may lead to discrepancies when results from different techniques are compared. 21.2.1.3. Quercetin. Quercetin is an abundant ﬂavonoid in the human diet that may contribute to the prevention of human disease [44] but also may be potentially harmful [45]. It consists of a trihydroxy chromone moiety linked by a single bond to pyrocatechol, responsible for the potent antioxidant ability. Upon excitation, two zwitterionic mesomers might be envisaged with quinoid structure (Figure 21.11). As suggested, HSA tightly binds quercetin in Site I, since ibuprofen, a marker ligand for Site II, does not displace it [46]. However, another study concluded that quercetin bound to both Site I and Site II [47]. Quercetin is optically inactive, so only its binding to the chiral host HSA can lead to the appearance of induced chirality. As the ﬁrst chiroptical study on quercetin binding proves [48], this is the case (Figure 21.12). Fatty acid-free HSA induces three CD bands, negative–positive–negative, in the region from 290 to 500 nm. Since quercetin binds to HSA in monomeric form, the source of Cotton effects is the nonplanar conformation of the molecule. The ortho hydrogens of the pyrocatechol and the C 3 hydroxyl force the attached ring out of the plane even in solution, leading to mirror image forms of which the HSA-binding site prefers only one. Upon addition of palmitic acid to quercetin–HSA solution, the longest-wavelength negative CE intensiﬁed and shifted to higher energies by 10 nm (417 → 407 nm) and the negative CD band at 307 nm also slightly increased (Figure 21.13). These changes indicate that palmitic acid, having several crystallographically well-deﬁned binding sites [49] (cf. Section 21.2.3), enhances the binding of quercetin to HSA just as observed for warfarin [8]. Addition of salicylate to the quercetin–HSA complex did not alter the position and amplitude of the negative CE at 416 nm but slightly decreased the intensity in the absorption spectrum. Thus, quercetin does not bind to subdomain IIIA, but rather to the hydrophobic cavity of subdomain IIA, which is large enough [50] to accommodate two ligand molecules. This explains why salicylate has little effect on the spectral properties of the quercetin–HSA complex. A recent investigation [51] on quercetin

673

PROBING HSA AND AGP DRUG-BINDING SITES

Figure 21.11. The chemical structure of quercetin and its mesomeric forms after excitation.

Wavenumber (10 3 cm2) 4

34.0

32.0

30.0

28.0

26.0

24.0

22.0

20.0

2

CD (mdeg)

0 −2

1

−4

2

−6

3

−8

4

−10

5

−12 300

330

360 λ (nm)

390

420

450

480

Figure 21.12. CD spectra of quercetin–HSA complexes at 37◦ C, pH 7.4. Quercetin/albumin molar ratios: 0.22:1; 0.36:1; 0.51:1; 0.65:1; and 0.80:1 for curves 1–5, respectively. (Reprinted with permission from reference 48, copyright 2003, Elsevier.)

674

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

2

CD (mdeg)

0 −2 −4 −6 −8 −10 1.2 Absorbance

1.0 0.8 0.6 0.4 0.2 0.0

300

330

360

390

420

450

480

λ (nm)

Figure 21.13. CD spectra of quercetin–HSA in the absence and in the presence of palmitic acid. HSA:quercetin:palmitic acid = 1:0.5:0 (solid line) and 1:0.5:2.5 (dotted line). (Reprinted with permission from reference 48, copyright 2003, Elsevier.)

binding complemented the above study with displacement of quercetin–HSA by warfarin. The induced CD spectrum of the quercetin–HSA system remains constant up to about [warfarin]/[quercetin] = 2, but for higher warfarin excess the bands attenuate. These indicate that the 1:1 quercetin–HSA complex is not disturbed up to warfarin/HSA = 2, and displacement starts to take place above this molar ratio. Although the titration was performed up to a molar ratio of 9, quercetin was not completely displaced from HSA, because its induced CD did not vanish (Figure 21.14). The reverse experiment of adding quercetin to the warfarin–HSA complex up to a molar ratio of quercetin/warfarin = 2.5 resulted in a monotonic increase of the induced CD intensity of quercetin. Another competition with ibuprofen led to increasing intensity of the induced Cotton effects of quercetin, instead of displacement, a possible sign of enhanced binding strength [51] (Figure 21.15). This study [51] led to the following conclusions: (1) Quercetin and warfarin share the same primary binding site on HSA, and (2) the simultaneous uptake of warfarin and quercetin may cause interference. Of these, the second conclusion is clinically relevant. Anticoagulant therapy should be strictly monitored in order to adjust medication dose and avoid the risk of bleeding. Since quercetin, as other ﬂavanols, has afﬁnity for the warfarin site, displacement from HSA may be harmful. Similar risk has been observed with sulﬁnpyrazone and phenylbutazone [52]. “It is nowadays very common to ﬁnd antioxidant-rich supplements, based on ﬂavonoids and notably quercetin. These are shelf products, which are widely advertised especially to prevent age-related disease” [51]. The ﬁrst conclusion conﬁrms the earlier result [48] and does not contradict the simultaneous binding of warfarin and quercetin in the large cavity of Site I at appropriate concentrations of the two ligands, as represented by Figure 21.16.

675

PROBING HSA AND AGP DRUG-BINDING SITES

6 W/Q 0

4

−1.1

0

2

4

6

8

10

θ (m°) at 415 nm

−2.2 −3.3

2

−4.4 −5.5

0 −2

Figure 21.14. Partial displacement of quercetin −4

from quercetin-HSA complex by increasing concentration of warfarin (cf. intensities at 355

−6

nm, and 415 nm) [51]. Arrows indicate increasing warfarin/quercetin ratios shown by the inset. (Reprinted with permission from reference 51,

340

360

380

400

420 λ (nm)

440

460

480

500

copyright 2010, Wiley.)

10

c

5

θ m°

0

−5

a

Figure 21.15. Net effect of −10

−15

competitor (c = a − b) on the CD

b

300

350

400

intensity of quercetin-HSA: a, increased by the addition of ibuprofen b (Reprinted with

450 λ (nm)

500

550

permission from reference 51, copyright 2010, Wiley.)

21.2.1.4. Curcumin. Curcumin, the coloring pigment in the powdered rhizome of Curcuma longa L., is chemically a β-diketone, diferuloyl methane, subject to keto–enol tautomerism (Figure 21.17). Although curcumin exhibits a wide range of pharmacological effects, it is practically insoluble in water at neutral pH. The ﬁrst report on the interaction of curcumin with HSA at pH 7.0 applying circular dichroism spectroscopy found very weak optical activity [53]. By extending the pH range up to 9.3 for curcumin/HSA = 1:1, exciton-coupled CD spectra have been observed with increasing intensity when the pH became more and more basic [54] (Figure 21.18).

676

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 21.16. Mutual positions of quercetin (molecular modeling [48]) and warfarin (X ray [8]) in the cavity of Site I subdomain IIA. (Reprinted with permission from reference 48, copyright 2003, Elsevier.) (See insert for color representation of the ﬁgure.)

Figure 21.17. Keto–enol tautomer structures of curcumin.

These spectra are in accord with a shift of the tautomer equilibrium (Figure 21.17) into the diketone side, thereby decreasing the conjugation between the feruloyl moieties and promoting a dissymmetric shape of the molecule. At neutral pH, curcumin exists practically in the enol form having extended conjugation over the molecule with a quasi-planar conformation. With increasing pH, both phenolic OH groups dissociate and tend to compensate the electron withdrawing effect of the β-diketone on the methylene group. This allows the molecule to deviate gradually from coplanarity: The feruloyl chromophores become more and more independent and produce exciton-coupled CD spectra [54]. According to the exciton chirality rule [21], the dissymetric conformation should be right-handed, as shown in Figure 21.19. Increasing amounts of curcumin added to HSA at pH 9 produce a set of exciton-coupled spectra (Figure 21.20), indicating the stoichiometry of the complex to be HSA/curcumin = 1:0.72 [54]. In other words, approximately 30% of bound curcumin molecules do not give a biphasic CD spectrum. Deviation of stoichiometry from the 1:1 ratio might indicate the existence of a second binding site in which the ligand molecule exhibits a weak CD spectrum [54].

677

PROBING HSA AND AGP DRUG-BINDING SITES

496.6 (Δε = + 32M−1cm−1)

60

CD (mdeg)

40 20 0 −20 −40

Absorbance

−60 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0

427 (Δε = −33 M−1cm−1) 6.91 7.38 7.68 7.97 8.31 8.57 8.78 9.00 9.32

433.8 (ε = 29000 M−1cm−1)

300 320 340 360 380 400 420 440 460 480 500 520 540 560 580 600 Wavelength (nm)

Figure 21.18. pH dependence of the CD and UV–vis spectra of curcumin–HSA complex. (Reprinted with permission from reference 54, copyright 2003, Elsevier.)

Figure 21.19. Right-handed chiral conformation of curcumin at pH 9. Dihedral angle between the feruloyl planes is θ ≈ 130◦ . Dotted arrows represent spatial orientations of the electric transition moments coupled excitonically. (Reprinted with permission from reference 54, copyright 2003, Elsevier.)

Comparable displacement of curcumin was found by warfarin (Site I marker) and ibuprofen (Site II marker), contrary to diazepam (Site II marker), a weak displacer [55]. In addition, palmitic acid was found a strong displacer of curcumin. As for ibuprofen (cf. Section 21.2.2.2), its displacing effect on warfarin–HSA was also observed [56]. These discrepancies point to the role of secondary binding sites and their interplay in

678

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

60

CD (mdeg)

40 20 0 −20 −40

Absorbance

−60 2.0 1.5 1.0 0.5 0.0 300 320 340 360 380 400 420 440 460 480 500 520 540 560 580 600 Wavelength (nm)

Figure 21.20. Exciton-coupled CD spectra at pH 9.0 of curcumin–HSA. Curves of increasing intensity correspond to curcumin/HSA molar ratios from 0.05 to 0.55 with 0.05 increments. (Reprinted with permission from reference 54, copyright 2003, Elsevier.)

displacement experiments [57]. One of the fatty acid-binding sites (FA6 in Figure 21.2) is positioned between subdomains IIA and IIB, where ibuprofen is also located. Thus, the binding sites of curcumin (for both ﬂat and bent conformations) can be visualized in Figure 21.21. Of the four positions found, it seems reasonable to locate the curcuminbinding site producing increasing Cotton effects in the pH range of 7.4–9.3 to be close to Site I at the frontier of subdomains IIA and IIB.

21.2.2. Examples for Site II Binding 21.2.2.1. Benzodiazepines. The substitution pattern of the seven-membered ring allows the existence of six different classes of benzodiazepines (BZs). Of these, the 1,4-class [58] gained such a dominance owing to the success of its representatives as anxiolytics in the 1970s that sometimes it is overlooked that other kinds also exist. While the 1,5-class (becoming important as cholecystokinin receptor (CCK)-B antagonist [59]) gains asymmetry only by its substitution, 1,4-benzodiazepines with asymmetrically positioned N atoms are always chiral owing to their nonplanarity [60]. 21.2.2.1.1. 1,4-Benzodiazepines. The principal member of this class is diazepam (Figure 21.22). The ﬁrst CD investigation on the binding of diazepam and related 1,4benzodiazepines (1,4-BZs) found that all of them bound to the same HSA-binding site [61], and this site was independent of the bilirubin site [62]. The source of the induced Cotton effect was found through comparison with 3-methyl-substituted derivatives [63]. The binding-induced Cotton effects of diazepam and DeMeD matched the solution spectra

679

PROBING HSA AND AGP DRUG-BINDING SITES

IB

IA

IIIB

site II IIA site I

IIIA

Figure 21.21. X-ray crystallographic structure of HSA [8] with curcumin molecules localized by docking. Subdomains are indicated. (Reprinted

IIB

with permission from reference 55, copyright 2003, Elsevier.) (See insert for color representation of the ﬁgure.)

of the corresponding S -3-methyl derivatives S -3-Me-D and S -3-Me-DeMeD, respectively [64] (Figure 21.23). Since the S -3-methyl derivatives exist in a single conformation [65], the CD spectrum indicated the preferred binding conformation of diazepam (Figure 21.24) [66]. The X-ray structure of Me2 D displays the characteristic structural details of one of the conformers [66] (Figure 21.25). Hence, all the above 1,4-BZs symmetrically substituted at C3 form a conformational racemate both in solution and in the crystalline phase [66]. Inversion of the diazepine ring—that is, ﬂipping of the C3 atom through the plane of the aromatic ring—can easily occur at room temperature [67] for both compounds, Me2 DeMeD and Me2 D, but becomes hindered when a t-butyl substituent appears at N1, making the confomers stable atropisomers [68]. The CD spectra of HSA-bound Me2 DeMeD and Me2 D [69] are shown in Figure 21.26. Saturation of diazepam at the 4–5 position creates a center of chirality at C5 (dihydrodiazepam, DHD), yielding a more ﬂexible hetero-ring. Resolution of the DHD racemate and carbamoylation of the optically pure enantiomers presented a puzzle: The levorotatory 5R-DHD ([α]L = −224) yielded dextrorotatory 5R-carbamoyl-DHD ([α]D = +618) [70]; hence carbamoylation caused inversion of the CD spectra [71] (Figure 21.27). Inversion of CD spectra reﬂected inversion of the hetero-ring conformation [71] (Figure 21.29). Binding of DHD enantiomers to HSA was found to prefer the 5S -(+) enantiomer [72]. 21.2.2.1.2. Conformational Recognition of 1,4-BZs. Realizing that the conformation is the crucial property of 1,4-BZ recognition by HSA, the idea of absolute conformation has been introduced [73]. This, however, did not achieve general approval. One of the bibles of stereochemistry [74] states: “.. absolute conformation is subsumed under the term absolute conﬁguration . . . Consequently, we do not use the term absolute conformation in this book.” But the community dealing with 1,4-BZs binding to HSA has accepted the concept. The original deﬁnition [73] of enantiomeric conformations

680

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 21.22. General formula of 1,4-BZs. R1

R2

R3

R4

H Me H

H H H

H H Me

H H H

Desmethyl diazepam (DeMeD) Diazepam S-3-Methyl-desmethyl-diazepam (S-3-Me-DeMeD)

H Me

Me Me

Me Me

H H

3,3-Dimethyl- desmethyl-diazepam (Me2 DeMeD) 3,3-Dimethyl-diazepam (Me2 D)

H H H

H H H

OH OAc OH

H H Cl

S-3-Oxazepama S-3-oxazepam acetate (S-3-OAc) S-3-lorazepama

H H Me

H H H

OMe OAc OAc

Cl Cl H

S-3-Lorazepam methylether (S-3-LoMe) S-3-Lorazepam acetate (S-3-LoAc) S-3-Temazepam acetate (S-3-Tac)

Me H H

H H H

OAc H Me

Cl Cl Cl

S-3-Methyl-lorazepam acetate (S-3-Me-LoAc) Clonazepam S-3-Methyl-clonazepam

a these enantiomers are stable when bound to HSA, but they are unstable in aqueous solution owing

to racemization.

was based on the torsion angle (O2–C2–C3–N4) selected for the diazepine ring according to its highest priority by the CIP convention [75]. Since this angle is negative for the HSA-preferred conformation (cf. Figure 21.25), it was named M (minus), while the “boat” standing on its bow was named P (plus). It was realized soon afterwards that among homochiral analogues “replacement of the (3S )-alkyl substituent by OH does not change the conﬁgurational descriptor (3S ) nor the conformational descriptor (M) along bond C3–N4; along bond C2–C3 it changes, however, formally to P [76].” Nevertheless, the original assignment was kept generally. BZ conformers interconverting in equilibria (Figure 21.24) [65, 66, 69, 77, 78], as well as those subject to inversion (Figure 21.28) [76] have been named accordingly. Agreeing with the necessity of assigning M and P descriptors to chiral ring conformers, one should concentrate on the selection of an internal torsion angle to which the assignment is related. Since the ring conformer is to be named irrespective of its substituents, CIP priority would dictate the choice of N1–C2–C3–N4 rather than C2–C3–N4–C5. But then, contrary to the original assignment [73] that has been followed for some decades [65, 66, 69, 77–79], the decisive torsion angle for both 3S -Me-,

681

PROBING HSA AND AGP DRUG-BINDING SITES

a 160

c

120

80 f b

ΔA x 105 d

40 d

e 0

c 40

a

80

Figure 21.23. CD spectra of 1,4-BZs: (a)

d

120

S-3-Me-DeMeD in buffer; (b) R-3-Me-DeMeD in buffer; (c) S-3-Me-DeMeD in buffer with HSA; (d) R-3-Me-DeMeD in buffer with HSA; (e) DeMeD in buffer with HSA; (f) diazepam in

b 160 250

300

350

λ (nm) axial

buffer with HSA. (Reprinted with permission from reference 64, copyright 1979, Elsevier.)

equatorial

R2

R3

N

0

R1

φ

N

φ

R1

Figure 21.24. Conformer equilibrium is shifted by

N 0 N

R3 axial

R2 equatorial

the conﬁguration of 3-methyl derivatives: S-3-methyl (R2 = H, R3 = Me) occupies equatorial position forming the boat-like conformation standing on its bottom, and R-3-methyl forms the boat upside down.

3S -OH-, and 3S -OAc-1,4-BZs were positive, and accordingly the assignment should be changed from M to P. This has already been done in a paper by one of the authors concerned [80] (Figure 21.29). This discrepancy should be clariﬁed in the future. 21.2.2.1.3. Allosteric Interaction Between Site I and Site II. Preference of high-afﬁnity HSA binding toward one of the conformers corroborates the early suggestion [61] that all 1,4-BZs bind to Site II, and they are independent of bilirubin binding [62]. A seminal paper applying CD called attention to the high stereoselectivity of oxazepam hemisuccinate binding to HSA [81]. Further studies on 3-acyloxy-1,4-BZs revealed that simultaneous Site I and Site II binding may lead to interaction [82]. Chiroptical spectroscopy was applied to detect optical activity in the ultraﬁltrate of HSA with racemic OAc [83] and LoAc [84] and allowed binding afﬁnity to be

682

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

16 17 3 2

9

N4

6

18

5

5A

7

Cl

N1

9A

8

O2

Figure 21.25. The shape of Me2 D shown in 15

10 14 11 13

12

the conformation preferred by HSA Site II (from atomic coordinates CB2 1EW [66]). Characteristic torsion angles are: N1–C2–C3–N4, +61.2◦ ; C2–C3–N4–C5, −66.2◦ ; C3–N4–C5–C5A, +0.6◦ ;

N4–C5–C5A–C9A, +40.3◦ ; C5–C5A–C9A–N1, +2.4◦ ; C9A–N1–C2–C3, +8.6◦ ; O2–C2–C3–N4, −118.0; O2–C2–C3–C17, −3.4◦ .

4b

50

40

4a

30 Δε

20

10

0

–10 250

270

290

310 λ (nm)

Ph 119.1

330

350

Figure 21.26. CD spectra of Me2 DeMeD (a) and Me2 D (b) bound to HSA [69]; absorption by the protein is strong below 250 nm.

683

PROBING HSA AND AGP DRUG-BINDING SITES

40 CH3 9 8

7

5

20

H

2

3C

B 5a

6

C

1

A CI

O

N

9a

R

4

CH

N Q

Δε

0

C

–20

1: R = H, Q = H 2: R = H, Q = CONH2 3: R = CH3, Q = H

–40

Figure 21.27. CD spectra of (+)-5S-1 (

200 220 240 260 280 300 λ (nm)

Chemica Acta.)

R

H N N

CI

), (−)-5S-2 (– · –), and (+)-(3S,5S)-3

(– – –). (Reprinted with permission from reference 71, copyright 1989, Croatica

CH3

O H

CH3 H

N

CI

N H

(+)-(5S)-1: R = H (+)-(3S,5S)-3: R = CH3

O

H

H

(—)-(5S)-2

CONH2

Figure 21.28. Inversion of DHD conformation upon carbamoylation at N4 [71]. (Reprinted with permission from reference 71, copyright 1989, Croatica Chemica Acta.)

Figure 21.29. Modiﬁed deﬁnition of conformer assignment for 1,4-BZs. (Reprinted with permission from reference 80, copyright 2001, Croatica Chemica Acta.)

684

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

determined. Dealing with racemic esters, chiroptical detection could advantageously be combined with chromatographic separation of the enantiomers on immobilized HSA serving as stationary phase [85]. The enantiomer bound more strongly to the HSA column proved for all the esters to be the dextrorotatory 3S -1,4-BZs [82–85]. Although LoAc bound more weakly than OAc to HSA, indicating decreased binding afﬁnity brought about by the ortho chlorine substituent [84, 86], it was LoAc that revealed allosteric interaction between Site I and Site II: In the presence of warfarin the binding of LoAc strongly increased. A thorough investigation disclosed the structural conditions governing this interaction [82]: (a) 2 -Cl elicits, while 1-Me prevents, allosteric interaction between BZs and warfarin; (b) the allosteric interaction is stronger for S -warfarin than for Rwarfarin; (c) the allosteric interaction with S -warfarin is inhibited by R-warfarin; (d) the allosteric interaction is speciﬁc for the S -LoAc enantiomer. A representative example is shown in Figure 21.30 [87]. The 3-S -oxy substituent is not a requirement for allosteric interaction. Clonazepam (cf. Figure 21.22), a 2 -chlorine derivative weakly bound by HSA without preference to either of its conformers, also exhibited mutually increased binding with S -warfarin [88]. 3-Me-clonazepam enantiomers also bind to HSA with low afﬁnity and without enantioselectivity. However, 3-S -Me-clonazepam bound strongly in the presence of S warfarin, and rac-warfarin displayed increased stereoselectivity in the presence of 3S -Me-clonazepam [89]. Moreover, bilirubin was also found to increase the binding of 3-S -Me-clonazepam [89], indicating that the allosteric interaction is not restricted to S -warfarin. In fact, S -phenprocoumon also exerts the phenomenon [90]. The increased stereoselectivity brought about by the interaction of Site I and Site II could be exploited for microscale resolutions [91, 92]. These ﬁndings gave additional support for both warfarin enantiomers binding at the same site (Site I). They also prove the independent binding of diazepam (Site II marker) and warfarin (Site I marker), since the 1-Me substituent prohibits interaction between the sites. 21.2.2.1.4. Additional BZs. The fundamental role of conformation in protein binding has been corroborated by the study of 2,3-BZs, as well. A representative compound

a

dpm/104

0 b 2 0 c

I 2 0

II V0

10 20 Fraction number

Δε (M–1cm–1)

2 10 0 –10 –20 30

200

250

300 λ (nm)

(A)

(B)

Figure 21.30. (A) radiochromatograms of rac-LoMe on HSA stationary phase: (a) elution by buffer; (b) elution by 10−4 M R-warfarin; (c) elution by 10−4 M S-warfarin yielding fractions I and II. (B) CD spectra of LoMe fractions I (broken line) and II (solid line) [87].

PROBING HSA AND AGP DRUG-BINDING SITES

Figure 21.31. The chemical structure of toﬁsopam; the center of chirality (*) is indicated.

of this family is the anxiolytic toﬁsopam used as a racemic mixture (Figure 21.31). Toﬁsopam contains two chiral elements, a chiral center and the chiral conformation of the diazepine ring. Accordingly, HSA binding distinguishes four different species in which optical rotation depends on the conformation. Inversion of the diazepine ring is slightly hindered by neighboring methyl and ethyl groups at positions 4 and 5, respectively. Nevertheless, in aqueous solution the conformer equilibrium is easily established [93] with t1/2 = 3 h. Studying differently substituted analogues, two types of preferences of the binding site could be established: Conformation P (deﬁned by the C1–N2–N3–C4 torsion angle) is preferred together with the quasi-equatorial orientation of the ethyl group [94]. Different conformers of the same enantiomer, behaving as diastereomers [95], can be separated by chromatography and detected by CD [96] (Figure 21.32). Although it is a 1,4-BZ, tiﬂuadom is a derivative with an unusual activity proﬁle. Unlike most benzodiazepines, tiﬂuadom has opioid activity [97]. It lacks the oxomoiety, thereby having a ﬂexible hetero-ring, and instead contains a chiral center at C2 (Figure 21.33). Unlike toﬁsopam, tiﬂuadom elutes from the HSA column in two fractions [98]. The CD spectra of the weakly bound enantiomer extracted from fraction I was analogous to those of 3-R-1,4-BZs (Figure 21.34). Consequently, the preference of HSA towards tiﬂuadom and diazepam conformations is most probably identical. 21.2.2.2. Nonsteroidal Anti-inﬂammatory Drugs. Nonsteroidal antiinﬂammatory drugs (NSAIDs) have analgesic and antipyretic effects and they reduce inﬂammation. This name covers several classes of chemical structures, so their classiﬁcation as Site II drugs is not justiﬁed. Historically, however, a characteristic representative of the aryl-propionic acids class, ibuprofen, was deﬁned [5] as a Site II marker owing to its high-afﬁnity binding to the diazepam site. Figure 21.35 displays examples of this group of drugs. Neither ibuprofen nor naproxen shows measurable induced Cotton effect, so they can be studied by the diazepam–albumin interaction, as a probe. It was found that both ibuprofen and naproxen reduced the ellipticity induced by diazepam, and the S enantiomers of both drugs bound stronger than their antipodes [99]. When ibuprofen was used in molar excess to HSA, secondary binding with lower afﬁnity occurred and Site I could also be affected [99]. Contrary to ibuprofen, carprofen induced Cotton effects upon HSA binding, but the shape of the spectrum depended on the carprofen:HSA molar ratio [100]. At equimolar conditions, carprofen bound with high afﬁnity to Site II; however at a high ratio to HSA, carprofen bound to Site I as well and its CD spectrum dramatically

685

686

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

31 30

R(M) [+]

25 S(M) [+]

CD

20

15

R(P) [–]

10 S(P) [–] 6 2

Abs

1.8

1.6

Figure 21.32. HPLC separation of toﬁsopam 1.4

0

20 30 Time [min]

10

40

diastereomers with CD detection. (Reprinted with permission from reference 96, copyright 1989, Taylor & Francis.)

N N

50

S H N

F O

Figure 21.33. The chemical structure of tiﬂuadom.

changed (Figure 21.36a). Ibuprofen at equimolar ratio to the HSA–carprofen complex effectively displaced carprofen from its high-afﬁnity binding site (Site II) and, when in excess, ibuprofen reproduced the CD spectra of carprofen characteristic of its low-afﬁnity binding (Site I), as seen in Figure 21.36b [101]. However, ibuprofen was also able to act as a Site I binder as indicated by the allosterically strengthened binding of S -LoAc by S -ibuprofen [102], the more strongly

687

PROBING HSA AND AGP DRUG-BINDING SITES

K (I) = 6.3 × 10–5 M–1; V0 = 49 ml VI = 570 ml VII = 913 ml

K (II) = 1.5 × 10–6 M–1.

I

II

VI

V0

VII

Tifluadom-I 10 x 0.1

Δε

5

0

–5

–10

200

250

300 λ (nm)

350

400

Figure 21.34. Elution of tiﬂuadom from HSA-sepharose column and CD spectrum of fraction I in ethanol (solid line) and cyclohexane (broken line); afﬁnity constants are indicated [98].

(a)

(b)

(c)

(d)

(e)

(f)

Figure 21.35. Selected NSAIDs: (a) ibuprofen; (b) naproxen; (c) carprofen; (d) diclofenac; (e) phenylbutazone; (f) tenoxicam.

688

2

2

0

0

–2

–2

θobs (mdeg)

θobs (mdeg)

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

–4

–4

–6

–6

–8

–8 250

300 350 Wavelength (nm) (a)

250

300 350 Wavelength (nm) (b)

Figure 21.36. CD spectra of the carprofen-HSA system. (a) Molar ratio of carprofen/HSA 0.5 (– – –), 1.0 (

), 3.0 (-.-.), 5.0 (— — —) [100]. (b) Molar ratio of ibuprofen to 1:1 carprofen–HSA

complex 0 ( ), 1.0 (– – –), 5.0 (— — —). (Reprinted with permission from reference 101, copyright 1993, Elsevier.)

bound enantiomer. This is fully in agreement with crystallographic studies [9], indicating a lower-afﬁnity binding site of ibuprofen at the border of subdomains IIA-IIB (Figure 21.2). It is again a manifestation that weaker binding could be important through its ability to induce afﬁnity at another binding site [89]. A ﬁnal example stresses the interplay between different binding sites. Tenoxicam at physiological pH binds strongly at Site I but also at Site II to a lesser extent [103]. Competition studies proved that tenoxicam and diazepam mutually increase the binding strength of each other, although they bind primarily to different sites [104].

21.2.3. Fatty Acid Binding Although fatty acids (FAs) are physiological ligands of HSA and their effect on both Site I and Site II were early detected, FA sites were not deﬁned [4] by the early classiﬁcation. Instead, fatty acid-free HSA has been used for binding studies. Location of FA sites became possible only by X-ray crystallographic analysis [8, 105] (Figure 21.37). FAs are poor chromophores, hence their binding sites could only be investigated indirectly. It was found that the increase in the 315-nm induced ellipticity of HSA–warfarin complex observed in the presence of long-chain FAs was due to a modiﬁcation in the conformation of the warfarin-binding site caused by the FAs [106]. This interpretation was also proved by X-ray crystallographic analysis [8, 9] (Figure 21.38). Recently, even the relative afﬁnities of the seven identiﬁed FA sites could be estimated: FA sites 2, 4, and 5 bind FA with high afﬁnity, while sites 1, 3, 6, and 7 exhibit low afﬁnity for fatty acids [107]. Based on the location of FA sites, a direct chiroptical spectroscopic study could be performed [108]. Crocetin, a carotenoid dicarboxylic acid component of saffron extract, has a strong absorption owing to its conjugated chain. The molecule lacks chirality and cannot be deformed. Similarly to other carotenoids, its hydrophobic character is dominant. Crocetin binds to HSA and induces biphasic Cotton effects in the visible

689

PROBING HSA AND AGP DRUG-BINDING SITES

1 5

7 2 4 3 6 C16:0

Figure 21.37. Structure of HSA complexed with seven palmitic acid molecules. (Reprinted with permission from reference 105, copyright 2000, Elsevier.) (See insert for color representation of the ﬁgure.)

(a)

(b)

Figure 21.38. Conformational changes in warfarin binding (Site I) as a result of fatty acid binding. (a) Helices h2 and h3 are shown by light shades for defatted and by dark shades for myristate bound HSA. (Reprinted with permission from reference 8, copyright 2001, American Society for Biochemistry and Molecular Biology.) (b) The volume of Site I in defatted HSA is depicted by a light brown semitransparent surface that becomes expanded upon myristate binding (blue semitransparent surface), the red arrows point to structural changes associated with fatty acid binding. (Reprinted with permission from reference 9, copyright 2005, Elsevier.) (See insert for color representation of the ﬁgure.)

region, indicating exciton interaction and implying that at least two crocetin molecules bind to HSA because such a chiroptical signal cannot be expected from a single crocetin molecule [108] (Figure 21.39). The maximal exciton intensity corresponded to a crocetin/HSA molar ratio of 4 to 5. Bound crocetin is displaced by palmitic acid as indicated by the concentration-dependent decrease of CD peak intensity [108]. Considering the location of FA sites, the source of exciton interaction has been assigned to neighboring FA3–FA4 sites where two molecules of crocetin could favorably be accommodated. Their position is displayed in Figure 21.40. Although this ﬁgure shows the location of crocetin molecules in two dimensions, they are not accommodated in a single plane, which would cancel chiral interaction of their

690

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

O OH HO O (a)

12 8

CD (mdeg)

4 0 Crocetin/HSA molar ratios a = 10.2/1 b = 6.9/1 c = 5.2/1 d = 4.2/1 e = 2.1/1 f = 1/0.9

–4 –8 –12

b c d

a

e f

–16 a

Absorbance

0.8

b

c

d e

0.6 f 0.4

0.2

0.0 300

350

450

400

500

550

(b)

Figure 21.39. The chemical structure of crocetin (a) and the chiroptical spectra of crocetin–HSA interaction (b) at molar ratios indicated. (Reprinted with permission from reference 108, copyright 2001, Elsevier.)

electric dipole transition moments. The hydrophobic binding of the two crocetins to HSA FA3 and FA4 sites, reinforced by hydrogen bonding of carboxylate groups to Ser342, Arg348 and Tyr411, Ser489, Ser 419, Thr422, respectively, effectively immobilizes the chromophores and represents a simple case of supramolecular chirality [109].

21.3. α1 -ACID GLYCOPROTEIN (AGP) AGP is a minor protein component of blood: its concentration varies between 25 and 150 μM, with higher values in case of diseases [110]. AGP is composed of a single

691

PROBING HSA AND AGP DRUG-BINDING SITES

Figure 21.40. Two crocetin molecules ﬁtted to FA3 and FA4 sites; the negative exciton dictates the horizontal crocetin molecule to be behind the slanting one. (Reprinted with permission from reference 108, copyright 2001, Elsevier.) (See insert for color representation of the ﬁgure.)

protein chain of 183 amino acid units [111] containing two disulﬁde bridges (linking Cys pairs 5–147 and 72–165) [112] and ﬁve Asn units with attached bi-, tri-, and tetra-antennary glycan chains [113] ending in sialic acid residues. The carbohydrate content of AGP amounts to 45% (w/w) and its structure characterizes certain illnesses [114]. Despite its much lower concentration, AGP may represent the major drug-binding protein in human plasma for β-blockers [115] and psychotropic drugs [116]. There are three phenotypes of AGP (A, F1, S) encoded by two genes [117]. F1 (the most important variant induced by diseases [118]) and S are coded for by alleles of the same gene and only differ in a single amino acid codon (Q20R) [119], while F1 and A differ in 21 amino acid units and show different drug-binding properties [120]. Chromatographic separation of AGP yielded two fractions [121]: one contained both variants F1 (migrating fast electrophoretically) and S (slowly migrating), in the following distinguished by F1-S, while the other contained variant A. The separation allowed to select markers for these variants [122]. Their complete genetic sequence is given in reference 123. Comparison of the F1-S and A variants was modeled by color coding the electrostatic potentials [124] (Figure 21.41). X-ray crystallographic investigation has long been hindered because the high carbohydrate content prevented crystallization of AGP. Finally, the reﬁned crystal structure

Figure 21.41. Electrostatic potentials of genetic variants. (a) lysophospholipid ligand binding at the surface of variant F1-S. (b) variant a. (Reprinted with

(a)

(b)

permission from reference 124, copyright 2006, ACS.) (See insert for color representation of the ﬁgure.)

692

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 21.42. Crystal structure of variant F1-S. Side chains of N-linked glycosylation sites (Asn15, Asn38, Asn54, Asn75, and Asn85) and Trp25, as well as N and C terminals, are indicated. (Reprinted with permission from reference 125, copyright 2008, Elsevier.)

˚ resolution has been determined [125]. It of the unglycosylated F1-S variant at 1.8-A contains a large cavity bordered by eight β-strands (A-H) with four loops connecting them (Figure 21.42). A bound Tris molecule inside the cavity is shown. For variant A, no X-ray crystallographic structure is available so far (see note at the end of chapter).

21.3.1. Stereoselective Binding to AGP Studies on AGP binding and the application of CD spectroscopy started before separated variants became available. The ﬁrst indication of enantioselective binding by AGP, a sign of speciﬁc interaction, was related to propranolol [126] and was soon conﬁrmed [127]. Further importance of this study was (a) the demonstration of opposite stereoselectivity in the binding of (+)- and (−)-propranolol to AGP and HSA and (b) the proof that in plasma the preferred binding of (−)-propranolol by AGP predominates [126]. Inverse stereoselectivity in the binding of acenocoumarol to HSA and to AGP was demonstrated by the CD spectra of unbound ligands [38] as shown in Figure 21.44. It was also found that the selectivity by AGP binding strengthened in the sera of sick persons [128].

Figure 21.43. The chemical formula of S-(−)-propranolol.

693

PROBING HSA AND AGP DRUG-BINDING SITES

Earlier estimations predicted low stereoselectivity for AGP binding [129, 130]. However, enantioselectivity of AGP for vinca alkaloids was found to be exceedingly high [131].

21.3.2. Examples of Binding to AGP Genetic Variants The ﬁrst supposition that warfarin labels only variant F1-S [122] was conﬁrmed, and acridine orange 10-dodecyl bromide (AODB) (Figure 21.45) was found selective for variant A [132]. On separated variant A, the selectivity of AODB has been proved [133] (Figure 21.46). The induced biphasic CD curve of AODB-AGP is the manifestation of chiral exciton interaction between at least two dye molecules bound to the protein in left-handed helical arrangement [134]. Similar interpretation of a right-handed induced exciton spectrum was put forward for acridine orange binding to DNA [135]. (b)

(a)

1 Δε, M−1cm−1

Δε, M−1cm−1

10

0 240

320

λ, nm

0 280

320

360 λ, nm

−10 −1

Δε, M−1cm−1

(c)

1

0 280

320

360 λ, nm

−1

Figure 21.44. CD spectra indicating the stereoselectivity of HSA and AGP. (a) R-acenocoumarol in ethanol (dashed line) and in pH 7.4 Ringer buffer. (b) Ultraﬁltrate of rac-acenocoumarol–HSA; (c) Ultraﬁltrate of rac-acenocoumarol–AGP. (Reprinted with permission from reference 38, copyright 1989, Elsevier.)

(a)

Figure 21.45.

(b)

(c)

(d)

Variant A selective ligands. (a) acridine orange dodecyl bromide, AODB; (b) chlorpromazine; (c) imipramine; (d) protriptyline.

694

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

14 12 10 8 6 4 2 CD (mdeg)

0 −2 −4 −6 −8 −10

AGP/A + AODB

−12

AGP/F1-S + AODB

−14 −16 −18

Figure 21.46. Induced ellipticities of AODB

−20

binding by AGP variants; F1-S is contaminated by 4% of variant A; cAODB = cF1-S = cA = 30 mM.

400

420

440

460

480

500

520

Wavelength (nm)

540

560

(Reprinted with permission from reference 133, copyright 2004, Elsevier.)

The most spectacular induced ellipticity of AGP binding was produced by dicoumarol (Figure 21.47) [136]. The intensity of the CD spectra depended on pH, temperature, and the presence of other drugs, like chlorpromazine and protriptyline, that characteristically did not displace dicoumarol but dramatically changed its spectrum [137]. The characteristic polyphasic CD spectra of dicoumarol and its change by tricyclic drugs is due to binding to variant F1-S [133], as shown in Figure 21.48. Investigation of dicoumarol derivatives revealed that reversal occurred up to propylidenebis 4-hydroxycoumarin, but was absent for benzylidenebis 4-hydroxycoumarin, and two ligands were indicated to be accommodated by the large AGP F1-S-binding site [138]. A closer look at the dicoumarol spectra reversed by chlorpromazine (Figure 21.47c) or by imipramine (Figure 21.48b) reveals that around 300 nm, no reversal occurs, while the intense peaks change signs. For the structure of dicoumarol in solution the existence of intramolecularly hydrogen-bonded chiral conformers was suggested [139] and the chiroptical spectrum has been interpreted as the selective recognition of one of the conformers, by analogy to the case of bilirubin–HSA interaction (cf. Section 21.2.1.1) [133]. The phenomenon has been studied further with deramciclane and its analogues (Figure 21.49) [140]. Deramciclane (DER) is an anxiolytic agent selectively binding to variant A [140]. Although DER is chiral, its CD spectrum appears around 220 nm and does not interfere with dicoumarol spectra above 240 nm. As with imipramine (Figure 21.48b), DER caused reversal of the dicoumarol–AGP F1-S spectrum in a concentration-dependent fashion. It was also conﬁrmed that only intensity changes occur in the middle of the

695

PROBING HSA AND AGP DRUG-BINDING SITES

3.0 2.0

0 θobs (mdeg)

θobs (deg × 10−3)

+1.0 0.3

−1.0

0.6

−2.0

1.0 2.0

−3.0 −4.0

4.0

+3.0 +2.0

0 −1.0 3 −2.0

2

−3.0

1

+1.0 0 −1.0 −2.0

250

250 270 290 310 330 350

1 2 3

θobs (mdeg)

+2.0

1.0

300

350

250

300

Wavelength (nm)

Wavelength (nm)

Wavelength (nm)

(a)

(b)

(c)

350

Figure 21.47. Induced CD spectra of dicoumarol binding to AGP. (a) concentration dependence, drug-to-protein ratios indicated. (Reprinted with permission from reference 136, copyright 1988, Kumamoto University.) (b) Temperature dependence: 1, 10◦ C; 2, 25◦ C; 3, 40◦ C. (c) Effect of chlorpromazine (- - -) and protriptyline (– – –) on dicoumarol/AGP (solid line). (Reprinted with

0.6 AGP/F1-S AGP/A AGP/F1-S UV

0.5 0.4 0.3 0.2

CD (mdeg)

6 5 4 3 2 1 0 −1 −2 −3 −4 −5 −6 250 260

Absorbance

CD (mdeg)

permission from reference 137, copyright 1991, Elsevier.)

0.1 0.0 270 280 290 300 310 320 330 340 350 360 370 Wavelength (nm)

6 5 4 3 2 1 0 −1 −2 −3 −4 −5 −6 250 260 270 280 290 300 310 320 330 340 350 360 370 Wavelength (nm)

(a)

(b)

Figure 21.48. Induced CD spectra of dicoumarol binding to F1-S and A genetic variants of AGP. (a) the polyphasic spectrum is due to F1-S, while noise is recorded only on variant A, cdicumarol = cF1-S = cA = 30 mM. (b) reversal of the F1-S-induced ellipticity by imipramine, cdicumarol = cF1-S = cimipramine = 30 mM. (Reprinted with permission from reference 133, copyright 2004, Elsevier.)

O

1 N R

R2

R1 N

O

R2

Figure 21.49. The structures of deramciclane (DER) (1), its DER: R 1 = R2 = Me 1: R 1 = H; R 2 = Me 2: R 1 = R2 = H

Enantiomer of DER 3: R 1 = R2 = Me

desmethyl analogue (2), and its mirror-image enantiomer (3). (Reprinted with permission from reference 140, copyright 2010, Elsevier.)

696

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

spectrum without alteration of sign around 290 nm (Figure 21.50) [140]. The effect of desmethyl-analogue 2 was essentially the same as that of DER; however, the opposite enantiomer of DER (3) effectively abolished the induced CD spectra, leaving the peaks around 300 nm practically unaffected (Figure 21.51) [140]. The spectra modiﬁed by 3 cannot be interpreted by a conformational change. The “irregular” middle portion of the dicoumarol ICD spectra was simulated by the additive contributions of two Gaussian components shifted by partial deprotonation of dicoumarol [133]. Dicoumarol is subject to a tautomer equilibrium between 4-hydroxy coumarin and 2-hydroxy chromone through a ketolactone structure [139] (Figure 21.52). Coumarin and chromone would be associated with different electric dipole transition moments. Which one is dominant within the binding site is not known. Further uncertainty comes from the possibility that the two halves of dicoumarol might have different structures. Hence, it is not yet possible to offer an adequate interpretation to the observed phenomena. Diazepam binds to AGP with inverse stereoselectivity [141] compared to HSA. Proof was obtained by chiroptical spectra (Figure 21.53). The conformational preference of diazepam by AGP could be assigned to variant F1-S; but in the case of Me2 DeMeD (cf. Figure 21.22), the preference by variant A was about threefold higher (Fig. 54). Nevertheless, similarly to HSA, the binding of S -3-LoAc was preferred by both variants, with higher selectivity again by variant A (not shown). The preferred conformer of diazepam could reasonably be docked [125] into the crystal structure of AGP variant F1 (Figure 21.55). Several studies aimed at identifying structural reasons for selectivity of AGP variants. Part of the efforts concentrated on assigning the role of tryptophan residues (W25, W122, W160) in speciﬁc binding interactions. First W160 was indicated to be responsible for the binding of an anticancer drug [142]. Progesterone binding was related to W122 [143]. Then W25 was assigned as the key residue being responsible for exciton-type ICD bands

Figure 21.50. Induced CD spectra of 1:1 dicoumarol/AGP F1-S (20 μM) modiﬁed by DER. The induced CD peaks around 275 nm and 325 nm are inverted by increasing concentration of DER, while no inversion takes place around 295–300 nm. Arrows indicate increasing DER concentrations. (Reprinted with permission from reference 140, copyright 2010, Elsevier.)

697

PROBING HSA AND AGP DRUG-BINDING SITES

Figure 21.51. Induced CD spectra of 1:1 dicoumarol/AGP F1-S (20 μM) modiﬁed by DER analogues 2 and 3. Incresasing concentration of 3 (arrow) abolishes the induced spectra to a large extent. Reprinted with permission from reference 140, copyright 2010, Elsevier.)

Figure 21.52. Tautomer equilibria between coumarin and chromone structures.

[144]. Trp25 is found at the bottom of the cavity (cf. Figure 21.43) and its accessibility is somewhat obstructed by the side chain of Tyr110 [125]. The side chain of Trp122 points outward from the binding site, just underneath the tip of loop 4 [125]. A further site-directed mutagenesis study indicated Glu92 to be responsible for the selectivity of variant A [123]. Quite recently, a C149R mutant of variant A was prepared and found to be equivalent to F1-S in certain binding interactions [145]. Despite the application of the most up-to-date techniques, these proposals may refer to particular cases. Until the complete picture of recognition unfolds, separated variants and selective probes should be used for characterization of the protein species. Here follows the example of imatinib, an important drug. Imatinib mesylate is a selective tyrosine kinase inhibitor successfully applied for the treatment of chronic leukemia. Its binding in blood plasma—mainly due to AGP—is strong enough to affect antitumor activity [146]. Imatinib contains ﬁve isolated rings lacking any chiral center (Figure 21.56). Its binding to native AGP (containing both variants), as well as to variants F1-S and A, has been studied by chiroptical spectroscopy (Figure 21.57) [147]. As seen, binding-induced spectra of imatinib on native AGP and

698

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 21.53. Induced CD spectra of diazepam and derivatives binding to AGP (solid line) and HSA (dashed line). Key: 1, diazepam; 2, DeMeD; 3, Me2 D; 4, Me2 DeMeD. For abbreviations see Figure 21.22. (Reprinted with permission from reference 141, copyright 2007, Elsevier.)

Figure 21.54. Induced CD spectra of diazepam (1) and Me2 DeMeD (4) binding to AGP variants. (Reprinted with permission from reference 141, copyright 2007, Elsevier.)

699

PROBING HSA AND AGP DRUG-BINDING SITES

Figure 21.55. Preferred conformer of diazepam docked into the crystal structure of AGP F1. Hydrogen bonds of the carbonyl oxygen to Glu64 and Gln66, as well as contacts of the ring nitrogens with Arg90 and Tyr127, are indicated. (Reprinted with permission from reference 125, copyright 2008, Elsevier.) (See insert for color representation of the ﬁgure.)

Figure 21.56. The chemical formula of imatinib mesylate.

those on variant F1-S are almost identical, while variant A makes a negligible contribution. Since variant F1-S is induced in the acute phase (for ill patients), this result is in accord with observations on the decrease of therapeutic activity in the presence of elevated AGP levels in blood plasma [146].

700

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

250.2

CD (mdeg)

10

[native AGP] = 51 μM

[AGP/F1-S] = 34 μM

[AGP/A] = 28 μM

0 −10 −20 −30

306.2

[L]/[P] 0.12 0.25 0.37 0.49 0.62 0.74

(a)

[L]/[P] 0.20 0.29 0.39 0.49 0.58 0.78 0.97

(b)

[L]/[P] 0.21 0.42 0.63 0.84 1.04

(c)

Figure 21.57. Induced CD spectra of imatinib mesylate binding to native AGP (a), and to variants AGP F1-S (b) and AGP A (c). (Reprinted with permission from reference 147, copyright 2006, Elsevier.)

21.4. CONCLUSION The concept of binding sites—emerging from hundreds of thousands of experiments on thousands of drugs—proved to be useful for classiﬁcation of individual studies and for creating a blueprint for further investigations. As shown, some drugs bind to multiple sites and the assignment then refers to the site of highest afﬁnity. Further complications can be seen when the low-afﬁnity sites interact with high-afﬁnity ones by modifying protein conformation and demonstrate that these sites are not entirely independent. The value of chiroptical spectroscopy mainly relates to identifying the source of chirality through which ligand conformation and its interaction with protein environment plays an important role. These can lead to determining even geometrical prameters, especially from exciton-coupled spectra. Nevertheless, CD spectroscopy alone without the knowledge accumulated by other techniques would not be powerful enough for characterizing binding sites. As a complementary technique, it offers useful contributions to discovering phenomena important theoretically, as well as for practical application in medicine. Note: After the completion of this chapter, the following paper appeared having basic relevance to the subject: Nishi K. et al. STRUCTURAL INSIGHTS INTO DIFFERENCES IN DRUG BINDING SELECTIVITY BETWEEN TWO FORMS OF HUMAN α1-ACID GLYCOPROTEIN GENETIC VARIANTS, THE A AND F1*S FORMS at http://www.jbc.org/cgi/doi/10.1074/jbc.M110.208926

ACKNOWLEDGMENT I thank my co-workers for their continuing interest spanning a period of more than 30 years.

REFERENCES 1. F. Zsila, Z. Bik´adi, I. Fitos, M. Simonyi, Current Drug Discovery Technologies 2004, 1 , 133–153. 2. G. A. Ascoli, E. Domenici, C. Bertucci, Chirality 2006, 18 , 667–679. 3. T. Pieters, Jr., All about Albumin, Academic Press, New York, 1996.

PROBING HSA AND AGP DRUG-BINDING SITES

4. Sudlow, D. J. Birkett, D. N. Wade, Mol. Pharmacol . 1976, 12 , 1052–1061. 5. I. Sj¨oholm, B. Ekman, A. Kober, I. Ljungsted- PAAhlman, B. Seiving, T. Sj¨odin, Mol. Pharmacol . 1979, 16 , 767–777. 6. P. P. Minghetti, D. E. Ruffner, W. J. Kuang, O. E. Dennison, J. W. Hawkins, W. G. Beattie, G. A. Dugaiczyk, J. Biol. Chem. 1986, 261 , 6747–6757. 7. X. M. He, D. C. Carter, Nature 1992, 358 , 209–215. 8. I. Petitpas, A. A. Bhattacharya, S. Twine, M. East, S. Curry, J. Biol. Chem. 2001, 276 , 22804–22809. 9. J. Ghuman, P. A. Zunszain, I. Petitpas, A. A. Bhattacharya, M. Otagiri, S. Curry, J. Mol. Biol . 2005, 353 , 38–52. 10. M. D. Maines, FASEB J . 1988, 2 , 2557–2568. 11. J. Jacobsen, FEBS Lett., 1969, 5 , 112–114. 12. J. Fog, E. Jellum, Nature 1963, 198 , 88–89. 13. R. Brodersen, H Floodgaard, J. Kragh-Hansen, Acta Chem. Scand . 1967, 21 , 2284–2285. 14. G. Blauer, T. E. King, J. Biol. Chem. 1970, 245 , 372–381. 15. G. Blauer, D. Harmatz, A. Naparstek, FEBS Lett. 1970, 9 , 53–55. 16. G. Blauer, D. Harmatz, J. Snir, Biochim. Biophys. Acta 1972, 278 , 68–88. 17. O. J. M. Bos, J. F. A. Labro, J. E. Fischer, J. Wilting, L. H. M. Janssen, J. Biol. Chem. 1989, 264 , 953–959. 18. R. Brodersen, CRC Crit. Rev. Clin. Lab. Sci . 1979, 305–399. 19. J. Jacobsen, R. Brodersen, J. Biol. Chem. 1983, 258 , 6319–6326. 20. D. A. Lightner, M. Reisinger, G. L. Landen, J. Biol. Chem. 1986, 13 , 6034–6038. 21. Harada, N., and Nakanishi, K. Circular Dichroic Spectroscopy—Exciton Coupling in Organic Stereochemistry, University Science Books, Mill Valley, CA, 1983. 22. D. A. Lightner, J. K. Gawronski, K. Gawronska, J. Am. Chem. Soc. 1985, 107 , 2456–2461. 23. D. A. Lightner, J. K. Gawronski, W. M. Wijekoon, J. Am. Chem. Soc. 1987, 109 , 6354–6362. 24. I. Sj¨oholm, M. Michaelsson, Clin Chim Acta 1974, 52 , 187–196. 25. C. E. Petersen, C-E. Ha, K. Harohalli, J. B. Feix, N. V. Bhagavan, J. Biol. Chem. 2000, 28 , 20985–20995. 26. A. Knudsen, A. O. Pedersen, R. Brodersen, Arch. Biochem. Biophys. 1986, 244 , 273–284. 27. R. Brodersen, J. Clin. Invest. 1974, 54 , 1353–1364. 28. F. Ebbesen, A. Knudsen, Eur. J. Pediatr. 1992, 151 , 910–912. 29. R. Brodersen, L. Stern, Acta Pediatrica 1990, 79 , 12–19. 30. S. Onishi, S. Itoh, K. Isobe, M. Ochi, T. Kunikata, T. Imai, Biochem. J . 1989, 257 , 711–714. 31. C. Cooper-Peel, R. Brodersen, A. Robertson, Pharmacol Toxicol . 1996, 79 , 297–299. 32. C. Diot, Y. Kibleur, L. Desfrere, Early Hum. Dev . 2010, 86 , 315–317. 33. L. Trynda-Lemiesz, Bioorg. Med. Chem. 2004, 12 , 3269–3275. 34. W. F. Trager, Coumarin anticoagulants, in Problems and Wonders of Chiral Molecules, M. Simonyi, eds., Akad´emiai Kiad´o, Budapest, 1990. 35. J. H. Perrin, D. A. Nelson, Life Sci ., 1972, 11 , 277–283. 36. O. J. M. Bos, J. P. M. Remijn, M. J. E. Fischer, J. Wilting, L. H. M. Janssen, Biochem. Pharmacol . 1988, 37 , 3905–3909. 37. N. A. Brown, E. Jahnchen, W. E. M¨uller, Mol. Pharmacol . 1977, 13 , 70–79. 38. I. Fitos, J. Visy, A. Magyar, J. Kajt´ar, M. Simonyi, Biochem. Pharm. 1989, 38 , 2259–2262. 39. E. J. Valente, E. C. Lingafelter, W. R. Porter, W. J. Trager, J. Med. Chem. 1977, 20 , 1489–1493. 40. E. J. Valente, W. R. Porter, W. J. Trager, J. Med. Chem. 1978, 21 , 231–234.

701

702

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

41. E. J. Valente, W. J. Trager, J. Med. Chem. 1978, 21 , 141–143. 42. C. Bertucci, A. Canepa, G. A. Ascoli, L. F. Lopes Guimaraes, G. Felix, Chirality 1999, 11 , 675–679. 43. M. Dockal, D. C. Carter, F. R¨uker, J. Biol. Chem. 1999, 274 , 29303–29310. 44. E. Middleton, C. Kandaswami, T. C. Theoharides, Pharmacol. Rev . 2000, 52 , 673–751. 45. M. I. Kaldas, U. K. Walle, H. van der Woude, J. M. McMillan, T. Walle, J. Agric. Food Chem. 2005, 53 , 4194–4197. 46. D. W. Boulton, U. K. Walle, T. Walle, J. Pharm. Pharmacol . 1998, 50 , 243–249. 47. A. Sytnik, I. Litvinyuk, Proc. Natl. Acad. Sci. USA 1996, 93 , 12959–12963. 48. F. Zsila, Z. Bik´adi, M. Simonyi, Biochem. Pharmacol . 2003, 65 , 447–456. 49. A. A. Bhattacharya, T. Gr¨une, S. Curry, J. Mol. Biol . 2000, 303 , 721–732. 50. U. Kragh-Hansen, Mol. Pharmacol . 1988, 34 , 160–171. 51. L. Di Bari, S. Ripoli, S. Pradhan, P. Salvadori, Chirality 2010, 22 , 593–596. 52. S. Harder, P. Th¨urmann, Clin Pharmacokinet. 1996, 30 , 416–444. 53. A. C. P. Reddy, E. Sudharshan, A. G. A. Rao, B. R. Lokesh, Lipids 1999, 34 , 1025–1029. 54. F. Zsila, Z. Bik´adi, M. Simonyi, Biochem. Biophys. Res. Commun. 2003, 301 , 776–782. 55. F. Zsila, Z. Bik´adi, M. Simonyi, Tetrahedron: Asymmetry 2003, 14 , 2433–2444. 56. I. Fitos, J. Visy, M. Simonyi, J. Hermansson, Chirality 1999, 11 , 115–120. 57. M. H. Rahman, T. Maruyama, T. Okada, T. Imai, M. Otagiri, Biochem. Pharmacol . 1993, 46 , 1733–1740. 58. L. H. Sternbach, R. I. Fryer, W. Metlesics, E. Reeder, G. Sach, G. Saucy, A. Stempel, J. Org. Chem. 1962, 27 , 3788–3796. 59. G. Curotto, D. Donati, G. Pentassuglia, A. Ursini, Bioorg. Med. Chem. Lett. 1995, 5 , 3011–3016. 60. L. H. Sternbach, F. D. Sancilio, J. F. Blount, J. Med. Chem. 1974, 374–377. 61. T. Sj¨odin, N. Roosdorp, I. Sj¨oholm, Biochem. Pharmacol . 1976, 25 , 2131–2140. 62. R. Brodersen, T. Sj¨odin, I. Sj¨oholm, J. Biol. Chem. 1977, 252 , 5067–5072. 63. V. Sunjic, F. Kajfez, I Stromar, N. Blazevic, D. Kolbah, J. Heterocycl. Chem. 1973, 10 , 591–599. 64. T. Alebic-Kolbah, F. Kajfez, S. Rendic, V. Sunjic, A. Konowal, G. Snatzke, Biochem. Pharmacol . 1979, 28 , 2457–2464. 65. E. Decorte, R. Toso, T. Fajdiga, G. Comisso, F. Moimas, A. Sega, V. Sunjic, A. Lisini, J. Heterocycl. Chem. 1983, 20 , 1321–1327. ¨ os, Bioorg. 66. M. Simonyi, G. Maksay, I. Kov´acs, Z. Tegyey, L. P´ark´anyi, A. K´alm´an, L. Otv¨ Chem. 1990, 18 , 1–12. 67. P. Linscheid, J.-M. Lehn, Bull. Soc. Chim. France, 1967, 992–997. 68. N. W. Gilman, P. Rosen, J. V. Earley, C. Cook, L. J. Todaro, J. Amer. Ch. Soc. 1990, 112 , 3969–3978. ¨ os, Archiv der Pharmazie 1986, 319 , 69. I. Fitos, M. Simonyi, Z. Tegyey, M. Kajt´ar, L. Otv¨ 744–749. ¨ ogdi, L. Kisfaludy, Tetrahedron Lett. 70. M. Czugler, A. K´alm´an, J. R¨ohricht, M. Low, L. Ur¨ 1977, 917–920. ´ 71. M. Kajt´ar, J. Kajt´ar, J. R¨ohricht, J. G. Angy´ an, Croat. Chem. Acta, 1989, 62 , 245–265. 72. I. Fitos, M. Simonyi, Experientia 1983, 39 , 591–592. 73. A. Konowal, G. Snatzke, T. Alebic-Kolbah, F. Kajfez, S. Rendic, V. Sunjic, Biochem. Pharmacol . 1979, 28 , 3109–3113. 74. E. L. Eliel, S. H. Wilen, Stereochemistry of Organic Compounds, John Wiley & Sons, New York, 1994, p. 1009.

PROBING HSA AND AGP DRUG-BINDING SITES

75. R. S. Cahn, C. Ingold, V. Prelog, Angew. Chem. 1966, 78 , 413–447. 76. G. Snatzke, A. Konowal, A. Sabljic, N. Blazevic, V. Sunjic, Croat. Chem. Acta 1982, 55 , 435–455. 77. E. Decorte, R. Toso, A. Sega, V. Sunjic, Z. Ruzic-Toros, B. Kojic-Prodic, N. Bresciani-Pahor, G. Nardin, L. Randaccio, Helv. Chim. Acta 1981, 64 , 1145–1149. 78. A. Avdagic, A. Lesac, Z. Majer, M. Holl´osi, V. Sunjic, Helv. Chim. Acta 1998, 81 , 1567–1582. 79. A. Avdagic, A. Lesac, V. Sunjic, Tetrahedron 1999, 55 , 1407–1416. 80. M. Majeric-Elenkov, D. Ziher, A. Visnjevac, Z. Hamersak, B. Kojic-Prodic, V. Sunjic, Croat. Chem. Acta 2001, 74 , 707–724. 81. W. E. M¨uller, U. Wollert, Mol. Pharmacol . 1975, 11 , 52–60. 82. I. Fitos, Z. Tegyey, M. Simonyi, I. Sj¨oholm, T. Larsson, C. Lagercrantz, Biochem. Pharmacol . 1986, 35 , 263–269. 83. M. Simonyi, I. Fitos, Z. Tegyey, J.C.S. Chem.Comm. 1980 1105–1106. 84. M. Simonyi, I. Fitos, J. Kajt´ar, M. Kajt´ar, Biochem. Biophys. Res. Commun. 1982, 109 , 851–857. ¨ os, J. Kajt´ar, M. Kajt´ar, J. Chromatogr. 1983, 85. I. Fitos, M. Simonyi, Z. Tegyey, L. Otv¨ 494–498. 86. R. W. Lucek, C. B. Coutinho, Mol. Pharmacol . 1976, 12 , 612–619. 87. I. Fitos, Z. Tegyey, M. Simonyi, M. Kajt´ar, in Bio-Organic Heterocycles—1986 , H. C. van der Plas, M. Simonyi, F. C. Alderweildt, and J. A. Lepoivre, eds., Elsevier, Amsterdam, 1986. 88. I. Fitos, M. Simonyi, J. Chromatogr. A, 1988, 450 , 217–220. 89. I. Fitos, J. Visy, A. Magyar, J. Kajt´ar, M. Kajt´ar, Chirality 1990, 2 , 161–166. 90. I. Fitos, M. Simonyi, Chirality 1992, 4 , 421–423. 91. E. Domenici, C. Bertucci, P. Salvadori, I. W. Wainer, J. Pharm. Sci . 1991, 80 , 164–166. 92. I. Szinai, M. Simonyi, I. Fitos, Z. Tegyey, A. Magyar, Eur. J. Drug Metab. Pharmacokin. 1990, 15 , 103–107. 93. M. Simonyi, I. Fitos, Biochem. Pharmacol . 1983, 32 , 1917–1920. 94. J. Visy, M. Simonyi, Chirality 1989, 1 , 271–275. 95. I. Fellegv´ari, J. Visy, K. Valk´o, T. L´ang, M. Simonyi, J. Liquid Chromatogr. 1989, 12 , 2719–2732. 96. F. Zsila, A. Gergely, P. Horv´ath, G. Sz´asz, J. Liq. Chrom. Rel. Technol . 1999, 22 , 713–719. 97. D. R¨omer, H. H. B¨uscher, R. C. Hill, R. Maurer, T. J. Petcher, H. Zeugner, W. Benson, E. Finner, W. Milkowski, P. W. Thies, Nature 1982, 298 , 759–760. 98. M. Simonyi, J. Visy, I. Kov´acs, M. Kajt´ar, in Bio-Organic Heterocycles, H. C. van der Plas, ¨ os, and M. Simonyi, eds., Akad´emiai Kiad´o-Elsevier, Budapest-Amsterdam, 1984. L. Otv¨ 99. V. K. Cheruvallath, C. M. Riley, S. R. Narayanan, S. Lindenbaum, J. H. Perrin, J. Pharm. Biomed. Anal . 1997, 15 , 1719–1724. 100. M. H. Rahman, T. Maruyama, T. Okada, K. Yamasaki, M. Otagiri, Biochem. Pharmacol ., 1993, 11 , 1721–1731. 101. M. H. Rahman, T. Maruyama, T. Okada, T. Imai, M. Otagiri, Biochem. Pharmacol ., 1993, 11 , 1733–1740. 102. I. Fitos, J. Visy, M. Simonyi, J. Hermansson, Chirality 1999, 11 , 115–120. 103. F. Bree, P. Nguyen, S. Urien, P. Riant, E. Albengres, H. Fenner, J.-P. Tillement, Fundam. Clin. Pharmacol . 1989, 3 , 267–279. 104. F. Br´ee, S. Urien, P. Nguyen, J-P. Tillement, A. Steiner, C. Vallat-Molliet, B. Testa, J. Visy, M. Simonyi, J. Pharm. Pharmacol . 1993, 45 , 1050–1053.

703

704

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. 125. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139.

A. A. Bhattacharya, T. Grune, S. Curry, J. Mol. Biol . 2000, 303 , 721–732. B. Sebille, N. Thuaud, J-P. Tillement, J. Brienne, Int. J. Biol. Macromol . 1984, 6 , 175–178. J. R. Simard, P. A. Zunszain, J. A. Hamilton, S. Curry, J. Mol. Biol . 2006, 361 , 336–351. F. Zsila, Z. Bik´adi, M. Simonyi, Tetrahedron: Asymmetry 2001, 12 , 3125–3137. M. Simonyi, Z. Bik´adi, F. Zsila, J. Deli, Chirality 2003, 15 , 680–698. P. A. Routledge, W. W. Stargel, G. S. Wagner, D. G. Shand, Br. J. Clin. Pharmacol . 1980, 9 , 438–440. K. Schmid, H. Kaufmann, S. Isemura, F. Bauer, J. Emura, T. Motoyama, M. Ishiguro, S. Nanno, Biochemistry 1973, 12 , 2711–2724. K. Schmid, W. B¨urgi, J. H. Collins, S. Nanno, Biochemistry 1974, 13 , 2694–2697. M. J. Ttreuheit, C. E. Costello, H. B. Halsall, Biochem J . 1992, 283 , 105–112. K. N´emeth, T. Kremmer, L. Kocsis, J. Visy, Chromatographia 2009, 69 , 1307–1313. M. Lemaire, J-P. Tillement, Biochem. Pharmacol . 1982, 31 , 359–365. M. L. Kornguth, L. G. Hutchins, B. S. Eichelman, Biochem. Pharmacol . 1981, 30 , 2435–2441. L. Dente, U. R¨uther, M. Tripodi, E. F. Wagner, R. Cortese, Genes Dev . 1988, 2 , 259–266. L. Tomei, C. B. Eap, P. Baumann, L. Dente, Hum. Genet. 1989, 84 , 89–91. F. Herve, G. Caron, J.-C. Duche P. Gaillard, N. A. Rahman, A. Tsantili-Kakoulidou, P. A. Carrupt, P. d’Athis, J.-P. Tillement, B. Testa, Mol. Pharmacol . 1998, 54 , 129–138. F. Herve, E. Gomas, J-C. Duche, J-P. Tillement, Br. J. Clin. Pharmacol . 1993, 36 , 241–249. F. Herve, E. Gomes, J.-C. Duche, J-P. Tillement, J. Chromatogr. B , 1993, 615 , 47–57. F. Herve, J-C. Duche, P. D’Athis, C. Marche, J. Barre, J.-P. Tillement, Pharmacogenetics 1996, 6 , 403–415. K. Nishi, M. Ueno, Y. Murakami, N. Fukunaga, T. Akuta, D. Kadowaki, H. Watanabe, A. Suenaga, T. Maruyama, M. Otagiri, J. Pharm. Sci . 2009, 98 , 4316–4326. P. A. Ojala, M. Hermansson, M. Tolvanen, K. Polvinen, T. Hirvonen, U. Impola, M. Jauhiainen, P. Somerhajru J, Parkkinen, Biochemistry 2006, 45 , 14021–14031. D. L. Sch¨onfeld, R. B. G. Ravelli, U. M¨uller, A. Skerra, J. Mol. Biol . 2008, 384 , 393–405. U. K. Walle, T, Walle, S. A. Bai, L. S. Olanoff, Clin. Pharmacol. Ther. 1983, 34 , 718–723. F. Albani, R. Riva, M. Contin, A. Baruzzi, Br. J. Clin. Pharmacol . 1984, 18 , 244–246. I. Fitos, J. Visy, M. Simonyi, J. Hermansson, Chirality 1993, 5 , 346–349. W. E. M¨uller, in Drug Stereochemistry Analytical Methods and Pharmacology, I. W. Wainer, D. E. Drayer, eds., Marcel Dekker, New York, 1988. F. Brunner, W. E. M¨uller, J. Pharm. Pharmacol . 1987, 39 , 986–990. I. Fitos, J. Visy, M. Simonyi, J. Hermansson, J. Chromatogr. A 1992, 609 , 163–171. T. Maruyama, M. Otagiri, A. Takadate, Chem. Pharm. Bull . 1990, 38 , 1688–1691. I. Fitos, J. Visy, F. Zsila, G. M´ady, M. Simonyi, Biochem. Pharmacol . 2004, 67 , 679–688. K. Murakami, K. Mizuguchi, Y. Kubota, Y. Fujisaki, Bull. Chem. Soc. Japan 1986, 59 , 3393–3397. D. Fornasiero, T. Kurucsev, J. Phys. Chem. 1981, 85 , 613–618. M. Otagiri, R. Yamamichi, T. Imai, Y. Imamura, A. Takadate, Chem. Pharm. Bull . 1988, 36 , 4958–4962. M. Otagiri, T. Miyoshi, R. Yamamichi, T. Maruyama, J. H. Perrin, Biochem. Pharmacol . 1991, 42 , 729–733. T. Miyoshi, R. Yamamichi, T. Maruyama, A. Takadate, M. Otagiri, Biochem. Pharmacol . 1992, 43 , 2161–2167. D. W. Hutchinson, J. A. Tomlinson, Tetrahedron 1969, 25 , 2531–2537.

PROBING HSA AND AGP DRUG-BINDING SITES

140. I. Fitos, J. Visy, M. Simonyi, G. M´ady, F. Zsila, Biochim. Biophys. Acta 2010, 1800 , 367–372. 141. I. Fitos, J. Visy, F. Zsila, G. M´ady, M. Simonyi, Bioorg. Med. Chem. 2007, 15 , 4857–4862. 142. M. Katsuki, V. T. G. Chuang, K. Nishi, K. Kawahara, H. Nakayama, N. Yamaotsu, S. Hirono, M. Otagiri, J. Biol. Chem. 2005, 280 , 1384–1391. 143. J. R. Albani, Carbohydrate Res. 2006, 341 , 2557–2564. 144. F. Zsila, Y. Iwao, Biochim. Biophys. Acta 2007, 1770 , 797–809. 145. K. Nishi, N. Fukunaga, T. Ono, T. Akuta, N. Yumita, H. Watanabe, D. Kadowaki, A. Suenaga, T. Maruyama, M. Otagiri, Drug Metab. Pharmacokinet. 2010, 25 , 200–207. 146. C. Gambacorti-Passerini, M. Zucchetti, D. Russo, R. Frapolli, M. Verga, S. Bungaro, L. Tornaghi, F. Rossi, P. Pioltelli, E. Pogliani, D. Alberti, G. Corneo, M. D’Incalci, Clin. Cancer Res. 2003, 9 , 625–632. 147. I. Fitos, J. Visy, F. Zsila, G. M´ady, M. Simonyi, Biochim. Biophys. Acta 2006, 1760 , 1704–1712.

705

22 CONFORMATIONAL STUDIES OF BIOPOLYMERS, PEPTIDES, PROTEINS, AND NUCLEIC ACIDS. A ROLE FOR VIBRATIONAL CIRCULAR DICHROISM Timothy A. Keiderling and Ahmed Lakhani

22.1. INTRODUCTION Vibrational circular dichroism (VCD) is CD—that is, differential absorption of leftand right-circularly polarized light (A = AL − AR ), of molecular vibrational transitions which are typically measured in the infrared (IR) region of the spectrum—and has been the subject of many reviews [1–24]. VCD samples the same transitions as in IR spectra but has a different intensity distribution, since VCD arises from chiral interactions of vibrations of molecular bonds (nuclear motions). While absolute conﬁguration determination has evolved to be a major application of VCD [3, 7, 25], this is not an important question for biopolymers, since amino acids in proteins and ribose rings in nucleic acids have a known conﬁguration. In biopolymers, the relative conformation of successive residues in a chain often dominates structural studies and determines its secondary structure (helical, extended, etc.). Since this relative orientation affects the mechanical coupling of molecular vibrations as well as their aggregate transition dipole moments, it impacts both IR and VCD. The important coupled modes can be thought of as the polymer IR chromophores and differ from those which might be important in the monomer VCD. Thus, for example, amino acid VCD is not very relevant for protein analyses, nor is that of an isolated amide function, but VCD of model peptides, especially of amide groups, are the basis of structure–spectra correlations. Consequently, one of the major applications of IR and VCD spectra in biopolymer structural studies has been determination of the dominant or of average fractional secondary structure content (most often of peptides and proteins, but occasionally for nucleic acids) [8, 9, 13, 15, 17, 18, 21, 22, 26–46]. Contributions to the spectrum from different secondary structural types tend to overlap; thus, these Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

707

708

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

studies can typically only yield fractional components (such as % helix, etc.). IR and VCD have some advantages and disadvantages as compared to other spectral techniques for secondary structure analyses, but are best used in conjunction with electronic or UVCD (ECD), Raman, and magnetic resonance (NMR) to exploit complementary structural sensitivities available. This focus on average secondary structure is a consequence of modest vibrational interactions that alter shapes but do not resolve components of those spectra. Consequently, IR and VCD do not yield site-speciﬁc information without selective isotopic substitution [17, 22, 47–52]. X-ray crystallography or NMR spectroscopy do provide sitespeciﬁc structural information, but both require extended (time-consuming) data analyses and involve complex interpretive processes. They are also normally limited to relatively small, soluble or crystallizable biomolecules and require a substantial amount of sample. Furthermore, neither one is well-suited for rapid real-time measurements for analyses of dynamic structures or of conformations undergoing fast changes. The dominant technique for secondary structure analyses, particularly of peptides and proteins, has been far-UV ECD of π –π ∗ and n –π ∗ transitions of the amide group in peptides [53–60] (Chapters 14 and 15, this volume) or the bases in nucleic acids [61–64] (Chapter 17, this volume). ECD’s sensitivity to molecular structure, resulting from conformationally determined changes in sign and bandshape, and its ability to study relatively small amounts and low concentrations of sample give it an advantage. Since ECD bands are often unresolved, and those corresponding to different conformations are totally overlapped, the usual interpretive methods employ bandshape pattern-recognitionbased algorithms that are either qualitative in nature or dependent on a statistical ﬁt to a set of (typically protein) spectra which provide a structural reference set [29, 53, 57, 58, 60, 65–70]. IR and Raman analyses of secondary structure historically took a different approach [71–78], due to the natural resolution of the spectrum into contributions from vibrational modes characteristic of different bond types in the molecule. The focus has been the assignment of bands or their component frequencies to various secondary structural types, as discussed in several reviews [31, 48, 78–82]. Most protein analyses used the amide I band (C=O stretch, ∼1610–1690 cm−1 ) with some application of the amide II (NH deformation plus C–N stretch, ∼1500–1550 cm−1 ) in IR spectra and especially the amide III (oppositely phased N–H deformation plus C–N stretch, ∼1330–1220 cm−1 ) with Raman methods. In nonaqueous media, the amide A (N–H stretch, ∼3300 cm−1 ) can be useful for determining the extent of H-bonding. For DNA/RNA the typical bands used for IR studies are C=O stretches and the coupled in-plane base deformation modes (whose most intense bands have frequencies varying from ∼1690 to 1620 cm−1 , depending on the base and the level of H-bond formation, duplex or single-strand), the phosphate, asym and sym –PO2 − , modes at ∼1250 and 1070 cm−1 , respectively, and some selected ribose deformations [9, 83–99]. A major factor in single-strand to duplex (or triple-strand) formation is the change in vibrational frequency of the base modes that directly reﬂect the cross-strand H-bond formation. The correct tautomeric forms of the bases were also identiﬁed by vibrational spectral patterns [100]. By contrast, the phosphate modes are quite stable, with some complexity from overlap and mixing with ribose (C–C and C–O) modes, which can shift the patterns. These features and those characteristic of ribose conformational variation have been used to identify the polymeric conformation of various nucleic acids [9, 83, 85, 101–103]. Further vibrational spectroscopic applications for nucleic acids

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

have used Raman spectroscopy, focused on the more intense modes of the polarizable bases [104–109]. Since both IR and Raman techniques give rise to single-signed spectral bandshapes that are effectively just the dispersed sum of the contributions from all the component transitions, and components of amide transitions differ in frequency by only relatively small intervals compared to the bandwidths, the bandshapes for different proteins are very similar. The same is true for the –PO2 − modes in nucleic acids, but the base modes in DNAs or RNAs are quite sequence-speciﬁc, since the various bases have different characteristic dipole-allowed modes [86–88, 92]. However, due to the high signal-to-noise ratio (S/N) of Fourier transform IR (FTIR), resolution enhancement using second-derivative or Fourier self-deconvolution (FSD) techniques [73, 110, 111] can add strength to the interpretation. Since it is computational spectral enhancement and not based on physical interactions, FSD can be misused [30, 31, 112]. Solvent as well as secondary structural segment nonuniformity and end effects have an impact on frequencies, dispersing the contribution of each component over a signiﬁcant spectral range, leading to real difﬁculties with the simplifying assumptions typically used for protein band assignments [74, 113–117]. For example, most methods depend on an assumption that the dipole strengths (extinction coefﬁcients) of all the residues are the same, regardless of their conformation, while they in fact vary [31, 118, 119]. Nonetheless, surprisingly accurate secondary structure analyses have appeared. Bandshape-based analyses, similar to those used for ECD, have also been applied to FTIR and Raman spectra with reasonable success [29, 32, 74, 76, 77, 120–125]. The desire to combine the sensitivity of ECD bandshapes to structure with the natural resolution of vibrational band components led to the development of vibrational CD (VCD) (Chapter 5, Volume 1) and its counterpart, Raman optical activity (ROA) (Chapter 6, Volume 1; and Chapter 23, this volume). These measure the same transitions as IR and Raman, but have a conformationally dependent sign variation, resulting in distinct bandshapes with an enhanced resolution of contributions having a physical basis. Only VCD will be addressed here, but other reviews dealing with ROA are widely available [5, 126–129]. The key impetus for moving to the vibrational region of the spectrum is its distribution of multiple resolved transitions, which are characteristic of localized parts of the molecule. By probing their chirality, VCD measurements can expose the distinct stereochemical sensitivity of these vibrational modes [4, 5, 15–17 20–22, 33, 35, 126, 130–134]. VCD chromophores are the bonds themselves whose transitions involve excitations of stretching and bond angle deformation. Often the most intense transitions in IR correspond to motions of planar parts of the molecule, which will be locally achiral and thus give rise to little intrinsic VCD. However, the secondary structure in a polymer can create a chiral interaction between these achiral repeating segments, and the resulting VCD spectral bandshape for these modes thus will reﬂect the polymeric structural character. VCD does have reduced S/N compared to IR and some added interpretive difﬁculty. Developments in the theoretical capability for prediction of VCD spectra have enhanced interpretive methods for small molecules to a level that is demonstrably superior to that for ECD spectra [2, 135–137]. We have been able to extend such reliable calculational methods to moderately large peptides, with some assumptions [49, 138–143], and others have done so for nucleic acids and small proteins [144–147]. These methods originally involved transfer of force ﬁeld (FF) parameters, but with improvements in computers and algorithms, IR, Raman, VCD, and ROA spectra can be computed directly at least for sizeable peptides [148–151].

709

710

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Experimentally, instrumentation has reached a stage where VCD spectra for most molecular and many complex systems of interest (even in multiple phases) can be measured under various sampling conditions [1, 17, 152–169]. Most VCD studies of biomolecules are done in aqueous solutions, the condition of prime biological interest, and require somewhat high concentrations, as compared to ECD, but of the same order as typical of NMR. Often D2 O-based solvents are used to shift the strongly interfering HOH bend down from 1650 cm−1 (location of the peptide amide I and vicinity of some C=O stretches in nucleic acids) and improve S/N as well as permit study of more dilute solutions, all at the cost of added complications due to H–D exchange. (ROA has less H2 O interference, but its measurement tends to demand higher concentrations [128, 129, 170].) Sampling conditions for obtaining experimental VCD spectra of protein and peptide samples are similar to those used in FTIR studies, with the important exception that the data are differential spectra of much smaller amplitude (A is the order of 10−4 of the sample absorbance, A) and thus lower S/N. Similarly, as compared to ECD, VCD requires higher sample concentrations, due to lower dipole strength for vibrations, and shorter paths, due to solvent interference. Consequently, to obtain quality VCD spectra, longer data collection times are required than for FTIR or ECD, and specially designed instruments are used. Instrumentation and sampling methods for biomolecular VCD are summarized in the next section and are more fully discussed in separate reviews [4, 5, 22, 35, 126, 153–155, 160] and various instrument development papers [156, 157, 159, 163, 166, 171, 172]. Theoretical techniques for simulation of small molecule VCD are also the focus of several previous reviews [1, 2, 5, 12, 25, 33, 130, 132, 135–137, 173] and will not be covered in detail here. A survey of methods for biopolymer VCD spectral simulation is presented in a following section [140, 141]. On the other hand, qualitative analyses of secondary structure for any size biopolymer can be easily done utilizing the VCD bandshape and its frequency position, assuming there is a dominant uniform structural type [15, 17, 18, 161]. For small, fully solvated biopolymers, the frequency shifts due to the inhomogeneity of the secondary structure pose problems for frequency-based analyses (FTIR, Raman). However, VCD bandshapes are conserved, since they arise from interactions between local modes; and they shift with the absorption bands, permitting relatively simple analyses, when correlated to the absorbance, often in terms of local helicity or type of coupling along the backbone, especially for nucleic acids. For globular proteins, qualitative estimations of structure can be of interest for determining the dominant fold type, but quantitative estimates of fractional secondary structure content are usually of more interest. Such methods for analysis of protein VCD spectra follow the quantitative methods established for ECD analyses [53, 54, 58, 60, 65, 68–70, 174] and employ spectral decompositions and ﬁts to a training set of protein spectra [26, 27, 29, 175–179]. In summary, unlike ECD, VCD can be used to correlate data for several different spectrally resolved features; and, unlike IR and Raman spectroscopies, each of these features will have at its source a physical dependence on stereochemistry. The price of this added sensitivity is somewhat lower S/N as compared to IR, need for higher concentration than for ECD, and different solvent interferences. Use of a combination of these methods in analyzing a biomolecular structure can compensate for the weaknesses of each, providing a balance between accuracy and reliability [17, 21, 26, 30]. The prime questions in the VCD ﬁeld relate to application and interpretation of the resulting spectra. In our biomolecular work, different types of spectral data are used to place bounds on or add credence to the reliability of structural inferences that might be drawn

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

from any one method alone. In particular, use of static structures obtained from X ray and NMR provide reference points for native states and for computations of spectra for ideal structures allowing analyses of equilibrium results. Techniques such as CD and ﬂuorescence can also follow rapid changes in tertiary structure, and those of IR, VCD, Raman and ECD can follow dynamic variations in secondary structure under various conditions. Correlating spectral results from different sources leads to interpretability of dynamic changes in biochemical processes as well as folding in protein and nucleic acid systems. Modeling the spectra for the ensemble of conformations resulting from such changes is more challenging, but coupling QM and MD methods is leading to some useful insight [148, 149, 180–183].

22.2. EXPERIMENTAL METHODS Extending optical activity measurements into the mid-IR necessitates special design considerations, in that the rotational strengths of vibrational transitions as detected in VCD are much weaker than those of electronic transitions detected in ECD, and IR sources are weaker and detectors are noisier than those available for UV–vis studies. Similarly, since VCD is a differential IR technique, its S/N can never approach that of FTIR, which represents a summed response. Several research groups have developed instrumentation that makes the measurement of VCD reasonably routine over much of the IR region [153–157, 159, 160, 171, 172, 184–190]. Commercial FTIR vendors now provide high-quality VCD accessories [191–197] or stand-alone VCD instruments that have exceptional S/N and can be modiﬁed to have very good baseline characteristics [153, 172, 198–201]. In this section, general VCD instrument designs are brieﬂy summarized and compared, but more detail is available in separate reviews dedicated to instrumentation and technique issues [22, 35, 126, 154, 155, 160, 202].

22.2.1. Fourier Transform and Dispersive Instruments VCD instruments are normally modiﬁcations of a dispersive IR or an FTIR spectrometer to incorporate time-varying modulation of the polarization state of the light and detection of the resulting intensity oscillation. Various instruments have been described in the literature in detail [1, 5, 154, 155, 160, 203], and detailed reviews contrasting designs are available [22, 155]. Recently, nonlinear IR-based approaches to CD have been developed [204–206] (Chapter 8, Volume 1). While these are currently beyond the scope of this review, they offer the future potential for obtaining very fast timescale dynamic VCD spectra to allow fast biochemical processes like protein folding events to be followed. Two methods, dispersive and Fourier transform (FT), are used to encode the optical frequencies. Dispersive VCD instruments use a grating monochromator that must scan through the wavelength region of interest, recording the spectrum sequentially. By comparison, FT-VCD instruments use a Michelson interferometer and collect all frequencies simultaneously (multiplex advantage). Dispersive instruments use analog time constants and ﬁltering to improve S/N, as well as averaging of sequential spectral collections, while FT-VCD normally average many (rapid-scan) interferograms to gain better S/N. Both types of instrument work well for their designed purposes. For conformationally constrained molecules in nonaqueous solutions, FT-VCD can produce high-quality spectra over the entire mid-IR, typically between 2000 cm−1 and 700 cm−1 , with good resolution (typically 4 cm−1 ) in a reasonable time, due to its multiplex advantage. Thus, since most

711

712

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

VCD applications are for small molecules and absolute stereochemistry determinations, FT-VCD has come to dominate the market, much as has FTIR. Initial VCD instruments were dispersive, but after the development of FT-VCD, these could not compete for covering a broad wavenumber region, due to the slow scans required. However, a dispersive instrument can be optimized for high efﬁciency and response in a speciﬁc spectral region, to measure VCD for a single band with very high S/N and a good-quality baseline at modest resolution (∼8–10 cm−1 ). Since this is often just what is needed for biopolymer studies, particularly of peptides and proteins, dispersive VCD still has a select role. Both styles of VCD instrument normally use a blackbody light source, typically a heated ceramic rod, for broad spectral coverage. The wavelength-encoded output is passed through (a) a wire grid polarizer to linearly polarize the light and (b) a photoelastic modulator (PEM), typically made of ZnSe, to phase retard it, yielding elliptically polarized light whose degree of circularity varies with IR wavelength. The beam then passes through the sample and onto the detector, typically a liquid N2 -cooled Hg1−x (Cd)x Te (MCT) photoconducting diode. The optical design imposes two modulations on the signal: (a) one at low frequencies created either by a chopper or by variations in the interference caused by the scanning mirror motion (varying path difference) and (b) another at higher frequency produced by the PEM. These modulations are electronically separated by ﬁlters into two channels for the transmitted intensity (Itrans ) of instrument and sample and for the polarization modulation intensity (Imod ) due to the sample VCD. Lock-in ampliﬁers demodulate both signals in a dispersive instrument, but just Imod in an FT-VCD while the computer digitizes and Fourier transforms Itrans . The raw VCD signal is the ratio, Imod /Itrans . Since VCD is a differential absorbance measurement, A = AL − AR , it is necessary to ratio these two intensities to obtain a signal proportional to A and normalize for instrument transmission. In the limit of small A values, Imod /Itrans = (1.15A)J1 (α0 )gI ,

(22.1)

where J1 (α0 ) is the ﬁrst-order Bessel function at α0 (λ), the maximum phase retardation (expressed as an angle) of the modulator which depends on the wavelength λ, and gI is an instrument gain factor. Calibration to eliminate the gain factors and the modulator λ dependence is typically done using a birefringent plate and polarizer pair or by measuring the VCD of a known sample [22, 152, 154, 155, 157, 158, 189]. Baseline correction and spectral averaging or smoothing can complete processing, and spectra can be converted to molar quantities; for example, ε = A/bc, where b is the pathlength in centimeters and c is the concentration in moles per liter, as desired. Dispersive VCD. An improved dispersive instrument has been constructed at UIC [171] around a 0.3-m-focal-length, ∼f /4 monochromator (Acton Research, SpectraPro2300i) operated in ﬁrst order (∼5μ, long-wave pass ﬁlter). A water-cooled, 2500 K carbon rod provides IR radiation that is focused onto the entrance slit [155, 184] and passed through a chopper to modulate the transmission allowing detection with an MCT detector. A schematic of our design is shown in Figure 22.1a, which is optimized for biopolymer VCD spectra and is a reﬁnement of historical designs in our and other labs [22, 152, 155, 160, 184, 186–188, 190, 207, 208]. For applications in the amide I and II regions, a BaF2 -substrate wire grid polarizer (Cambridge Physical Sciences) and a 57-kHz CaF2 PEM (Hinds Instruments) are placed before the sample in a weakly focused beam (BaF2 lens) that exits the monochromator.

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

Spherical Mirror (M)

713

C-rod source (S)

Mechanical chopper (C) Flat mirror (M)

G

L1 F

P PEM SC

L2

D

0.3 m, f/4 Monochromator

Figure 22.1. Schematics of the main optical components of (a) dispersive VCD and (b) FT-VCD instruments. These

(a) FTIR Schematic FM

M

include in common: (S) light source, (M) mirrors, (F) ﬁlter, (P) wire grid polarizer, (PEM) photoelastic modulator, (SC)

S A

sample cell, and (L) lens; with a (C) rotating wheel chopper (150 Hz),

External VCD Bench MM D M

M

BS

L

F

SC

External

Flip Mirror

P PEM

M (b)

L

D

monochromator and (G) grating (150 grooves/mm) for dispersive VCD; and the interferometer containing (FM) ﬁxed mirror; (MM) moving mirror; and (BS) beam splitter for FT-VCD.

Following the sample, an f /1 BaF2 lens focuses the light onto a narrow-band (cutoff at 8μ), high-D∗ MCT detector (Infrared Associates). (Previous, still operational designs measured VCD to <900 cm−1 with ZnSe modulators and broader-band MCT detectors, made as small arrays stacked to match the slit image [22, 155].) High sensitivity in the near-IR (∼5000–1900 cm−1 ) is possible with an InSb photovoltaic detector and use of a higher groove density grating [155, 203, 209–211]. To process the signal, a lock-in ampliﬁer converts the transmitted intensity in phase with the chopping frequency, ωC , to a DC voltage, Vtrans , and a second one measures the polarization modulation intensity, phase referenced to the PEM frequency, ωPEM . Since the VCD depends on the light level, the Imod lock-in signal, Vout , can be demodulated again with a third lock-in referenced to the chopper, ωC , yielding Vmod . Both Vmod and Vtrans , are digitized (we use the A/D channels in the Stanford SR830 lock-in which measures Imod ) and ratioed in the controlling computer (with a LabVIEW software routine). This processing method, ﬁrst utilized by Diem and co-workers [160, 186, 188, 190], replaces an analogue normalization method [152, 155, 207, 208] and has advantages of more dynamic range (enabling VCD measurement for highly absorbing samples, A > 1.0) and simultaneous collection of absorbance and VCD in each scan. FT-VCD. FTIR-VCD spectrometers are commercially available from a number of vendors, of which BioTools, Bruker, and Jasco are three examples, and have also been

714

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

constructed in several labs [5, 153, 155–157, 159, 172, 189, 191–193, 202, 212–221]. These can be conﬁgured as accessories to a conventional FTIR or as stand-alone instruments. The UIC FT-VCD was originally assembled around a Digilab FTS-60A FTIR [155–157, 159, 189], but is not used for biomolecular VCD in our lab, so it will not be described in any detail. Other very successful instruments have been conﬁgured around a number of different FTIRs [185, 213, 217–220]; and some of them, including some commercial designs, have yielded quite reliable and high S/N VCD spectra for biomolecules. The heart of these instruments is an FTIR, typically of rapid-scan design, including a coated KBr beam splitter and a glowbar or other ceramic source. The FT-modulated beam is passed through a polarizer (wire grid) and PEM (ZnSe) and on to a sample as a collimated or weakly focused beam [159, 189, 214], after which it is focused onto an MCT detector [5, 22, 155, 202]. Use of lenses (ZnSe or BaF2 ) in the optical train can have some advantage for control of artifacts. The optics are functionally the same for most instruments, as schematically indicated in Figure 22.1b, but have some variations in sampling and electronics. The output of the detector is ﬁltered and split, allowing the low-frequency modulation to be digitized for a conventional interferogram of the transmission and allowing the highfrequency components to be demodulated with a lock-in using a τ ∼ ms time constant to shift the carrier, ωM , to DC such that its sidebands provide an interferogram of the polarization modulated signal. Both the transmission and modulated signals are processed in the FT computer, often using a transferred phase correction, and ratioed to give the VCD spectrum [155, 157, 159, 189, 202, 217]. (Some instruments have been modiﬁed for the near IR, with good results, but require processing of higher frequencies or use of slower scan speeds [157, 163, 166, 213–216, 220].) Modiﬁcations have appeared, with varying results. Digital signal processing can eliminate the lock-in; however, in our hands it gave little advantage but added a cost in terms of slow data collection with step scanning [156, 157, 220, 222]. With newer instrumentation advances, an alternative method might prove advantageous. Due to the high FTIR light intensity levels, MCT detectors and preamps can saturate, leading to a nonlinear response. Optical ﬁlters (e.g. 1900, cm−1 cutoff low-pass) are used to isolate the mid-IR spectral region of interest; and for aqueous, biological samples, the spectral band pass is additionally strongly limited at lower wavenumbers by the solvent absorbance, so that detector saturation is not a signiﬁcant problem even with high source powers and large apertures. A dual source design, offered by BioTools, making use of corner-cube reﬂectors, offers an alternative method that cancels most of the centerburst and alleviates part of the saturation [172]. Alternatively, MCT detectors with nonlinear corrections are available [223, 224]. A dual modulator system, designed by Naﬁe and available from BioTools, can eliminate many baseline artifacts arising from birefringence in the sample path [153, 212]. Comparison. While FT-VCD has many potential advantages for the study of biopolymer systems, in practice, the relatively broad bands and the restriction to measurement of data only in the spectral windows of water offset the throughput and multiplex advantages of FTIR. This can result in dispersive VCD being more useful in practice, since it focuses on a single spectral band or narrow region of interest (especially for proteins) and, with proper optical design, can have very high throughput, thereby offsetting both main advantages of FTIR [22, 155, 157, 171, 225]. In summary, modern FTIR-based VCD instruments are the method of choice for VCD over a wide spectral range, typically for studies of small molecules in organic

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

715

solvents, but for measurement of a single broad band, dispersive instruments (e.g., the new UIC instrument [171]) still have an advantage. A comparison of VCD measured for bovine serum albumin (BSA), in the same total time and for the same resolution, for the amide I mode on both a modern FT-VCD instrument and our dispersive instrument, are compared in Figure 22.2. Since the dispersive instrument samples the signal at least every cm−1 , the spectra sometimes appear to have rapid noise ﬂuctuations which can be smoothed out with no loss of resolution, but this was not needed for the BSA spectrum in Figure 22.2 [22, 156, 171]. By contrast, the 8-cm−1 resolution FT-VCD often results in a smoother representation, but the ﬂuctuations look like real spectral bands, so a proper comparison requires analysis of the ﬂuctuations in the noise traces, as shown at the top of Figure 22.2. These conﬁrm the S/N advantage of the improved dispersive method. The small difference in overall shape may indicate a residual baseline or absorbance artifact or may be due to resolution. Nonetheless, dispersive VCD spectra with τ ∼ 10 s and resolution of ∼10 cm−1 result in slow scans (15–30 min for the amide I and II) and are usually averaged for a number of repeated scans [49, 50, 171, 186, 188, 226, 227]. FTIR-VCD spectra can sample a much wider spectral region in a short time (mirror scan), but require

Noise (×105)

4 3 2 1 0 −1 12

New VCD (7 cm−1) New VCD (8 cm−1) (c)

Chiral-IR

New VCD (7 cm−1)

10

ΔA × 105

8 6 4 2

(b) New VCD (8

cm−1)

0 −2 −4

Chiral-IR

Figure 22.2. Comparison with FTIR absorbance (a) of VCD (b) measured for BSA in D2 O, using the same

Absorbance

0.5

(a)

0.4 0.3 0.2

concentration and pathlength and the same total amount of time, on an FT-VCD (BioTools Chiral-IR) and on the UIC dispersive instrument. Upper traces (c) indicate noise levels. The resolution of the FT-VCD appears higher than for the 8-cm−1 dispersive spectra, and the remeasured

0.1 0.0 1700

1650 1600 Wavenumber (cm−1)

7-cm−1 VCD partially corrects for lost resolution, as seen in the positive component (it had no further change at 6 cm−1 ).

716

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

extensive averaging to obtain adequate S/N for aqueous phase biopolymers. Modern rapid-scan FTIRs efﬁciently sample the interferogram, so there is no real cost for measuring larger spectral ranges, but higher resolution requires longer collection times and yields lower S/N. It can be noted that dispersive data are intuitively easier to interpret, particularly if something is wrong, such as spurious signals (artifacts or noise) [171, 188]. either type of VCD measurement is now compatible with monitoring fast kinetic events. Due to its intrinsically weak intensity, VCD is subject to artifacts that must be corrected by careful baseline subtraction. A racemic sample of the same material would provide an ideal baseline determination, but that is not normally practical for biopolymers. Satisfactory corrections can often be obtained from VCD of the same sample cell containing just solvent, particularly for short-path, aqueous solutions. While baseline and artifact variation due to change of sample cell can be an issue, especially for FT-VCD instruments, we have found that homemade cells consisting of two thin CaF2 windows separated with a spacer and clamped in a metal holder typically work interchangeably in our dispersive instrument (Figure 22.3a) [228]. This design can be thermally controlled by ﬁtting it into a jacketed holder whose temperature is set by ﬂow from a thermostatted bath (Figure 22.3b). Finally, most artifacts (false signals) can be minimized by careful optical adjustments, and residual baseline characteristics can normally be corrected by subtraction of a spectrum obtained with a blank.

out

in

Thermocouple

Teflon spacer

Brass body

Teflon Holder

Teflon Holder

Outer window

Outer window

Screw

O-rings

O-rings

Screw Windows (a)

(b)

Figure 22.3. (a) Schematic of demountable IR/VCD cell, with CaF2 windows separated by a Teﬂon spacer and sealed in a brass ring assembly, shown blown apart to indicate assembly. (b) Thermal control is done by inserting the cell in a double-walled brass holder having outer windows to even out temperature variation and control condensation. Flow from a thermostatted bath through in/out ports (top) is used to control temperature, as monitored at the sample with a thermocouple probe.

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

22.2.2. Sampling Considerations for Biopolymer VCD Aqueous environments, ideal for biomolecules, pose difﬁculties because H2 O has strong fundamental IR transitions that directly overlap regions of interest such as the N–H and C=O stretches. Consequently, peptide and protein amide I (primes indicate N-deuterated amides) VCD at ∼1650 cm−1 , dominated by the C=O stretch, are normally measured in D2 O, as are base-centered C=O VCD in DNA and RNA. On the other hand, due to their contributions from N–H deformations, the amide II at ∼1550 cm−1 and amide III at ∼1300 cm−1 in peptides and the –PO2 − at 1250 and 1070 cm−1 in nucleic acids are best studied in H2 O where they have less interference and no isotope shift. VCD samples in D2 O can typically be prepared at concentrations in the range of 5–50 mg/mL, depending on the pathlength and S/N needed. For our cells, an aliquot (typically 20–40 μL) is placed between a pair of CaF2 or BaF2 windows (25-mm diameter, 2 mm thick) separated by a 25- to 100-μm spacer (e.g., Teﬂon), as shown in Figure 22.3a. Many commercial cell designs are available, including special windows ground to avoid the need for spacers [229], and reﬁllable designs can be used to eliminate variations in pathlength between baseline and sample. The design is not critical if the windows are strain-free and pass the entire beam without any obstruction. For studies in H2 O, higher concentrations (up to 100 mg/mL, but sample volume of ∼10 μL) and shorter pathlengths (6–15 μm) are needed for amide I VCD (can be longer for just amide II). For a 6 μm pathlength, water has an absorbance of ∼0.8 at 1650 cm−1 , which causes a loss in S/N for protein or peptide VCD but, in our hands, no dramatic increase in artifacts [162, 226]. Other popular IR sampling techniques such as dried ﬁlms and mulls of powders are not widely used for VCD and require care in preparation. Solid samples, even ﬁlms, can have residual birefringence that can lead to artifacts, especially for biomolecular samples. Spectra of ﬁlms and mulls were measured early on with varying results [230–233], but more recent studies have suggested that use of the dual modulation technique [153] may permit reproducible measurements of VCD for solid-phase samples [167, 169, 234]. Even VCD of peptide and protein ﬁbrils have been shown to be diagnostically useful [164, 165, 235–237]. VCD sampling methods center on use of conventional liquid transmission cells. While ATR cells are very useful for FTIR studies of solids, the reﬂections can phase retard the polarization, analogous to a birefringence, making the polarization ill-deﬁned in the sample and VCD challenging. Perhaps multiple-modulation techniques may make reﬂection-based-VCD feasible in the future [153]. Ideally, VCD should be plotted in molar units, such as ε and ε, as is done with ECD measurements, but concentration and pathlengths can be inaccurate for IR-based biomolecular experiments, so VCD is sometimes normalized to the absorbance for comparison of magnitudes. This is only a ﬁrst-order concentration correction, since amides in different peptide conformations [229] or local environments will have different molar extinction values, as will bases with and without H-bonds in DNA. Although ECD spectra are normally obtained with dilute samples, for direct comparison of conformational properties, ECD can be measured to ∼200 nm on the VCD samples themselves, using short-path CaF2 cells [190, 227, 238, 239].

22.3. BRIEF SURVEY OF THEORETICAL BIOPOLYMER VCD Simulations of biopolymer vibrational spectra are often based on exciton coupling, whereby vibrations of local modes on equivalent residues are coupled via transition

717

718

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

dipoles to yield a band of overlapping contributions representing a characteristic vibrational motion, such as the amide I or II [240]. These same couplings lead to VCD in the IR as was modeled in the earliest efforts to compute VCD [241–244]. Dipole-coupling (DC)-based exciton models of VCD were revived as the extended coupled oscillator (ECO) model for oligomers by Diem and co-workers [245, 246]. By comparison to more exact theoretical methods, DC-based approaches have been shown to be most reliable for weakly interacting (nonbonded) high dipolar strength vibrations [247]. Thus, this simple approach actually works best for describing VCD of the largest molecules we have studied, focusing on the characteristic local modes in DNA [92, 95, 245, 246], and is somewhat less useful for short-range interactions in peptides [140, 248, 249], although DC can provide a reasonable approximation for spectra of β-sheet structures [143, 249–252], and many groups use empirical methods as a basis for spectral interpretation [253–256]. Use of quantum mechanical force ﬁelds and ab initio calculation of the magnetic and electric transition dipole moments, parameterized as the atomic polar and axial tensors (APT and AAT, respectively), are the basis of modern VCD spectral simulations. These computations normally involve use of moderately large basis sets and the magnetic ﬁeld perturbation method (MFP) of Stephens and co-workers [2, 132, 137, 257, 258] to evaluate the AAT or magnetic transition dipole matrix elements. These calculations were initially done at the Hartree–Fock (HF) level and the resultant force ﬁelds (FF) were often scaled to get reasonable frequencies, but most labs now use density functional theory (DFT) methods to achieve improved FFs, incorporate partial correction for correlation, and reduce computational expense [2, 136, 137, 258–263]. Even DFT computations with reasonable basis sets remain formidable for large biomolecules, but these size-based barriers are dropping fast with continued computational and algorithmic developments. Early on, we reported model HF calculations for dipeptides [264], as well as extensions to DFT methods for tri- and longer oligopeptides [139, 140, 142]. To model even longer peptides, we have exploited Bour’s property tensor method to transfer ab initio FF, APT, and AAT tensor values obtained from calculations on smaller molecules to simulate spectra for larger polymers [138, 248]. These latter approaches have been successfully used to explain spectra for experimentally accessible oligopeptides [49, 51, 117, 140, 141, 146, 182, 249, 251, 265–268] and even oligonucleotides [144–146]. Tests of the transfer methods have veriﬁed their value and pointed to problems, particularly for nonrepeating structures [182, 269]. With modern computers it has become possible to calculate spectra directly for experimentally relevant, sizeable peptides with irregular structures [49, 148–150 270–273].

22.4. NUCLEIC ACID VCD SPECTRA AND APPLICATIONS VCD couples separate probing of functional groups in the polynucleotide, which is possible using vibrational spectra [72, 83, 85, 108, 109, 274–278], to the stereochemical sensitivity to stacking and polymer-repeat interactions, which is typical of CD. In large part the VCD patterns originate from dipole coupling, which also gives rise to ECD [63]. Most VCD studies of DNA and RNA have developed and then used empirical correlations of bandshape with helical conformation (handedness) [9, 18, 92–95, 98, 99, 145, 245, 246, 279–296], but some theoretical modeling is available [92, 94, 144–146, 246, 297].

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

22.4.1. Empirical RNA and DNA VCD, Correlation with Helical Form and Applications RNAs and DNAs in buffered (D2 O) solution (10–40 mg/mL and 50-μm path) have relatively weak absorbance yet intense VCD, due to high A/A for a variety of modes in duplex structures [84]. Phosphate modes are more intense absorbers and generally are studied in H2 O with shorter paths [92]. VCD can access in-plane base deformation modes in the range of 1800–1400 cm−1 to study interbase stacking interactions, phosphate P–O stretches in the 1250- to 1000-cm−1 region to sense backbone stereochemistry, and coupled C–H or C–O motions to monitor the ribose conformation. The VCD of the sugar-centered modes has proven only marginally useful, due to their having little spectral deﬁnition in the C–H region, where such characteristic sugar modes are more isolated, and due to overlap of the C–O stretches with other nucleic acid modes (–PO2 − ) in the mid-IR and difﬁculties going to much lower frequencies [18]. Other diagnostic modes have been less studied, but some have been assigned in efforts to theoretically model oligonucleotide VCD [96, 144–146]. ROA studies have also addressed DNA conformational variations and have utilized sugar-based modes with more success [127, 298, 299]. Single-stranded RNA samples give rise to a positive VCD couplet in the in-plane base deformation region [84], typically centered over the most intense band lying between 1600 and 1700 cm−1 , which, except for A (adenine), arises from C=O stretching. Dinucleotides and random copolymers have similar but weaker VCD, and monophosphates have little or no detectable VCD. Duplex RNA and DNA have quite intense VCD patterns, which arise from interaction through the polymer fold and a common bandshape resulting from the right-handed helical twist [18, 84, 92–94]. For example, in the base region, VCD for a synthetic duplex RNA, such as poly(rI) · poly(rC), is virtually the same as found for B-form duplex DNA—for example, d(G-C)·d(G-C) (Figure 22.4a)—and both have stronger intensity but the same shape as mixed sequence, natural A- and Bform DNA as well as t-RNA (Figure 22.4b). VCD of t-RNA is broader and weaker due to its heterogeneous structure, and its temperature dependence shows less stability [18, 93, 94]. Z-form DNA (Figure 22.4a, bottom) has the opposite sign pattern primarily due to its left-handed twist. The A-form base modes (dehydrated DNA or in RNA) shift to higher wavenumbers and have a sharper and more intense high-frequency VCD couplet, more like that of RNA duplex VCD than the B-form (Figure 22.4b). The overall proﬁle and sign patterns are maintained in this B-to-A structural transition [93], including sensitivities to base content. Poly(rA)·poly(rU) provides an exception to this simple pattern, since uracil, U, has two C=O oscillators and A has none. Furthermore, its VCD abruptly changes at 55◦ C to an alternate pattern due to formation of a triple helical form, polyU∗ polyA•polyU (the ∗ indicates Hoogsteen base pairing and the • indicates Watson–Crick pairing) [95, 288]. At higher temperatures this complex melts to single strands and the spectrum changes again. This intermediate triple-helical VCD spectral form has general characteristics that can be found in all pyrimidine–purine–pyrimidine, A(T)U-based DNA, RNA, or mixed triplexes (Figure 22.5) and provides a unique diagnostic for triplex conformation, even for triplexes formed as C•G ∗ C+ [95]. The difference in the poly(dA–dT) and poly(dG–dC) type VCD patterns [93] means that base deformation VCD, though dominated by C=O stretching, is sequence-dependent and can be used to qualitatively determine DNA base content as shown in Figure 22.6a for a series of DNAs with varying G•C contents [92]. By contrast, the VCD intensity and shape is virtually constant for the sym –PO2 − mode of these same molecules

719

720

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

B-DNA r(I)r(C) ΔA

A-DNA

ΔA d(GC)

t-DNA d(GC)/alc

1750

1700

1650

1600

Figure 22.4. Typical nucleic acid VCD patterns in the base deformation region for duplex (a) synthetic A-form RNA (r(I)r(C)) and B-form DNA (d(G–C)) in D2 O

1750

1700

1650

Frequency (cm−1)

Frequency (cm−1)

(a)

(b)

1600

compared to Z-form d(G–C) in alcohol (left-handed helical) and (b) comparison of natural B- and A-form DNA (calf-thymus) with a t-RNA (partially A-form).

40 UAU ΔA TAT

20

TAU

CIC+ 0

CGC+ −20

Figure 22.5. Overlapped VCD spectra of triplex confomations: (top to bottom) rU∗rA•rU (in D2 O), dT∗dA•dT (in D2 O), rU∗ dA•dT (in 0.1 M NaCl/D2 O),

1700

Wavenumber (cm−1)

1600

rC+ ∗ rI•rC (in D2 O at pD 5.6), and rC+ ∗ rG•rC (in D2 O at pD 5.6). A units are in 10−5 absorbance.

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

20 20

1

10

2

10 ΔA 0 ΔA

−10

3

0

4

−10

0.5

5 A −20

0.0

1600

1700

1150

1100

1050

Frequency (cm−1)

Frequency (cm−1)

(a)

(b)

1000

Figure 22.6. (a) Sequence-dependent VCD and absorption of B-form DNA in the base deformation, C=O stretching, region for poly(dG-dC)·poly(dG–dC) (long dash, 1); M. lysodeikticus, 72% GC (medium dash, 2); calf thymus, 44% GC (dots, 3); C. perfringens, 26% GC (dashed-dot, 4); poly(dA–dT)∗ poly(dA–dT) (solid line, 5). (b) Sequence-independent VCD spectra in the phosphate region for the same B-form DNAs. A in units of 10−5 absorbance.

(Figure 22.6b). RNA and DNA differ in the –PO2 − region, due to contributions from overlapping ribose modes in RNA that are not present in deoxyribose (DNA) [92, 98, 281–285, 287]. The VCD spectra of poly(dG–dC) and related DNA oligomers in their B and Z forms have distinctly different bandshapes for the base stretching modes (Figure 22.7A) [18, 92, 94, 97, 297]. Both are dominated by a VCD couplet associated with the highestfrequency intense base modes, but these IR bands are signiﬁcantly shifted in frequency and have opposite VCD sign patterns (relative to the absorbance maximum), which reﬂects the handedness of their duplex helices (Figure 22.7a). Z-form base deformation VCD resulting from either Na+ or alcohol-induced B-to-Z transitions have very similar spectra [94], and Z-form –PO2 − VCD is also a couplet shape, opposite in sign to that of A and B form in the symmetric –PO2 − stretching region (Figure 22.7b), as can be predicted by DC modeling (Figure 22.7c, see below) [92, 94, 97, 245]. VCD has been used to assign conformations in a number of systems. B-form poly(dI–dC) has an inverted near-UV ECD pattern presumably due to methylation of the base, which was once misinterpreted as being due to Z form [300], but its VCD has deﬁnitive right-handed B-form spectral patterns [94]. Other applications have encompassed triplex studies that identiﬁed structural phase transitions and characterized the single-, double-, and triple-strand equilibria in DNA, RNA, and mixed species of A–T(U) strands [95, 288]. Quadruplexes, and the unique base interactions forming them, have additionally been studied with VCD at some level [33, 99, 228].

721

722

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

10

20

0.05

5

10 ΔA

ΔA 0

Δe

0

−5

−10

−0.05

−10

1000

0.5 0.5 A

A

0.0

500 e

0.0 1700 1600 Frequency (cm−1) (a)

0.00

0 1150 1100 1050 1000 Frequency (cm−1) (b)

1100 1000 Frequency (cm−1) (c)

Figure 22.7. (a) VCD (top) and IR absorption (bottom) spectra (a) of B- (dots) and Z-form (solid) DNA for poly(dG–dC) in the base deformation region, (b) The same for the phosphate, sym PO2 − , region, and (c) sym PO2 − VCD spectra (top) computed with a DC model and corresponding IR (bottom) for A- (dashed), B- (solid), and Z-form (dots) DNA using the (dG–dC)5 structure. A in units of 10−5 absorbance.

Some oligonucleotide studies have established constraints on the possible conformations that can exist for di- and tetra-nucleotides [279, 280]. Other studies, led by the Wieser group, focused on the interaction of metal ions with the bases and effects of changing the –PO2 − shielding in DNAs, using VCD to probe the likely binding sites and explore their conformational impact [9, 281–283, 285, 287, 290, 292, 295]. Finally, drug–DNA and protein (or peptide)–DNA interactions have been preliminarily explored in work by both the Wieser and Urbanova labs using VCD to study ligand-binding and sequence effects [33, 145, 284, 289, 291, 294].

22.4.2. Modeling of Nucleic Acid VCD The clear dependence of DNA VCD on helicity is indicative of the role of dipolar coupling (DC) in the stacking interaction between bases and the coupling between the phosphate groups, which is why model calculations, using the simple coupled-oscillator concept as a basis, have had some success in interpreting those DNA spectra [92, 94–96, 245, 246, 279, 280, 297, 301]. The phosphate –PO2 − modes are simplest to model with DC interactions, which yield predicted VCD characterized by a strong couplet for the symmetric stretch at 1070 cm−1 , which reﬂects the helicity (Figure 22.7c), and weak VCD for the asymmetric –PO2 − stretch at 1250 cm−1 [92, 94]. The sym –PO2 − modes, directed perpendicular to the helical axis, directly sense the twist of the backbone, while the asym modes are less useful, being more parallel to the axis. Overlap with ribose modes (locally chiral but uniform in conﬁguration and conformation) could complicate the sym –PO2 − region, but the observed patterns do not imply strong coupling to other modes. The DNA or RNA base modes themselves are more difﬁcult to model with DCbased methods because they are delocalized over the bases. The C=O stretch contribution

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

dominates the most intense and highest frequency mode in most bases (except A), but there is considerable mixing of local C=C and C=N motions. DC models adequately represent poly(dG–dC) base mode VCD spectra, since both bases have a single C=O, but conversely DC does not do well for dA–dT (or rA–rU) due to the large variation in modes between the bases (A has no C=O, but T or U have two each) and the strong coupling between the two C=O groups in U and T. Diem and co-workers [297, 301] attempted to compensate for this by extending the DC approach to multiple nondegenerate oscillators. Subsequently, an extension of the DeVoe model of ECD [302] was applied to DNA VCD with some success and showed that the triplex VCD spectra could be explained with a combination of Watson–Crick and Hoogsteen base pairing [303, 304]. Efforts to extend quantum mechanical determinations of FF, APT, and AAT values computed for segments of DNAs to simulate spectra for entire oligomers have provided new insight into the nature of the vibrational modes and their local coupling [97, 144–146]. Bour and co-workers [305] have used the idea of transfer of local modes to a polymer structure, but have carried it further by dividing the duplex molecule into base pairs and sugar–phosphate dimers computed on a DFT level with added corrections for two base pairs calculated for a dinucleotide on the PM3 level. For the single-strand model, two stacked bases plus a sugar–phosphate dimer were all calculated at the DFT level and were then coupled (transferred) into the full oligomer (Figure 22.8). Here the problem of which interactions to compute at the DFT level and which to leave at the approximate DC level, as would be needed in an ONIOM approach [270], and how to model solvent effects are signiﬁcant, and both remain subjects for continuing studies.

22.5. PEPTIDE VCD STUDIES The most extensive VCD applications in the biopolymer ﬁeld have been for peptides, both small oligomers and larger polymers, which have been the product of work from many labs following an initial report on helical poly-benzyl-glutamate ∼30 years ago [306]. VCD of peptides has proven to have heightened discrimination of secondary structures, adaptability to theoretical analyses at several levels, and a selectivity not seen in ordinary vibrational spectra. Peptide VCD also has the longest history, broadest expanse (results now coming from three continents), and the most direct and useful theoretical analyses for any of the biopolymers. Since several polypeptides have welldeﬁned established structures, they provided the ﬁrst entr´ee of VCD in the biopolymer ﬁeld. Development of an interpretive basis followed methods for IR and ECD analyses of proteins. VCD of α-helical polypeptides in nonaqueous solvents comprised the ﬁrst measurements [306–308], followed by β-sheet examples, which were constrained by limited peptide solubility [231–233, 238, 309, 310]. VCD studies of aqueous peptides focus on (a) the amide I band, ∼1650 cm−1 , which is partially overlapped by the HOH band, but measurable with short pathlengths, and (b) the amide II and III modes (∼1550–1500 and 1350–1250 cm−1 , respectively) measureable with longer paths. Since high concentrations are often not compatible with avoiding aggregation, particularly for extended peptides, VCD studies often focus on study of isolated molecular species and just the amide I vibration (C=O stretch after H/D exchange) for peptides in D2 O. The amide II (in D2 O) is strongly shifted and altered in character, and, due to solvent interference, the amide A (N–D stretch) and III vibrations are not detectable with VCD. Spectra over wider regions, including the amide A and I, II, and III regions have been measured for peptides in non-aqueous solution, when such conditions are relevant to the questions being posed.

723

724

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(rA)8

1690

1064 1107 1083 1087 1124 1100

Calculation 860

977

863

1025

1569

Experiment 813

1564

1106 1084

1633

1696

1740

Experiment

1121 1095 1075

1092

500

1689 1669 1631

976 926

e

Experiment

0

Absorption

1000

1027 988 957 912 864

1518

1572

1664

1627

1000

Calculation 892 857

1574

e

1088 1057

Absorption

VCD

1066 1053

1705 1677

-0.5

1633 1617

0.0 2000

1628

1740

1697 1665 1636

0.0

Δe

1623

1102

1572

Experiment

1692

VCD

0.2

Calculation

0.5

1086

1623 1620

Δe

Calculation

996

1574

0.4

1092 1055 1041 974

1612

Double-Stranded

1635

Single-Stranded

0

1750 1700 1650 1600 1550 1500 1100 1000 900 800

Wavenumber (cm−1)

1800 1700

1600 1500 1100 1000

900

800

Wavenumber (cm−1)

Figure 22.8. Calculated VCD and absorption spectra and the corresponding experimental VCD and absorption spectra in the base deformation and phosphate region for (left) single-stranded octamer (rA)8 compared to spectra of poly(rA) and (right) for double-stranded octamer (rA)8 ·(rU)8 compared to spectra of poly(rA)·poly(rU). [Adapted from calculations by Valery Andrushchenko, Hal Wieser, and Petr Bour, with permission (V. Andrushchenko, H. Wieser, P. Bour, J. Phys. Chem. B 108, 3899 (2004)[305]).]

22.5.1. Peptide VCD Empirical Correlation with Secondary Structure Right-handed α-helices yield VCD consisting of a negative couplet in the amide A VCD (∼3300 cm−1 , positive bias), a positive couplet (+ then −, with increasing frequency) amide I (∼1655 cm−1 , negative bias), a broad distorted negative amide II band with a maximum lower in frequency (at ∼1520 cm−1 ) than the absorption maximum (∼1550 cm−1 ), and net positive VCD in the general amide III region. A prime example of this behavior was found for poly-γ -benzyl-l-glutamate (PBLG) [306, 311] in CDCl3 , which, due to its very long persistence length, has one of the highest intensity and narrowest band width α-helical VCD we have measured (Figure 22.9a). The same (though weaker and broader) bandshape patterns are found for numerous peptides composed of α-helical segments of varying lengths, even in water (Figure 22.9b) [49, 50, 140, 226, 238, 265, 268, 306–308 311–313]. Deuteration of the amide N–H (normally as a consequence of D2 O solvation) changes the shape of the α-helical amide I VCD (as shown for poly-l-lys, PLL, at low pH and 5◦ C in Figure 22.10a, which has some β-sheet component) to a three peaked (−,+,−) pattern (amide I ) [49, 162, 238, 265, 307] and shifts the negative amide II VCD to below 1450 cm−1 (amide II ) with loss of intensity [307, 314, 315]. The β-sheet and coil forms have been shown to have distinctly different amide I VCD spectra from that of the α-helix, being weak and predominantly negative (note scale change for PLL spectra at low pH after heating, Figure 22.10b). These characteristics provide consistent spectra-conformation correlations for a variety of polypeptides [19, 134, 238, 251, 266, 267, 309, 310, 312, 316, 317]. (It might be noted that the Figure 22.10

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

10 VCD

10

725

VCD

ΔA × 105

0 −10

0

−20 −30

−10

1.0

1 IR

TFA

persistence length α-helical polypeptide in nonaqueous solution (CDCl3 ) and (b)

0.5

1700

1500

1600 Wavenumber

1700

(cm−1)

1500

rescaled to match an amide I peak absorbance of 1.0.

(b)

2

10

5

5

ΔA x 105

0 VCD

0 −2

VCD

1.0

1.0 Absorbance

IR 0.5

0.0 1700

1650

1600

Wavenumber (cm−1) (a)

−5

VCD

−15 1.0

IR

0.5

0.0

1750

0

−10

−4

−10

Absorbance

ΔA x 105

1600 Wavenumber (cm−1)

(a)

−5

Ac–(AAKAA)4 –GY–NH2 , an α-helical oligopeptide in H2 O. For comparison, the VCD spectra were

0

0.0

Absorbance

(top) and its corresponding FTIR (bottom) spectra of (a) poly-γ -benzyl-glutamate, a high

ΔA x 105

Absorbance

Figure 22.9. Amide I and II VCD IR

1750 1700 1650 1600 Wavenumber (cm−1) (b)

IR 0.5

0.0 1750 1700 1650 1600 Wavenumber (cm−1) (c)

Figure 22.10. VCD (top) and IR absorption (bottom) spectra in the amide I region for poly-Llysine (PLL) in D2 O (a) at pH ∼11, mostly α-helical, (b) at pH ∼11, heated to 65◦ C for ∼20 min and recooled, β-sheet, and (c) at neutral pH, disordered or ‘‘random coil’’ (PPII-like). For comparison, all spectra were normalized to A = 1 for the amide I peak absorbance.

spectra are much higher quality than we have previously published for β-sheets, due to remeasurement with our new dispersive instrument.) Widely split amide I IR absorbance features at ∼1615 cm−1 and 1690 cm−1 are characteristic of an antiparallel β-sheet conformation and normally indicate aggregation or formation of extended sheets [30, 31, 251, 318, 319]. This pattern is highly sensitive to sample conditions and can develop a

726

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

strong couplet shape under gel-like conditions [238, 320]. Such intense VCD patterns have more recently been recognized as additionally being characteristic of peptides and proteins that form cross-β ﬁbril-type structures and have been interpreted in terms of local structure [164, 165, 237]. Duplex DNA and high persistence-length helices (such as PBLG in CDCl3 ) also have strong VCD (but less dramatic than for cross-β examples), suggesting that coherence or in-phase coupling of vibrations over longer distances in the polymeric medium may be the source of such intensity. Oligopeptides that adopt an apparently β-structure in nonaqueous solution evidence an absorbance maximum at a higher frequency (∼1635 cm−1 ) than in these β-sheet aggregates and have a VCD whose shape is more variable and often quite weak [51, 233, 249, 251, 267, 316, 317, 321, 322]. Polypeptide β-sheet structures also yield medium to weak amide II VCD with a negative couplet shape and generally negative amide III VCD [162, 177, 315]. A number of oligopeptide VCD studies [227, 268, 323–326] have established that, while the 310 helix (Figure 22.11, top) has VCD of the same overall sign pattern as the α-helix (Figure 22.11, bottom), the 310 helix is distinguishable by its weaker amide I VCD couplet shape and sharper more intense amide II. This provides a means of differentiating these two related helical structures and identifying conformations in mixed structures, such as might be present in the (Aib–Ala)4 octamer (Figure 22.11, middle), which VCD shows to be more 310 -like [268, 325]. While ECD patterns have been proposed to distinguish α- and 310 -helices [327, 328], those results are dependent on the fraction of α, α-dialkyl residues employed to stabilize the 310 -helix, which leads to ambiguity [227, 329–331]. Characterizations of the VCD for other minor secondary structure variants, such as β-bend ribbons, alternate cis-trans dl-proline oligomers, parallel versus antiparallel strands and various turns, have also been proposed [233, 253, 267, 332–339]. The “random coil” form of polypeptides turns out to have a surprisingly intense negative couplet amide I VCD (at ∼1645 cm−1 , Figure 22.10c), which is opposite in sign to the α-helix VCD pattern and is relatively insensitive to deuteration [19, 134, 309,

Aib5-Leu-Aib2

Ac-(Aib)8-NH2

Figure 22.11. Comparison of experimental and computed VCD spectra

dash: in TFE So Solid:ininCDCl CDCl 33

Ac-(Aib-Ala)4-NH2

(Met2-Leu)8 Ac-(Ala)6-NH2

Δe (amide)

(A ib -A la )44 (Aib-Ala)

in amide I and amide II region for three model 310 - and α-helical peptides having different fractions of α-methylated residues. (Left) spectra measured in CDCl3 solution: 310 -helical, Aib5 –Leu–Aib2 (top) and (Aib–Ala)4 (middle, data also shown for TFE solution, black dashed line), with α-helical (Met2 –Leu)8 (bottom). (Right) simulated 310 -helical Ac–(Aib)8 –NH2 (top) and Ac–(Aib–Ala)4 –NH2 (middle), with α-helical Ac–(Ala)6 –NH2 (bottom). Vertical

1700

1600

Wavenumber (cm−1)

1500

1700

1600

1500

Wavenumber (cm−1)

lines indicate computed rotational (R) strengths, and envelopes are the sum of lorentzian shapes scaled to the R values.

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

727

312, 340–342]. For various coil-like peptides, the VCD decreases in amplitude as the temperature is increased, as the oligopeptide length is decreased, and also changes with solvent conditions, all of which suggest an underlying structure [19, 50, 134, 340]. Poly-l-proline (poly-Pro II) in water has a left-handed helical form (31 -helix) composed of trans amides that is termed PPII and has a characteristic negative couplet amide I VCD shape correlated to an IR band with a weak high-frequency shoulder (Figure 22.12a). The VCD spectra of poly-l-lysine (PLL, Figure 22.10c) or poly-lglutamic acid (PGA, Figure 22.12b) at neutral pH, which are traditionally termed “random coils” due to their lack of long-range order, are identical in shape to that of PPII, but are lower in magnitude for VCD and shifted up in frequency [19, 50, 340]. This VCD pattern provided evidence supporting the proposal that random coil peptides had local order related to PPII, originally called an “extended helix” [343], and has subsequently become an important tool in ensuing studies of structure in unordered proteins and peptides (see applications section) [134]. Proline forms tertiary rather than secondary amide bonds, so its amide I mode has a lower frequency than do most peptides. PPII structures, like β-sheet strands, have amide C=O groups directed out from the helix axis—in this case to enhance H-bonding to water, rather than to another strand as in β-sheets. Since both forms have left-handed twisted strands, the VCD of “coils” and twisted β-structures are similar, although they are distinguished by frequency and detailed pattern. A large range of peptide lengths have been studied and the VCD patterns described above are maintained, but increasing length in a uniform structure leads to sharpening of the observed bands in both IR and VCD, as seen in Figure 22.9a, since the intensity tends to be associated with only a few exciton components of the dispersed coupled modes [138–140, 143, 251, 252, 268, 344]. The close relationship with theoretical models generally lends strength to VCD in terms of structure analyses, and it is most evident for peptides due to their smaller size and more uniform structures.

22.5.2. Peptide VCD Applications An important aspect of VCD is its resolution of contributions due to the amide group (the repeating structural element) from those originating in the side chains or other parts

2

1

ΔA

ΔA 0

0

x 104

x 104

−2

−1

COO-

Figure 22.12. IR absorption and VCD

Amide I 1

spectra for (a) poly-L-proline II (PPII) and (b)

1

A

poly-L-glutamic acid (PGA) in D2 O solution at neutral pH. The spectra are normalized to the area of the amide I absorbance band

A

0

(corrected for baseline). Note that the VCD intensity of PPII is about twice that of PGA.

0 1700

1600

1500

1700

1600

Frequency (cm−1)

Frequency (cm−1)

(a)

(b)

1500

The intense band at ∼1560 cm−1 is due to the –COO− stretch and has no VCD, due to disorder.

728

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

of the molecule. This enabled secondary structure determination studies of peptides containing aromatic amino acids and nonproteinic residues with chromophoric side chains [194, 312, 345–348]. Similarly, we were able to establish the helical handedness of and propose a turn-like structure for related short oligopeptides containing Cα,α -disubstituted residues bridged by binaphthyl and biphenyl side chains (Figure 22.13) in a situation where the side chain dominated the UV ECD, completely obscuring any amide contribution [347, 349]. Other small cyclic and turn peptides have been similarly studied with VCD, again avoiding interference with aromatic transitions, such as can occur in ECD [148, 149, 350]. Early on and continuing, some labs studied very short peptides, including individual amino acids as well as di- and tripeptides and sought to derive structural insight and deduce intramolecular H-bond patterns from their VCD [161, 253, 337, 338, 346, 351–357]. Most short peptides have limited structural stability, although many studies have appeared in recent years using combinations of Raman, IR and VCD (as well as other methods) to show a propensity for partial (fractional population) structure even in tripeptides [253, 338]. By contrast, Pron oligomers offer a system that can form relatively stable helical structures for short oligomers and have high persistence lengths for poly(Pro), both of which show a very characteristic VCD for the dominant PPII conformation [50, 320, 340]. The fact that random-coil polypeptides as well as many disordered oligopeptides have IR and VCD bandshapes of the same shape but smaller amplitude and shifted frequency as compared to the amide I VCD pattern characteristic of PPII [50, 340, 341, 358] (Figure 22.12) has supported an earlier proposal, based on ECD and IR studies, that the “coil” conformation is related to PPII [19, 309, 340, 343]. The VCD length-dependence and intensity data suggest that the short-range conformation in such “coil” structures favors a local left-handed twist, similar to that in a twisted β-strand, and corresponds to a broad minimum of the Ramachandran map. This sense of there being local PPII structure for disordered peptides and proteins has come into wide acceptance [253, 338, 359, 360]. Studies of the PPII nature of short peptide structures has become a minor industry using a wide variety of physical methods. VCD has the advantage of being sensitive to shorter-range interactions which can highlight local structures such as turns and nonrepeats as well as the residual local population noted above in conformationally disordered peptides. Several reports of cyclic peptide VCD have appeared [21, 148, 266, 267, 316, 334–336 361–365]. These often have unique bandshapes that do not follow the extended patterns presented above and must be analyzed using theoretical modeling. VCD and IR can also distinguish PPII from the PPI structure, a right-handed helix of cis amides, and from mixed structures such as are formed with d,l-Pro mixed oligomers [332, 333, 358]. VCD was used to identify an intermediate, not detected with ECD, during the mutarotation of poly(pro) from the PPI to PPII conformation [358], by utilizing the frequency resolution and short-range conformational sensitivity of VCD. Identiﬁcation of an intermediate with multiple cis–trans junctions was enabled by study of alternating d,l-proline oligomers which sampled both helicities and established cis –trans-Pro IR and VCD characteristics [333]. Furthermore, VCD showed that these d,l-oligomers had a preference for a particular helical handedness depending on the chirality of the Cterminal residue. This presaged a similar effect, seen above for l-Val placed on the C- or N-termini of α, α-disubstituted residue sequences, as detected with VCD (Figure 22.13) [349]. VCD’s sensitivity to handedness and to 310 -helix formation has been used in a series of studies on peptides containing α, α-dialkylated residues [227, 268, 323–325,

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

2

729

Ac-(Bip)3-L-Val-OMe Boc-L-Val-(Bip)4-Ot Bu

0

ΔA x 105 −2

(R)

Figure 22.13. VCD (top) and IR absorption 0.4

(bottom) spectra over the spectral region 1775–1475 cm−1 for two peptides composed of biphenyl-bridged residues whose ECD is dominated by the biphenyl contribution.

X-HN CO -Y A 0.2

(Solid line) Tetrapeptide (Ac–(Bip)3 – L-Val–OMe) in a 46:11 (v/v) CDCl3 /TFE solvent mixture, resulting in a

0.0 1700

1600 Wavenumber (cm−1)

1500

left-handed (possibly 310 ) turn, and (dashed line) a pentapeptide (Boc– L-Val–(Bip)4 –OtBu) in CDCl3 solution, forming a right-handed 310 -helical segment.

327, 350, 366–368]. The usual clear preference of l-amino acids to form right-handed helices can be distorted for residues with α, α-disubstitution. Using VCD, blocked (αMe)Phe tetra- and pentapeptides were shown to form left-handed 310 -like helices in CDCl3 solution [368] in contrast to the right-handed forms seen for (α-Me)Val, Aib (i.e., (α,α-diMe)Gly), and Iva (isovaline, which has both α-Me and α-Et substitution) [227, 324]. Dehydro-phenylalanine-containing peptides were also shown by VCD to have a tendency to form 310 -helical conformations [161, 369, 370]. Alanine-rich peptides such as Ac-(AAKAA)n -GY-NH2 have a high propensity for helix formation at low temperatures, even in H2 O, with peptides having more than ∼16 residues. Their amide I frequencies occur at <1640 cm−1 in D2 O, a fact that was originally misinterpreted as being due to a 310 -helical structure to correlate with EPR results [371–373]. This frequency shift was gradually recognized as a solvation effect, after demonstration using VCD that they were indeed α-helical [49, 116, 226, 265]. Study of the temperature variation of the spectra of these model α-helical peptides and factor (or principal component) analysis [374] of their changing bandshapes showed that VCD, due to its short-range length dependence, was able to detect a third spectral component that represented an intermediate structure attributable to junctions between the central helical portion of the molecule and the steadily fraying ends [226]. VCD studies [265] of selectively isotopically labeled helical peptides, AcAAAA(KAAAA)3 Y-NH2 , with 13 C on the four Ala residues at the N-terminus or in the

730

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

center, had a positive couplet for the 13 C=O band at ∼1600 cm−1 and the expected α-helical bandshape for the 12 C=O band at ∼1640 cm−1 at 5◦ C (Figure 22.14a, solid line). However, when labeled at the four C-terminal Ala residues, only a weak negative shoulder is seen for the 13 C component (dashed line), which indicates that while the peptide is mostly helical at low temperatures, the C-terminal residues are frayed or disordered. At higher temperature, the pattern changes to one characteristic of a disordered structure for both the 12 C=O and 13 C=O modes and becomes identical for all labeling patterns (Figure 22.14b). Analysis of VCD frequency shifts and factor analyses of the bandshapes demonstrated that the ends unfold at lower temperatures than do the middle residues and that the C-terminal residues are substantially unwound even at low temperatures [265]. Partly from a basic structural modeling challenge and also in response to wide interest in amyloid diseases, many model β-sheet forming peptides have been prepared [375–390], and VCD studies on some of them have explored the spectral impact of various environments and forms of sheets [51, 164, 165, 231–233, 237, 238, 251, 266, 267, 316, 317, 321, 365, 391]. Although β-sheets have distinctive characteristic IR, their VCD is normally weaker in intensity, and developing a characteristic bandshape pattern by comparisons to models is complicated by the variety of ways and degree to which multiple strands can interact in forming a sheet structure from linear peptides. However,

12

T = 50°C Unlabeled N-terminus C-terminus

Unlabeled N-terminus C-terminus

8

8

4

4

0

0

4

4

0

0

−4

−4

−8

−8

1700

1650

1600

1550

1700

1650

1600

Wavenumber (cm−1)

Wavenumber (cm−1)

(a)

(b)

ΔAnorm (x 105)

ΔAnorm (x 105)

T = 5°C

Anorm (x 10)

Anorm (x 10)

12

1550

Figure 22.14. Experimental amide I’ (in D2 O) IR (top) and VCD (bottom) spectra for isotopically labeled peptides (Ac-AAAA(KAAAA)3 Y-NH2 ) with

13 C

on the amide C=O, measured (a) at low

◦

temperature (5 C) and (b) at higher temperature (50◦ C). Spectra for the unlabeled (U) are represented by a thin dash–dot–dot line, for the N-terminally labeled (N) by a solid line, and for the C-terminally labeled (C) peptide by a dashed line. Labels placed on the center residues give results qualitatively the same as for N, but for labels on the C terminus, the behavior is different, indicating local unfolding.

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

hairpins, or single chains folded to give sheet structure with a ﬁxed number of strands, give spectra that are better related to protein VCD results than to the multistranded polypeptide extended β-sheet VCD shown above. Interpretation of hairpin data is somewhat complicated by inclusion of a turn (or loop) between every two strands and by less uniformity (greater twist and fraying at the termini) than for more extended sheets. Nonetheless, we and others have measured VCD spectra for hairpins that qualitatively correlates with theoretical models. Initial attempts to make β-hairpins used sequences derived from proteins, but very few formed stable conformations. The β-subunit of human chorionic gonadotropin (hCG) has three hairpin segments that were thought to fold in an initial step (framework model), and misfolds were correlated with tumor function [392]. By study of separate sequences corresponding to each of the hairpins with ECD, FTIR and VCD, the stability of the βform of these peptides outside the protein was shown to be dependent on either high concentration or the presence of a charged micellar environment (e.g., with SDS) [322]. VCD distinguished the extended (PPII-like) structure of the sequence corresponding to hairpin H2-β from the antiparallel β-form that formed with the hairpin H1-β and H3-β sequences, and correlation of all the data led to a model for the folding mechanism [239, 322]. There are two popular means of stabilizing hairpins, either initiating the turn with residues that restrict conformation or by including hydrophobic cross-strand interactions to stabilize the strands. We have used both methods to form peptides in aqueous solution, and Polavarapu and co-workers have measured VCD of examples of the former in nonaqueous environments [51, 150, 266, 267, 316, 317, 365, 391, 393]. The VCD pattern seen in many of our hairpin examples is correlated to a broader amide I IR than seen for polypeptide β-sheets, but is normally characterized by a negative couplet centered on the lower-frequency stronger dipole amide I component at ∼1630–1640 cm−1 (Figure 22.15a). The observed hairpin VCD tends to reﬂect the turn VCD, based on model correlation and computation; but with the more extended dipole coupling, there are differences. The higher-frequency weak IR component at ∼1680–1690 cm−1 has little contribution to the VCD of most hairpins. Multiple-stranded β-sheet models have also been studied and used to demonstrate the difference in conformation among models. By use of the turn-initiating d-Pro–Gly sequence and/or incorporation of cross-stranded Trp–Trp interactions, we can make stable three-stranded structures with IR spectra like those of polypeptide β-sheets but having more VCD intensity, presumably due to the twist in the smaller designed peptides (Figure 22.15b) [394, 395]. Recent studies have noted that VCD intensities for aggregated peptides (and proteins) can become very large, which Lednev and co-workers have attributed to ﬁbril formation [164, 165]. The peptides develop an organized, ordered intermolecular structure so that the spectral transitions in a single sequence or strand are in-phase with those of many other strands, yielding a large enhancement. While such effects were seen earlier, such spectra were often previously discarded as evidence of aggregation, since the initial goal of most studies was to focus on monomeric molecules that one could model with theory or empirical correlation [238, 320, 348]. This sort of aggregate VCD also results when peptides develop gel forms [395]. Dzwolak and co-workers noted that insulin–dye complexes had CD characteristics that could change sign, depending on precipitation conditions to form ﬁbrils, and these indeed give oppositely signed VCD patterns (Ge et al., unpublished), which is consistent with results of Lednev and co-workers. More recently, we have measured PGA VCD in both ﬁlm and suspended ﬁbril forms (Figure 22.16) for varying mixtures of the l and d forms to show that the enhanced VCD gives a systematic progression of bandshape patterns reﬂecting loss of order [235, 237]. Such ﬁbril formation is also

731

732

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Se r +H N 3

H N

O O HN

N

0.0

H N Thr

O

Val

Glu

N H

Val Tyr

O

N OO H N

O

N H

Lys

O

O

Trp

H N

O H N

Glu N H

Thr

N H

O

NH

O NH 2 Glu

2

ΔA x 105

ΔA x 105

O

O

O

H N

4

0.4

−0.4 VCD −0.8

0 VCD

−2

−1.2 −1.6

−4 IR

0.4 Absorbance

Absorbance

Thr N H

N H

O H N

O

0.3

Trp Tyr

H N Lys

O

O

0.2

0.1

IR

0.3 0.2 0.1

0.0 1750 1725 1700 1675 1650 1625 1600 1575

0.0 1750 1725 1700 1675 1650 1625 1600 1575

Wavenumber (cm−1)

Wavenumber (cm−1)

(a)

(b)

Figure 22.15. VCD (top) and IR absorption (bottom) spectra in D2 O of (a) a β-hairpin derived from the Trpzip2 sequence with two Trps and two Tyr forming cross-strand hydrophobic stabilization of the hairpin conformation (SWTYENGKYTWK) and (b) of a designed three-stranded structure (SWTVED PGKYTYKGD PEVTWE) having both cross-strand aromatic interactions and D Pro–Gly turn stabilization. Schematics of the structures are shown above the spectra.

possible with shorter oligomers of Glu (Chi and Keiderling, unpublished). Polavarapu and co-workers took a different approach by noting that ﬁlm formation provided enhanced VCD signals and allowed detection of smaller quantities of analytes [167, 169].

22.5.3. Peptide VCD Theory As has been well established, the Stephens implementation of the ab initio quantum mechanical magnetic ﬁeld perturbation (MFP) method [3, 132, 136, 137, 257–259, 396, 397] can produce reliable theoretical simulations of the VCD of small molecules as long as the conformer distribution is manageable. In biopolymer applications, initial SCF-level (4-31G) and subsequent DFT (BPW91/6-31G∗∗ ) level computations for di- and tripeptide models, constrained to α-helix, β-sheet, 310 -helix, and PPII (left-handed 31 )-helix ϕ, ψ torsional angles, had qualitative success, which further improved on lengthening the strands computed [139, 264]. The modes from individual amides overlap in the amide I and II regions, but by assigning each transition a realistic bandshape scaled to the computed intensity, their sum yields a band envelope with comparatively useful patterns

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

20

733

R.C.PLGA Film: PLGA Suspension: PLGA

10

ΔA x 105

0

−10 −20

VCD

−30 −40 R.C.PLGA. Film: PLGA Suspension: PLGA

0.6

Absorbance

IR

Figure 22.16. Overlapped VCD (top) and IR absorption (bottom) spectra of poly-L-glutamic

0.4

acid (PGA) at neutral pH (solid line, representing the coil or PPII-like disordered form, the same conditions as in Figure 22.12b),

0.2

and low-pH PGA as a ﬁlm (dashed–dot line) and as a ﬁbril suspension (long dashed line), showing the enhanced VCD intensity for the

0.0 1750

1700

1650

1600

Wavenumber (cm−1)

1550

aggregate form. The intense band at ∼1730 cm−1 in the ﬁbrils and ﬁlm (which are low pH) is due to –COOH and shifts to ∼1560 cm−1 for –COO− in the higher pH solution spectra.

in qualitative agreement with our experimental results. The computed qualitative features, such as positive couplet amide I for α- and 310 -helix, and negative couplet amide I for 31 -helix (PPII-like), negative amide II for α- and 310 -helix, and negative couplet amide II VCD for β-sheet, are maintained for most levels of calculation and reﬂect experimental results. The degree of precision that can be obtained for helices is exempliﬁed by the comparison in Figure 22.11 of simulated and experimental VCD for related octapeptides in α-helical and 310 -helical conformers. An interpretive problem derives from the very large splitting computed between the amide I and II modes, with the amide I being ∼100 cm−1 too high, a property of these standard functionals (B3LYP is even worse) and basis sets, for calculations on peptides under vacuum conditions [117, 141, 148, 344, 398, 399]. While polarized continuum models (PCM) help reduce the amide I frequency, there is still a substantial difference from experiment, so that directed H-bonds are needed to properly simulate aqueous solution FF [117, 147, 249, 344, 399]. An empirical correction for the effect of water on the amide I FF proposed by Cho and co-workers [180, 181] was subsequently modiﬁed by Bour to compute the IR and VCD spectra of N -methyl acetamide and of a pentapeptide (β-turn) in explicit water solvent with modest success [399]. Despite these FF problems, simulated amide I and II VCD bandshapes, even with explicit solvent,

734

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

remain qualitatively the same as in vacuum, but are just shifted closer together, with the amide I moved signiﬁcantly down in wavenumber [Figure 22.17] [117, 249, 267, 269, 339, 344, 400]. Bour’s method of transferring FF, APT, and AAT parameters from smaller molecules (oligomers) to simulate vibrational spectra for a larger molecule [138] has enabled us to compute spectra for many larger oligomers with regular structures, up to ∼65 residues in one case [143]. As models for FF are extended to longer (and for β-sheets, more broadly interacting) model systems, the exciton coupling causes most of the intensity to appear in a single or few modes, much as expected for an inﬁnite repeating structure. This results in more intense, sharper IR bands, and similarly more intense VCD, which is reﬂected in the VCD observed for long-persistence-length helices and for highly interacting aggregates like ﬁbrils [140, 143, 249, 251, 268]. Systems with more irregular structures need spectral parameters from several small oligomers to obtain an appropriate model, which poses added challenges. Segmenting a structure, computing the parts, and reassembling have given useful results for hairpins [269, 339] and have been extended to describe a small protein [147, 401]. It is best to calculate the entire peptide, or as large a fragment as possible, based on an NMR structure [148–150, 391], but even then compensation for missing elements, such as solvent and side chain effects, is important.

2

(a)

(d)

(b)

(e)

(c)

(f)

0

−2

Δe′ (× 102 )

−4 2 0

−2 −4 2 0

−2 −4 1800 1700 1600 1500 1400 1300 1200 1800 1700 1600 1500 1400 1300 1200 Wavenumber (cm−1)

Figure 22.17.

Effect of solvent correction on simulated VCD spectra for 21-amide (Ac–Ala20 –NHCH3 ) peptides in (a, d) α-helical, (b, e) 310 -helical, and (c, f) 31 -helical conformations obtained by transfer of DFT calculated parameters from shorter fragments. Isolated (vacuum) peptides (left, a, b, c) compared to corresponding solvent perturbed peptides (right, d, e, f) using PCM plus one layer of explicit solvent. Vertical lines indicate rotational strengths, R,

for individual modes and envelopes represent the sum of lorentzian bands for each mode scaled to the R values. Note that there is little amide II change except for the 31 -helix which has solvent accessible C=O groups.

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

The sensitivity of IR and VCD to structure ultimately depends on coupling of repeated elements in the structure, such as the amides in peptides. Computational modeling can provide theoretical values for such coupling, but the modes are not experimentally resolved, so tests of the model can be difﬁcult. Isotope labeling can overcome this and provide the needed link between computation, experiment, and site-speciﬁc conformational interpretation of helical and hairpin IR and VCD spectra. For helices, experimental results show that just two labeled amide groups are detectable in a 25-residue peptide for the amide I IR as a band shifted ∼40 cm−1 down in frequency [21, 49, 50, 140, 326]. The important observation, seen identically in computations and experiments, is that for two sequential labels, the 13 C amide I is more intense and closer to the 12 C amide I than if the labels alternate—that is, are separated by one residue. For VCD, the sequentially labeled 13 C segment gives rise to the same bandshape pattern as do the 12 C residues; but if alternately labeled, the VCD for the 13 C amide I band changes sign (Figure 22.18) [49]. Both the IR and VCD effects are consequences of the change in sign for the coupling between labeled residues for these two patterns, consequently ﬂipping the VCD sign pattern by reversing the order of the in- and out-of-phase components and making the higher exciton component more intense in the sequential IR, and less in the alternate, which is the source of the observed wavenumber shift between the two for the aggregate 13 C band. Thus VCD provides a direct measure of the sign of the vibrational coupling and, with IR, offers a means to derive its magnitude. More recently we have identiﬁed similar behavior for 31 - (PPII) and 310 -helices; but since the coupling strengths in these structures are smaller, the effects are harder to discern and require a bandshape analysis [50, 326]. Again it was the combination of IR and VCD that brought out the results, aided by theoretical modeling of the detailed interactions. Raman can also identify coupling, but it effectively yields the same result as IR (with opposite sign for PPII). Coupled-oscillator calculations, which utilize only dipolar coupling (DC) to simulate VCD, yield poorer representations of peptide VCD [264]. DC has been argued to make a critical contribution to β-sheet spectra [80, 250], and in fact it can provide the correct overall pattern, but the interactions are too small if realistic dipoles are used [143, 249, 252]. DC modeling can be used to ﬁt the IR spectra for β-sheet peptides if a sufﬁcient number and type of interactions are included as empirical parameters [250]. Since the ab initio MFP computations of the VCD for just a constrained dipeptide system replicate so many of the observed VCD features, it is evident that the short-range contributions represented in such a molecule are very important to understanding longer peptide VCD. The property transfer method of Bour, when coupled to a sufﬁciently diverse library of smaller peptides to represent short-range structure–spectra parameter relations, might be able to model local interactions of relatively large, hetero-structure polypeptides and proteins in the future. Coupling residues with TDC for long-range interactions can certainly help. Such efforts might encompass more protein structures in the near future [147, 401] and, given our recent success with simulating spectra for homogeneous-structured peptides, are likely to be much more accurate than empirically based parameter transfer methods.

22.6. PROTEIN VCD Proteins differ from small peptides in terms of the degree of solvation of the amide links and the uniformity of the secondary structure segments developed. While a helix in a peptide may terminate in a large number of conformations representing the “fraying”

735

736

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a)

4

0

0

−2

VCD

VCD

Simulated Experimental

−4

(b) 4

Unlabeled 2LT 2L1S

−4 −8

(d)

Unlabeled 2LT 2L1S

8

IR

2

IR

4

e (x1 03)

A (x 10)

Unlabeled 2LT 2L1S

Δe (x 10)

ΔA (x 105)

2

(c)

Unlabeled 2LT 2L1S

0

0 1700 1650 1600 Wavenumber (cm−1)

1775

1725 1675 1625 Wavenumber (cm−1)

Figure 22.18. Experimental (left, a, b) and simulated (right, c, d) VCD (top, a, c) and IR (bottom, b,d) for the amide I (in D2 O) of α-helical (Ala-rich) 25-mer peptides including double 13 C isotopically labeled variants with different separations. Spectra for unlabeled peptide (dashed–dot line) are compared with that for peptides having adjacent 13 C labels (2LT, dashed line) and 13 C labels separated by one residue (2L1S, solid line). The calculated spectra (c, d) are for an ideal structure in vacuum and reproduce the 13 C=O mode contributions in IR and VCD. With solvent correction the shape of the 12 C=O modes better matches experiment [343].

of that segment, in a protein segment termination is likely to be relatively well-deﬁned or, at least, conformationally constrained in a turn or loop sequence. Furthermore, the segment lengths in a protein are typically rather short but determined by the entire protein fold rather than by an isolated segment’s structural thermodynamic stability. Thus interpretation of VCD for proteins has followed bandshape deconvolution methods referenced to a set of proteins with known structures as are traditionally used for ECD, IR, and Raman spectra [53, 60, 65, 66, 68–70, 74, 75, 121, 402–406]. Protein spectra retain a qualitative mapping onto peptide VCD, but vary in ways dependent on their structural differences. These primarily stem from environment effects and irregularity differences in globular proteins, which are even further differentiated for membrane proteins, due to their heterogenous solvation, in and out of the membrane. Such structural properties impact on different physical probes to a varying degree, so it is best to interpret protein VCD spectra by controlling the methodology used to maintain a consistent conformation with that derived from ECD and FTIR data of the same system. This will reduce the possibility of drawing nonsensical conclusions from analyses of these relatively simple and modestly resolved spectral bandshapes.

22.6.1. Qualitative Spectral Interpretations Most protein VCD has been obtained for the amide I (D2 O buffer), but studies in H2 O are also feasible and data sets are available (Figure 22.19) [26, 112, 162, 175, 177, 225,

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

H2O

20 MYO

I

II

D2O

737

I’

MYO

IMUN ΔA x 105

10 IMUN

Figure 22.19. Protein VCD in the amide I and LCF

II regions for samples at ∼100 mg/mL (10 μL volume, 6-μm path) in H2 O and the amide I at

LCF

0

CAS

∼30 mg/mL (40 μl volume, 50 μm path) in D2 O, for myoglobin (MYO, highly helical),

CAS

immunoglobulin (IMUN, high sheet content), lactoferrin (LCF, sheet-helix mixture), and casein (CAS, intrinsically disordered protein).

−10

1700

1600 1500 Wavenumbers (cm−1)

1600

These data are from our previous VCD spectrometer to allow comparisons of H2 O (left) and D2 O (right) spectra for a selection of proteins.

314, 315, 407, 408]. With our new dispersive instrument, the ability to obtain VCD in the presence of relatively large H2 O absorbance has improved over published accounts and that in Figure 22.19, and commercial FT-based VCD instruments have also enabled studies of less concentrated samples in H2 O [166]. Of the examples shown in Figure 22.19, myoglobin (MYO) has a very high fraction of α-helix, while immunoglobulin (IMUN) has substantial β-sheet component with little helix, and lactoferrin (LCF) has substantial contributions of each. These three proteins have overlapped ECD spectra (not shown), where the longer wavelength (∼220 nm) intensity depends strongly on helix content, which is a strength of ECD analyses for α-helical fractions [26, 56–58, 66, 409]. The FTIR spectra of these proteins (not shown) are also quite similar, the primary difference being an amide I frequency shift reﬂecting β-sheet content, which often appears as a distorted shape peaked at ∼1630 cm−1 . In the VCD spectra of the same three proteins, strong variations are seen in the bandshapes with signiﬁcant frequency shifts [27, 29, 162, 176, 177, 225, 408]. The highly helical MYO has an amide I VCD dominated by a positive couplet, with an added weak negative feature to low energy when N-deuterated (amide I ), matching the spectral pattern measured for model α-helical polypeptides (compare Figures 22.9, 22.10, and 22.19). By contrast, the IMUN amide I VCD is predominantly negative with the main feature falling between 1630 and 1640 cm−1 , and a couplet shape amide I when N-D, reﬂecting aspects of the poly-l-lysine antiparallel β-sheet VCD (Figure 22.10) but more like the hairpin results (Figure 22.15). Globular proteins with a mix of α and β components, such as LCF (Figure 22.19), have amide I and I VCD spectra resembling a linear combination of these two more limiting types. Natively disordered proteins such as CAS, as well as those denatured by thermal unfolding, show a weak and broadened amide I VCD couplet, qualitatively reﬂecting that of PGA or PLL, “coil-form” peptides (compare Figure 22.10, 22.12, and 22.19).

738

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Considering the amide II in H2 O, MYO (α) has an intense negative amide II VCD lower in wavenumber than the IR, while IMUN (β) has a weaker, negative couplet amide II VCD and LCF (αβ) has a mix of those forms [315]. These qualitatively match the characteristics seen in peptide model systems. Amide III VCD for α and β proteins is very broad, with helical and sheet proteins on the whole showing opposite sign patterns over the whole 1400 to 1250-cm−1 region, helices having positive VCD and sheets negative, but both with very weak intensity, as in the IR [177, 410, 411]. While VCD can access many transitions to yield added structural selectivity, VCD bandshape patterns vary less in both the amide II and III regions than for amide I spectra, resulting in the amide I having most of the variation and consequently being the band used for most analyses [26, 29].

22.6.2. Applications of VCD for Protein Structure A number of applications have been reported that apply VCD-based analyses to determination of conformational aspects of speciﬁc proteins. These have been the topics of previous reviews, and only a few will be noted here. At least in our laboratory, VCD has been found to be more broadly useful for peptide rather than protein applications, due to the potential for detailed theoretical correlation. There remains a place for protein VCD. In particular, its ability to probe environments (lipid membranes, aggregates, for example) not well suited to ECD continue to make it and IR very useful. Bovine α-lactalbumin (bLA) is a protein whose structure appears to be unusually malleable and as such has been the focus of many studies of what is termed the molten globule transition [412]. At low pH, bLA, as well as human and other variants, expands and loses tertiary structure, but VCD analysis showed that it gains α-helix content [28]. Although the bLA crystal structure is very similar to that of hen egg white lysozyme (HEWL), its ECD, FTIR, and VCD spectra are noticeably different [413]. The spectral results indicate that the bLA structure dynamically ﬂuctuates in solution, in contrast to the crystal, which may make it a better co-factor for β-galactosyl transferase. Cytochrome c (cyt c) is a small globular protein with a helical bundle fold, which has misfolded or intermediate folded states, as identiﬁed with various spectroscopic probes [414–416]. Thermal changes monitored with VCD distinguished between the various intermediate forms [417–419], showing the acid-denatured state to have a residual helical component that was reversibly lost on heating (Figure 22.20). The comparison of FTIR, VCD, ECD, and UV–vis absorbance and ﬂuorescence allows one to isolate various components of the folding change under these different perturbations. A common feature of misfolding is aggregation to form extended β-sheet structures, which are ﬂatter and more extended than the twisted β-sheets seen in globular proteins. An example is concanavalin A (con A) which has relatively ﬂat sheets in a tetrameric structure at neutral pH, but is monomeric at low pH. When heated, the con A aggregates, as seen by an IR band growing in at <1620 cm−1 and a shift in the VCD shape [420]. VCD is not normally the method of choice for studying aggregation, but loss of VCD intensity coupled to growth of the very low wavenumber amide I band in the IR is a useful combination of spectral tools for analysis. Insulin is a small protein with two separate chains coupled together in a highly helical native state that is only soluble in low pH environments and exists as a multimeric state in most cases. By change of pH, addition of salt, and increase in concentration, insulin can be induced to form ﬁbril structures, and these can have multiple forms [421–424]. Recently, it has been shown that the ﬁbrils give rise to unique and very intense VCD bands that can change sign, depending on growth conditions, to provide a measure of

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

2nd derivative

4 pH = 1.7

pH = 1.7

pH = 9.1

t = 40°C

4

0

t = 5°C

−4 −8

t = 30°C

t = 65°C

2

pH = 1.7

0.2

t = 75°C

0

0.0 0.6

Absorbance

t = 40°C

ΔA x 105

Absorbance

0.4

pH = 9.1

t = 70°C t = 85°C

0.4

−2

0.2 −4

0.0 1700

1600

1700 1600 1700 Wavenumbers (cm−1)

1600

Figure 22.20. Thermal denaturation of cytochrome c at pH 9.1 and 1.7 in D2 O monitored with FTIR (left) and VCD (center and right). The process at high pH leads to aggregation, evident in the growth of a 1615 cm−1 shoulder in the FTIR as the main band broadened and lost intensity (lower left). The transition to a disordered state occurred only above 75◦ C at high pH, as seen in the pH 9.1 VCD data (far right). Lowering pH also lowered the stability, leading to an unfolded state at ∼40◦ C, which was reversible; the original state could be recovered (center spectra). The second-derivative representation of the FTIR (top left) illustrate the changes and complexity more clearly. The low pH state, though expanded, retained much secondary structure as seen from comparison of the VCD at low and high pH (middle and right, top spectra).

local conformation or of inter-strand packing [164, 165]. These insulin results directly reﬂect phenomena leading to the intense VCD seen in model peptides such as poly-Lys and poly-Glu (Figure 22.16) when they form aggregates and gels [237]. An important question in protein structure and folding is the role of protein interaction with membranes. Here the problem often is one of overcoming the interference due to absorbance or scattering from the lipids and the self-assembled structures they can form (micelles and vesicles). This can be a particular issue for ECD, although some techniques have been proposed to circumvent the scattering problem [425–427]. We have carried out a number of studies using various spectroscopic techniques to demonstrate the unfolding and insertion mechanism following β-lactoglobulin (βLG) interaction with negatively charged membranes [428–430]. VCD can easily measure spectra of βLG both in the native state (mostly sheet-like with a small fraction of helix) and in a DMPG lipid-vesicle (more helical, with a small fraction of sheet) to differentiate the secondary structure β → α change on vesicle binding, as illustrated in Figure 22.21. Protein conformational changes upon addition of metal ions can often be important. If the salts added have Cl− counterions, their UV absorbance can pose a problem for CD, limiting the concentrations that can be used, but such ions have no impact on VCD measurement. While salt variations have been studied with a number of systems

739

740

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

2

ΔA x 105

1 0 −1

VCD

−2 −3 0.5

IR

BLG BLG + DMPG

Absorbance

0.4 0.3

Figure 22.21. Overlapped experimental VCD (top) and IR

0.2

(bottom) spectra of βLG in its native state (dashed line) and in a DMPG lipid vesicle, showing increased helicity (solid

0.1 0.0 1750

1700

1650

Wavenumber (cm−1)

1600

line) with a 1:6 ratio, both measured in phosphate/D2 O buffer at pH 4.6. The intense, resolved band at ∼1740 cm−1 in the IR is due to the phosphate ester and contributes negligible VCD with no interference at the amide I band.

[28, 342, 417, 429, 431], the conformational change of calmodulin to form a long helical segment upon addition of Ca+ is a good example [432]. More problematic is study of changes in a speciﬁc protein on protein–peptide or protein–protein interactions. Use of isotopic labeling, here global rather than speciﬁc, to shift the spectra of one interacting species away from that of the other can provide insight as to the separate changes in the two species detectable with VCD and IR, but such a method is not possible for ECD studies [432]. Further quantifying structural changes correlated to environmental perturbations (such as in Figure 22.20) can be done using difference spectra [28, 413, 431] or various bandshape deconvolution methods [73, 405]. Factor analysis (or, equivalently, principal component or singular value decomposition analyses) [374] of a set of such spectra obtained for varying degrees of the perturbation normally yields the average spectrum and the major linearly independent variation from it as the ﬁrst and second components. The shape of the second component can often be interpreted in terms of the dominant type of structural change in this process, and its loading will represent the extent of shift in the equilibrium [52, 226, 433]. An alternative approach is to use 2-D correlation analyses to help resolve the spectral consequences of the structural changes resulting as a function of perturbation [434].

22.6.3. Protein Secondary Structure Quantitative Analyses As noted above, analyses of protein VCD spectra in terms of the fractional components (FC) of their secondary structure has followed methods based on bandshape analyses, as used for ECD and for some FTIR protein secondary structure determinations [53, 54, 60, 70]. Our particular approach centered on the use of the principal component method

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

of factor analysis (PC/FA) [374, 435, 436] to characterize the protein spectra in terms of a weighted combination of a relatively small number of coefﬁcients [26, 27, 29, 175] and has previously been reviewed in depth [15–17, 34, 35, 175]. Consequently, we will provide only some overall general comments here and leave the details to previous works, since our methods have not changed for some time. In PC/FA, the ﬁrst component (or factor) is essentially an average and the second is the major deviation over the training set, with each successive component becoming less signiﬁcant, eventually representing just the spectral noise. These form orthogonal components in a function space of the spectra, but their orientations (i.e., composition) depend on the training set. The original spectra, θi , can be reconstructed as a linear combination of factors (or spectral components), j , and their loadings, αij , as θi = j αij j .

(22.2)

Truncating the sum to include just the factors containing structural information yields a more compact representation by eliminating those j corresponding to noise. We have applied this method to a number of different data sets—CD, FTIR, Raman, and VCD (in D2 O and H2 O)—for varying sizes of training sets. For example, amide I VCD (D2 O) spectra of 28 proteins [27] yielded six subspectral components of signiﬁcance that could accurately reconstruct the input spectra (with reduced noise). By contrast, the amide I + II VCD (H2 O) needed more components, as did the FTIR [29, 76, 77, 178], while ECD spectra (to 190 nm) were sufﬁciently reproduced with ﬁve components [174, 176]. The loadings of these spectral components for each protein provide a small set of variables characterizing that protein’s spectral bandshapes. We established regression relationships between the loadings and structural parameters of interest and tested them for correlation with speciﬁc structural components [26, 29, 53, 65, 68, 437]. Our approach was unusual in employing restricted multiple regression analyses (RMR) between the loadings (αij ) of the spectra and the fractional composition (FCς i ) values for the i proteins in the set, ς structural types (e.g., α-helix, β-sheet, etc.), and j subspectral coefﬁcients derived from PC/FA [26, 29, 175]. We applied the RMR method to several types of spectra and their combinations, all taken from our own libraries of spectra, which are available on request but are now old. Others have taken related approaches and have developed independent databases, some of which are available commercially (BioTools, Jupiter FL) [77, 438]. Other pattern recognition techniques could be applied to analysis of VCD data sets, and numerous methods have been developed for ECD and FTIR studies [54, 58, 60, 65, 69, 70, 177, 178, 314]. The PROTA software available from BioTools, Inc. uses an algorithm similar to that developed in and available from our labs but includes a more accessible program and interface. The RMR method has been tested extensively with protein amide I and II spectra measured in D2 O and H2 O with FTIR and VCD as well as for ECD data and has been found to work best when data from two independent spectral regions, IR and UV, are used in a combined analysis to predict FC values. For example, VCD analysis was better for β-sheet, but ECD was much better for α-helix [26, 176], so that combining them gives the best of both methods. An important observation was that the best secondary structure predictions are found to be correlated with only a few coefﬁcients [26, 29, 175]. Some ECD-based analyses have taken alternative approaches, selectively limiting the training set, cutting off the number of loadings, or optimizing the method for self-consistent prediction behavior [68, 77, 437, 438]. The interdependence of helix and sheet content can obscure tests of

741

742

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

prediction algorithms, as demonstrated using a neural network study of crystal structure data for ∼450 nonhomologous structures in the protein data bank (PDB) [17, 179, 439]. If the structural elements are correlated, a physical measurement can sense one and predict the other, but in a dependent manner. Outlier proteins, those not following the correlation pattern, will be predicted in error. Combination of the long-range dependence of ECD with the short-range sensitivity of vibrational spectra (FTIR or VCD) is the key to better secondary structure prediction. The combined analysis gains stability and has a signiﬁcant reduction in the errors associated with outliers. Other researchers using independent methods have found similar improvements for combining ECD and FTIR data, thereby also sensing long- and short-range interactions [77, 120]. The best of both methods results in superior α-helix prediction, dependent mostly on ECD, combined with the special β-sheet sensitivity of IR and VCD [26, 29]. Since the best RMR predictions of average secondary structure neglect some signiﬁcant components of the spectral data, there must be more information content potentially available in the optical spectra, especially ECD and VCD [26, 29, 175]. It is natural to assume that the short-range-dependent techniques such as VCD will sense the distortions characteristic of the conformation of residues at the ends of uniform segments of secondary structure in a different manner than does ECD with its longer range sensitivity. Consequently, methods have been developed to predict the number of segments of uniform structure (helices, sheet strands, etc.) in a globular protein, which is equivalent to knowing their average length, once one knows the FC values and the size of the protein [67, 440, 441]. Our approach used a matrix descriptor of the segment distribution and neural network analyses to develop a predictive algorithm from a training set of spectral data. VCD and FTIR obtained for the thermal denaturation of ribonuclease T1 were used to demonstrate that the helix segment is lost before the sheet segments on heating [407]. Similarly, the number of intramembrane helices in aquaporin was predicted correctly using FTIR data [442]. Others have used a simpler algorithm of predicting a fraction of distorted helix and sheet and assigning that fraction to a ﬁxed number of residues per segment to attain a measure of the average number of segments [67]. We have shown that FTIR data, incorporating systematic H–D isotope exchange, can be used to generate even better predicted FC values than possible with conventional ECD or VCD [314]. The perturbation of the spectral data set appears to distinguish helix and turn contributions and leads to a more stable analysis. Extension of this isotope approach to include kinetic H–D exchange data and VCD spectra obtained with our improved instrumentation may yield even further improvement. This same data set has also been used to predict fold types, at least at the level of intersegment contact maps, thereby yielding tertiary structure information [443]. In summary, the key to utilization of VCD or any spectroscopic technique for quantitative structure analysis is to establish “reliability.” We and others [53, 58, 65, 77, 444] have tested algorithms for interpretation of spectral data to determine their most useful applications. For example, we used a neural network to correlate ECD and VCD spectra for the protein training set discussed above [408]. The ECD could be predicted with reasonable accuracy from the amide I VCD, but the reverse was not true. Thus, despite its added noise, the VCD was shown to have a higher information content, presumably from its higher resolution and shorter length dependence. We also used 2-D correlation methods to show which regions of a given spectral data set are dependent on a speciﬁc secondary structure type by treating the α-helix or β-sheet content of proteins in a training set as the perturbation variable [434, 445, 446]. Hetero-correlation between different

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

types of spectra (e.g., ECD with IR and Raman, IR with Raman, VCD with IR, and VCD with Raman) using 2-D maps can help identify unique bands with their secondary structure source by using the more easily assigned method (e.g., ECD), to assign a more difﬁcult one (e.g., Raman).

22.7. SUMMARY: WHERE ARE WE NOW? The role of VCD as compared to its more widely used optical spectral counterparts, ECD and FTIR, for biomolecular structure studies was the focus of this chapter. Clearly, FTIR has signiﬁcant S/N advantages over VCD, but less clear conformational discrimination, and ECD has considerable sampling ﬂexibility, allowing study of more dilute solutions with little or no solvent interference, but has little resolution of overlapped contributions. Due to its very high S/N, FTIR can be enhanced by use of deconvolution or derivative techniques [30, 73, 111] and by nonlinear 2-D coherence methods [204, 254, 447] and it is capable of very fast timescale dynamics studies to monitor structural change and biomolecular functions [115, 151, 448–450]. However, interpretation of FTIR data depends on a frequency correlation of band features with secondary structure types, and these are subject to perturbation by the environment. Since VCD analysis depends on bandshape correlations, it partially decouples the vibrational spectra from this frequency dependence. However, for quantitative analyses, the PC/FA methods use the same frequency base for all proteins; thus frequency shifts due to environment will degrade accuracy. Perhaps a future method could correct for such VCD shifts by correcting with correlation to the FTIR. Raman spectral analyses parallel FTIR strengths and weaknesses, and good S/N is possible but normally this requires high sample concentrations. More vibrational modes are accessible due to the low scattering of water, for which the spectra can be corrected. These modes tend to favor local parts of the biomolecule (such as chromophores, disulﬁdes, or aromatic side chains) and thus can provide site-directed structural information. The signed aspect of optical activity data, with its direct dependence on structure, gives CD-based measurements an important dimension beyond frequency assignment or even other differential measurements. For proteins, the strong dependence of far-UV ECD bands on the α-helix contribution has been both a strength (quick helical analyses) and weakness (lack of sensitivity to sheets). Furthermore, interfering absorbances, particularly from aromatic groups, do impact ECD spectra but cause little problem for VCD [55]. Similarly in DNA, minor modiﬁcations of the bases can shift the π -electron states enough to seriously distort components of the spectrum, while VCD maintains a clear helical dependence, particularly in the –PO2 − modes [300]. Both aspects are exacerbated by the intrinsically low resolution of ECD. Thus use of VCD and ECD together accesses the beneﬁts of both techniques, despite the limitations of both. ROA offers spectra whose sign patterns are conformationally dependent as well, with the limitations and strengths noted above for Raman. ROA spectra tend to reﬂect local structure more than the extended coupling that dominates ECD and VCD of biopolymers, which suggests a complementarity of application and structural response [451]. Theoretical modeling of structure–spectra relationships will continue to grow as a major component of utilizing these techniques for biopolymer structure studies. Here VCD and IR have a major advantage in being ground-state properties that are intrinsically easier to compute. This has already shown its strength in small molecule absolute conﬁguration studies and will continue to develop in biopolymer work, with the largest impact being for peptides, due to the simpler modeling possible. An added advantage of

743

744

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

probing speciﬁc states of the vibrational potential surface is the ability to add site-speciﬁc resolution to the spectra–structure correlation study by means of isotopic substitution. This rediscovery of an old technique is affecting a wide range of studies, particularly involving dynamics. Finally, the new nonlinear methods offer a future possibility of using VCD along with IR to measure very fast dynamics for structural change in biopolymers [204–206]. Questions of biological structure are too important and too complex to be addressed by analysis of only a single technique having known (and sometimes hidden) limitations. The virtues of taking data from multiple spectroscopic techniques and ﬁnding the structural model that can satisfactorily encompass all the data obtained cannot be overemphasized. Each technique can probe the biomolecular structure with differing physical sensitivity and act as a control on the interpretation of the others. For some questions, VCD may give important insight, and for most questions it can offer critical data, even if only in a conﬁrmatory role as a supplement to IR, Raman, and ECD, but using all these methods together will add to the overall picture of biopolymer conformation and folding.

ACKNOWLEDGMENTS Development of biopolymer applications of VCD was originally supported in the UIC lab by a succession of grants from the National Institutes of Health, and subsequently peptide studies were continued under various grants from the National Science Foundation (currently: CHE 07-18543), Guggenheim Foundation, Humboldt Foundation, and Donors of the Petroleum Research Fund administered by the American Chemical Society. The results described here are the product of the dedicated work of a number of talented postdoctoral and graduate student coworkers, who have been cited and referenced. Their effort has enabled VCD to become a detailed, biomolecular structural tool. Many of the samples for special applications came to us through generous gifts and/or collaborations with co-authors from around the world, as cited in the publications, to whom we are most grateful.

REFERENCES 1. L. A. Naﬁe, G. S. Yu, X. Qu, T. B. Freedman, Faraday Discuss. 1994, 99 , 13–34. 2. P. J. Stephens, F. J. Devlin, C. S. Ashvar, C. F. Chabalowski, M. J. Frisch, Faraday Discuss. 1994, 99 , 103–119. 3. P. J. Stephens, F. J. Devlin, J.-J. Pan, Chirality 2008, 20 , 643–663. 4. M. Diem, Vibr. Spectra Struct. 1991, 19 , 1–54. 5. L. A. Naﬁe, Annu. Rev. Phys. Chem. 1997, 48 , 357–386. 6. L. A. Naﬁe, T. B. Freedman, Vibrational optical activity theory, in Circular Dichroism: Principles and Applications, N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, 2000, pp. 97–131. 7. L. A. Naﬁe, Nat. Prod. Commun. 2008, 3 , 451–466. 8. V. Andrushchenko, J. L. McCann, J. H. van de Sande, H. Wieser, Vib. Spectrosc. 2000, 22 , 101–109. 9. V. Andrushchenko, D. Tsankov, H. Wieser, J. Mol. Struct. 2003, 661–662, 541–560. 10. S. Mason, Enantiomer 1998, 3 , 283–297.

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

11. P. L. Polavarapu, J. He, Anal. Chem. 2004, 76 , 61A–67A. 12. P. L. Polavarapu, Vibrational spectra: Principles and Applications with Emphasis on Optical Activity, Elsevier, New York, 1998. 13. P. L. Polavarapu, C. Zhao, Fresenius’ J. Anal. Chem. 2000, 366 , 727–734. 14. T. Taniguchi, N. Miura, S. Nishimura, K. Monde, Mol. Nutr. Food Res. 2004, 48 , 246–254. 15. T. A. Keiderling, Peptide and protein conformational studies with vibrational circular dichroism and related spectroscopies, in Circular Dichroism: Principles and Applications, N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, pp. 621–666. 16. T. A. Keiderling, Protein structural studies using vibrational circular dichroism spectroscopy, in Spectroscopic Methods for Determining Protein Structure in Solution, H. Havel, ed., VCH Publishers, New York, 1995, pp. 163–169. 17. T. A. Keiderling, Vibrational circular dichroism applications to conformational analysis of biomolecules, in Circular Dichroism and the Conformational Analysis of Biomolecules, Ed. G. D. Fasman. Plenum, New York, 1996, pp. 555–598. 18. T. A. Keiderling, NATO ASI Ser., Ser. E , 1997, 342 , 299–317. 19. T. A. Keiderling, R. A. G. D. Silva, G. Yoder, R. K. Dukor, Bioorg. Med. Chem. 1999, 7 , 133–141. 20. T. A. Keiderling, Curr. Opin. Chem. Biol . 2002, 6 , 682–688. 21. T. A. Keiderling, P. Pancoska, Structural studies of biological macromolecules using vibrational circular dichroism, in Biomolecular Spectroscopy, Part B , R. E. Hester and R. J. H. Clarke, eds., Wiley, Chichester, 1993, pp. 267–315. 22. T. A. Keiderling, J. Kubelka, J. Hilario, Vibrational circular dichroism of biopolymers. Summary of methods and applications, in Vibrational Spectroscopy of Polymers and Biological Systems, M. Braiman and V. Gregoriou, eds., Taylor & Francis, Atlanta, 2006, pp. 253–324. 23. S. Abbate, E. Castiglioni, F. Gangemi, R. Gangemi, G. Longhi, Chirality, 2009, 21 , e242–e252. 24. J. Sadlej, J. C. Dobrowolski, J. E. Rode, Chem. Soc. Rev . 2010, 39 , 1478–1488. 25. T. B. Freedman, X. Cao, R. K. Dukor, L. A. Naﬁe, Chirality, 2003, 15 , 743–758. 26. P. Pancoska, E. Bitto, V. Janota, M. Urbanova, V. P. Gupta, T. A. Keiderling, Protein Sci . 1995, 4 , 1384–1401. 27. P. Pancoska, S. C. Yasui, T. A. Keiderling, Biochemistry 1991, 30 , 5089–5103. 28. T. A. Keiderling, B. Wang, M. Urbanova, P. Pancoska, R. K. Dukor, Faraday Discuss. 1994, 99 , 263–286. 29. V. Baumruk, P. Pancoska, T. A. Keiderling, J. Mol. Biol . 1996, 259 , 774–791. 30. W. Surewicz, H. H. Mantsch, D. Chapman, Biochemistry 1993, 32 , 389–394. 31. M. Jackson, H. H. Mantsch, Crit. Rev. Biochem. Mol. Biol . 1995, 30 , 95–120. 32. J. L. Arrondo, A. Muga, J. Castresana, F. M. Goni, Prog. Biophys. Mol. Biol . 1993, 59 , 23–56. 33. M. Urbanov´a, Chirality 2009, 21 , E215–E230. 34. T. A. Keiderling, Vibrational circular dichroism of proteins, polysaccharides, and nucleic acids, in Physical Chemistry of Food Processes. Advanced Techniques, Structures and Applications, I. C. Baianu, H. Pessen, and T. F. Kuminosinski, eds., Van Nostrand Reinhold, New York, 1993, pp. 307–337. 35. T. A. Keiderling, Vibrational circular dichroism of peptides and proteins: Survey of techniques, qualitative and quantiative analyses and applications, in Infrared and Raman spectroscopy of biological materials, H.-U. Gremlich and B. Yan, eds., Dekker, New York, 2001, pp. 55–100. 36. A. M. Polyanichko, H. Wieser, Vibrational circular dichroism and its applications to protein studies, in Methods in Protein Structure and Stability Analysis, Nova Science Publishers, Hauppauge, NY, 2007, pp. 267–302.

745

746

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

37. R. Mendelsohn, G. Mao, C. R. Flach, Biochim. Biophys. Acta Biomembr. 2010, 1798 , 788–800. 38. P. Carmona, M. Molina, A. Rodriguez-Casado, On the nucleic acid structures in triplexes and protein complexes as revealed by Raman and infrared spectroscopy, in Insights into Vibrational Spectroscopy of Nucleic Acids and Their Complexes, C. M. Muntean, ed., Transworld Research Network, Trivandrum, India, 2009, pp. 87–100. 39. M. Tsuboi, G. J. Thomas, Jr., Polarized Raman and polarized infrared spectroscopy of proteins and protein assemblies, in Methods in Protein Structure and Stability Analysis, V. N. Uversky and E. A. Permyakov, eds., Nova Science Publishers, Hauppauge, NY, 2007, pp. 153–194. 40. Y. S. Kim, R. M. Hochstrasser, J. Phys. Chem. B 2009, 113 , 8231–8251. 41. K. Gerwert, C. Koetting, Bioforum 2008, 31 , 36–38. 42. Z. Ganim, H. S. Chung, A. W. Smith, L. P. DeFlores, K. C. Jones, A. Tokmakoff, Acc. Chem. Res. 2008, 41 , 432–441. 43. J. L. R. Arrondo, I. Iloro, M. Garcia-Pacios, F. M. Goni, Springer Ser. Biophys. 2006, 10 , 73–88. 44. A. Barth, Biochim. Biophys. Acta, Bioenerg. 2007, 1767 , 1073–1101. 45. Y. Zhu, T. Wang, F. Gai, New challenges. Laser-induced T-jump method: A nonconvectional photoreleasing approach to study protein folding, in Dynamic Studies in Biology, M. Goeldner and R. Givens, eds., Wiley-VCH Verlag, 2005, pp. 461–478. 46. H. Fabian, D. Naumann, Methods (San Diego, CA, U. S.) 2004, 34 , 28–40. 47. S. M. Decatur, Acc. Chem. Res. 2006, 39 , 169–175. 48. H. H. Mantsch, D. Chapman, Infrared Spectroscopy of Biomolecules, Wiley-Liss, Chichester, UK, 1996. 49. R. Huang, J. Kubelka, W. Barber-Armstrong, R. A. G. D. Silva, S. M. Decatur, T. A. Keiderling, J. Am. Chem. Soc. 2004, 126 , 2346–2354. 50. H. Chi, A. Lakhani, A. Roy, M. Nakaema, T. A. Keiderling, J. Phys. Chem. B 2010, 114 , 12744–12753. 51. R. Huang, V. Setniˇcka, M. A. Etienne, J. Kim, J. Kubelka, R. P. Hammer, T. A. Keiderling, J. Am. Chem. Soc., 2007, 129 , 13592–13603. 52. R. A. G. D. Silva, J. Kubelka, S. M. Decatur, P. Bour, T. A. Keiderling, Proc. Natl. Acad. Sci. USA 2000, 97 , 8318. 53. W. C. Johnson, Annu. Rev. Biophys. Biophys. Chem. 1988, 17 , 145–166. 54. N. Sreerama, R. W. Woody, Circular dichroism of proteins and peptides, in Circular Dichroism: Principles and Applications, N. Berova, K. Nakanishi, R. W. Woody, eds., Wiley-VCH, New York, 2000, pp. 601–620. 55. R. W. Woody, Theory of circular dichroism of proteins, in Circular Dichroism and the Conformational Analysis of Biomolecules, G. D. Fasman, ed., Plenum Press, New York, 1996, pp. 25–67. 56. J. T. Yang, C. S. C. Wu, H. M. Martinez, Meth. Enzymol . 1986, 130 , 208–269. 57. M. Manning, J. Pharmaceut. Biomed. Anal . 1989, 7 , 1103–1119. 58. S. Y. Venyaminov, J. T. Yang, Determination of protein secondary structure, in Circular Dichroism and the Conformational Analysis of Biomolecules, G. D. Fasman, ed., Plenum Press, New York, 1996, pp. 69–107. 59. N. J. Greenﬁeld, Trends Anal. Chem., 1999, 18 , 236–244. 60. N. Sreerama, R. W. Woody, Anal. Biochem., 2000, 287 , 252–260. 61. W. C. Johnson, Determination of the conformation of nucleic acids by electronic CD, in Circular Dichroism and the Conformational Analysis of Biomolecules., G. D. Fasman, eds., Plenum Press, New York, 1996, pp. 433–468.

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

62. J. C. Maurizot, Circular dichroism of nucleic acids: Nonclassical conformations and modiﬁed oligonucleotides., in Circular Dichroism. Principles and Applications, 2nd ed., N. Berova, K. Nakanishi, and R. W. Woody, ed., Wiley-VCH, New York, 2000, pp. 719–739. 63. W. C. Johnson, CD of nucleic acids, in Circular Dichroism: Principles and Applications, N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, pp. 703–718. 64. D. M. Gray, CD of protein–nucleic acid interaction, in Circular Dichroism: Principles and Applications, N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, pp. 769–796. 65. N. Sreerama, R. W. Woody, J. Mol. Biol . 1994, 242 , 497–507. 66. W. C. Johnson, Meth. Biochem. Anal . 1985, 31 , 61–163. 67. N. Sreerama, S. Y. Venyaminov, R. W. Woody, Protein Sci . 1999, 8 , 370–380. 68. N. Sreerama, R. W. Woody, Anal. Biochem. 1993, 209 , 32–44. 69. L. Whitmore, B. A. Wallace, Nucl. Acids Res. 2004, 32 , W668–W673. 70. L. Whitmore, B. A. Wallace, Biopolymers 2008, 89 , 392–400. 71. H. H. Mantsch, H. L. Casel, R. N. Jones, Resolution enchancement of infrared spectra of biological systems, in Spectroscopy of Biological Systems, R. J. H. Clark and R. E. Hester, eds., Wiley, London, 1986, pp. 1–46. 72. F. S. Parker, Applications of Infrared, Raman and Resonance Raman Spectroscopy, Plenum, New York, 1983. 73. D. M. Byler, H. Susi, Biopolymers 1986, 25 , 469–487. 74. F. Dousseau, M. Pezolet, Biochemistry 1990, 29 , 8771–8779. 75. K. A. Oberg, J. M. Ruysschaert, E. Goormaghtigh, Eur. J. Biochem. 2004, 271 , 2937–2948. 76. D. C. Lee, P. I. Haris, D. Chapman, R. C. Mitchell, Biochemistry, 1990, 29 , 9185–9193. 77. R. Pribic, I. H. M. Van Stokkum, D. Chapman, P. I. Haris, M. Bloemendal, Anal. Biochem. 1993, 214 , 366–378. 78. P. I. Haris, D. Chapman, Fourier transform infrared spectrscopic studies of biomembrane systems, in Infrared Spectroscopy of Biomolecules, H. H. Mantsch and D. Chapman, eds., Wiley-Liss, Chichester, 1996, pp. 239–278. 79. A. T. Tu, Raman Spectroscopy in Biology, Wiley, New York, 1982. 80. S. Krimm, J. Bandekar, Adv. Protein Chem. 1986, 38 , 181–364. 81. S. Krimm, Peptides and Proteins, in Biological Applications of Raman Spectroscopy, Vol. 1: Raman Spectra and the Conformations of Biological Macromolecules, T. G. Spiro, ed., Wiley, New York, 1987, pp. 1–45. 82. J. A. Hering, P. I. Haris, FTIR Spectroscopy for analysis of protein secondary structure, in Biological and Biomedical Infrared Spectroscopy, A. Barth and P. I. Haris, eds., IOS Press, Amsterdam, 2009, pp. 129–167. 83. M. Tsuboi, T. Nishimura, A. Y. Hirakawa, W. L. Peticolas, Resonance Raman spectroscopy and normal modes of nucleic acid bases, in Biological Applications of Raman Spectroscopy, T. G. Spiro, ed., Wiley, New York, 1987, pp. 109–179. 84. A. Annamalai, T. A. Keiderling, J. Am. Chem. Soc. 1987, 109 , 3125–3132. 85. E. Taillandier, J. Liquier, Meth. Enzymol . 1992, 211 , 307–335. 86. A. Letellier, M. Ghomi, E. Taillandier, J. Biomol. Struct. Dyn. 1986, 3 , 671–687. 87. A. Letellier, M. Ghomi, E. Taillandier, J. Biomol. Struct. Dyn. 1987, 4 , 663–683. 88. A. Letellier, M. Ghomi, E. Taillandier, J. Biomol. Struct. Dyn. 1989, 6 , 755–768. 89. J. A. Mondragon-Sanchez, J. Liquier, R. H. Shafer, E. Taillandier, J. Biomol. Struct. Dyn. 2004, 22 , 365–373. 90. J. Liquier, F. Geinguenaud, T. Huynh-Dinh, C. Gouyette, E. Khomyakova, E. Taillandier, J. Biomol. Struct. Dyn. 2001, 19 , 527–534. 91. C. Dagneaux, J. Liquier, E. Taillandier, Biochemistry 1995, 34 , 16618–16623.

747

748

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

92. 93. 94. 95. 96. 97. 98. 99. 100. 101.

102.

103. 104. 105.

106.

107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121.

L. J. Wang, L. G. Yang, T. A. Keiderling, Biophys. J . 1994, 67 , 2460–2467. L. Wang, T. A. Keiderling, Biochemistry 1992, 31 , 10265–10271. L. Wang, T. A. Keiderling, Nucl. Acids Res. 1993, 21 , 4127–4132. L. J. Wang, P. Pancoska, T. A. Keiderling, Biochemistry 1994, 33 , 8428–8435. V. Maharaj, D. S. H. Van, D. Tsankov, A. Rauk, H. Wieser, Mikrochim. Acta, Suppl . 1997, 14 , 529–530. V. Andrushchenko, H. Wieser, P. Bour, J. Phys. Chem. B 2002, 106 , 12623–12634. M. L. Krasteva, D. Tsankov, d. S. van, J. H., H. Wieser, Bulg. Chem. Commun. 2005, 37 , 243–255. D. Tsankov, M. Krasteva, V. Andrushchenko, J. H. van de Sande, H. Wieser, Biophys. Chem. 2006, 119 , 1–6. H. T. Miles, Proc. Natl. Acad. Sci. USA 1961, 41 , 791. H. A. Tajmir-Riahi, C. N. N’Soukpoe-Kossi, D. Joly, Structural analysis of protein–DNA and protein–RNA interactions by FTIR spectroscopy, in Biological and Biomedical Infrared Spectroscopy, A. Barth and P. I. Haris, eds., IOS Press, Amsterdam, 2009, pp. 288–311. T. Theophanides, Metal ion nucleic acid interactions as studied by Fourier transform infrared spectroscopy, in Spectroscopy of Inorganic Bioactivators. Theory and Applications—Chemistry, Physics and Medicine, T. M. Theophanides, ed., Kluwer Academic Publishers, Amsterdam, 1989, pp. 265–272. A. G. Petrovic, P. L. Polavarapu, J. Phys. Chem. B 2008, 112 , 2255–2260. J. M. Benevides, S. A. Overman, G. J. J. Thomas, J. Raman Spectr. 2005, 36 , 279–299. W. L. Peticolas, M. Tsuboi, Raman spectroscopy of nucleic acids, in Infrared and Raman Spectroscopy of Biological Macromolecules, T. M. Theophanides, ed., Reidel, London, 1978, pp. 153–165. G. J. Thomas, Water as a bioactivator and probe of DNA structure: Investigation by laser Raman spectroscopy, in Spectroscopy of Inorganic Bioactivators. Theory and Applications—Chemistry, Physics and Medicine, T. Theophanides, ed., Kluwer Academic Publishers, Amsterdam, 1989, pp. 247–263. G. J. Thomas, Drug–nucleic acid interactions, in Spectroscopy of Biological Systems, R. J. H. Clark, R. E. Hester. Wiley-Interscience, New York, 1986, pp. 233–309. G. J. Thomas, Annu. Rev. Biophys. Biomolec. Struct. 1999, 28 , 1–27. G. J. Thomas, Jr., M. Tsuboi, Adv. Biophys. Chem. 1993, 3 , 1–70. J. K. Kauppinen, D. J. Moffatt, H. H. Mantsch, D. G. Cameron, Appl. Spectrosc. 1981, 35 , 271–276. W. K. Surewicz, H. H. Mantsch, Biochim. Biophys. Acta 1988, 952 , 115–130. P. Pancoska, L. Wang, T. A. Keiderling, Protein Sci . 1993, 2 , 411–419. S. Krimm, W. C. Reisdorf Jr., Faraday Disc. 1994, 99 , 181–194. N. N. Kalnin, I. A. Baikalov, S. Y. Venyaminov, Biopolymers 1990, 30 , 1273–1280. R. B. Dyer, F. Gai, W. H. Woodruff, Acc. Chem. Res. 1998, 31 , 709–716. S. Williams, T. P. Causgrove, R. Gilmanshin, K. S. Fang, R. H. Callender, W. H. Woodruff, R. B. Dyer, Biochemistry 1996, 35 , 691–697. D. R. Turner, J. Kubelka, J. Phys. Chem. B , 2007, 111 , 1834–1845. S. Y. Venyaminov, N. N. Kalnin, Biopolymers, 1990, 30 , 1259–1271. H. H. J. De Jongh, E. Goormaghtigh, J. M. Ruysschaert, Anal. Biochem., 1996, 242 , 95–103. R. W. Sarver, W. C. Kruger, Anal. Biochem. 1991, 199 , 61–67. R. W. Williams, A. K. Dunker, J. Mol. Biol . 1981, 152 , 783–813.

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

122. E. Goormaghtigh, FTIR data processing and analysis tools, in Biological and Biomedical Infrared Spectroscopy, A. Barth and P. I. Haris, eds., IOS Press, Amsterdam, 2009, pp. 104–128. 123. Y. Ozaki, S. Sasic, Two-dimensional correlation spectroscopy of biological and polymeric materials, in Vibrational Spectroscopy of Polymers and Biological Systems, M. Braiman and V. Gregoriou, eds., Taylor & Francis, Atlanta, 2006, pp. 153–214. 124. I. Noda, Two-dimensional correlation spectroscopy in vibrational spectroscopy, in Handbook of Vibrational Spectroscopy, P. R. Grifﬁths and J. M. Chalmers, eds., John Wiley & Sons, Chichester, 2002, pp. 2135–2172. 125. J. L. Arrondo, F. M. Goni, Prog. Biophys. Mol. Biol . 1999, 72 , 367–405. 126. P. L. Polavarapu, Vibr. Spectra Struct. 1989, 17B , 319–342. 127. L. D. Barron, L. Hecht, E. W. Blanch, A. F. Bell, Prog. Biophys. Mol. Biol . 2000, 73 , 1–49. 128. L. D. Barron, L. Hecht, Vibrational Raman optical activity: From fundamentals to biochemical applications, in Circular Dichroism: Principles and Applications, N. Berova, K. Nakanishi, and R. W. Woody, eds., Wiley-VCH, New York, 2000, pp. 667–701. 129. L. D. Barron, Curr. Opin. Struct. Biol . 2006, 16 , 638–643. 130. P. L. Polavarapu, Chem. Rec. 2007, 7 , 125–136. 131. L. A. Naﬁe, Raman optical activity, in Modern Nonlinear Optics, Part 3, M. Evans and S. Kielich, eds., Wiley, New York, 1994, pp. 105–206. 132. P. J. Stephens, M. A. Lowe, Annu. Rev. Phys. Chem. 1985, 36 , 213–241. 133. T. B. Freedman, L. A. Naﬁe, Top. Stereochem. 1987, 17 , 113–206. 134. T. A. Keiderling, Q. Xu, Adv. Protein Chem. 2002, 62 , 111–161. 135. T. B. Freedman, L. A. Naﬁe, Theoretical formalism and models for vibrational circular dichroism intensity, in Modern Nonlinear Optics, M. Evans and S. Kielich, eds., Wiley, New York, 1994, pp. 207–263. 136. P. J. Stephens, F. J. Devlin, A. Aamouche, Determination of the structures of chiral molecules using vibrational circular dichroism spectroscopy, in Chirality: Physical Chemistry, ACS Symposium Series, J. M. Hicks, ed., Oxford University Press, Washington. D.C., 2002, pp. 18–33. 137. P. J. Stephens, F. J. Devlin, Vibrational circular dichroism, in Continuum Solvation Models in Chemical Physics: From Theory to Applications, B. Mennucci and R. Cammi, eds., John Wiley & Sons, Chichester, 2007, pp. 180–205. 138. P. Bour, J. Sopkova, L. Bednarova, P. Malon, T. A. Keiderling, J. Comput. Chem. 1997, 18 , 646–659. 139. P. Bour, J. Kubelka, T. A. Keiderling, Biopolymers 2000, 53 , 380–395. 140. J. Kubelka, R. A. G. D. Silva, P. Bour, S. M. Decatur, T. A. Keiderling, Chirality in peptide vibrations. Ab initio computational studies of length, solvation, hydrogen bond, dipole coupling and isotope effects on vibrational CD, in The Physical Chemistry of Chirality, ACS Symposium Series, J. M. Hicks, ed., Oxford University Press, New York, 2002, pp. 50–64. 141. J. Kubelka, P. Bour, T. A. Keiderling, Quantum mechanical calculations of peptide vibrational force ﬁelds and spectral intensities, in Biological and Biomedical Infrared Spectroscopy, Ed. A. Barth and P. I. Haris, eds., IOS Press, Amsterdam, 2009, pp. 178–223. 142. P. Bour, J. Kubelka, T. A. Keiderling, Biopolymers 2002, 65 , 45–59. 143. J. Kubelka, T. A. Keiderling, J. Am. Chem. Soc. 2001, 123 , 12048–12058. 144. V. Andrushchenko, P. Bouˇr, Chirality 2010, 22 , E96–E114. 145. V. Andrushchenko, H. Wieser, P. Bour, J. Phys. Chem. A 2007, 111 , 9714–9723. 146. P. Bour, V. Andrushchenko, M. Kabel´ac, V. Maharaj, H. Wieser, J. Phys. Chem. B , 2005, 109 , 20579–20587. 147. J. A. Grahnen, K. E. Amunson, J. Kubelka, J. Phys. Chem. B 2010, 114 , 13011–13020.

749

750

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

148. P. Bour, J. Kim, J. Kapitan, R. Hammer, P., R. Huang, L. Wu, T. A. Keiderling, Chirality 2008, 20 , 1104–1119. 149. J. Hrud´ıkov´a, J. Kapit´an, V. Baumruk, R. P. Hammer, T. A. Keiderling, P. Petr Bouˇr, J. Phys. Chem. A 2010, 114 , 7642–7651. 150. R. Huang, L. Wu, D. McElheny, P. Bour, A. Roy, T. A. Keiderling, J. Phys. Chem. B 2009, 113 , 5661–5674. 151. K. Hauser, O. Ridderbusch, C. Krejtschi, A. Hellerbach, R. Huang, T. A. Keiderling, J. Phys. Chem.B , 2010, 114 , 11628–11637. 152. L. A. Naﬁe, T. A. Keiderling, P. J. Stephens, J. Am. Chem. Soc. 1976, 98 , 2715–2723. 153. L. A. Naﬁe, Appl. Spectrosc., 2000, 54 , 1634–1645. 154. P. L. Polavarapu, Fourier transform infrared vibrational circular dichroism, in Fourier Transform Infrared Spectroscopy, J. R. Ferraro and L. Basile, eds., Academic Press, New York, 1985, pp. 61–96. 155. T. A. Keiderling, Vibrational circular dichroism. Comparison of technique and practical considerations, in Practical Fourier Transform Infrared Spectroscopy, K. Krishnan, and J. R. Ferraro, eds., Academic Press, San Diego, 1990, pp. 203–284. 156. J. Hilario, D. Drapcho, R. Curbelo, T. A. Keiderling, Appl. Spectrosc. 2001, 55 , 1435–1447. 157. B. Wang, T. A. Keiderling, Appl. Spectrosc. 1995, 49 , 1347–1355. 158. P. Malon, T. A. Keiderling, Appl. Spectrosc. 1996, 50 , 669–674. 159. P. Malon, T. A. Keiderling, Appl. Spectrosc. 1988, 42 , 32–38. 160. M. Diem, Application of infrared CD to the analysis of the solution conformation of biological molecules, in Techniques and Instrumentation in Analytical Chemistry, N. Purdie and H. G. Brittain, eds., Elsevier, Amsterdam, 1994, pp. 91–130. 161. T. B. Freedman, L. A. Naﬁe, T. A. Keiderling, Biopolymers 1995, 37 , 265–279. 162. V. Baumruk, T. A. Keiderling, J. Am. Chem. Soc. 1993, 115 , 6939–6942. 163. X. Cao, R. D. Shah, R. K. Dukor, C. Guo, T. B. Freedman, L. A. Naﬁe, Appl. Spectrosc. 2004, 58 , 1057–1064. 164. D. Kurouski, R. A. Lombardi, R. K. Dukor, I. K. Lednev, L. A. Naﬁe, Chem. Commun. 2010, 46 , 7154–7156. 165. S. Ma, X. Cao, M. Mak, A. Sadik, C. Walkner, T. B. Freedman, I. K. Lednev, R. Dukor, L. A. Naﬁe, J. Am. Chem. Soc. 2007, 129 , 12364–12365. 166. S. Ma, T. B. Freedman, R. K. Dukor, L. A. Naﬁe, Appl. Spectrosc. 2010, 64 , 615–626. 167. G. Shanmugam, P. L. Polavarapu, Chirality 2009, 21 , 152–159. 168. A. G. Petrovic, P. L. Polavarapu, J. Phys. Chem. B 2006 110 , 22826–22833. 169. P. Zhang, P. L. Polavarapu, Appl. Spectrosc. 2006, 60 , 378–385. 170. F. Zhu, N. W. Isaacs, L. Hecht, L. D. Barron, Structure 2005, 13 , 1409–1419. 171. A. Lakhani, P. Malon, T. A. Keiderling, Appl. Spectrosc. 2009, 63 , 775–785. 172. L. A. Naﬁe, H. Buijs, A. Rilling, X. Cao, R. K. Dukor, Appl. Spectrosc. 2004, 58 , 647–654. 173. R. D. Shah, L. A. Naﬁe, Curr. Opin. Drug Discov. Devel . 2001, 4 , 764–775. 174. J. P. Hennessey, W. C. Johnson, Biochemistry 1981, 20 , 1085–1094. 175. P. Pancoska, E. Bitto, V. Janota, T. A. Keiderling, Faraday Discuss. 1994, 99 , 287–310. 176. P. Pancoska, T. A. Keiderling, Biochemistry 1991, 30 , 6885–6895. 177. B. I. Baello, P. Pancoska, T. A. Keiderling, Anal. Biochem. 1997, 250 , 212–221. 178. S. Wi, P. Pancoska, T. A. Keiderling, Biospectroscopy 1998, 4 , 93–106. 179. P. Pancoska, Circular dichroism in analysis of biomolecules, in Encyclopedia of Analytical Chemistry, R. A. Meyers, ed., John Wiley & Sons, Chichester, 2000, pp. 244–383. 180. J.-H. Choi, S. Hahn, M. Cho, Int. J. Quantum Chem., 2005, 104 , 616–634. 181. J.-H. Choi, J.-S. Kim, M. Cho, J. Chem. Phys., 2005, 122 , 174903–174913.

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

182. 183. 184. 185. 186. 187. 188. 189. 190. 191. 192. 193. 194. 195. 196. 197. 198. 199. 200. 201. 202. 203. 204. 205. 206. 207. 208. 209. 210. 211. 212.

213. 214. 215. 216.

J. Kim, R. Huang, J. Kubelka, P. Bouˇr, T. A. Keiderling, J. Phys. Chem. B 2006, 110 , 23590. J. Kim, T. A. Keiderling, J. Phys. Chem. B 2010, 114 , 8494–8504. C. N. Su, V. Heintz, T. A. Keiderling, Chem. Phys. Lett. 1981, 73 , 157–159. C. C. Chen, P. L. Polavarapu, S. Weibel, Appl. Spectrosc. 1994, 48 , 1218–1223. M. Diem, G. M. Roberts, O. Lee, O. Barlow, Appl. Spectrosc. 1988, 42 , 20–27. R. Devlin, P. J. Stephens, Appl. Spectrosc. 1987, 41 , 1142–1144. P. Xie, M. Diem, Appl. Spectrosc. 1996, 50 , 675–680. R. K. Yoo, B. Wang, P. V. Croatto, T. A. Keiderling, Appl. Spectrosc. 1991, 45 , 231–236. G. Yoder, Vibrational circular dichroism of selected model peptides. Ph.D. thesis, University of Illinois at Chicago, Chicago, IL, 1997. V. Setnicka, M. Urbanova, S. Pataridis, V. Kral, K. Volka, Tetrahedron Asymm. 2002, 13 , 2661–2666. V. Setnicka, M. Urbanova, V. Kral, K. Volka, Spectrochim. Acta Part A 2002, 58 , 2983–2989. M. Urbanova, V. Setnicka, K. Volka, Chemicke Listy 2002, 96 , 301–304. M. Urbanova, V. Setnicka, V. Kral, K. Volka, Biopolymers 2001, 60 , 307–316. M. Urbanova, V. Setnicka, K. Volka, Chirality 2000, 12 , 199–203. T. Kuppens, W. Langenaeker, J. P. Tollenaere, P. Bultnick, J. Phys. Chem. A 2003, 107 , 542–553. J. Dobler, N. Peters, C. Larsson, A. Bergman, E. Geidel, H. Huhnerfuss, J. Mol. Struct. (Theochem), 2002, 586 , 159–166. T. Tanaka, K. Inoue, T. Kodama, Y. Kyogoku, T. Hayakawa, H. Sugeta, Biospectroscopy, 2001, 62 , 228–234. S. Shin, A. Hirakawa, Y. Hamada, Enantiomer 2002, 7 , 191–196. C. S. Ashvar, F. J. Devlin, P. J. Stephens, J. Am. Chem. Soc. 1999, 121 , 2836–2849. P. K. Bose, P. L. Polavarapu, Carbohyd. Res. 1999, 322 , 135–141. L. A. Naﬁe, M. Citra, N. Ragunathan, G.-S. Yu, D. Che, Tech. Instrum. Anal. Chem. 1994, 14 , 53–89. T. A. Keiderling, Appl. Spectrosc. Rev . 1981, 17 , 189–226. M. Bonmarin, J. Helbing, Opt. Lett.,, 2008, 33 , 2086–2088. H. Rhee, Y.-G. June, J.-S. Lee, K.-K. Lee, J.-H. Ha, Z. H. Kim, S. J. Jeon, M. Cho, Nature 2009, 458 , 310–313. H. Rhee, S.-S. Kim, S.-J. Jeon, M. Cho, ChemPhysChem 2009, 10 , 2209–2211. G. A. Osborne, J. C. Cheng, P. J. Stephens, Rev. Sci. Instrum. 1973, 44 , 10–15. J. C. Cheng, L. A. Naﬁe, P. J. Stephens, J. Opt. Soc. Am. 1975, 65 , 1031–1035. T. A. Keiderling, P. J. Stephens, Chem. Phys. Lett. 1976, 41 , S. Abbate, G. Longhi, S. Boiadjiev, D. A. Lightner, C. Bertucci, P. Salvadori, Enantiomer 1998, 3 , 337–347. E. Castiglioni, F. Lebon, G. Longhi, S. Abbate, Enantiomer 2002, 7 , 161–173. L. A. Naﬁe, R. K. Dukor, The use of dual polarization modulation in vibrational circular dichroism spectroscopy, in The Physical Chemistry of Chirality, ACS Symposium Series, J. M. Hicks, ed., Oxford University Press, New York, 2002, pp. 79–88. F. Long, T. B. Freedman, R. Hapanowicz, L. A. Naﬁe, Appl. Spectrosc. 1997, 51 , 504–507. F. Long, T. B. Freedman, T. J. Tague, L. A. Naﬁe, Appl. Spectrosc. 1997, 51 , 508–511. L. A. Naﬁe, R. K. Dukor, J. R. Roy, A. Rilling, X. L. Cao, H. Buijs, Appl. Spectrosc. 2003, 57 , 1245–1249. M. Niemeyer, G. G. Hoffmann, B. Schrader, J. Mol. Struct. (Theochem) 1995, 349 , 451–454.

751

752

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

217. C. A. McCoy, J. A. de Haseth, Appl. Spectrosc. 1988, 42 , 336–341. 218. D. Tsankov, T. Eggimann, H. Wieser, Appl. Spectrosc. 1995, 49 , 132–138. 219. R. W. Bormett, G. D. Smith, S. A. Asher, D. Barrick, D. M. Kurz, Faraday Discuss. 1994, 99 , 327–339. 220. C. Marcott, A. E. Dowrey, I. Noda, Appl. Spectrosc. 1993, 47 , 1324–1328. 221. P. Malon, T. A. Keiderling, Appl. Optics 1997, 36 , 6141–6148. 222. D. L. Drapcho, R. Curbelo, E. Y. Jiang, R. A. Crocombe, W. J. McCarthy, Appl. Spectrosc. 1997, 51 , 453–460. 223. L. Shao, P. R. Grifﬁths, Anal. Chem. (Washington, DC, U. S.) 2008, 80 , 5219–5224. 224. R. L. Richardson, Jr., H. Yang, P. R. Grifﬁths, Appl. Spectrosc. 1998, 52 , 572–578. 225. P. Pancoska, S. C. Yasui, T. A. Keiderling, Biochemistry 1989, 28 , 5917–5923. 226. G. Yoder, P. Pancoska, T. A. Keiderling, Biochemistry 1997, 36 , 15123–15133. 227. G. Yoder, A. Polese, R. A. G. D. Silva, F. Formaggio, M. Crisma, Q. B. Broxterman, J. Kamphuis, C. Toniolo, T. A. Keiderling, J. Am. Chem. Soc. 1997, 119 , 10278–10285. 228. L. Wang, Vibrational circular dichroism studies of nucleic acid conformations. Ph.D. thesis, University of Illinois at Chicago, Chicago, IL, 1993. 229. S. Y. Venyaminov, N. N. Kalnin, Biopolymers 1990, 30 , 1243–1257. 230. M. Diem, H. Khouri, L. A. Naﬁe, J. Am. Chem. Soc. 1979, 101 , 6829–6837. 231. A. C. Sen, T. A. Keiderling, Biopolymers 1984, 23 , 1533–1545. 232. U. Narayanan, T. A. Keiderling, G. M. Bonora, C. Toniolo, Biopolymers 1985, 24 , 1257–1263. 233. U. Narayanan, T. A. Keiderling, G. M. Bonora, C. Toniolo, J. Am. Chem. Soc. 1986, 108 , 2431–2437. 234. D. M. Tigelaar, W. Lee, K. A. Bates, A. Sapringin, V. N. Prigodin, X. Cao, L. A. Naﬁe, M. S. Platz, A. J. Epstein, Chem. Mater. 2002, 14 , 1430–1438. 235. W. Dzwolak, Chirality 2010, 22 , E154–E160. 236. W. Dzwolak, V. Smirnovas, R. Jansen, R. Winter, Protein Sci . 2004, 13 , 1927–1932. 237. W. Dzwolak, A. Fulara, A. Lakhani, T. A. Keiderling, 2011, 115 , 11010–11016. 238. V. Baumruk, D. F. Huo, R. K. Dukor, T. A. Keiderling, D. Lelievre, A. Brack, Biopolymers 1994, 34 , 1115–1121. 239. R. A. G. D. Silva, S. A. Sherman, F. Perini, E. Bedows, T. A. Keiderling, J. Am. Chem. Soc. 2000, 122 , 8623–8630. 240. I. Tinoco, Radiation Res. 1963, 20 , 133. 241. C. W. Deutsche, A. Moscowitz, J. Chem. Phys. 1968, 49 , 3257. 242. C. W. Deutsche, A. Moscowitz, J. Chem. Phys. 1970, 53 , 2630. 243. J. Snir, R. A. Frankel, J. A. Schellman, Biopolymers 1974, 14 , 173. 244. G. Holzwarth, I. Chabay, J. Chem. Phys. 1972, 57 , 1632. 245. M. Gulotta, D. J. Goss, M. Diem, Biopolymers 1989, 28 , 2047–2058. 246. W. Zhong, M. Gulotta, D. J. Goss, Biochemistry 1990, 29 , 7485–7491. 247. P. Bour, T. A. Keiderling, J. Am. Chem. Soc. 1992, 114 , 9100–9105. 248. P. Bour, Computational Study of the Vibrational Optical Activity of Amides and Peptides. Ph.D. thesis, Academy of Science of the Czech Republic Prague, Czech Republic, 1993. 249. P. Bour, T. A. Keiderling, J. Phys. Chem. B , 2005, 5348–5357. 250. J. W. Brauner, C. Dugan, R. Mendelsohn, J. Am. Chem. Soc. 2000, 122 , 677–683. 251. J. Kubelka, T. A. Keiderling, J. Am. Chem. Soc. 2001, 123 , 6142–6150. 252. J. Kubelka, J.-H. Kim, P. Bour, T. A. Keiderling, Vibr. Spectrosc. 2006, 42 , 63.

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

253. R. Schweitzer-Stenner, F. Eker, K. Griebenow, C. Xiaolin, L. A. Naﬁe, J. Am. Chem. Soc. 2004, 126 , 2768–2776. 254. A. W. Smith, A. Tokmakoff J. Chem. Phys. 2007, 126 , 045109/01–045109/11. 255. J. Wang, R. M. Hochstrasser, J. Chem. Phys. 2004, 297 , 195–219. 256. H. Maekawa, M. DePoli, C. Toniolo, N. H. Ge, J. Am. Chem. Soc. 2009, 131 , 2042. 257. P. J. Stephens, J. Phys. Chem. 1985, 89 , 748. 258. P. J. Stephens, F. J. Devlin, C. F. Chabalowski, M. J. Frisch, J. Phys. Chem. 1994, 98 , 11623–11627. 259. P. J. Stephens, C. S. Ashvar, F. J. Devlin, J. R. Cheeseman, M. J. Frisch, Mol. Phys. 1996, 89 , 579–594. 260. F. J. Devlin, P. J. Stephens, J. Am. Chem. Soc. 1994, 116 , 5003. 261. P. J. Stephens, K. J. Jalkanen, F. J. Devlin, C. F. Chabalowski, J. Phys. Chem. 1993, 97 , 6107. 262. F. J. Devlin, P. J. Stephens, J. R. Cheeseman, M. J. Frisch, J. Phys. Chem. 1997, 101 , 9912–9924. 263. F. J. Devlin, P. J. Stephens, J. R. Cheeseman, M. J. Frisch, J. Phys. Chem. 1997, 101 , 6322–6333. 264. P. Bour, T. A. Keiderling, J. Am. Chem. Soc. 1993, 115 , 9602–9607. 265. R. A. G. D. Silva, J. Kubelka, S. M. Decatur, P. Bour, T. A. Keiderling, Proc. Natl. Acad. Sci. USA 2000, 97 , 8318–8323. 266. J. Hilario, J. Kubelka, F. A. Syud, S. H. Gellman, T. A. Keiderling, Biopolymers (Biospectroscopy) 2002, 67 , 233–236. 267. J. Hilario, J. Kubelka, T. A. Keiderling, J. Am. Chem. Soc. 2003, 125 , 7562–7574. 268. J. Kubelka, R. A. G. D. Silva, T. A. Keiderling, J. Am. Chem. Soc. 2002, 124 , 5325–5332. 269. P. Bour, T. A. Keiderling, J. Phys. Chem. B 2005, 232687–23697. 270. R. Wieczorek, J. J. Dannenberg, J. Phys. Chem. B 2008, 112 , 1320–1328. 271. R. Viswanathan, J. J. Dannenberg, J. Phys. Chem. B 2008, 112 , 5199–5208. 272. Y.-f. Chen, R. Viswanathan, J. J. Dannenberg, J. Phys. Chem. B 2007, 111 , 8329–8334. 273. J. R. Cheeseman, P. J. Stephens, unpublished results, 2010, 274. E. Taillandier, J. Liquier, M. Ghomi, J. Mol. Struct. 1989, 214 , 185–211. 275. J. Anastassopoulou, T. Theophanides, Aquation of metal ions and infrared and Raman spectra of metal complexes and their interaction with DNA components, in Insight into Vibrational Spectroscopy of Nucleic Acids and Their Complexes, Transworld Research Network, Kerala, India, 2009, pp. 47–65. 276. M. Banyay, M. Sarkar, A. Graslund, Biophys. Chem. 2003, 104 , 477–488. 277. J. Liquier, E. Taillandier, Infrared spectroscopy of nucleic acids, in Infrared Spectrsocopy of Biomolecules, Wiley-Liss, Chichester, 1996, pp. 131–158. 278. M. Tsuboi, Infrared and Raman Spectroscopy, Academic Press, New York, 1974. 279. S. S. Birke, M. Diem, Biophys. J . 1995, 68 , 1045–1049. 280. S. S. Birke, M. Moses, B. Kagalovsky, D. Jano, M. Gulotta, M. Diem, Biophys. J . 1993, 65 , 1262–1271. 281. V. Andrushchenko, Z. Leonenko, D. Cramb, H. van de Sande, H. Wieser, Biopolymers 2001, 61 , 243–260. 282. V. Andrushchenko, H. van de Sande, H. Wieser, Biopolymers 2003, 69 , 529–545. 283. V. Andrushchenko, J. H. van de Sande, H. Wieser, Biopolymers 2003, 72 , 374–390. 284. D. Tsankov, B. Kalisch, H. V. de Sande, H. Wieser, J. Phys. Chem. B 2003, 107 , 6479–6485. 285. V. Andrushchenko, Y. Blagoi, J. H. van de Sande, H. Wieser, J. Biomol. Struct. Dyn. 2002, 19 , 889–906.

753

754

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

286. V. V. Andrushchenko, J. H. van de Sande, H. Wieser, Vib. Spectrosc., 1999, 19 , 341–345. 287. V. V. Andrushchenko, J. H. van de Sande, H. Wieser, S. V. Kornilova, Y. P. Blagoi, J. Biomol. Struct. Dyn. 1999, 17 , 545–560. 288. L. Yang, T. A. Keiderling, Biopolymers 1993, 33 , 315–327. 289. A. Cupane, A. Polyanichko, H. Wieser, Spectroscopy (Amsterdam) 2010, 24 , 239–244. 290. A. M. Polyanichko, E. V. Chikhirzhina, V. V. Andrushchenko, V. I. Vorob’ev, H. Wieser, Biopolymers, 2006, 83 , 182–192. 291. J. Novy, M. Urbanova, K. Volka, Collect. Czech. Chem. Commun. 2005, 70 , 1799–1810. 292. A. M. Polyanichko, E. V. Chikhirzhina, V. V. Andrushchenko, H. Wieser, V. I. Vorob’ev, Bioﬁzika 2005, 50 , 810–817. 293. A. Polyanichko, H. Wieser, Biopolymers 2005, 78 , 329–339. 294. J. Novy, M. Urbanova, K. Volka, J. Mol. Struct. 2005, 748 , 17–25. 295. A. M. Polyanichko, V. V. Andrushchenko, E. V. Chikhirzhina, V. I. Vorob’ev, H. Wieser, Nucl. Acids Res. 2004, 32 , 989–996. 296. A. T. Krummel, M. T. Zanni, J. Phys. Chem. B 2006, 110 , 24720. 297. T. Xiang, D. J. Goss, M. Diem, Biophys. J . 1993, 65 , 1255–1261. 298. A. F. Bell, L. Hecht, L. D. Barron, J. Am. Chem. Soc. 1998, 120 , 5820–5821. 299. A. M. Smith, R. F. Collins, R. V. Ulijn, E. Blanch, J. Raman Spectrosc. 2009, 40 , 1093–1095. 300. J. C. Sutherland, K. P. Grifﬁn, Biopolymers 1983, 22 , 1445–1448. 301. S. S. Birke, I. Agbaje, M. Diem, Biochemistry 1992, 31 , 450. 302. H. DeVoe, J. Phys. Chem. 1971, 75 , 1509–1515. 303. B. D. Self, D. S. Moore, Biophys. J . 1997, 73 , 339–347. 304. B. D. Self, D. S. Moore, Biophys. J . 1998, 74 , 2249–2258. 305. V. Andrushchenko, H. Wieser, P. Bour, J. Phys. Chem. B 2004, 108 , 3899. 306. R. D. Singh, T. A. Keiderling, Biopolymers 1981, 20 , 237–240. 307. A. C. Sen, T. A. Keiderling, Biopolymers 1984, 23 , 1519–1532. 308. B. B. Lal, L. A. Naﬁe, Biopolymers 1982, 21 , 2161–2183. 309. M. G. Paterlini, T. B. Freedman, L. A. Naﬁe, Biopolymers 1986, 25 , 1751–1765. 310. S. C. Yasui, T. A. Keiderling, J. Am. Chem. Soc. 1986, 108 , 5576–5581. 311. P. Malon, R. Kobrinskaya, T. A. Keiderling, Biopolymers 1988, 27 , 733–746. 312. S. C. Yasui, T. A. Keiderling, Biopolymers 1986, 25 , 5–15. 313. S. C. Yasui, T. A. Keiderling, R. Katachai, Biopolymers 1987, 26 , 1407–1412. 314. B. Baello, P. Pancoska, T. A. Keiderling, Anal. Biochem. 2000, 280 , 46–57. 315. V. P. Gupta, T. A. Keiderling, Biopolymers, 1992, 32 , 239–248. 316. C. Zhao, P. L. Polavarapu, C. Das, P. Balaram, J. Am. Chem. Soc. 2000, 122 , 8228–8231. 317. R. Mahalakshmi, G. Shanmugam, P. L. Polavarapu, P. Balaram, Chembiochem 2005, 6 , 2152–2158. 318. D. M. Byler, J. M. Purcell, SPIE Fourier Transform. Spectrosc. 1989, 1145 , 539–544. 319. A. H. Clark, D. H. P. Saunderson, A. Sugget, Int. J. Pept. Protein Res. 1981, 17 , 353–364. 320. R. K. Dukor, Vibrational Circular Dichroism of Selected Peptides, Polypeptides and Proteins. Ph.D. thesis, University of Illinois at Chicago, Chicago, IL, 1991. 321. V. Setnicka, R. Huang, C. L. Thomas, M. A. Etienne, J. Kubelka, R. P. Hammer, T. A. Keiderling, J. Am. Chem. Soc. 2005, 127 , 4992–4993. 322. R. A. G. D. Silva, S. A. Sherman, T. A. Keiderling, Biopolymers 1999, 50 , 413–423. 323. S. C. Yasui, T. A. Keiderling, F. Formaggio, G. M. Bonora, C. Toniolo, J. Am. Chem. Soc. 1986, 108 , 4988–4993. 324. S. C. Yasui, T. A. Keiderling, G. M. Bonora, C. Toniolo, Biopolymers 1986, 25 , 79–89.

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

325. R. A. G. D. Silva, J. Kubelka, T. A. Keiderling, Biopolymers 2002, 65 , 229–243. 326. A. Lakhani, A. Roy, M. de Poli, M. Nakaema, C. Toniolo, T. A. Keiderling, J. Phys. Chem. B , submitted. 327. C. Toniolo, F. Formaggio, S. Tognon, Q. B. Broxterman, B. Kaptein, R. Huang, V. Setnicka, T. A. Keiderling, I. H. Iain H. McColl, L. Hecht, L. D. Barron, Biopolymers 2004, 75 , 32–45. 328. M. Crisma, W. Bisson, F. Formaggio, Q. B. Broxterman, C. Toniolo, Biopolymers 2002, 64 , 236–245. 329. S. Aravinda, N. Shamala, P. Balaram, Chem. Biodivers 2008, 5 , 1238–1262. 330. T. S. Sudha, E. K. Vijayakumar, P. Balaram, Int. J. Pept. Protein Res., 1983, 22 , 464–468. 331. B. V. Prasad, P. Balaram, CRC Crit. Rev. Biochem. 1984, 16 , 307–348. 332. G. Yoder, T. A. Keiderling, F. Formaggio, M. Crisma, C. Toniolo, Biopolymers 1995, 35 , 103–111. 333. W. Mastle, R. K. Dukor, G. Yoder, T. A. Keiderling, Biopolymers 1995, 36 , 623–631. 334. H. R. Wyssbrod, M. Diem, Biopolymers 1992, 31 , 1237. 335. P. Xie, Q. Zhou, M. Diem, Faraday Discuss. 1995, 99 , 233–244. 336. P. Xie, Q. Zhou, M. Diem, J. Am. Chem. Soc. 1995, 117 , 9502–9508. 337. W. M. Zuk, T. B. Freedman, L. A. Naﬁe, Biopolymers 1989, 28 , 2025–2044. 338. F. Eker, X. Cao, L. Naﬁe, R. Schweitzer-Stenner, J. Am. Chem. Soc., 2002, 124 , 14330–14341. 339. J. Kim, J. Kapitan, A. Lakhani, P. Bour, T. A. Keiderling, Theor. Chem. Acc. 2006, 119 , 81–97. 340. R. K. Dukor, T. A. Keiderling, Biopolymers 1991, 31 , 1747–1761. 341. R. K. Dukor, T. A. Keiderling, V. Gut, Int. J. Pept. Protein Res. 1991, 38 , 198–203. 342. Q. Xu, T. A. Keiderling, Macromol. Symp. 2005, 220 , 17–31. 343. M. L. Tiffany, S. Krimm, Biopolymers 1972, 11 , 2309–2316. 344. J. Kubelka, R. Huang, T. A. Keiderling, J. Phys. Chem. B 2005, 8231–8243. 345. S. C. Yasui, T. A. Keiderling, Vibrational circular dichroism of polypeptides with aromatic side chains, in Peptides: Chemistry and Biology, Proceedings, 10th American Peptides Symposium, G. R. Marshall, ed., ESCOM, Leiden, 1988, pp. 90–92. 346. T. B. Freedman, A. C. Chernovitz, W. M. Zuk, G. Paterlini, L. A. Naﬁe, J. Am. Chem. Soc. 1988, 110 , 6970–6974. 347. J. P. Mazaleyrat, K. Wright, A. Gaucher, M. Wakselman, S. Oancea, F. Formaggio, C. Toniolo, V. Setnicka, J. Kapitan, T. A. Keiderling, Tetrahed. Asym. 2003, 14 , 1879–1893. 348. S. C. Yasui, T. A. Keiderling, M. Sisido, Macromolecules 1987, 20 , 403–2406. 349. J.-P. Mazaleyrat, K. Wright, A. Gaucher, M. Toulemonde, M. Wakselman, S. Oancea, C. Peggion, F. Formaggio, V. Setnicka, T. A. Keiderling, C. Toniolo, J. Am. Chem. Soc. 2004, 126 , 12874–12879. 350. J. Wang, D. McElheny, Y. Fu, G. Li, J.-H. Kim, Z. Zhou, L. Wu, T. A. Keiderling, R. P. Hammer, Pept. Sci. (Biopolymers) 2009, 92 , 452–464. 351. M. G. Paterlini, T. B. Freedman, L. A. Naﬁe, Y. Tor, A. Shanzer, Biopolymers 1992, 32 , 765–782. 352. A. C. Chernovitz, T. B. Freedman, L. A. Naﬁe, Biopolymers 1987, 26 , 1879–1900. 353. M. Diem, J. Am. Chem. Soc. 1988, 110 , 6967–6970. 354. M. Diem, O. Lee, G. M. Roberts, J. Phys. Chem. 1992, 96 , 548–554. 355. G. M. Roberts, O. Lee, J. Calienni, M. Diem, J. Am. Chem. Soc. 1988, 110 , 1749–1752. 356. O. Lee, G. M. Roberts, M. Diem, Biopolymers 1989, 28 , 1759–1770. 357. M. Miyazawa, Y. Kyogoku, H. Sugeta, Spectrochim. Acta. 1994, 50 , 1505–1511.

755

756

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

358. R. K. Dukor, T. A. Keiderling, Biospectroscopy 1996, 2 , 83–100. 359. Z. Shi, N. R. Kallenbach, R. W. Woody, Adv. Protein Chem. 2002, 62 , 163–240. 360. Z. S. Shi, C. A. Olson, G. D. Rose, R. L. Baldwin, N. R. Kallenbach, Proc. Natl. Acad. Sci. USA 2002, 99 , 9190–9195. 361. E. Vass, U. Strijowski, K. Wollschlager, I. M. Mandity, G. Szilvagyi, M. Jewginski, K. Gaus, S. Royo, Z. Majer, N. Sewald, M. Hollosi, J. Pept. Sci . 16 , 613–620. 362. P. Xie, M. Diem, J. Am. Chem. Soc. 1995, 117 , 429–437. 363. P. Malon, P. Pirkova, W. Mastle, T. A. Keiderling, Vibrational circular dichroism of cyclo (tril-2-azetidine carbonyl), in Peptides 1992: Proceedings of 22th European Peptide Symposium, Escom, Leyden, 1993, p. 499. 364. F. Wang, C. Zhao, P. L. Polavarapu, Biopolymers, 2004, 75 , 85–93. 365. L. Wu, Thermodynamic and spectral studies of structure and stability for related beta-hairpin peptides. Ph.D. thesis, University of Illinois at Chicago, Chicago, IL, 2010. 366. K. Wright, R. Anddad, J.-F. Lohier, V. Steinmetz, M. Wakselman, J.-P. Mazaleyrat, F. Formaggio, C. Peggion, M. De Zotti, T. A. Keiderling, R. Huang, C. Claudio Toniolo, Eur. J. Org. Chem. 2008, 1224–1241. 367. A. Moretto, M. Crisma, F. Formaggio, B. Kaptein, Q. B. Broxterman, T. A. Keiderling, C. Toniolo, Biopolymers (Peptide Sci.) 2007, 88 , 233–238. 368. G. Yoder, T. A. Keiderling, F. Formaggio, M. Crisma, Tetrahedron Asymm. 1995, 6 , 687–690. 369. M. J. Citra, M. G. Paterlini, T. B. Freedman, A. Fissi, O. Pieroni, Proc. SPIE 1993, 2089 , 478–479. 370. M. J. Citra, M. G. Paterlini, T. B. Freedman, A. Fissi, O. Pieroni, Proc. SPIE-Int. Soc. Opt. Eng. 1993, 2089 , 478–479. 371. S. M. Miick, G. Martinez, W. R. Fiori, A. P. Todd, G. L. Millhauser, Nature 1992, 359 , 653–655. 372. G. L. Millhauser, Biochemistry 1995, 34 , 3873–3877. 373. G. L. Millhauser, C. J. Stenland, P. Hanson, K. A. Bolin, F. J. M. Vandeven, J. Mol. Biol . 1997, 267 , 963–974. 374. E. R. Malinowski, Factor Analysis in Chemistry, Wiley, New York, 1991. 375. S. H. Gellman, Curr. Opin. Struct. Biol . 1998, 2 , 717–725. 376. S. H. Gellman, R. T. Raines, Curr. Opin. Chem. Biol , 2002, 727–728. 377. J. W. Kelly, Curr. Opin. Struct. Biol . 1998, 8 , 101–106. 378. J. M. Langenhan, I. A. Guzei, S. H. Gellman, Angew. Chem. Int. Ed. Engl . 2003, 42 , 2402–2405. 379. M. Ramirez-Alvarado, T. Kortemme, F. J. Blanco, L. Serrano, Bioorg. Med. Chem. 1999, 7 , 93–103. 380. H. E. Stanger, F. A. Syud, J. F. Espinosa, I. Giriat, T. Muir, S. H. Gellman, Proc. Natl. Acad. Sci. USA 2001, 98 , 12015–12020. 381. M. Lopez de la Paz, E. Lacroix, Ramirez-Alvarado, S. M., L., J. Mol. Biol . 2001, 229–246. 382. M. S. Searle, B. Ciani, Curr. Opin. Struct. Biol . 2004, 14 , 458–464. 383. R. M. Hughes, M. L. Waters, Curr. Opin. Struct. Biol . 2006, 16 , 514–524. 384. C. Das, V. Nayak, S. Raghothama, P. Balaram, J. Pept. Res. 2000, 56 , 307–317. 385. F. A. Syud, H. E. Stanger, H. S. Mortell, J. F. Espinosa, J. D. Fisk, C. G. Fry, S. H. Gellman, J. Mol. Biol . 2003, 326 , 553–568. 386. E. K. Koepf, H. M. Petrassi, M. Sudol, J. W. Kelly, Protein Sci . 1999, 8 , 841–853. 387. H. L. Schenck, S. H. Gellman, J. Am. Chem. Soc. 1998, 120 , 4869–4870. 388. T. Kortemme, M. Ramirez-Alvarado, L. Serrano, Science 1998, 281 , 253–256.

C O N F O R M AT I O N A L S T U D I E S O F B I O P O LY M E R S , P E P T I D E S , P R O T E I N S , A N D N U C L E I C A C I D S

389. A. G. Cochran, N. J. Skelton, M. A. Starovasnik, Proc. Natl. Acad. Sci. USA 2001, 98 , 5578–5583. 390. C. D. Tatko, M. L. Waters, J. Am. Chem. Soc. 2002, 124 , 9372–9373. 391. L. Wu, D. McElheny, R. Huang, T. A. Keiderling, Biochemistry 2009, 48 , 10362–10371. 392. R. W. Ruddon, S. A. Sherman, E. Bedows, Protein Sci ., 1996, 5 , 1443–1452. 393. A. G. Petrovic, P. L. Polavarapu, R. Mahalakshmi, P. Balaram, Chirality 2009, 21 , E76–E85. 394. S. V. Kuznetsov, J. Hilario, T. A. Keiderling, A. Ansari, Biochemistry 2003, 42 , 4321–4332. 395. H. Chi, T. A. Keiderling, (unpublished), 2010, 396. P. J. Stephens, J. Phys. Chem. 1987, 91 , 1712. 397. R. D. Amos, K. J. Jalkanen, P. J. Stephens, J. Phys. Chem. 1988, 92 , 5571. 398. J. Kubelka, T. A. Keiderling, J. Phys. Chem. A 2001, 105 , 10922–10928. 399. P. Bour, T. A. Keiderling, J. Chem. Phys. 2003, 119 , 11253–11262. 400. J. Kubelka, IR and VCD spectroscopy of model peptides. Theory and experiment. Ph.D. thesis, University of Illinois at Chicago, Chicago, IL, 2002. 401. K. E. Amunson, L. Ackels, J. Kubelka, J. Am. Chem. Soc. 2008, 130 , 8146–8147. 402. R. W. Williams, Meth. Enzymol . 1986, 130 , 311–331. 403. P. I. Haris, D. Chapman, Biopolymers 1995, 37 , 251–263. 404. G. Vedantham, H. G. Sparks, S. U. Sane, S. Tzannis, T. M. Przybycien, Anal. Biochem. 2000, 285 , 33–49. 405. W. Surewicz, H. H. Mantsch, Infrared absorption methods for examining protein structure, in Spectroscopic Methods for Determining Protein Structure in Solution, VCH, Deerﬁeld Beach, FL, 1996, pp. 135–162. 406. E. Goormaghtigh, V. Cabiaux, J.-M. Ruysschaert, Subcell. Biochem. 1994, 23 , 405–450. 407. P. Pancoska, H. Fabian, G. Yoder, V. Baumruk, T. A. Keiderling, Biochemistry 1996, 35 , 13094–13106. 408. P. Pancoska, V. Janota, T. A. Keiderling, Appl. Spectrosc. 1996, 50 , 658–668. 409. C. T. Chang, C. S. C. Wu, J. T. Yang, Anal. Biochem. 1978, 91 , 13–31. 410. K. Kaiden, T. Matsui, S. Tanaka, Appl. Spectrosc. 1987, 41 , 180–184. 411. F. Fu, D. B. DeOliveira, W. R. Trumble, H. K. Sarkar, B. R. Singh, Appl. Spectrosc. 1994, 48 , 1432–1440. 412. K. Kuwajima, FASEB J . 1996, 10 , 102–109. 413. M. Urbanova, R. K. Dukor, P. Pancoska, V. P. Gupta, T. A. Keiderling, Biochemistry 1991, 30 , 10479–10485. 414. V. E. Bychkova, A. E. Dujsekina, S. I. Klenin, Biochemistry 1996, 35 , 6058–6063. 415. Y. O. Kamatari, T. Konno, M. Kataoka, K. Akasaka, J. Mol. Biol . 1996, 259 , 512–523. 416. S. R. Yeh, S. W. Han, D. L. Rousseau, Acc. Chem. Res. 1998, 31 , 727–736. 417. Q. Xu, T. A. Keiderling, Biopolymers (Biospectroscopy) 2004, 73 , 716–727, 418. Q. Xu, T. A. Keiderling, Protein Sci . 2004, 2949–2959. 419. Q. Xu, T. A. Keiderling, Proteins: Struct. Funct. Bioinform. 2006, 63 , 571–580. 420. Q. Xu, T. A. Keiderling, Biochemistry 2005, 7976–7987. 421. A. Loksztejn, W. Dzwolak, J. Mol. Biol . 2010, 395 , 643–655. 422. A. Fulara, S. Wojcik, A. Loksztejn, W. Dzwolak, J. Phys. Chem. B 2008, 112 , 8744–8747. 423. A. Loksztejn, W. Dzwolak, J. Mol. Biol . 2008, 379 , 9–16. 424. W. Dzwolak, Biochim. Biophys. Acta, Proteins Proteomics 2006, 1764 , 470–480. 425. B. A. Wallace, J. G. Lees, A. J. W. Orry, A. Lobley, R. W. Janes, Protein Sci . 2003, 12 , 875–884. 426. N. A. Swords, B. A. Wallace, Biochem. J . 1993, 289 , 215–219.

757

758

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

427. 428. 429. 430. 431. 432. 433. 434.

435. 436. 437. 438. 439. 440. 441. 442. 443.

444.

445. 446. 447. 448. 449. 450. 451.

B. A. Wallace, D. Mao, Anal. Biochem. 1984, 142 , 317–328. N. Ge, X. Zhang, T. A. Keiderling, Biochemistry 2010, 49 , 8831–8838. X. Zhang, N. Ge, T. A. Keiderling, Biochemistry 2007, 46 , 5252–5260. X. Zhang, T. A. Keiderling, Biochemistry 2006, 45 , 8444–8452. M. Urbanova, T. A. Keiderling, P. Pancoska, Bioelectrochem. Bioenerg. 1996, 41 , 77–80. A. A. Pandyra, A. P. Yamniuk, V. V. Andrushchenko, H. Wieser, H. J. Vogel, Biopolymers, 2005, 79 , 231–237. S. D. Stelea, T. A. Keiderling, Biophys. J . 2003, 83 , 2259–2269. J. Kubelka, P. Pancoska, T. A. Keiderling, Protein spectra-structure relationship by novel modiﬁcation of 2D correlation analysis. VCD applications, in Spectroscopy of Biological Molecules: New Directions, J. Greve, G. J. Puppels, and C. Otto., eds., Kluwer Academic Publishers, Dordrecht, 1999, pp. 67–68. R. J. Rummel, Applied Factor Analysis, Northwestern University Press, Evanston, IL, 1970. P. Pancoska, I. Fric, K. Blaha, Collect. Czech. Chem. Commun. 1979, 44 , 1296–1312. P. Manavalan, W. C. Johnson, Jr., Anal. Biochem. 1987, 167 , 76–85. I. H. M. van Stokkum, H. J. W. Spoelder, M. Bloemendal, R. van Grondelle, F. C. A. Groen, Anal. Biochem. 1990, 191 , 110–118. P. Pancoska, M. Blazek, T. A. Keiderling, Biochemistry 1992, 31 , 10250–10257. P. Pancoska, V. Janota, T. A. Keiderling, Anal. Biochem. 1999, 267 , 72–83. P. I. Haris, Biochem. Soc. Trans. 2010, 38 , 940–946. V. Cabiaux, K. Oberg, P. Pancoska, T. Walz, P. Agre, A. Engle, Biophys. J . 1997, 73 , 406–417. B. Baello, Protein secondary structure and tertiary structure studies using vibrational spectroscopy and hydrogen exchange. Ph.D. thesis, University of Illinois at Chicago Chicago 2000. A. J. P. Alix, Molecular modeling of globular proteins, in Biomolecular Structure and Dynamics, G. Vergoten and T. Theophanides, eds., Kluwer Acadamic Press, Amsterdam, 1997, pp. 121–150. P. Pancoska, J. Kubelka, T. A. Keiderling, Appl. Spectrosc. 1998, 53 , 655–665. J. Kubelka, P. Pancoska, T. A. Keiderling, Appl. Spectrosc. 1998, 53 , 666–671. J. Wang, J. Chen, R. M. Hochstrasser, J. Phys. Chem. B 2006, 110 , 7545–7555. S. H. Brewer, B. Song, D. P. Raleigh, R. B. Dyer, Biochemistry 2007, 46 , 3279–3285. K. Hauser, C. Krejtschi, R. Huang, L. Wu, T. A. Keiderling, J. Am. Chem. Soc. 2008, 130 , 2984–2992. R. Callender, R. B. Dyer, Curr. Opin. Struct. Biol . 2002, 12 , 628–633. C. N. Tam, P. Bour, T. A. Keiderling, J Am. Chem. Soc. 1997, 119 7061–7064.

23 STRUCTURE AND BEHAVIOR OF BIOMOLECULES FROM RAMAN OPTICAL ACTIVITY Laurence D. Barron and Lutz Hecht

23.1. INTRODUCTION The techniques of X-ray crystallography and nuclear magnetic resonance (NMR) spectroscopy dominate structural biology due to their ability to reveal the details of biomolecular structure at atomic resolution. However, there are limitations to their applicability. For example, many proteins are difﬁcult to crystallize while many others have structures too large to be solved by current NMR methods. While not providing information at atomic resolution, chiroptical spectroscopic techniques are nonetheless valuable in biomolecular science since they can be applied routinely to a wide range of biological systems and provide information on their structure and behavior [1–3]. A powerful chiroptical spectroscopy for the study of biomolecules is Raman optical activity (ROA), which measures a small difference in the intensity of vibrational Raman scattering from chiral molecules in right- and left-circularly polarized incident light or, equivalently, the intensity of a small circularly polarized component in the scattered light using incident light of ﬁxed polarization [1–8] (Figure 23.1). The ﬁrst and second experiments are called incident circular polarization (ICP) and scattered circular polarization (SCP) ROA, respectively. ROA, ﬁrst observed by Barron et al. in 1972 [9] using the ICP strategy and conﬁrmed by Hug et al. in 1975 [10], has been developed to the point where it is now an incisive probe of the structure and behavior of biomolecules in aqueous solution [11–17]. The power of ROA in this area derives from the fact that, like the complementary technique of vibrational circular dichroism (VCD) (Chapter 5), Volume 1, it is a form of vibrational optical activity and so is sensitive to chirality associated with Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

759

760

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

R

L

F

ω

H

CI

Br

IR − IL ≠ 0 (a) ICP ROA ω − ων F

ω

H

CI Br

R L IR − IL ≠ 0

(b) SCP ROA

ω − ων

Figure 23.1. Two equivalent ROA experiments in transparent Stokes vibrational Raman scattering at angular frequency, ω − ωv , in incident light of angular frequency ω. (a) ICP ROA measures IR − IL , where IR and IL are the scattered intensities in right- and left-circularly polarized incident light, respectively. (b) SCP ROA measures IR − IL , where IR and IL are the intensities of the rightand left-circularly polarized components, respectively, of the scattered light using incident light of ﬁxed polarization (shown here as unpolarized). A positive value of IR − IL corresponds to a small degree of right-circular polarization in the scattered light.

all the 3N − 6 fundamental molecular vibrational transitions, where N is the number of atoms. Raman spectroscopy itself provides molecular vibrational spectra by means of inelastic scattering of visible light. During the Stokes Raman scattering event, the interaction of the molecule with the incident visible photon of energy ω, where ω is its angular frequency, can leave the molecule in an excited vibrational state of energy ωv , with a concomitant energy loss, and hence a shift to lower angular frequency, ω − ωv , of the scattered photon. Therefore, by analyzing the scattered light with a visible spectrometer, a complete vibrational spectrum may be obtained. Conventional Raman spectroscopy has several favorable characteristics that have led to many applications in biochemistry [18, 19]. In particular, the complete vibrational spectrum from ∼50 to 4000 cm−1 is accessible on one simple instrument, and both H2 O and D2 O are excellent solvents for Raman studies. ROA is able to build on these advantages by adding an extra sensitivity to three-dimensional structure through its dependence on absolute handedness, which opens a new window on the structure and behavior of biomolecules. Its simple routine application to aqueous solutions, with no restrictions on the size of the biomolecules, makes ROA ideal for studying many timely problems. Only brief outlines of theory and experiment are given below since detailed accounts are provided in Chapters 6 and 24 of Volume 1.

761

S T R U C T U R E A N D B E H AV I O R O F B I O M O L E C U L E S F R O M R A M A N O P T I C A L A C T I V I T Y

23.2. OUTLINE OF ROA THEORY 23.2.1. ROA Observables The fundamental scattering mechanism responsible for ROA was discovered by Atkins and Barron [20], who showed that interference between the light waves scattered via the molecular polarizability and optical activity tensors of the molecule yields a dependence of the scattered intensity on the degree of circular polarization of the incident light and to a circular component in the scattered light. Barron and Buckingham [21] subsequently developed a more deﬁnitive version of the theory and introduced the dimensionless circular intensity difference (CID) = (I R − I L )/(I R + I L )

(23.1)

as an appropriate experimental quantity, where I R and I L are the scattered intensities in right- and left-circularly polarized incident light, respectively. ROA measurements may be performed using several different experimental conﬁgurations [2]. In particular, the scattering angle can be varied, with the backward direction being the most important for studies of biomolecules in aqueous solution (vide infra). In terms of the electric dipole–electric dipole molecular polarizability tensor ααβ and the electric dipole–magnetic dipole and electric dipole–electric quadrupole optical and Aαβγ , ICP CIDs associated with forward (0◦ ) and backward activity tensors Gαβ ◦ (180 ) scattering for an isotropic collection of chiral molecules with dimensions much smaller than the wavelength of the incident light are [2] ◦

(0 ) = ◦

4[45αG + β(G )2 − β(A)2 ] , c[45α 2 + 7β(α)2 ]

(180 ) =

24[β(G )2 + 13 β(A)2 ] . c[45α 2 + 7β(α)2 ]

(23.2a)

(23.2b)

The various tensor component products have been averaged over all orientations of the scattering molecule to generate collections of products that are invariant to axis rotations. Speciﬁcally, α = 13 ααα = 13 (αxx + αyy + αzz ),

(23.3a)

= 13 (Gxx + Gyy + Gzz ) G = 13 Gαα

(23.3b)

are the isotropic invariants, and β(α)2 = 12 (3ααβ ααβ − ααα αββ ), 2

β(G ) =

1 2 (3ααβ Gαβ

− ααα Gββ ),

β(A)2 = 12 ωααβ εαγ δ Aγ δβ

(23.3c) (23.3d) (23.3e)

are the anisotropic invariants in which the tensor components may be referred to molecule-ﬁxed axes. All these invariants are independent of the choice of origin, and each is accessible to measurement. We are using a Cartesian tensor notation in which a repeated Greek sufﬁx denotes summation over the three components, and εαβγ is the third-rank antisymmetric unit tensor.

762

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

It was mentioned above that, as well as the circular intensity difference, ROA is also manifest as a small circularly polarized component in the scattered beam [2, 4, 5, 20, 22, 23]. Within the far-from-resonance approximation, measurement of this circular component (SCP ROA) as (IR − IL )/(IR + IL ), where IR and IL denote the intensities of the right- and left-circularly polarized components, respectively, of the scattered light, provides equivalent information to the CID measurement (ICP ROA). These results apply speciﬁcally to Rayleigh (elastic) scattering. For Raman (inelastic) scattering the same basic CID expressions apply but with the molecular property tensors replaced by corresponding vibrational Raman transition tensors between the initial and ﬁnal vibrational states nv and jv . Thus ααβ , etc., are replaced by jv |ααβ (Q)|nv , etc., where ααβ (Q), etc., are effective polarizability and optical activity operators that depend parametrically on the normal vibrational coordinates Q so that, within the Placzek polarizability theory of the Raman effect [24], the ROA intensity depends on products /∂Q)0 and (∂ααβ /∂Q)0 εαγ δ (∂Aγ δβ /∂Q)0 , where the subscript such as (∂ααβ /∂Q)0 (∂Gαβ 0 indicates the value for the equilibrium structure. At transparent frequencies, Stokes and anti-Stokes ROA CIDs are identical [25]. For the case of a molecule composed entirely of idealized axially symmetric achiral bonds, for which β(G )2 = β(A)2 and αG = 0 [2, 22, 26], ROA is generated exclusively by anisotropic scattering and the forward and backward CID expressions (23.2a) and (23.2b) reduce to [2, 27] ◦

(0 ) = 0, ◦

(180 ) =

32β(G )2 . c[45α 2 + 7β(α)2 ]

(23.4a) (23.4b)

Since conventional Raman intensities are the same in forward and backward scattering, backscattering boosts the ROA signal relative to the background Raman intensity and is therefore the best experimental strategy for most ROA studies of biomolecules in aqueous solution due to the weak signals and high backgrounds [27, 28].

23.2.2. Calculation of ROA Spectra Calculation of ROA spectra is an important aspect of the technique since simulations of observed spectra can provide the complete solution structure (conformation, absolute conﬁguration, conformational populations). Calculations of the ROA observables, which are usually based on the Placzek approximation, can proceed in several ways. Models of ROA exist such as the bond-polarizability (valence-optical) model in which the molecule is decomposed into bonds or groups supporting local internal vibrational coordinates [2, 29]. However, due to the approximations inherent in these models, such calculations do not reproduce experimental data well. Models can, nonetheless, provide physical insight into the generation of ROA [2], which is often not transparent in the computationally superior ab initio approach, now the method of choice. Since the property tensors Gαβ and Aαβγ responsible for ROA are time even, there is no fundamental problem in ROA theory analogous to that arising in VCD due to the time-odd nature of the magnetic dipole , and moment operator [30]. An ab initio method, based on (a) calculations of ααβ , Gαβ Aαβγ in a static approximation due to Amos [31] and (b) how these property tensors vary with the normal vibrational coordinates, was developed by Polavarapu in the late 1980s [32, 33]. Although the ﬁrst tranche of such calculations of ROA spectra did not reach the high levels of accuracy now attainable, they nonetheless proved valuable. For

S T R U C T U R E A N D B E H AV I O R O F B I O M O L E C U L E S F R O M R A M A N O P T I C A L A C T I V I T Y

example, the absolute conﬁguration of the archetypal chiral molecule CHFClBr, which had resisted assignment for over a century, was reliably determined from a comparison of the experimental and ab initio theoretical ROA spectra [34]. Subsequent approaches produced signiﬁcant improvements in the quality of ab initio ROA calculations. By including basis sets containing moderately diffuse p-type orbitals on hydrogen atoms, Zuber and Hug [35] demonstrated that ab initio ROA calculations of a similar high level of quality to those of VCD may be achieved. In many systems, ranging from small chiral organic molecules to proteins, it is found that vibrations with large contributions from C–H and N–H deformations often generate large ROA signals (vide infra). Interest in ROA calculations among theoreticians is growing, with a plethora of publications providing further reﬁnements and applications to ever-larger systems. Currently, ROA calculations may be performed using the DALTON [36] and Gaussian [37] software packages: The latest 09 version of Gaussian implements an analytic timedependent protocol for the calculation of the ROA property-tensor derivatives, resulting in an order-of-magnitude increase in speed of the calculations. Recent reviews of ROA computations may be found in Chapter 24 of Volume 1 and in references 38 and 39. As an example of what may currently be achieved, Figure 23.2 presents experimental and simulated Raman and ROA spectra of 1-phenylethanol. This was taken from a recent study that revisited the samples used for the ﬁrst observations of ROA, namely both enantiomers of 1-phenylethanol and 1-phenylethylamine, using state-of-theart instrumentation and calculations [40]. The experimental backscattered SCP Raman and ROA spectra of both enantiomers of 1-phenylethanol are displayed in Figure 23.2a, with the corresponding simulated Raman and ROA spectra of the (+)-(R)-enantiomer, which closely reproduce the experimental spectra, displayed in Figure 23.2b. It is often necessary to allow some degree of conformational freedom in order to simulate the observed Raman and ROA bandshapes, something especially important for many biomolecules [41, 42]. In the present case, a Boltzmann average over several hundred rotameric conformations of the phenyl and –OH groups was taken. Clearly, assignments of absolute conﬁgurations are completely secure from ROA results of this quality.

23.3. EXPERIMENTAL Since most ROA intensity is maximized in backscattering, a backscattering geometry has proved essential for the routine measurement of ROA spectra of biomolecules in aqueous solution. Backscattering ROA spectra may be acquired using both the ICP and SCP measurement strategies, although the designs of the corresponding instruments are completely different. Both ICP and SPC ROA spectra are presented in this chapter. A backscattering ICP measurement strategy was used in our Glasgow laboratory for some years and provided a large number of biomolecular ROA spectra. The Glasgow backscattering ICP ROA instrument is described in detail elsewhere and compared with other related instruments [43]. Essentially, a visible laser beam is weakly focused into the sample solution contained in a small rectangular fused quartz cell. The cone of backscattered light is reﬂected off a 45◦ mirror, which has a small central hole drilled to allow passage of the incident laser beam, through an edge ﬁlter to remove the Rayleigh line and into the collection optics of a single grating spectrograph containing a volume holographic transmission grating with a back-thinned charge-coupled device (CCD) detector, allowing the full spectral range to be measured in a single acquisition. To measure the small ROA signals, the spectral acquisition is synchronized with an electro-optic modulator used to switch the state of polarization of the incident laser beam between right and

763

764

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a)

(b)

Wavenumber (cm−1)

Figure 23.2. (a) Experimental backscattered SCP Raman (IR + IL ) and ROA (IR − IL ) spectra of the (+)-(R)- and (−)-(S)-enantiomers of 1-phenylethanol displayed as solid and dotted lines, respectively, and recorded on the ChiralRAMAN instrument. (b) Boltzmann average of simulated spectra for the (+)-(R)-enantiomer. (Adapted from reference 40.)

left circular at a suitable rate. Spectra are displayed in analog-to-digital converter units as a function of the Stokes Raman wavenumber shift with respect to the exciting laser line. Typical laser power at the sample is ∼700 mW and sample concentrations of proteins, polypeptides, and nucleic acids are ∼30–100 mg/mL while those of intact viruses are ∼5–30 mg/mL. Under these conditions ROA spectra over the range ∼600–1700 cm−1 may be obtained in ∼5–24 h for proteins and nucleic acids and ∼1–4 days for intact viruses. Although ROA spectra may be recorded down to ∼200 cm−1 on favorable samples such as carbohydrates, spectra below ∼600 cm−1 can be unreliable from highly scattering samples like proteins due to offsets associated with the intense Rayleigh wing. Measurements may be performed over the temperature range of ∼0–60◦ C by directing dry air downwards over the sample cell from a device used to cool protein crystals in X-ray diffraction experiments, in order to study dynamic behavior. Although ICP ROA instruments have established the value of ROA in biomolecular science, a completely new design of ROA instrument with signiﬁcant advantages accruing from the use of the SCP strategy has been developed by Hug [5, 44] (described in Chapter 6), Volume 1. In particular, “ﬂicker noise” arising from dust particles traveling through the laser beam, density ﬂuctuations in the sample, laser power ﬂuctuations, and so on, are eliminated since the intensity difference measurements required to extract the circularly polarized components of the scattered beam are taken between two orthogonal components of the scattered light measured simultaneously during the same acquisition period. The ﬂicker noise consequently cancels out, resulting in greatly superior signalto-noise characteristics. A commercial instrument based on this new design that also incorporates an artefact suppression protocol, which greatly facilitates the routine acquisition of reliable ROA spectra [45], is available (the ChiralRAMAN from BioTools, Inc.). Typical protein ROA spectra measured on this instrument are displayed in subsequent sections. The fact that they are of even higher quality than those acquired previously on our ICP Glasgow instruments, having been recorded in ∼2–5 h compared with ∼10–20 h using a sample volume of ∼30 μL (half that needed previously) with ∼40% less laser

S T R U C T U R E A N D B E H AV I O R O F B I O M O L E C U L E S F R O M R A M A N O P T I C A L A C T I V I T Y

power, demonstrates just how fast and reliable ROA measurements have now become. The ChiralRAMAN instrument extends biomolecular ROA data acquisition routinely to the low-wavenumber region ∼200–600 cm−1 , even for proteins, thereby opening a new spectral window on structure.

23.4. ROA OF BIOMOLECULES 23.4.1. General ROA spectra may be measured routinely on the central molecules of life, namely proteins, carbohydrates, nucleic acids, and viruses, all in aqueous solution to reﬂect their natural biological environment. References 11–13 and 14–17 review applications based on ROA data from the Glasgow home-built backscattering ICP and commercial SCP instruments, respectively. The normal modes of vibration of biomolecules can be highly complex, with contributions from vibrational coordinates within both the backbone and side chains. ROA spectroscopy is able to provide more informative, less complex spectra than conventional infrared or Raman since the largest signals are often associated with vibrational coordinates that sample the most rigid and chiral parts of the structure. These are usually located within the backbone and often give rise to ROA band patterns characteristic of the backbone conformation. Polypeptides in the standard conformations deﬁned by characteristic Ramachandran φ, ψ angles found in secondary, loop, and turn structure within proteins are particularly favorable in this respect since signals from the peptide backbone usually dominate the ROA spectrum. In contrast, the parent conventional Raman spectrum of a protein is dominated by bands arising from the amino acid side chains which often obscure the peptide backbone bands. Carbohydrate ROA spectra are similarly dominated by signals from skeletal vibrations, in this case centered on the constituent sugar rings and the connecting glycosidic links. Although the parent Raman spectra of nucleic acids are dominated by bands from the intrinsic base vibrations, their ROA spectra tend to be dominated by bands characteristic of the stereochemical dispositions of the bases with respect to each other and to the sugar rings, together with signals from the sugar-phosphate backbone. The timescale of the Raman scattering event (∼3.3 × 10−14 s for a vibration with Stokes wavenumber shift ∼1000 cm−1 excited in the visible) is much shorter than that of the fastest conformational ﬂuctuations. The ROA spectrum is therefore a superposition of “snapshot” spectra from all the distinct conformations present in the sample at equilibrium. Since ROA observables depend on absolute chirality, there is a cancellation of contributions from quasi-enantiomeric structures, which can arise as mobile structures explore the range of accessible conformations. These factors result in ROA exhibiting an enhanced sensitivity to the dynamic aspects of biomolecular structure, making it a new source of information on order–disorder and other types of structural transitions. In contrast, observables that are “blind” to chirality, such as conventional Raman band intensities, are generally additive and therefore less sensitive to conformational mobility. Ab initio ROA computations are starting to make an impact on studies of the aqueous solution conformations of biomolecules. For example, an analysis of the conformational space of zwitterionic l-alanine revealed that shapes of Raman and ROA bands are to a large extent determined by rotation of NH3 + , COO− , and CH3 groups and hence that it is essential to take into account dynamic factors for successful simulations [41]. And by

765

766

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

transferring molecular property tensors from smaller fragments to the whole structure, together with optimization in normal coordinates, the inﬂuence of side chains on Raman and ROA spectra of poly(l-proline) was simulated by averaging different proline ring conformations [42].

23.4.2. ROA Signatures of Polypeptide and Protein ROA Spectra Vibrations of the backbone in polypeptides and proteins are usually associated with three main regions of the Raman spectrum [11, 19, 46]. These are the backbone skeletal stretch region ∼870–1150 cm−1 originating mainly in Cα –C, Cα –Cβ and Cα –N stretch coordinates; the extended amide III region ∼1230–1350 cm−1 assigned mostly to the in-phase combination of the in-plane N–H deformation with the Cα –N stretch together with Cα –H deformations; and the amide I region ∼1630–1700 cm−1 which arises mostly from the C=O stretch. The extended amide III region is particularly important for ROA studies because the coupling between N–H and Cα –H deformations is sensitive to geometry and generates a rich and informative ROA band structure. Bands in the amide II region ∼1510–1570 cm−1 , which originate in the out-of-phase combination of largely the in-plane N–H deformation with a small amount of the Cα –N stretch, are not usually observed in the conventional (nonresonance) Raman spectra of peptides and proteins but sometimes make weak contributions to the ROA spectra which are enhanced in D2 O. Side-chain vibrations also generate many characteristic Raman bands: Although these are often less prominent in ROA spectra due to conformational freedom which can suppress the ROA intensities, a few side-chain vibrations, especially those associated with tryptophan and phenylalanine, do generate useful ROA signals [11–13]. Poly(l-lysine) adopts well-deﬁned conformations under certain conditions of temperature and pH and has long been used as a model for the spectroscopic identiﬁcation of secondary structure in proteins [47]. Poly(l-lysine) at alkaline pH has neutral side chains and so is able to support α-helical conformations stabilized both by internal hydrogen bonds and by hydrogen bonds to the solvent; whereas poly(l-lysine) at neutral and acidic pH has charged side chains that repel each other, thereby encouraging a disordered structure [48]. The backscattered ICP Raman and ROA spectra of these samples are shown in Figures 23.3a and 23.3b, respectively [49, 50]. Figure 23.3c shows the spectra of the corresponding β-sheet conformation obtained by heating the α-helical sample [50]. There are clearly many differences between the ROA spectra of the α-helical, disordered, and β-sheet conformations of poly(l-lysine); this enables ROA to easily distinguish between them. As illustrated in Figure 23.4, a clear distinction is also apparent between the ROA spectra of proteins with folds dominated by α-helix, like human serum albumin (HSA), β-sheet, like human immunoglobulin, and disordered structure, like bovine β-casein. The corresponding MOLSCRIPT diagrams [51] of the ﬁrst two are displayed for convenience. Although similar overall to the corresponding poly(l-lysine) ROA spectra, there are some differences, especially for β-sheet. Some of the differences may be attributed to sidechain contributions, which can be more prominent in homopolypeptides than in proteins where contributions from the many different side chains tend to cancel in certain spectral regions. The low-wavenumber region ∼200–600 cm−1 , now accessible in polypeptides and proteins thanks to the ChiralRAMAN instrument, remains relatively unexplored. This covers the S–S stretch region ∼500–550 cm−1 from disulphide bridges in proteins, together with part of the region where modes such as helix breathing, torsions, and skeletal

S T R U C T U R E A N D B E H AV I O R O F B I O M O L E C U L E S F R O M R A M A N O P T I C A L A C T I V I T Y

767

Figure 23.3. Backscattered ICP Raman (IR + IL ) and ROA (IR − IL ) spectra of (a) α-helical, (b) disordered, and (c) β-sheet poly(L-lysine), in H2 O and recorded on the home-built Glasgow

Wavenumber (cm−1)

instrument.

deformations occur [46]. We have so far observed only weak ROA signals in the S–S stretch region ∼500–550 cm−1 : this may be because any such ROA is generated mainly through isotropic scattering, rather than through the anisotropic scattering mechanism that dominates ROA in backscattering [Eq. (23.2b)], in which case S–S stretch ROA might be best accessed through measurements in the forward direction which is dominated by isotropic scattering [Eq. (23.2a)]. An ROA study of a model hinge peptide containing two disulﬁde bonds similarly failed to identify a clear signature of the S–S stretch [52], but a study of α-cyclodextrin bridged by various patterns of disulﬁde bonds identiﬁed ROA bands originating in S–S and C–S stretches, with quantum-mechanical calculations indicating that the sense of the disulﬁde twist follows the sign of the measured S–S ROA band [53].

768

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Natively unfolded

Wavenumber (cm −1)

Figure 23.4. Backscattered SCP Raman and ROA spectra of (a) human serum albumin, (b) human immunoglobulin G, and (c) bovine β-casein, together with MOLSCRIPT diagrams of the ﬁrst two, in aqueous solution and recorded on the ChiralRAMAN instrument.

23.4.2.1. α-Helix. Polypeptides in model α-helical conformations, as well as proteins rich in α-helix, show a highly characteristic ROA band pattern [11–17, 54]. The example of α-helical poly(l-lysine) is shown in Figure 23.3a, and that of the mainly α-helical protein HSA is shown in Figure 23.4a. A similar pattern is observed in the ROA spectra of two examples displayed in Figure 23.5, namely the alanine-rich peptide AAKAAAAKAAAAKAAAAKAGY-NH2 (AK21) [55], and the ﬁlamentous bacterial virus Pf1 [56]. Focusing initially on HSA, individual ROA bands characteristic of αhelix include the broad positive band peaking at ∼935 cm−1 and the couplet centered at ∼1103 cm−1 negative at low wavenumber and positive at high, both signals being assigned to skeletal stretch modes of the peptide backbone; the positive bands at ∼1300 cm−1 and 1340 cm−1 assigned to extended amide III type modes; and the couplet centered at ∼1650 cm−1 negative at low wavenumber and positive at high arising from amide I modes. Similar bands are evident in the ROA spectra of other α-helical proteins. Likewise for viruses with α-helical coat protein folds such as Pf1 (Figure 23.5b) and other ﬁlamentous bacterial viruses [56–58], the major coat proteins of which have extended α-helix folds, as well as tobacco mosaic virus (TMV) and similar viruses [58–60], the coat proteins of which have helix bundle folds. We generally observe the HSA type of overall ROA band pattern, albeit with small wavenumber shifts in the constituent bands, in heteropolypeptide α-helical sequences as

S T R U C T U R E A N D B E H AV I O R O F B I O M O L E C U L E S F R O M R A M A N O P T I C A L A C T I V I T Y

IR + IL

(a) AK21

5.8 × 107 1309

IR − IL

ROA

1344 1651

0 1627

1280

3.5 × 104

IR + IL

(b) bacteriophage Pf1

1.1 × 108

IR − IL

ROA

1128

1343

1299

1666

0 1322 1.4 × 10 800

1097

5

1000

1200 1400 Wavenumber (cm –1)

1633

1600

Figure 23.5. Backscattered ICP Raman and ROA spectra of (a) the alanine-rich peptide AK21 and (b) ﬁlamentous bacteriophage Pf1, in aqueous solution and recorded on the home-built Glasgow instrument. Adapted from [54].

in proteins; whereas in α-helical homopolypeptides the backbone skeletal stretch region is sometimes rather different, presumably due to contributions from the particular side chains (in heteropolypeptides the ROA contributions from side chains appear to largely cancel each other out). But in all cases the general patterns in the extended amide III and the amide I regions are similar, consistent with the dominance of peptide backbone rather than side-chain modes in these regions. The strong sharp positive ROA band at ∼1340 cm−1 disappears immediately when the peptide, protein or virus is dissolved in D2 O. This indicates ﬁrst that N–H deformations of the peptide backbone make a signiﬁcant contribution to the generation of the ∼1340 cm−1 ROA band because the corresponding N–D deformations contribute to normal modes in a spectral region several hundred wavenumbers lower; and second that, in proteins and viruses, the corresponding sequences are exposed to solvent, rather than being buried in hydrophobic regions where amide protons can take months or even

769

770

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

years to exchange. Although the positive ∼1300 cm−1 α-helix ROA band also changes in D2 O, again suggesting some involvement of N–H deformations, in systems containing a signiﬁcant amount of α-helix in a protected hydrophobic environment quite a lot of intensity is often retained. Mainly for these reasons, the positive ∼1300 and 1340 cm−1 ROA bands have been assigned to extended amide III type modes of unhydrated and hydrated forms of α-helix, respectively [11, 54]. An especially interesting aspect of hydrated α-helix is its enhanced susceptibility to unfolding, which ROA studies have already suggested may be primarily to the poly(lproline) (PPII)-type helix rather than the random coil, and which may have implications for amyloid ﬁbril formation [55] (vide infra). 23.4.2.2. β-Sheet. Considerably more diversity is found in β-sheet than in α-helix structures, with a variety of twisted and curved surfaces that depend on the number and length of the constituent strands and on the environment of the sheet [61–63]. Although there are assignments for β-sheet bands in conventional infrared and Raman spectra and in UVCD and VCD, these techniques have difﬁculty in discriminating between the different structural types of β-sheet found in proteins. However, ROA shows enhanced discrimination with respect to different types of β-sheet due to the large number of wellresolved bands in protein ROA spectra. A case in point is the ability of ROA to easily distinguish between parallel and antiparallel β-sheet (vide infra). The Raman and ROA spectra of a polypeptide in a model β-sheet conformation, those of poly(l-lysine), are displayed in Figure 23.3c; while Figure 23.6 shows examples of the spectral band patterns observed for proteins containing distinct types of β-sheet [50]. According to the Protein Data Bank (PDB) X-ray crystal structure 1nls, jack bean concanavalin A contains 44.7% β-strand in antiparallel sheets and belongs to the mainly β class with a jelly roll β-sandwich fold [63]; structure 1sca of subtilisin Carlsberg contains 17.2% β-strand mostly in the form of parallel sheet plus 29.6% α-helix and belongs to the αβ class; structure 1dab of P.69 pertactin contains 52.9% β-strand in parallel sheets and belongs to the mainly β class with a β-helix fold [63]; and structure 2 ms2 of the MS2 capsid protein subunit contains 46.5% β-strand in up-and-down antiparallel sheets plus 16.3% α-helix and belongs to the αβ class with an unusual fold seen in the peptide binding domain of a class 1 major histocompatibility antigen [63]. The extended amide III region of the ROA spectra of β-sheet proteins is particularly informative. The negative ROA bands at ∼1219–1247 cm−1 in the four proteins in Figure 23.6 are assigned to β-structure [11]. Since β-turn vibrations have been assigned to the ∼1260 to 1295-cm−1 region in the conventional Raman spectra of proteins [64], the positive band at ∼1295 cm−1 in the ROA spectra of concanavalin A and the MS2 capsid are assigned to the β-turns connecting some of the strands within the multistranded antiparallel β-sheet that each contains, and the positive ∼1289-cm−1 ROA band in pertactin is assigned to the β-turns connecting the strands in the parallel β-sheet making up the β-helix. In combination with the lower-wavenumber negative β-structure bands just described, this serves to generate a large ROA couplet that is probably characteristic of the β-hairpin motif [63] (or, more rarely, the β-helix). Negative ROA bands in the range ∼1340–1380 cm−1 also appear to originate in similar β-turns, examples being evident in the ROA spectra of concanavalin A, pertactin, and the MS2 capsid in Figure 23.6. ROA band patterns similar to those exhibited by concanavalin A and the MS2 capsid in the region ∼1220–1380 cm−1 are observed in all the proteins containing multistranded up-and-down antiparallel β-sheet that we have studied, including immunoglobulin (Figure 23.4b).

S T R U C T U R E A N D B E H AV I O R O F B I O M O L E C U L E S F R O M R A M A N O P T I C A L A C T I V I T Y

Wavenumber (cm−1)

Figure 23.6. Backscattered ICP Raman and ROA spectra, together with MOLSCRIPT diagrams, of (a) jack bean concanavalin A, (b) subtilisin Carlsberg, (c) P.69 pertactin, and (d) the empty bacteriophage MS2 protein capsid, in aqueous solution and recorded on the home-built Glasgow instrument. (Adapted from reference 50.)

771

772

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

These β-turn assignments would explain the absence of similar positive bands in the range ∼1260–1295 cm−1 and negative bands in the range ∼1340–1380 cm−1 in the ROA spectrum of subtilisin Carlsberg since the parallel sheet in this protein is based on the beta–alpha–beta motif [63] in which the parallel strands are connected by α-helix loops. These bands are absent in the ROA spectra of other proteins we have studied having parallel β-sheet based on the beta–alpha–beta motif. Hence ROA appears to be unique in its ability to distinguish parallel from antiparallel sheet since the presence of bands characteristic of signiﬁcant amounts of both α-helix and β-sheet, but the absence of bands characteristic of β-turns and hairpin bends is diagnostic of parallel β-sheet (provided that β-helix, which has its own characteristic ROA band patterns, can be ruled out). The extended amide III region of the ROA spectrum of β-sheet poly(l-lysine) displayed in Figure 23.3c is dominated by a large couplet, negative at ∼1218 and positive at ∼1260 cm−1 and similar to that observed in the β-sheet proteins but shifted by ∼30 cm−1 to lower wavenumber. Amide III bands from β-sheet in conventional Raman spectroscopy are assigned to the region ∼1230–1245 cm−1 [18, 19], an example being the strong band at ∼1239 cm−1 in the parent Raman spectrum of β-sheet poly(l-lysine), which coincides with the center of the large poly(l-lysine) ROA couplet discussed above. In the ROA spectrum of β-sheet poly(l-lysine) in D2 O, both this ROA couplet and its associated parent Raman band are absent, but a new large couplet positive at ∼981 and negative at ∼1019 cm−1 is present that is associated with two new Raman bands at similar wavenumbers [50]. This observation conﬁrms that N–H deformations make signiﬁcant contributions to the generation of this ROA couplet, with corresponding N–D deformations contributing to the new ROA couplet at lower wavenumber. The ROA spectrum of β-sheet poly(l-lysine) also displays a large negative band at ∼1351 cm−1 assigned to β-turns. We therefore envisage β-sheet poly(l-lysine) to contain up-and-down β-sheet based on the hairpin motif. In the amide I region a large couplet, negative at ∼1658 cm−1 and positive at ∼1677 −1 cm , in the ROA spectrum of concanavalin A in Figure 23.6a is another signature of β-sheet and can be distinguished easily from a similar amide I couplet produced by αhelix, which typically occurs at ∼10 cm−1 lower (vide supra). A similar large couplet is observed in the ROA spectrum of the MS2 capsid. This correlates with β-sheet amide I bands in conventional Raman spectroscopy which occur in the range ∼1665–1680 cm−1 [18, 19]. However, the corresponding amide I couplet in pertactin is unusual in that it displays a much larger negative component that has not been observed in the ROA spectrum of any other β-sheet protein to date and may provide a unique signature of the β-helix fold. The negative–positive–negative–positive band pattern observed in the range ∼1600–1690 cm−1 in the ROA spectrum of β-sheet poly(l-lysine) displayed in Figure 23.3c is rather different from the band patterns observed in this region in the ROA spectra of the β-sheet proteins displayed in Figure 23.6 and may be characteristic of extended ﬂat multistranded β-sheet, the wavenumber range being similar to that for the amide I vibrations of such structures [65]. The suppressed intensity of the amide I ROA couplet centered at ∼1672 cm−1 in poly(l-lysine) relative to the large couplet centered at ∼1665 cm−1 in concanavalin A, for example, parallels the weak amide I VCD observed for such model β-sheet structures compared with the large VCD signals seen in typical β-sheet proteins [65]. This may result from the “planar” nature of the constituent strands within the ﬂat multistranded β-sheet supported by the polypeptide for which the intrinsic skeletal chirality and, hence, ROA and VCD intensities are smaller

S T R U C T U R E A N D B E H AV I O R O F B I O M O L E C U L E S F R O M R A M A N O P T I C A L A C T I V I T Y

than for the twisted strands present in the more irregular β-sheet structures found in typical native proteins [65]. The positive ROA band at ∼1564 cm−1 in β-sheet poly(l-lysine) is assigned to the amide II vibration. It disappears when β-sheet poly(l-lysine) is prepared in D2 O solution, with a new positive ROA band assigned to the amide II appearing at ∼1478 cm−1 together with other changes, conﬁrming that N–H deformations are heavily involved [50]. This ROA band may sometimes be confused with that from the W3-type vibration of the indole ring in tryptophan side chains which occurs in the range ∼1545–1560 cm−1 (vide infra). It is possible that an unusually large positive amide II ROA band is characteristic of the extended multistranded ﬂat type of β-sheet found in β-sheet polypeptides. 23.4.2.3. 310 -Helix and Left-Handed α-Helix. The identiﬁcation of bands characteristic of 310 -helix in protein ROA spectra has proved elusive. One problem is the absence of suitable model 310 -helical peptide states. Right-handed 310 -helical conformations are supported by peptides constructed from amino acid residues in which the Cα hydrogen atom is replaced by a group such as methyl [66], but not by peptides constructed from “conventional” residues with Cα –H groups. The absence of the Cα –H deformations in Cα -substituted peptides means that any associated ROA band structure in the extended amide III region could be very different from that observed in typical proteins. This is born out by the ROA spectrum of a water-soluble 310 -helical heptapeptide [67]. Although showing signiﬁcant extended amide III ROA bands in the form of a large couplet, negative at ∼1312 cm−1 and positive at ∼1333 cm−1 which will be useful for characterizing the 310 -helical conformation in Cα -substituted peptides, this band pattern is rather different to anything observed in typical proteins. Poly(β-benzyl l-aspartate) is thought to take up a left-handed α-helical conformation in CHCl3 [68]. The backscattered Raman and ROA spectra of this system are shown in Figure 23.7b [54]. It is reassuring that the ROA spectrum is similar to that of righthanded α-helical poly(γ -benzyl l-glutamate) in CHCl3 [54], shown in Figure 23.7a, in the extended amide III and amide I regions, but with all the signs reversed! This result is so clear-cut that ROA will be valuable for identifying the left-handed α-helical conformation in more general peptide and protein systems. 23.4.2.4. Side Chains. Bands from side chains are usually not very prominent in the ROA spectra of polypeptides and proteins. Side-chain differences may be responsible for some of the small variations observed in the characteristic helix, sheet, and turn ROA bands. There are, however, several distinct regions where side-chain vibrations appear to be largely responsible for the observed ROA features. In particular, ROA bands in the range ∼1400–1480 cm−1 originate in CH2 and CH3 side-chain deformations and also in tryptophan vibrations; ROA bands in the range ∼1545–1560 cm−1 originate in tryptophan vibrations; and some ROA bands in the range ∼1600–1630 cm−1 originate in vibrations of aromatic side chains, especially tyrosine [11]. Also the ring breathing mode of the aromatic ring in phenylalanine, which generates a strong band in the conventional Raman spectrum at ∼1000 cm−1 and a strong ROA band in small chiral molecules such as 1-phenylethanol (Figure 23.2), may be associated with ROA bands observed in this region in some proteins. The absolute stereochemistry of the tryptophan conformation, in terms of the sign and magnitude of the torsion angle χ 2,1 around the bond connecting the indole ring to the Cβ atom, may be obtained from the ∼1545 to 1560 cm−1 tryptophan ROA band, assigned to a W3-type vibration of the indole ring [19]. This was discovered from observations

773

774

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Figure 23.7. Backscattered ICP Raman and ROA spectra of (a) α-helical poly(γ -benzyl and (b) poly(β-benzyl L-aspartate),

L-glutamate)

in CHCl3 and recorded on the home-built Glasgow instrument. (Adapted from reference 54). The poly(β-benzyl L-aspartate) α-helix is

Wavenumber (cm−1)

left-handed!.

of tryptophan W3 ROA bands with equal magnitudes but opposite signs in two different ﬁlamentous bacterial viruses, M13 and H75, with coat protein subunits containing a single tryptophan (Figure 23.8), which suggested that the tryptophans adopt quasi-enantiomeric conformations in the two viruses [57]. Since the magnitude of the angle χ 2,1 may be deduced from the W3 Raman band wavenumbers [19], it was possible to obtain both the sign and magnitude of this angle from the ROA spectrum [57], something usually only obtainable from high-resolution X-ray protein crystal structures. This result was conﬁrmed by a recent ab initio computational study of the ROA spectra expected from modes of the indole ring in tryptrophan for a range of torsion angles [69]. The W3 ROA band may also be used as a probe of conformational heterogeneity among a set of tryptophans in disordered protein sequences since cancellation from ROA contributions with opposite signs will result in a loss of ROA intensity. Examples are observed in molten globule states of equine lysozyme [70] and human lysozyme [55]. Tryptophan ROA is similar in this respect to the near UVCD bands from aromatic side chains, which disappear when tertiary structure is lost on partial denaturation [71]. These two techniques provide complementary perspectives because ROA probes the intrinsic skeletal chirality of the tryptophan side chain whereas UVCD probes the chirality in the general environment of the chromophore. 23.4.2.5. Polyproline (II) Helix: A Careful Disorderliness. Although originally deﬁned for the conformation adopted by polymers of l-proline, the PPII helix can be supported by amino acid sequences other than those based on l-proline and has been

775

S T R U C T U R E A N D B E H AV I O R O F B I O M O L E C U L E S F R O M R A M A N O P T I C A L A C T I V I T Y

IR + I L

M13

1.3 × 107 1308

ROA

1345

IR − I L

1120

1660 1555

N 2

0

3 c2,1 1644 1.7 × 10

1089 4

1633

β

α

C

c1 N

IR + IL

H75

2.8 × 107

IR − IL

ROA

1121

1296

1344 1665

0 1092

1244

2.3 × 104 800

1400 1000 1200 Wavenumber (cm−1)

1550 1600

Figure 23.8. Backscattered ICP Raman and ROA spectra of ﬁlamentous bacteriophages M13 and H75 in aqueous solution and recorded on the home-built Glasgow instrument. The corresponding tryptophan W3 ROA bands (shaded) at ∼1555 and 1550 cm−1 have opposite signs, indicating quasi-enantiomeric tryptophan conformations with respect to the torsion angle χ 2,1 . (Adapted from reference 57.)

recognized as a common structural motif within the longer loops in the X-ray crystal structures of many proteins [72]. It consists of a left-handed helix with threefold rotational symmetry in which the φ,ψ angles of the constituent residues are restricted to values around −75◦ , +145◦ , corresponding to a region of the Ramachandran surface adjacent to the β-region. The extended nature of the PPII helix precludes intrachain hydrogen bonding, with protein X-ray crystal structures [72] and modeling studies [73] indicating that the structure is stabilized instead by main chain hydrogen bonding with water molecules and side chains. This produces a ﬂexible, adaptable structure that is becoming

776

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

increasingly recognized as a major conformational element of disordered polypeptides and unfolded proteins in aqueous solution [74, 75] and which may be important in the functional role of intrinsically unfolded sequences [76]. It can be distinguished from random coil in polypeptides using UVCD [74–76], VCD [77, 78], infrared and Raman [79], and UV resonance Raman [80]. However, these techniques in isolation have difﬁculty in identifying PPII helix when other conformational elements are present. ROA has proved especially useful in identifying PPII helix [81], even in proteins, and is proving valuable for the study of PPII structure and behavior in unfolded and partially folded proteins. Originally, the ROA band assignments for PPII were based on dominant features, such as the strong positive extended amide III band at ∼1319 cm−1 and the weaker positive amide I band at ∼1673 cm−1 , observed in the ROA spectra of disordered poly(l-lysine) (Figure 23.3b), and similar bands in poly(l-glutamic acid) [11, 12], relying mainly on the UVCD and VCD evidence that these polypeptide states contain large amounts of PPII [74–78]. Although the lack of main-chain hydrogen bonding makes NMR determinations of PPII structure difﬁcult, a study using a combination of NMR and UVCD has demonstrated convincingly that a seven-residue alanine peptide adopts predominantly the PPII conformation in aqueous solution [82]. Independently, a Raman, FTIR, and VCD study revealed that cationic tetra-alanine (Ala4 ) also adopts predominantly the PPII conformation in aqueous solution [79]. A subsequent ROA study of these two alanine peptides, along with Ala2 , Ala3 , Ala5 , and Ala6 , corroborated the earlier ROA band assignments for PPII structure and monitored PPII formation as a function of chain length [81], conﬁrming that it becomes fully established at Ala4 . Conﬁdence in the ROA band assignments of PPII structure has been reinforced by an ab initio simulation of the ROA spectrum of the model oligomer formyl-Ala10 -amide constrained to the PPII-helical geometry [83]. This simulated ROA spectrum is displayed in Figure 23.9b, together with the experimental ROA spectrum of acetyl-OOAla7 OOamide (O = l-ornithine) in Figure 23.9a. This peptide is very similar to that mentioned above that has been found to take up a predominantly PPII-helical conformation in aqueous solution [82]. The strong positive ∼1319-cm−1 ROA band, highly characteristic of PPII structure, in the experimental ROA spectrum of acetyl-OOAla7 OO-amide is well-reproduced in the simulated spectrum. The corresponding normal mode calculation provides an assignment of this band to mainly Cα –H deformations coupled along the entire molecule. The positive amide I peak at ∼1677 cm−1 , characteristic of PPII helix, which sometimes appears as part of a couplet with a small negative band at lower wavenumber, is also well-reproduced by the simulation in which it is generated by amide I vibrations coupled along the entire molecule. Most of the other bands in the experimental ROA spectrum of acetyl-OOAla7 OO-amide are also well-reproduced in the simulated spectrum of formyl-Ala10 -amide in the PPII-helical conformation, which provides assignments to mostly Cα –H deformations mixed with CH3 group vibrations. Furthermore, from X-ray crystal structures of liganded α-type human estrogen receptor, a PPII-helical sequence DAEPPILYSEY has been identiﬁed in the ligand-binding domain of the receptor: The ROA spectrum of the corresponding peptide acetyl-DAEPPILSEYamide is very similar to those in Figure 23.9, with UVCD and differential scanning calorimetry indicating a PPII content ∼85%, the same proportion as observed in the X-ray structures [84]. The manner in which hydration stabilizes the PPII helical conformation in a plastic adaptable manner is especially important since it enables “disordered” sequences in

S T R U C T U R E A N D B E H AV I O R O F B I O M O L E C U L E S F R O M R A M A N O P T I C A L A C T I V I T Y

Wavenumber (cm −1)

Figure 23.9. (a) Backscattered ICP ROA spectrum of the alanine-based peptide acetyl-OOAla7 OOamide (O = L-ornithine) in aqueous solution recorded on the home-built Glasgow instrument. (b) Ab initio simulated ROA spectrum of formyl-Ala10 -amide constrained to the PPII-helical geometry. (Adapted from reference 83.)

folded and natively unfolded proteins to perform many essential functions [75, 76]. Furthermore, hydrated PPII structure may facilitate protein folding since it pre-organizes the unfolded state, thereby lowering the entropy and reducing the conformational space to be searched [74]. Hence the dictum “there are some enterprises in which a careful disorderliness is the true method” (Herman Melville, Moby Dick ) seems to be just as applicable to protein structure, folding, and function as it was to chasing whales around the globe! A careful disorderliness may also be important for the immune response. This idea arose from a recent ROA study of PPII structure in poly(l-lysine) dendrigrafts (DGLs), which revealed that although generation 1 supported predominantly the PPII conformation, the PPII content steadily decreased with increasing generation, with a concomitant increase in other backbone conformations [85]. This behavior may be due to increasing crowding of the lysine side chains, together with suppression of backbone hydration, with increasing branching. Suppression of the PPII content of DGLs with increasing branching could be associated with their nonimmunogenic properties (http://www.colcom.eu). Intrinsic disorder is known to be crucial in the immune response: Short disordered peptides are good antigens, whereas long disordered regions and intrinsically disordered proteins initiate only weak immune responses or are completely nonimmunogenic [86, 87]. The careful disorderliness of a relatively short sequence supporting PPII structure may be associated with the ability to readily adapt to other conformations required by the structure of the antigen binding site. The conformational heterogeneity of the more random structure found in the short separate lysine sequences within DGLs, which would also have reduced mobility due to steric crowding, may prevent them from adapting so readily. 23.4.2.6. Multivariate Analysis of Protein ROA Spectra. The large number of resolved structure-sensitive bands in protein ROA spectra makes them suitable for the application of multivariate analysis (pattern recognition) techniques to extract structural

777

778

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

information. Useful structural relationships among proteins may be obtained by analyzing their ROA spectra using the method of principal component analysis (PCA) (e.g., references 12–14 and 50). The more advanced multivariate analysis method called nonlinear mapping (NLM) was found to yield even better results [88]. Multivariate analysis techniques start by considering each digitized spectrum to be a vector from the origin to a point in a multidimensional space, with the axes representing the digitized wavenumbers. Points close to or distant from each other in this multidimensional space are then taken to represent similar or dissimilar spectra, respectively. Like PCA, NLM seeks to create a lower-dimensional space in which the relative positions of the points approximately preserve the relationships between the spectra, thereby providing a more easily comprehended representation. The advantage of NLM over PCA is that it aims to represent the relationships between all spectra rather than just describing the gross overall variance (which can lead to poor representation of some sample members), so that characterization of spectral and hence structural similarities is optimized. A two-dimensional (2D) NLM plot for a set of 80 polypeptide, protein, and virus ROA spectra in aqueous solution shows excellent clustering corresponding to different types of structure [88]. Clusters corresponding to the following structural classes are observed: all α, mainly α, αβ, mainly β, all β, mainly disordered/irregular, and all disordered/irregular. It was also found that mapping into three dimensions has the advantage of separating distinct clusters that are otherwise superimposed and therefore indistinguishable in the 2D NLM plot [88]. The average standardized ROA spectra of the polypeptides and proteins falling within each structure class in the 2D NLM plot are presented in Figure 23.10. Each has several distinct features characteristic of the type of structure. The peak wavenumbers of some of the more prominent ROA bands in the average spectra of the all α, all β, and all-disordered/irregular classes have been marked in the top three spectra of Figure 23.10 and reinforce the validity of the earlier assignments. The average all-disordered/irregular ROA spectrum is similar to that of PPII helix. As well as structure classes, analytical methods have also been applied to determine quantitative secondary structure content. Applying the partial least squares algorithm with ﬁvefold cross-validation to 44 ROA and 24 Raman protein spectra from the same dataset as that used for the PCA and NLM analyses outlined above, secondary structure contents could be determined with considerably better accuracy than any other spectroscopic method, including UVCD, yet reported [89]. 23.4.2.7. Unfolded Proteins. The study of proteins that are unfolded in their native functional states is a burgeoning new area of protein science that ROA, together with NMR, has helped to establish. Such “intrinsically disordered” or “natively unfolded” proteins are now recognized as constituting an important structural class that have a variety of important functions [90, 91]. ROA has proved useful in several recent studies of natively unfolded proteins of biological and physiological signiﬁcance. The ROA spectra of bovine milk caseins (the example of β-casein being displayed in Figure 23.4c) [92], several wheat prolamins [93], and human recombinant synuclein and tau brain proteins [92], some of which are associated with neurodegenerative disease, were all found to be dominated by a strong positive band at ∼1316–1322 cm−1 assigned to PPII helix. Although most natively unfolded proteins have ROA spectra that are very similar to the average mainly disordered/irregular ROA spectrum at the bottom of Figure 23.10, there are differences of detail which reﬂect differences in residue composition and minor differences in structural elements. It should not be thought that natively unfolded proteins are always devoid of extensive amounts of

S T R U C T U R E A N D B E H AV I O R O F B I O M O L E C U L E S F R O M R A M A N O P T I C A L A C T I V I T Y

4

1344

all a 1125

2

1662

1302

0 −2 4

1632

1092 all b

1311 1288 1318

2

1674

0 −2 4

1242 all disordered/irregular

1356 1318

1654

1677

2 0 −2 4

1263 ab

2 0 −2 4

mainly a

2 0 −2 4

mainly b

2 0 −2 4

mainly disordered/irregular

2 0 −2 800

1000

1400 1200 Wavenumber (cm−1)

1600

Figure 23.10. Averages of the standardized backscattered ROA spectra (ICP and SCP) in aqueous solution for the seven main protein structure classes within a set of 80 polypeptide, protein and virus ROA spectra. (Adapted from reference 88.) The vertical axis represents arbitrary intensities standardized to put the spectra on an equal footing.

secondary structure; for example, as well as a large amount of PPII, ROA demonstrates that A-gliadin also contains a large amount of hydrated α-helix [93]. A recent study of residual structure in disordered peptides and unfolded proteins was carried out via multivariate analysis of ROA spectra [83] and revealed striking differences between the structural characteristics of natively unfolded proteins and proteins unfolded by denaturation. The former cluster in the mainly disordered/irregular region of the NLM plot and contain a signiﬁcant amount of PPII structure; the latter appear in other regions

779

780

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

and contain signiﬁcant amounts of (a) β-sheet in the case of reduced proteins and (b) α-helix in the case of acid molten globules. Many proteins exist in native states which are hybrid structures comprising a compact folded domain and a domain which can range from a semi-compact molten globule-like structure to a completely unfolded structure. A famous example is the prion protein, the cellular form of which has a compact folded C-terminal α-helical domain and a long unfolded N-terminal tail [94]. Such systems present a serious challenge to conventional spectroscopic techniques since these have difﬁculty in studying the structures of the folded and unfolded domains simultaneously in the intact protein. In work that will serve as a paradigm for such studies, ROA has been used to simultaneously probe the structures of both the folded and unfolded domains of the ovine prion protein [95]. Figure 23.11a displays the backscattered Raman and ROA spectra of the full-length ovine prion protein PrP25−233 and Figure 23.11b that of the truncated protein PrP94−233 from which most of the N-terminal tail has been removed. The sharp positive band at ∼1315 cm−1 assigned to PPII structure that is present in the ROA spectrum of the full-length protein is absent from that of the truncated protein, together with negative and positive bands at ∼1257 and 1296 cm−1 , respectively, that are assigned to β-turns. This suggests that the 23–89 fragment is composed mainly of PPII-helical segments interspersed with β-turns, and that little PPII is present in the C-terminal structured domain. These structural elements are associated with copper binding to the N-terminal region and may have a functional role in copper metabolism [96]. Viral coat proteins sometimes possess long disordered sequences together with compact folded domains. ROA has identiﬁed large amounts of PPII structure in several cases. One example is tobacco rattle virus (TRV), which has a similar rod-shaped particle structure to TMV but which shows a strong positive ROA band at ∼1315 cm−1 not shown by TMV [59]. This may arise from PPII in surface-exposed sequences of the coat proteins, which are thought to have a helix bundle fold. These PPII sequences may serve to ﬁll the extra volume required by the larger diameter of the TRV particle relative to that of TMV, and they may be associated with its transmission by nematodes. Another example is satellite tobacco mosaic virus, a small icosahedral virus, which shows a strong positive ROA band at ∼1316 cm−1 [58]. This may originate in a PPII structure in the long disordered N-terminal strand of the coat proteins, all having the same jelly roll β-sandwich fold; this structure projects into the virus interior where it interacts with the RNA. 23.4.2.8. Protein Misfolding and Disease: Amyloid Fibril Formation. The mobile regions within native proteins, partially denatured protein states such as molten globules, and natively unfolded proteins underlie many of the protein misfolding diseases [97]. Many of these diseases involve amyloid ﬁbril formation, as in amyloidosis from mutant human lysozymes, neurodegenerative diseases such as Parkinson’s and Alzheimer’s due to the ﬁbrillogenic propensities of α-synuclein and tau respectively, and the prion encephalopathies such as scrapie, BSE, and new variant CJD where amyloid ﬁbril formation is triggered by exposure to the amyloid form of the prion protein. It has been suggested that PPII helix is the “killer conformation” in some of these amyloid diseases [55]. This was prompted by an ROA study of human lysozyme, the thermal denaturation behavior of which is different from that of hen lysozyme in that it supports a partially folded molten globule state at low pH and elevated temperatures. Incubation at 57◦ C and pH 2.0, under which conditions the molten globule state is the most highly populated, induces the formation of amyloid ﬁbrils. ROA revealed that the α-domain of the protein is partially destabilized, with the formation of PPII structure at the expense of hydrated α-helix [55].

S T R U C T U R E A N D B E H AV I O R O F B I O M O L E C U L E S F R O M R A M A N O P T I C A L A C T I V I T Y

1619

IR + IL

(a) ovine PrP25-233

2.0 × 109 1296

IR − IL

ROA

1315

1676

1339

1556

0

7.5 × 105

1230

1257

1442

1616

1636

IR + IL

(b) ovine PrP94–233

3.4 × 109 1300

IR − IL

ROA

1119

1308 1340

1671 1557

0 1228

1.1 × 106 800

1000

1200

1441

1242

1400

1617

1639

1600

Wavenumber (cm−1)

Figure 23.11. Backscattered ICP Raman and ROA spectra of (a) ovine PrP25−233 and (b) ovine PrP94−233 in aqueous solution and recorded on the home-built Glasgow ICP instrument. (Adapted from reference 95.)

PPII appears to be a favorable conformation for amyloid ﬁbril formation for the following reasons. Elimination of water molecules between extended PPII chains having hydrated backbone C=O and N–H groups to form β-sheet hydrogen bonds is a favorable thermodynamic process [98]. Since PPII chains are close in conformation to β-strands, they would be expected to readily undergo this type of aggregation with each other and with established β-sheet. A later study using conventional techniques corroborated this idea, but without explicitly identifying the extended polypeptide chains as PPII-type structure [99]. Subsequent ROA work on the brain proteins α-synuclein and tau, which have a propensity to form the ﬁbrils associated with Parkinson’s and Alzheimer’s disease, respectively, revealed that these natively unfolded proteins consist largely of PPII helical sequences [92]. Although disorder of the PPII type may be essential for the formation of regular ﬁbrils, the presence of a large amount of PPII structure does not necessarily impart a

781

782

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

ﬁbrillogenic character since not all PPII-rich nonregular protein structures form amyloid ﬁbrils. For example, the casein milk proteins and the brain proteins β- and γ -synuclein show little propensity for amyloid ﬁbril formation and are not associated with disease, yet their natively unfolded structures are based largely on PPII sequences as in α-synuclein [92], which is highly amyloidogenic. A more complete understanding of the ﬁbrillogenic propensity of a particular sequence, PPII or otherwise, requires knowledge of the various physicochemical properties of the constituent residues [100]. In particular, a combination of high net charge and low mean hydrophobicity has been shown to be an important prerequisite for protein sequences to remain natively unfolded [101]. These considerations suggest that production of PPII in sequences in destabilized protein states for which evolution has not provided a propensity to remain unfolded are likely to confer amyloidogeneticity, as in destabilized human lysozyme. Prion proteins are associated with a variety of neurodegenerative diseases, known as transmissible spongiform encephalopathies (TSE) [102]. Further understanding requires more detailed knowledge of the solution structures of prion proteins and the conformational changes associated with the possible different transformation pathways of the cellular into the scrapie (amyloid ﬁbril) form that are involved in the infectious, genetic, and sporadic versions. The contribution of the N-terminal disordered tail to prion function and misfunction is attracting increasing attention. As mentioned above, ROA has demonstrated that it contains much PPII structure that may be associated with its copper-binding ability [95]. A subsequent ROA study revealed dramatic differences in the inﬂuence of divalent copper and manganese ions on prion protein folding [103], showing that there is ample scope for these and other metal ions to be involved in misfunction and associated TSE.

23.4.3. Carbohydrates Carbohydrates in aqueous solution give rich and informative ROA band structures over a wide range of the vibrational spectrum. Their complex and highly coupled normal modes generate strong ROA bands that produce patterns characteristic of the various types of structural units (sugar rings and glycosidic links) that are usually much easier to interpret than the parent Raman band patterns. The ROA of a range of monosaccharides, including aldose, pentose, and ketose sugars, have been surveyed [104–106] and provide a database of spectral assignments for the key stereochemical features. It has been demonstrated that the ROA spectra contain information on the ring conformation, the relative disposition of OH groups around the ring, the absolute conﬁguration and axial or equatorial orientation of groups attached to the anomeric carbon, and the exocyclic CH2 OH conformation. In structural glycobiology, much importance is placed on the conformation of the C–O–C glycosidic link, formed upon condensation of two monosaccharide units, as it is the most important single determinant of di-, oligo-, and polysaccharide conformation [107]. A study of disaccharides containing d-glucose residues revealed that ROA bands originating in the vibrations of the glycosidic link could be readily identiﬁed [108]. Figure 23.12 displays the ROA spectra of d-maltose and d-cellobiose containing α(1–4) and β(1–4) linkages, respectively. ROA bands assigned to vibrations involving the glycosidic linkages are shaded. Those in the range ∼375–475 cm−1 originate in skeletal deformations and torsions; those in the range ∼850–950 cm−1 are associated with bending and stretching coordinates of the glycosidic link together with C–H deformations, with a large couplet in d-maltose but d-cellobiose showing little ROA here.

S T R U C T U R E A N D B E H AV I O R O F B I O M O L E C U L E S F R O M R A M A N O P T I C A L A C T I V I T Y

Wavenumber (cm −1)

Figure 23.12. Backscattered ICP Raman and ROA spectra of D-maltose and D-cellobiose in aqueous solution, recorded on the home-built Glasgow instrument.

The sensitivity of ROA to glycosidic linkage conformation in large structures was demonstrated by a study of the polysaccharides laminarin and pullulan [109]. Laminarin is composed mainly of d-glucose units joined through β(1–3) links and is known to form triple helical structures in aqueous solution. Comparison of the ROA spectrum of laminarin with that of its constituent disaccharide d-laminaribiose reveals striking differences. Most notably, large changes in the region ∼1050–1150 cm−1 , especially a sign change in a sharp strong couplet centered at ∼1120 cm−1 , were observed and ascribed to changes in the glycosidic link conformation on adopting an ordered helical conformation. In contrast, it was found that the unordered polysaccharide pullulan exhibits an ROA spectrum that appears to be a sum of contributions from the constituent repeating units. Similarly, a recent ROA study of heparin glycosaminoglycans indicates that these polysaccharides do not adopt strongly helical conformations in solution [110]. The cyclodextrins are interesting samples for ROA studies. These are cyclic oligosaccharides containing six, seven, or eight d-glucose residues joined through α(1–4) links and labeled with the preﬁx α, β or γ respectively. They are stabilized by a ring of intramolecular hydrogen bonds and have the shape of a hollow truncated cone with a hydrophobic center, which is responsible for their renowned ability to form inclusion complexes with a wide variety of guest molecules. Cyclodextrins exhibit an enormous (-value ∼30 times larger) glycosidic ROA couplet centered at ∼918 cm−1 as compared with that observed in the corresponding α(1–4) linked disaccharided-maltose [108]. ROA

783

784

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

was used to study the dynamics of the cyclodextrin system by monitoring changes in the glycosidic couplet brought about by changes in environmental factors such as solvent and temperature and by the formation of host–guest inclusion complexes [111]. Simulations have shown that this couplet originates mainly in C–H deformations coupled around the cyclodextrin ring via the hydrogen bonded network [112].

23.4.4. Glycoproteins Intact glycoproteins provide excellent ROA spectra with clear bands originating in both the polypeptide and carbohydrate components [11]. This should be especially valuable in view of the central importance of glycoproteins in biochemistry and the biopharmaceutical industry, along with the fact that they are difﬁcult to study using X-ray crystallography or solution NMR [107]. Figure 23.13 displays the examples of bovine α1 -acid glycoprotein (AGP), bovine mucin, and yeast invertase, which are N-linked, O-linked, and N-linked respectively. Only AGP shows bands that can be clearly assigned to polypeptide secondary structure (shaded); most of the other bands in all three glycoproteins probably originate in the carbohydrate. The ROA spectrum of AGP has been discussed in detail [113], with bands assigned to (a) N , N -diacetylchitobiose in the pentasaccharide core and (b) β-sheet polypeptide structure that may originate in a lipocalin-type fold. The absence of any clear bands from polypeptide secondary structure in the ROA spectra of mucin and invertase indicates that their high levels of glycosylation (∼50%) hold the polypeptide in a ﬁxed completely disordered conformation. Much of the ROA band pattern of invertase originates in the mannose residues and their associated glycosidic links within the glycan chains of this high-mannose glycoprotein.

23.4.5. Nucleic Acids Although ROA studies of nucleic acids are not as advanced as for proteins, the results so far are promising. ROA is sensitive to three different sources of nucleic acid chirality: the chiral base-stacking arrangement of intrinsically achiral base rings, the chiral disposition of the base and sugar rings with respect to the C–N glycosidic link, and the inherent chirality associated with the asymmetric centres of the sugar rings. Studies on pyrimidine nucleosides [114] and synthetic polyribonucleotides [115] have provided a basis for the interpretation of ROA spectra of DNA and RNA. The Raman and ROA spectra of calf thymus DNA, and of phenylalanine-speciﬁc transfer RNA (tRNAPhe ) in the presence and absence of Mg2+ ions, are shown in Figure 23.14 [116]. ROA bands in the region ∼900–1150 cm−1 originate in vibrations of the sugar rings and phosphate backbone. The region ∼1200–1550 cm−1 is dominated by normal modes in which the vibrational coordinates of the base and sugar rings are mixed. ROA band patterns in this sugar-base region appear to reﬂect the mutual orientation of the two rings and the sugar ring conformation. The region ∼1550–1750 cm−1 contains ROA bands characteristic of the bases and their stacking arrangements. Although the ROA spectra of the DNA and the two RNAs are similar, important differences of detail exist. The main differences originate in the DNA taking up a B-type double helix in which the sugar puckers are mainly C2 -endo and the RNAs taking up A-type double-helical segments where the sugar puckers are mainly C3 -endo. There are smaller differences between the two RNA spectra, and these are most apparent in the sugarphosphate region ∼900–1150 cm−1 . It is known that Mg2+ ions are necessary to hold RNAs in their speciﬁc tertiary folds. As illustrated in Figure 23.14, for tRNAPhe in the

S T R U C T U R E A N D B E H AV I O R O F B I O M O L E C U L E S F R O M R A M A N O P T I C A L A C T I V I T Y

IR + I L

(a) bovine AGP

6 × 109 ROA

IR − IL

3 × 105

I R + IL

(b) bovine mucin

6 × 109 0

IR − IL

ROA

0 1 × 105

IR + IL

(c) yeast invertase

5 × 109

IR − IL

0

ROA

0 2 × 105 400

600

800

1000

Wavenumber

1200

1400

1600

(cm−1)

Figure 23.13. Backscattered SCP Raman and ROA spectra of the intact glycoproteins (a) bovine AGP, (b) bovine mucin, and (c) yeast invertase, all in aqueous solution and recorded on the ChiralRAMAN instrument.

presence of Mg2+ this is a compact L-shaped form, whereas in the absence of Mg2+ the tRNAPhe adopts an open cloverleaf secondary structure [117]. The Mg2+ -free tRNAPhe shows a strong negative–positive–negative ROA triplet (shaded) at ∼992, 1048, and 1091 cm−1 which is very similar to that found in A-type polyribonucleotides [115] and is assigned to the C3 -endo sugar pucker. This signature is weaker and more complex in the ROA spectrum of the Mg2+ -bound sample, suggesting a wider range of sugar puckers associated with the loops and turns that characterize the tertiary structure of the folded form. Unique ROA ﬁngerprints of several distinct RNA structural motifs have been reported [118].

785

786

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Wavenumber (cm −1)

Figure 23.14. Backscattered ICP Raman and ROA spectra of (a) calf thymus DNA, (b) Mg2+ -bound tRNAPhe and its associated L-shaped tertiary fold, and (c) Mg2+ -free tRNAPhe and its associated open clover-leaf secondary structure, in aqueous solution and recorded on the home-built Glasgow instrument. (Adapted from reference 116.)

23.4.6. Viruses Knowledge of the structure of viruses at the molecular level is essential for understanding their modus operandi . However, the application of key structural biology techniques such as X-ray crystallography or ﬁber diffraction is often hampered by practical difﬁculties. Conventional Raman is valuable in studies of intact viruses at the molecular level because it is able to simultaneously probe both the protein and nucleic acid constituents [19, 119]. The additional incisiveness of ROA, which may be applied to most types of virus including ﬁlamentous, helical, rod-shaped, and icosahedral [58], further enhances the value of Raman spectroscopy in structural virology.

S T R U C T U R E A N D B E H AV I O R O F B I O M O L E C U L E S F R O M R A M A N O P T I C A L A C T I V I T Y

The ﬁrst virus ROA spectra were reported for ﬁlamentous bacteriophages [56]. As discussed in Section 23.4.2.1, the data facilitated the identiﬁcation of ROA bands associated with unhydrated and hydrated α-helix since large amounts of both types are present in the overlapping extended helical coat proteins in the intact viruses. ROA has also proved useful in the comparison of the helix bundle coat protein structure of the rodshaped viruses potato virus X (PVX) and narcissus mosaic virus (NMV) with that of TMV [60]. In Section 23.4.2.4, it was mentioned that the discovery that tryptophan absolute stereochemistry could be determined from the signs of the tryptophan W3 ROA bands was also made from ROA data on ﬁlamentous bacteriophages. Section 23.4.2.4 also alludes to signiﬁcant amounts of PPII structure detected by ROA in several viruses which may have functional signiﬁcance. To illustrate the power of ROA in structural virology, Figure 23.15 displays a set of spectra measured on cowpea mosaic virus (CPMV) in aqueous solution [58]. This virus has a nucleic acid genome consisting of two different RNA molecules called RNA-1 and RNA-2 which are separately encapsidated in identical icosahedral protein shells, the structure of which is known from X-ray crystallography. Figure 23.15a illustrates how the icosahedral capsid is constructed from 60 copies of an asymmetric unit made up of three different protein domains A, B, and C, each of which has a similar structure with the same jelly roll β-sandwich fold illustrated in Figure 23.15b. Virus preparations can be separated into empty protein capsids, capsids containing RNA-1 and capsids containing RNA-2. The top panel of Figure 23.15c shows the Raman and ROA spectra of the empty protein capsid, the band patterns being characteristic of the jelly roll β-sandwich fold of the individual protein domains (compare with the ROA spectrum of concanavalin A in Figure 23.6a). The middle panel shows the spectra of the capsid containing RNA-2, with bands from the nucleic acid now evident in addition to those from the protein. The bottom panel shows the spectra obtained by subtracting the top spectrum from the middle spectrum. The difference ROA spectrum looks similar to those of typical RNA molecules such as the Mg2+ -free tRNAPhe in Figure 23.14c, and it is taken as originating mainly in the viral RNA. Details such as the shaded triplet similar to that in the ROA spectrum of Mg2+ -free tRNAPhe and characteristic of the C3 -endo sugar pucker, reﬂect the single-stranded A-type helical conformation of the RNA-2 packaged in the core. This work provided new information on the RNA structure of CPMV since the nucleic acid was not observed in the X-ray crystal structure [120]. Hence new information about both the protein and nucleic acid constituents of an intact virus may be deduced from ROA data!

23.5. FUTURE DIRECTIONS ROA should be particularly valuable for the determination of protein structure and function in the post-genomic era, especially for the many proteins speciﬁed by a genome, be they folded, unfolded, or partially unfolded, which are inaccessible to X-ray and NMR methods. ROA will be useful even for those proteins that do crystallize since it provides fold information, albeit not at atomic resolution. Chirality is a burgeoning topic in supramolecular chemistry and nanoscience [121]. The homochirality of biological macromolecules and their many diverse structural and functional roles provide important lessons for chiral nanoscience, which is also developing its own themes with regard to synthetic supramolecular systems and nonaqueous solvents not encountered in the chemistry of life. Applications of chiroptical spectroscopies in nanoscience are already important and will continue to grow (Chapter 24, this

787

788

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(c)

Empty protein capsid

IR + IL

(a)

5.4 × 109

A

B

IR − IL

C

0 7.0 × 105

IR + IL

Capsid containing RNA-2

6.2 × 109 A

IR − IL

(b) 0 6.9 × 105

B IR + IL

Viral RNA-2

C

IR − IL

1.6 × 109 0 4.8 × 105 800

1000

1200

1400

1600

Wavenumber (cm−1)

Figure 23.15. (a) Icosahedral capsid of cowpea mosaic virus. (b) Asymmetric unit comprising three different protein domains, each having the same fold represented as MOLSCRIPT diagrams. (c) Backscattered ICP Raman and ROA spectra measured in aqueous solution on the home-built Glasgow instrument of the empty protein capsid (top pair), the intact capsid containing RNA-2 (middle pair), and the difference spectra obtained by subtracting the top from the middle spectra to reveal the spectra of the viral RNA-2 (bottom pair). (Adapted from reference 2.)

volume). ROA is especially promising in this respect since it is able to determine the absolute handedness and conformational details of synthetic chiral macromolecules and supramolecular structures in solution by means of theoretical simulations of observed spectra, a recent example being the determination of the solution structure of a synthetic β-peptide from ab initio simulations of experimental ROA spectra [122]. There also appears to be no upper size limit to the structures that may be studied, examples being the intact viruses [58] and lysine dendrigraft generations [85] discussed earlier. Application of two-dimensional spectroscopic correlation methods, based on linear relationships between spectral data obtained under a perturbing inﬂuence such as change of temperature or pH, has the potential to increase the amount of information that may be extracted from ROA spectra. Spectral resolution is enhanced by spreading bands over

S T R U C T U R E A N D B E H AV I O R O F B I O M O L E C U L E S F R O M R A M A N O P T I C A L A C T I V I T Y

a second dimension that can provide, inter alia, unambiguous band assignments. Studies of the α-helix to β-sheet transition of poly(l-lysine) as a function of temperature [123], as well as studies of the α-helix to disordered transition in poly(l-glutamic acid) as a function of pH [124], have demonstrated the value of two-dimensional ROA measurements. Although resonance ROA is still in its infancy [125, 126], it has the potential to boost the intensity by several orders of magnitude, thereby enabling much more dilute samples to be studied. Resonance and pre-resonance ultraviolet ROA measurements on biomolecules, using UV lasers or perhaps synchrotron beams for the light source, could be especially valuable. An ICP ROA instrument with both backscattering and right-angle scattering capability using UV excitation at 244 nm from a frequency-doubled continuous argon-ion laser has recently been built in Glasgow and is starting to provide interesting data on both small chiral organic molecules and biomolecules [127]. A sufﬁcient body of experimental data has been accumulated and analyzed to demonstrate that ROA can provide new and incisive information about chiral molecular systems complementary to that obtained from other techniques. The many applications of ROA to biomolecular science that are outlined in this chapter provide the merest glimpse of what is now possible.

ACKNOWLEDGMENTS We thank the EPSRC and BBSRC for research grants, and we are grateful to the many students and collaborators who have contributed to the Glasgow biomolecular ROA program over many years.

REFERENCES 1. N. Berova, K. Nakanishi, R. W. Woody, eds., Circular Dichroism: Principles and Applications, 2nd ed., Wiley-VCH, New York, 2001. 2. L. D. Barron, Molecular Light Scattering and Optical Activity, 2nd ed., Cambridge University Press, Cambridge, 2004. 3. P. L. Polavarapu, Vibrational Spectra: Principles and Applications with Emphasis on Vibrational Optical Activity, Elsevier, Amsterdam, 1998. 4. L. A. Naﬁe, Annu. Rev. Phys. Chem. 1997, 48 , 357–386. 5. W. Hug, Handbook of Vibrational Spectroscopy, Vol. 1, J. M. Chalmers and P. R. Grifﬁths, eds., John Wiley & Sons, Chichester, 2002, pp. 745–758. 6. L. D. Barron, L. Hecht, I. H. McColl, E. W. Blanch, Mol. Phys. 2004, 102 , 731–744. 7. L. D. Barron, F. Zhu, L. Hecht, G. E. Tranter, N. W. Isaacs, J. Mol. Struct. 2007, 834–836 , 7–16. 8. L. D. Barron, A. D. Buckingham, Chem. Phys. Lett. 2010, 492 , 199–213. 9. L. D. Barron, M. P. Bogaard, A. D. Buckingham, J. Am. Chem. Soc. 1973, 95 , 603–605. 10. W. Hug, S. Kint, G. F. Bailey, J. R. Scherer, J. Am. Chem. Soc. 1975, 97 , 5589–5590. 11. L. D. Barron, L. Hecht, E. W. Blanch, A. F. Bell, Prog. Biophys. Mol. Biol . 2000, 73 , 1–49. 12. L. D. Barron, E. W. Blanch, L. Hecht, Adv. Protein Chem. 2002, 62 , 51–90. 13. L. D. Barron, E. W. Blanch, I. H. McColl, C. D. Syme, L. Hecht, K. Nielsen, Spectroscopy 2003, 17 , 101–126. 14. F. Zhu, N. W. Isaacs, L. Hecht, L. D. Barron, Structure, 2005, 13 , 1409–1419. 15. F. Zhu, N. W. Isaacs, L. Hecht, G. E. Tranter, L. D. Barron, Chirality 2006, 18 , 103–115.

789

790

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

16. L. D. Barron, Curr. Op. Struct. Biol . 2006, 16 , 638–643. 17. L. D. Barron, F. Zhu, L. Hecht, N. W. Isaacs, Methods in Protein Structure and Stability Analysis, V. N. Uversky and E. A. Permyakov, eds., Nova Science Publishers, New York, 2007, pp. 27–68. 18. P. R. Carey, Biochemical Applications of Raman and Resonance Raman Spectroscopies, Academic Press, New York, 1982. 19. T. Miura, G. J. Thomas, Jr., Subcellular Biochemistry, Vol. 24: Proteins: Structure, Function and Engineering, B. B. Biswas and S. Roy, eds., Plenum Press, New York, 1995, pp. 55–99. 20. P. W. Atkins, L. D. Barron, Mol. Phys. 1969, 16 , 453–466. 21. L. D. Barron, A. D. Buckingham, Mol. Phys. 1971, 20 , 1111–1119. 22. L. D. Barron, A. D. Buckingham, Annu. Rev. Phys. Chem. 1975, 26 , 381–396. 23. L. A. Naﬁe, T. B. Freedman, Chem. Phys. Lett. 1989, 154 , 260–266. 24. D. A. Long, The Raman Effect, John Wiley & Sons, Chichester, 2002. 25. L. D. Barron, Mol. Phys. 1976, 31 , 1929–1931. 26. L. D. Barron, A. D. Buckingham, J. Am. Chem. Soc. 1974, 96 , 4769–4773. 27. L. Hecht, L. D. Barron, W. Hug, Chem. Phys. Lett. 1989, 158 , 341–344. 28. L. Hecht, L. D. Barron, Appl. Spectrosc. 1990, 44 , 483–490. 29. J. R. Escribano, L. D. Barron, Mol. Phys. 1988, 65 , 327–344. 30. L. D. Barron, A. D. Buckingham, Acc. Chem. Res. 2001, 34 , 781–789. 31. R. D. Amos, Chem. Phys. Lett. 1982, 87 , 23–26. 32. P. K. Bose, L. D. Barron, P. L. Polavarapu, Chem. Phys. Lett. 1989, 155 , 423–429. 33. P. L. Polavarapu, J. Phys. Chem. 1990, 94 , 8106–8112. 34. J. Costante, L. Hecht, P. L. Polavarapu, A. Collet, L. D. Barron. Angew. Chem. Int. Ed. Engl . 1997, 36 , 885–887. 35. G. Zuber, W. Hug, J. Phys. Chem. A 2004, 108 , 2108–2118. 36. Dalton, a molecular electronic structure program, Release 2.0, 2005 ( http://www.kjemi.uio.no/software/dalton/dalton.html.). 37. Gaussian 09, M. J. Frisch et al., Gaussian, Inc., Wallingford CT, 2009. 38. K. Ruud, A. J. Thorvaldsen, Chirality 2009, 21 , E54–E67. 39. M. Pecul, Chirality 2009, 21 , E98–E104. 40. J. Kapit´an, C. Johannessen, P. Bouˇr, L. Hecht, L. D. Barron, Chirality 2009, 21 , E4–E12. 41. J. Kapit´an, V. Baumruk, V. Kopeck´y Jr., P. Bouˇr, J. Phys. Chem. A 2006, 110 , 4689–4696. 42. J. Kapit´an, V. Baumruk, V. Kopeck´y Jr., R. Pohl, P. Bouˇr, J. Am. Chem. Soc. 2006, 128 , 13451–13462. 43. L. Hecht, L. D. Barron, E. W. Blanch, A. F. Bell, L. A. Day, J. Raman Spectrosc. 1999, 30 , 815–825. 44. W. Hug, G. Hangartner, J. Raman Spectrosc. 1999, 30 , 841–852. 45. W. Hug, Appl. Spectrosc. 2003, 57 , 1–13. 46. A. T. Tu, Adv. Spectrosc. 1986 13 , 47–112 47. B. Davidson, G. D. Fasman, Biochemistry 1967, 6 , 1616–1629. 48. P. R. Bergethon, The Physical Basis of Biochemistry, Springer-Verlag, New York, 1998. 49. G. Wilson, L. Hecht, L. D. Barron, J. Chem. Soc. Faraday Trans. 1996, 92 , 1503–1510. 50. I. H. McColl, E. W. Blanch, A. C. Gill, A. G. O. Rhie, M. A. Ritchie, L. Hecht, K. Nielsen, L. D. Barron, J. Am. Chem. Soc. 2003, 125 , 10019–10026. 51. P. J. Kraulis, J. Appl. Crystallogr. 1991, 24 , 946–950. 52. J. Kapit´an, V. Baumruk, H. Hulanˇcov´a, P. Maloˇn, Vib. Spectrosc. 2006, 42 , 88–92.

S T R U C T U R E A N D B E H AV I O R O F B I O M O L E C U L E S F R O M R A M A N O P T I C A L A C T I V I T Y

53. P. Maloˇn, L. Bedn´arov´a, M. Straka, L. Krejˇc´ı, L. Kumprecht, T. Kraus, M. Kub´anˇ ov´a, V. Baumruk, Chirality, 2010, 22 , E47–E55. 54. I. H. McColl, E. W. Blanch, L. Hecht, L. D. Barron, J. Am. Chem. Soc. 2004, 126 , 8181–8188. 55. E. W. Blanch, L. A. Morozova-Roche, D. A. E. Cochran, A. J. Doig, L. Hecht, L. D. Barron, J. Mol. Biol . 2000, 301 , 553–563. 56. E. W. Blanch, A. F. Bell, L. Hecht, L. A. Day, L. D. Barron, J. Mol. Biol . 1999, 290 , 1–7. 57. E. W. Blanch, L. Hecht, L. A. Day, D. M. Pederson, L. D. Barron, J. Am. Chem. Soc. 2001, 123 , 4863–4864. 58. E. W. Blanch, L. Hecht, C. D. Syme, V. Volpetti, G. P. Lomonossoff, L. D. Barron, J. Gen. Virol . 2002, 83 , 2593–2600. 59. E. W. Blanch, D. J. Robinson, L. Hecht, L. D. Barron, J. Gen. Virol . 2001, 82 , 1499–1502. 60. E. W. Blanch, D. J. Robinson, L. Hecht, C. D. Syme, K. Nielsen, L. D. Barron, J. Gen. Virol . 2002, 83 , 241–246. 61. F. R. Salemme, Prog. Biophys. Mol. Biol . 1983, 42 , 95–133. 62. C. L. Nesloney, J. W. Kelly, Bioorg. Med. Chem. 1996, 4 , 739–766. 63. C. Branden, J. Tooze, Introduction to Protein Structure, 2nd ed., Garland Publishing, New York, 1999. 64. E. Vass, M. Holl´osi, F. Besson, R. Buchet, Chem. Rev . 2003, 103 , 1917–1954. 65. J. Kubelka, T. A. Keiderling, J. Am. Chem. Soc. 2001, 123 , 12048–12058. 66. C. Toniolo, M. Crisma, F. Formaggio, G. Valle, G. Cavicchioni, G. Pr´ecigoux, A. Aubry, J. Kamphuis, Biopolymers, 1993, 33 , 1061–1072. 67. C. Toniolo, F. Formaggio, S. Tognon, Q. B. Broxterman, B. Kaptein, R. Huang, V. Setnicka, T. A. Keiderling, I. H. McColl, L. Hecht, L. D. Barron, Biopolymers, 2004, 75 , 32–45. 68. J. A. Schellman, C. Schellman, The Proteins, 2nd ed. H. Neurath, ed., Academic Press, New York, 1964, pp. 1–32. 69. C. R. Jacob, S. Luber, M. Reiher, Chem.Phys.Chem. 2008, 9 , 2177–2180. 70. E. W. Blanch, L. A. Morozova-Roche, L. Hecht, W. Noppe, L. D. Barron, 2000, Biopolymers 2000, 57 , 235–248. 71. R. W. Woody, A. K. Dunker, Circular Dichroism and the Conformational Analysis of Biomolecules, G. D. Fasman, ed., Plenum Press, New York, 1996, pp. 109–157. 72. A. A. Adzhubei, M. J. E. Sternberg, J. Mol. Biol . 1993, 229 , 472–493. 73. T. P. Creamer, M. N. Campbell, Adv. Protein Chem. 2002, 62 , 263–282. 74. Z. Shi, R. W. Woody, N. R. Kallenbach, Adv. Protein Chem. 2002, 62 , 163–240. 75. B. Bochicchio, A. M. Tamburro, Chirality 2002, 14 , 782–792. 76. G. Siligardi, A. F. Drake, Biopolymers 1 995, 37 , 281–292. 77. R. K. Dukor, T. A. Keiderling, Biopolymers 1991, 31 , 1747–1761. 78. T. A. Keiderling, Q. Xu, Adv. Protein Chem. 2002, 62 , 111–161. 79. R. Schweitzer-Stenner, F. Eker, K. Griebenow, X. Cao, L. A. Naﬁe, J. Am. Chem. Soc. 2004, 126 , 2768–2776. 80. S. A. Asher, A. V. Mikhonin, S. Bykov, J. Am. Chem. Soc. 2004, 126, 8433–8440. 81. I. H. McColl, E. W. Blanch, L. Hecht, N. R. Kallenbach, L. D. Barron, J. Am. Chem. Soc. 2004, 26 , 5076–5077. 82. Z. Shi, C. A. Olsen, G. D. Rose, R. L. Baldwin, N. R. Kallenbach, Proc. Natl. Acad. Sci. USA 2002, 99 , 9190–9195. 83. F. Zhu, J. Kapit´an, G. E. Tranter, P. D. A. Pudney, N. W. Isaacs, L. Hecht, L. D. Barron, Proteins 2008, 70 , 823–833.

791

792

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

84. J. Kapit´an, D. Gallo, N. Goasdou´e, M. Nicaise, M. Desmadril, L. Hecht, G. Leclercq, L. D. Barron, Y. Jacquot, J. Pept. Sci . 2009, 15 , 455–464. 85. C. Johannessen, J. Kapit´an, H. Collet, A. Commeyras, L. Hecht, L. D. Barron, Biomacromolecules 2009, 10 , 1662–1664. 86. V. N. Uversky, C. J. Oldﬁeld, A. K. Dunker, J. Mol. Recognit. 2005, 18 , 343–384. 87. H. Xie, S. Vucetic, L. M. Iakoucheva, C. J. Oldﬁeld, A. K. Dunker, V. N. Uversky, Z. Obradovich, J. Proteome Res. 2007, 6 , 1882–1898. 88. F. Zhu, G. E. Tranter, N. W. Isaacs, L. Hecht, L. D. Barron, J. Mol. Biol . 2006, 363 , 19–26. 89. M. N. Kinalwa, E. W. Blanch, A. J. Doig, Anal. Chem. 2010, 82 , 6347–6349. 90. P. Tompa, Structure and Function of Intrinsically Disordered Proteins, CRC Press, Boca Raton, FL, 2009. 91. A. K. Dunker, I. Silman, V. N. Uversky, J. L. Sussman, Curr. Op. Struct. Biol . 2008, 18 , 756–764. 92. C. D. Syme, E. W. Blanch, C. Holt, R. Jakes, M. Goedert, L. Hecht, L., L. D. Barron, Eur. J. Biochem. 2002, 269 , 148–156. 93. E. W. Blanch, D. D. Kasarda, L. Hecht, K. Nielsen, L. D. Barron, Biochemistry 2003, 42 , 5665–5673. 94. K. W¨uthrich, R. Riek, Adv. Protein Chem. 2001, 57 , 55–82. 95. E. W. Blanch, A. C. Gill, A. G. O. Rhie, J. Hope, L. Hecht, K. Nielsen, L. D. Barron, J. Mol. Biol . 2004, 343 , 467–476. 96. G. L. Millhauser, Annu. Rev. Phys. Chem. 2006, 58 , 299–320. 97. C. M. Dobson, Seminars Cell Develop. Biol . 2004, 15 , 3–16. 98. Y. C. Sekharudu, M. Sundaralingham, Water and Biological Macromolecules: E. Westhof, ed., CRC Press, Boca Raton, Fl., 1993, pp. 148–162. 99. M. F¨andrich, V. Forge, K. Buder, M. Kittler, C. M. Dobson, S. Diekmann, Proc. Natl. Acad. Sci. USA 2003, 100 , 15463–15468. 100. F. Chiti, M. Stefani, N. Taddei, G. Ramponi, C. M. Dobson, Nature 2003, 424 , 805–808. 101. V. N. Uversky, J. R. Gillespie, A. L. Fink, Proteins 2000, 41 , 415–427. 102. S. B. Prusiner, Proc. Natl. Acad. Sci. USA 1998, 95 , 13363–13383. 103. F. Zhu, P. Davies, A. R. Thompsett, S. M. Kelly, G. E. Tranter, L. Hecht, N. W. Isaacs, D. R. Brown, L. D. Barron, Biochemistry 2008, 47 , 2510–2517. 104. Z. Q. Wen, L. D. Barron, L. Hecht, J. Am. Chem. Soc. 1993, 115 , 285–292. 105. A. F. Bell, L. D. Barron, L. Hecht, Carbohydr. Res. 1994, 257 , 11–24. 106. A. F. Bell, L. Hecht, L. D. Barron, Spectrochim. Acta A 1995, 51 , 1367–1378. 107. M. R. Wormald, A. J. Petrescu, Y.-L. Pao, A. Glithero, R. A. Dwek, Chem. Rev . 2002, 102 , 371–386. 108. A. F. Bell, L. Hecht, L. D. Barron, J. Am. Chem. Soc. 1994, 116 , 5155–5161. 109. A. F. Bell, L. Hecht, L. D. Barron, J. Raman Spectrosc. 1995, 26 , 1071–1074. 110. T. R. Rudd, R. Hussain, G. Siligardi, E. A. Yates, Chem. Commun. 2010, 46 , 4124–4126. 111. A. F. Bell, L. Hecht, L. D. Barron, Chem. Eur. J . 1997, 3 , 1292–1298. 112. J. Kapit´an, C. Johannessen, L. Hecht, L. D. Barron, in preparation. 113. F. Zhu, N. W. Isaacs, L. Hecht, L. D. Barron, J. Am. Chem. Soc. 2005, 127 , 6142–6143. 114. A. F. Bell, L. Hecht, L. D. Barron, J. Chem. Soc. Faraday Trans. 1997, 93 , 553–562. 115. A. F. Bell, L. Hecht, L. D. Barron, J. Am. Chem. Soc. 1997, 119 , 6006–6013. 116. A. F. Bell, L. Hecht, L. D. Barron, J. Am. Chem. Soc. 1998, 120 , 5820–5821. 117. W. Saenger, Principles of Nucleic Acid Structure, Springer-Verlag, New York, 1984. 118. A. J. Hobro, M. Rouhi, G. L. Conn, E. W. Blanch, Vib. Spectrosc. 2008, 48 , 37–43.

S T R U C T U R E A N D B E H AV I O R O F B I O M O L E C U L E S F R O M R A M A N O P T I C A L A C T I V I T Y

119. G. J. Thomas, Jr., Annu. Rev. Biophys. Biomol. Struct. 1999, 28 , 1–27. 120. T. Lin, Z. Chen, R. Usha, C. V. Stauffacher, J.-B. Dai, T. Schmidt, J. E. Johnson, Virology 1999, 265 , 20–34. 121. D. B. Amabilino, ed., Chirality at the Nanoscale, Wiley-VCH, Weinheim, 2009. 122. J. Kapit´an, F. Zhu, L. Hecht, J. Gardiner, D. Seebach, L. D. Barron, Ang. Chem. Int. Ed. Engl . 2008, 47 , 6392–6394. 123. L. Ashton, L. D. Barron, B. Czarnik-Matusewicz, L. Hecht, J. Hyde, E. W. Blanch, Mol. Phys. 2006, 104 , 1429–1455. 124. L. Ashton, L. D. Barron, L. Hecht, J. Hyde, E. W. Blanch, Analyst 2007, 132 , 468–479. 125. L. A. Naﬁe, Theor. Chem. Acc. 2008, 119 , 39–55. 126. L. Jensen, J. Autschbach, M. Krykunov, G. C. Schatz, J. Chem. Phys. 2007, 127 , 134101. 127. J. Kapit´an, L. D. Barron, L. Hecht, in preparation.

793

24 OPTICAL ROTATION, ELECTRONIC CIRCULAR DICHROISM, AND VIBRATIONAL CIRCULAR DICHROISM OF CARBOHYDRATES AND GLYCOCONJUGATES Tohru Taniguchi and Kenji Monde

24.1. INTRODUCTION Since the discovery of optical activity in a sucrose solution by Biot in the early nineteenth century, chiroptical spectroscopy has played a pivotal role in carbohydrate science. Basic information on the structures of carbohydrates—such as structures of the eight aldohexoses elucidated by Emil Fischer and the presence of anomeric isomers found by his pupils—was obtained by optical rotation experiments a century ago [1]. Since then, much effort has been made to correlate carbohydrate structures with optical rotation data. In the 1960s, electronic circular dichroism (ECD) was developed and was applied to a variety of carbohydrates with the aim of establishing spectra–structure relationships. However, advances in NMR gradually overshadowed these techniques in glycoscience, and little attention is now paid to these methodologies, despite the fact that chiroptical techniques have several advantages over NMR. The recent revival of vacuum–UV circular dichroism (VUVCD) applications, development of analytical methodologies using normal UV–vis CD, commercialization of vibrational circular dichroism (VCD) and Raman optical activity (ROA) instruments, and accessibility of algorithms for calculating chiroptical properties in readily available software have been generating a new trend in chiroptical analysis of carbohydrates, although these new approaches need more systematic studies. In this chapter, we brieﬂy summarize the applications of optical rotation and circular dichroism (CD) spectroscopy to structural analysis of carbohydrates and glycoconjugates with CD being divided into VUVCD, UV–vis CD, and VCD. Although we have attempted to avoid redundant description of out-of-date examples, we decided to include several old ﬁndings for each technique. This is partly because such ﬁndings are still the newest for each technique, especially for optical rotation, which has made Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

795

796

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

little progress in glycoscience in recent years, and also because there has been no review article dealing with a comparative description of optical rotation, ECD and VCD. Most importantly, some of the old approaches to analysis of sugar structure were inventive and involved sophisticated theory and comprehensive data, and we therefore hope that these examples, as well as recent reports, will encourage further applications of chiroptical spectroscopies to carbohydrates and elevate their status to a necessary tool in the future ﬁeld of glycoscience. Although the ROA technique has recently been extensively applied to carbohydrates, it is outside the scope of this chapter and readers should refer to Chapter 23, this volume.

24.2. BIOLOGICAL AND STRUCTURAL BASIS OF GLYCOCONJUGATES The biological roles of carbohydrates were believed to be merely as energy sources and structural materials until a few decades ago, but it has been discovered that they are responsible for various biological phenomenon, including immunity, cell proliferation, coagulation, cell maintenance and disease, and that a lack of some glycosyltranferases is even lethal or teratogenic [2]. Sugars on proteins and lipids can mediate intermolecular interactions that are essential in cell–small-molecule, cell–matrix, cell–cell, and cell–virus interactions (Figure 24.1a). Therefore, elucidation of the glycan structures that govern such interactions has been indispensible in the study of glycoconjugates. Glycans may also change the secondary structure of aglycans (e.g., proteins), though their role is often just stabilization of the entire glycoconjugate. In any case, it is sometimes necessary to consider the structures of glycoconjugates as an entity, since the conformations of the glycan and aglycan in a glycoconjugate can be signiﬁcantly different from those of the glycan and aglycan existing separately. Similar to glycoproteins and glycolipids, natural products can also be modiﬁed with sugars. The pharmacokinetic signiﬁcance of such glycosylation is often ambiguous, but it may change the stability, localization and target of the natural product (Figure 24.1b). The ﬁnding that glycosylation can render or alter the potency of drugs and antibiotics has increased the demand for an analytical method for such glycoconjugate entities in modern pharmaceutical science. Polysaccharides and their glycoconjugates form the cell wall and matrix (Figure 24.1c). Because their linear structures as well as conformations must be closely linked to their functions, structural analysis of these polysaccharides should be useful in, for example, development of antibiotics. Structural analysis of carbohydrates, however, has not been as well established as that of the other two classes of biomacromolecules (i.e., nucleic acids and proteins) due to their structural complexity. Figure 24.2a shows the structure of the representative carbohydrate monomer d-glucose. Methods for analyzing carbohydrates should be capable of distinguishing monosaccharides that sometimes differ only in the stereochemistry of one among four asymmetric centers, as in the case of discrimination of d-glucose and d-galactose. The variety of monosaccharides can be further widened by substitution of hydroxyl groups with acetoamide and carboxylic acid groups or reduction to methylene groups. Figure 24.2b shows the structures of monosaccharides that are commonly found in nature. When a monosaccharide is attached to an aglycan or another sugar through a glycosidic linkage, its anomeric conﬁguration is ﬁxed to either α or β. Anomeric conﬁguration regulates recognition, stability, and secondary structure; and, because of its importance, the determination and regulation of anomeric stereochemistry have been the central focuses of glycoscience. If the coupling partner is also a sugar, the partner can use

797

C A R B O H Y D R AT E S A N D G LY C O C O N J U G AT E S

(a)

(b)

cell

OH

OH HO HO

O

O

O

bacteria virus

proteins

Me HO

(c) lipopolysaccharide

O OH

naringin

HO

O

OH

OH

O

O Me OH

O

MeO

daunomycin

OH

Me

O

H

O NH OH 2

HO HO O

HO Me H2N

glycoprotein glycolipid

OH

Me

O

O

OH Cl H N

N O H

HOOC

peptidoglycan

O

HO O

OM

Cl

O

HN

cell membrane

vancomycin

O

O

N H

H N O

N H

O

Me NH

O

H

NH2 HO

OH OH

IM

Figure 24.1. The role of carbohydrates in biological systems. (a) Sugars existing on cell surfaces in the form of glycoproteins and glycoconjugates mediate various cell–cell, cell–pathogen, and cell–secreted-molecule interactions. (b) Sugars on small molecules can dramatically change their pharmacological properties. The structures of naringin (a blood-lipid-lowering agent), daunomycin (anticancer drug), and vancomycin (antibiotic) are shown. (c) Polysaccharides are the major component of the cell wall of bacteria, plants, and so on, in addition to functioning for energy storage in animals. The ﬁgure shows the cell wall structure of a gram-negative bacterium. The periplasmic space between the IM (inner membrane) and OM (outer membrane) is loosely ﬁlled with peptidoglycan, whereas the cell surface is covered with lipopolysaccharides.

any of its hydroxyl groups to bind with the anomeric carbon of the original sugar, which exponentially increases the number of possible stereoisomers. For example, there are 11 possible linking patterns for glucobiose, some of which are shown in Figure 24.2c. Moreover, the presence of multiple hydroxyl groups is the source of further complication that does not reside in nucleic acids and proteins, since it yields a branched oligosaccharide structure (e.g., sialyl Lewis X and xanthan). Therefore, the sequencing of carbohydrates should determine the types of monomers, the anomeric conﬁgurations, and the pattern of glycosidic linkages. Since oligosaccharides exist mainly as glycoconjugates, it is more desirable if the method can analyze the structure of the aglycan as well. Meanwhile, conformational studies of carbohydrates and glycoconjugates present special difﬁculties. Unlike nucleic acids and proteins, whose secondary structures can be routinely studied by electronic circular dichroism (ECD), there is no simple classiﬁcation for the secondary structures of carbohydrates, nor is there an analytical method. Consequently, analysis of the higher-order structure of a carbohydrate, such as distinction between a random coil and an ordered structure, distinction between a right-helical sense and left-helical sense, and distinction between single-stranded and double-stranded helix, has been made by using a combination of various methods. One approach to study the higher-order structure of a carbohydrate is to deduce it from conformational information at a monomer level that involves ring conformations, orientations of hydroxyl and terminal hydroxymethyl groups, and dihedral angles of the covalent bonds in glycosidic linkages. For pyranose ring conformations, the most frequently encountered are the two chair forms 4 C1 and 1 C4 (Figure 24.3a). d-sugars and l-sugars normally adopt 4 C1 and 1 C4 conformations, respectively, but exceptions are not uncommon:

798

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a)

(c) (b)

(d)

Figure 24.2. Structures of various classes of carbohydrates. (a) Structure of D-glucose in Fischer and Haworth projections. The anomeric conﬁguration of a ring form of pyranose takes either α or β. (b) structures of monosaccharides commonly found in nature. (c) Structures of disaccharides and an oligosaccharide, sialyl Lewis X. All of the disaccharides shown here are composed of two glucose units. (d) Structures of representative polysaccharides. Sialyl Lewis X and xanthan (highlighted in shadow) have a branched saccharide structure.

l-arabinose takes a 4 C1 conformation, whereas l-iduronic acid can exist as an equilibrium mixture of several conformations including 4 C1 and 1 C4 . The dihedral angles of glycosidic linkage are also very informative in modeling a higher-order structure of polysaccharides. In a glycosidic bond with a hydroxyl group on a methine carbon, two linkage dihedral angles are considered: φ = H1–C1–O1–Cn and ψ = C1–O1–Cn–Hn (Figure 24.3b). In the case in which a glycosidic linkage is formed with a terminal hydroxymethyl group, as in isomaltose and gentiobiose, one more dihedral angle, between C5 and C6, is taken into account. A comprehensive understanding of the higher-order structures of carbohydrates will require the development of a convenient and universal analytical method, an equivalent of ECD spectroscopy for nucleic acids and proteins. Considering its superior sensitivity to conﬁguration and conformation, chiroptical spectroscopy is a promising tool for carbohydrate structural analysis. However, few practical methodologies using chiroptical spectroscopy for carbohydrates have been

799

C A R B O H Y D R AT E S A N D G LY C O C O N J U G AT E S

(a)

Figure 24.3. Various conformations in sugars. (a) The 4 C1 and 1 C4 conformations of a pyranose ring are interconvertable by ring-ﬂipping. Although most sugars exclusively take either of these forms, L-iduronic acid may

(b)

exist as an equibrium mixture of these conformations. (b) Scheme of the two dihedral angles of a glycosidic linkage of cellulose.

established for the following reasons. As described above, the degree of structural complexity of a carbohydrate is much higher than that of other biomolecules. Moreover, since sugars encompass various levels of conformation, interpretation of the data obtained must be done with great care. For example, as discussed later, the chiroptical property of some polysaccharides is known to change upon complexation with certain cations. However, whether the change is due to alteration in the secondary structure or only at the monomeric level should be evaluated with other experimental results, such as measurement of the corresponding monomer in the same condition, or data obtained by NMR or molecular mechanics. A simple sugar is an assembly of hydroxyl and ether functional groups, and hence observed spectra are the result of densely overlapping electronic or vibrational transitions, thus making data interpretation arduous. The use of UV–vis CD technique for naturally occurring carbohydrates is limited because a simple carbohydrate lacks UV–chromophores, except in the vacuum UV (Figure 24.4). The poor availability of carbohydrate samples has restricted systematic measurements of carbohydrates and has thus hampered the establishment of reliable correlations of observed chiroptical data and sugar structures. Although a number of sugars are currently commercially available, it must be cautioned that a sample of an anomeric mixture is

50000

40000

C–O (n → σ*)

3600 3200 2800

−1 1800 1600 1400 1200 1000 ν[cm ]

(C–C, C–O, C–N) O–H, N–H C–H

C–N (n → σ*)

H

C=O (π → π*, n → π*)

λ[nm] 160

200 ECD

240

(C=O)

3000 VCD (hydrogen stretching)

6000

10000 VCD (mid-IR)

Figure 24.4. Electronic and vibrational transitions in carbohydrate CD spectra. A simple sugar is composed solely of hydrogen, carbon, and oxygen atoms that are connected to each other by single bonds, and therefore the accessible spectral region is limited. Additional signals can be observed for substituted sugars, containing a carboxylic acid or acetamide functional group, in the regions shown by gray bars.

800

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

not appropriate for establishing empirical spectra–structures relationships since a pair of anomeric isomers exhibits completely different chiroptical properties (vide infra). Support from organic synthetic chemists may be necessary for future studies in this ﬁeld. Due to the hydrophilicity and ﬂexibility of a carbohydrate, reliable theoretical calculation of its chiroptical property has been challenging. For instance, it has been almost impossible for measurements on a carbohydrate to avoid using a coordinating solvent (such as water), which can change spectral properties, and therefore the calculation should be capable of handling solvation effects appropriately. Meanwhile, because the orientation of hydroxyl and hydroxymethyl groups may greatly inﬂuence chiroptical data [3–6], the population of each conformer should be accurately estimated. These difﬁculties, in addition to the delayed understanding of the importance of carbohydrates in biological systems, have limited the application of chiroptical spectroscopy to carbohydrates. However, several practical methods for structural analysis of sugar have been developed, initially for optical rotation and ECD and recently for VCD, although the generality of some of these methods remains to be validated. The usefulness of density functional theory (DFT) and time-dependent DFT (TDDFT) in establishing spectra-structure relationships has been proven recently [5, 6]. In order to establish further analytical approaches based on chiroptical spectroscopies, not only do more VUVCD and VCD studies need to be conducted but also (TD)DFT calculations should shed light on old methodologies that use optical rotation and UV–vis CD. In the remainder of this chapter, we describe applications of optical rotation and CD spectroscopy to structural analysis of sugars, with the hope that this chapter will stimulate further studies in this ﬁeld.

24.3. OPTICAL ROTATION Although the relationship between magnitude of optical rotation and structure of a carbohydrate has been ambiguous, optical rotation was the most powerful technique in structural analysis of carbohydrates in the pre-NMR era. Among various empirical approaches based on optical rotation, Hudson’s rule is the most well-known and widely used. Hudson proposed that the [α]d value of a sugar ring can be approximated as the sum of the contributions from each asymmetric center [7]. According to his rule, the α and β-stereochemistries of C-1 in a d-sugar contribute dextrorotatorily and levorotatorily, respectively. In fact, an α-d-anomer normally shows more dextrorotatory power than that of the corresponding β-d-anomer (see Figure 24.5). Whiffen [8] and Brewster [9] invented advanced empirical approaches that consider the relative orientations of adjacent substituents. Brewster’s prediction model divides unsubstituted carbohydrates into several components, each of which has an empirical partial optical rotation constant. These rules were successful in predicting the optical rotation of some sugars with a higher accuracy. Yamana [10] and Lemieux [11] developed further sophisticated rules by evaluating the inﬂuence of solvation on sugar conformations. Some of these approaches were proven to be applicable also for conformational analysis of polysaccharides. For example, the [α]d values of the double helix and random coil states of ι-carrageenan (Figure 24.2d) were quantitatively predicted by combining Rees’ equation, which takes the inﬂuence of φ and ψ into account, and Brewster’s model [12]. Relying solely on a single-point value, rules for [α]D are subject to exceptions and are not widely used nowadays. Aiming at improving the accuracy of structural prediction, optical rotatory dispersion (ORD) has been examined. The ORD curves of methyl α- and β-d-glycopyranosides of several

801

C A R B O H Y D R AT E S A N D G LY C O C O N J U G AT E S

0 a

7000 −500 6000

−1000

c

d

−1500 Molar rotation

Molar rotation

5000

b

4000

3000

−2000

−2500 a

2000 b 1000

−3000 −4000

c d

−6000

e 0 −200

200 250 300 350 400 450 500 550 600 l (mμ)

−8000

200 250 300 350 400 450 500 550 600 l (mμ)

Figure 24.5. Optical rotatory dispersion curves for several methyl glycopyranosides of α- (left) and β- (right) anomers. Left: (a) methyl α-D-galactopyranoside, (b) methyl α-D-glucopyranoside, (c) methyl α-D-xylopyranoside, (d) methyl α-D-mannopyranoside, and (e) methyl α-D-arabinopyranoside. Right: (a) methyl β-D-glucopyranoside, (b) methyl β-D-glucopyranoside, (c) methyl β-D-xylopyranoside, and (d) methyl β-D-arabinopyranoside.

representative sugars are shown in Figure 24.5. These data and the measurement of other sugars led to the estimation of contributions of the stereochemistries of each asymmetric center to the optical rotations at 200 nm, as well as in the longer-wavelength region [13]. Still, most of the empirical approaches for optical rotation are valid only for sugars without UV chromophores and, moreover, the empirical values in these rules take no account of spectroscopic origin. The octant rule was applied for ORD curves of chromophoric 2-acetamide-2-deoxy sugars in order to analyze preferred orientations of the amide group [14]. Although not systematized, optical rotation was also used to study structural changes of polysaccharides. For example, a single-wavelength optical rotation measurement, in this case at 578 nm, was performed in a study of the reversible order–disorder transitions of carrageenan that accompany the sol–gel transition, which was monitored as a temperature-dependent sigmoidal curve [15]. Similarly, the dynamics of the disorder–order transition of xanthan (Figure 24.2), an extracellular bacterial polysaccharide, was analyzed by a stopped-ﬂow optical rotation measurement. Figure 24.6 shows the optical rotation values of xanthan at various temperatures in the absence and presence of KCl and shows the ﬁrst-order rate constant of the coil–helix transition upon the salt addition. This experiment revealed not only that xanthan maintains an ordered structure when K+ is present but also that the transition from coil to helix involves multiple

802

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

0

Disorder −20

Salt Jump 2

−40

1

In k1

Optical rotation (° × 103)

3

Order

−60

0

and analysis of dynamics by stopped-ﬂow optical rotation measurement. Equilibrium optical rotation

−1

values at 365 nm were plotted for xanthan (0.5% w/v) in distilled water (closed triangle) and in 0.5 M KCl aqueous solution (open triangle). The ordered

−2

conformation is stabilized at a lower temperature or in the presence of potassium cations and can be induced by rapid addition of salt (‘‘salt jump’’). The ﬁrst-order

−80

300

320 T (K)

340

Figure 24.6. Monitoring of conformational changes

rate constants for the salt-induced transition from a totally or partially disordered state to an ordered state are also shown (closed circle).

steps [16]. The formation of an inclusion complex of amylose (Figure 24.2d) with the salts of fatty acids was also studied by using ORD [17]. Despite its convenience, optical rotation is rarely used now for conformational studies of carbohydrates.

24.4. ELECTRONIC CIRCULAR DICHROISM: VUVCD Since CD signals are directly associated with sugar chromophores, the interpretation of CD data is more straightforward than that of optical rotation. However, carbohydrates were not considered as good analytes for ECD spectroscopy at ﬁrst because electronic transitions of simple carbohydrates lie below 200 nm—beyond the range of ordinary ECD spectrometers. Studies on unsubstituted sugars require a VUVCD spectrometer. The ﬁrst CD study on unsubstituted carbohydrate, reported in 1968, used an ECD spectrometer without special equipment for VUVCD and measured aqueous solutions of various monosaccharides down to 190 nm [18]. The resultant CD data presented the beginning of the ﬁrst Cotton effect between 190 and 200 nm that was positive for d-glucose and negative for d-galactose, without band extrema. Nelson and Johnson obtained CD spectra of sugars down to 165 nm by using a prototype VUVCD spectrometer, revealing that the ﬁrst extremum of an aqueous solution of d-glucose occurs at 169.5 nm [19]. In their following work on anomerically pure monosaccharides, they found that a difference in anomeric conﬁguration drastically alters the overall CD properties, emphasizing the importance of CD measurements for separate anomers. The anomeric stereochemistry of d-glucose inﬂuenced the band intensity and the position of band extrema as shown

803

C A R B O H Y D R AT E S A N D G LY C O C O N J U G AT E S

in Figure 24.7a. Meanwhile, an anomeric pair of d-galactose showed CD curves having opposite signs (Figure 24.7b) [20, 21]. The sign of the ﬁrst CD band was proposed to be related to the conformation of the C6 hydroxymethyl group, which is strongly inﬂuenced by the stereochemistry of C4 [18,20–22]. Arndt and Stevens extended the measurement down to 140 nm by using ﬁlm samples. Due to the absence of solvation effects, the spectral shapes obtained for ﬁlms are different from those for solutions, but these data were still informative in understanding the nature of sugar electronic transitions (vide infra) [23]. Matsuo and Gekko have recently been using a synchrotron-radiation VUVCD spectrophotometer for measurement of various carbohydrates down to 160 nm [22, 24]. Although these studies have demonstrated high sensitivity of VUVCD technique to the structures of carbohydrates, establishment of spectra–structure relationships awaits future extensive work. Reliable interpretation of CD data requires a more comprehensive assignment of each absorption band. Arndt and Stevens [23] deepened the understanding of transitions observed in the VUVCD region. They proposed that the lower-energy transitions observed in the 160- to 190-nm region originate from a nonbonding oxygen-centered lone-pair orbital and terminate in a valence orbital, a Rydberg orbital, or an admixture of both. Recent advances in TDDFT calculation should expand the potential of VUVCD spectroscopy. Although a TDDFT approach to CD spectra of carbohydrates has yet to be explored, theoretical calculation of methyl 2-deoxy-d-xylopyranoside indicated the involvement of a larger number of electronic transitions in sugar VUVCD spectra (unpublished). Notwithstanding the lack of a detailed understanding of sugar electronic transition, VUVCD has been useful in obtaining structural information on sugars with a more complex structure. VUVCD measurements of various d-glucobioses revealed that a difference in linking patterns alters spectral shapes [22]. In a conformational study of α1–4 glucan oligomers and amylose, comparison of their VUVCD spectra indicated that the conformation of the interior subunits is approximately the same in oligomers and in amylose [25]. Similar to the optical rotation experiments in Figure 24.6, VUVCD can be used to monitor conformational changes in polysaccharides, such as the thermally induced sol-gel transition in agarose (Figure 24.2e) [26]. The structures of glycosaminoglycans have also been studied by using VUVCD spectroscopy [24, 27]. For example, the ring

4

1

3

ΔE 0

2

−1

1

−2

ΔE 0

−3

160

180 λ (nm) (a)

200

160

180 λ (nm)

200

(b)

Figure 24.7. VUVCD spectra of D2 O solutions of (a) D-glucopyranose and (b) D-galactopyranose of the α-D-anomer (solid line), the β-D-anomer (dashed line), an anomeric equlilibrium mixture (dot-and-dashed line), and a calculated equilibrium sugar (dotted).

804

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

conformation of l-idulonic acid in dermatan sulfate (Figure 24.2) was investigated by extensive use of the octant rule for its VUVCD spectrum [27]. It should be mentioned that a VUVCD spectrum is valuable even for substituted sugars that exhibit CD bands in the normal UV region: The former can strengthen a conclusion drawn from the latter and can add extra structural information. Nevertheless, measurement in the vacuum–UV region has been uncommon because of the rarity of VUVCD spectrophotometers. However, considering that the theoretical detection limit of some commercially available state-of-the-art ECD spectrometers is beyond 170 nm, VUVCD experiments using an ordinary ECD spectrometer may become possible in the near future.

24.5. ELECTRONIC CIRCULAR DICHROISM: UV–VIS CD 24.5.1. Derivatized Sugars As discussed in the previous section, an intact unsubstituted sugar is not suitable for a UV–vis CD experiment. One approach for studying such a sample is to introduce chromophores by chemical modiﬁcation and observe CD signals that are directly related to sugar stereochemistry. Alternatively, coordinating chromophoric molecules can be added to the sugar sample, with carbohydrate chiral information being transferred to the chromophoric addend through, for example, a reversible covalent bond or hydrogen bond. A carbohydrate is optimal for these manipulations because it possesses multiple hydroxyl groups. Since the introduction of chromophores can enhance CD signals, these methodologies are also useful for sugars that originally have chromophores, such as carboxylic acid and acetamide. The use of the exciton chirality method (see Chapter 4, this volume) can facilitate the interpretation of CD spectra and also enhance the signal amplitude, in which the sign and intensity of exciton coupled CD reﬂects the relative orientation and distance of the alcohols. Observation of exciton-coupled CD signals requires the introduction of two or more chromophores into alcohols. The utility of sugar’s polyol structure for the exciton chirality method has been demonstrated by Nakanishi’s seminal work. His work showed not only that tetrakis-p-bromobenzoates of pyranosides exhibit characteristic CD curves depending on the sugar type [28] but also that modiﬁcation using two different chromophores gives rise to CD spectra that are characteristic of each substitution pattern and type of pyranose [29]. Furthermore, Nakanishi invented a microscale analytical method for determination of the sugar type and the linkage positions of oligosaccharides. This protocol involves introduction of ﬁrst chromophores to free hydroxyls, cleavage of glycosidic linkages, introduction of second chromophores to the exposed hydroxyls, CD measurement of the HPLC-puriﬁed sugars, and comparison of the data obtained with those in a CD database [30]. This procedure was successfully applied to structural elucidation of the pentasaccharide of digitonin and the tetrasaccharide of sarasinoside C1 using less than 0.3 mg of each sample [31]. Glycosylated natural products, especially ones obtained from microorganisms, can contain sugars of the opposite conﬁguration to that found in higher-order animals. Although such sugar conﬁguration can be determined by comparing the optical rotation of sugars cleaved from the aglycan with authentic data (for example, reference 32), the UV–vis CD technique is much more sensitive and informative. Kuroyanagi and coworkers determined the absolute conﬁguration of sugars in lancemasides, triterpene glycosides, by a series of acid methanolysis of sugars, per-O-p-bromobenzoylation, CD

805

C A R B O H Y D R AT E S A N D G LY C O C O N J U G AT E S

measurement and comparison of the spectra with reported data [33]. A similar procedure was employed for absolute conﬁguration determination of the constituent sugar residues of a bacterial polysaccharide obtained from Streptococcus sanguis by using perO-benzoylation instead of per-O-p-bromobenzoylation [34]. The absolute conﬁguration of the sugar part of ganefromycin α was assigned by sophisticated use of the exciton chirality method [35]. The conformation of modiﬁed sugars has been studied by UV–vis CD spectroscopy. Although their conformational information may provide an insight into the conformation of the unmodiﬁed sugars [36], it is well known that such derivatization can strikingly perturb the secondary structure [37]. Therefore, the primary objective of these studies is to elucidate the conformation of the derivatized saccharide itself. For example, the higher-order structure of 2,3,6-tris-modiﬁed amylose derivatives that were prepared for the chiral stationary phase for HPLC was examined by UV–CD spectroscopy [38]. As another example, the exciton chirality method was used to estimate the population of gg, gt, tg rotamers around the C5–C6 bond of derivatized sugars [39, 40]. Introduction of chromophores onto carbohydrates by utilizing supramolecular chemistry is more convenient, since many of these methods require only mixing a sugar sample and an achiral coordinating reagent. Any chromophoric coordinating agent would show induced CD signals whose shape is characteristic of the nature of both elements. The coordinating additives range from simple metal cations [41] to more complicated molecules such as a hydrogen-bonding oligomer that takes a chiral helical conformation upon binding with sugars [42]. Striegler and co-workers [43, 44] have developed binuclear copper (II) complexes such as 1 (Figure 24.8) that can discriminate the types of monosaccharide and disaccharide. Compound 1 adopts various binding modes depending on the nature of the sugar, thus giving rise to characteristic CD spectral shapes, including shifts in absorption extrema. A supramolecular approach can be highly versatile because the absorption and selectivity of such additives can be modulated. A recently reported compound 2 can

OMe

H N

H N N

Cu

O

2+ B

O

Cu N

HO

B OH HO

1

HN HN

NH

O HN

HN O

Et

MeO [CH2CH2O]3

O O

N

R

NH

R

HN

−

N

NH

O

N [OCH2CH2]3-OMe N Br ⊕ N

CH2OCH2CH2COO−

CH2OCH2CH2COO− CH2OCH2CH2COO−

2

O OH B−

OMe OMe

M

HN

HN

O

N ⊕ − Br

O O O

O O HO B

(HO)2B R NH

O R

OH

3

O

O R

OMe

(M = H2, Cu or Zn)

Et

−

B(OH)2

4

Figure 24.8. Examples of achiral chromophoric additives that exhibit induced CD signals upon interaction with sugars. The scheme of complexation of the diboronic acid 3 and a sugar is shown in the box.

806

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

accommodate disaccharides in the space between two terphenyl planes with a remarkable selectivity and afﬁnity that are close to those of a lectin, a sugar-binding protein, and exhibit sugar-speciﬁc CD spectra upon binding [45]. Shinkai and co-workers have been extensively using diboronic acids as an excitontype CD reporter group for sugar stereochemical information. Diboronic acid 3 reversibly forms small cyclic complexes with monosaccharides, upon which the sugar chirality is transferred to the diboronic acid as illustrated in the scheme in Figure 24.8 [46]. Some of the complexes of 3 with a set of monosaccharides and disaccharides yielded a CD couplet having various magnitudes and signs, where d-sugars, except d-galactose, presented a positive couplet, with the band extrema of the ﬁrst and second Cotton effects at 205 and 190 nm, respectively [47]. Shinkai has developed several other diboronic acids that have a larger spacer between the two boronic acid moieties. Among them, 4 forms CD-active complexes with a broad range of oligosaccharides including sialyl Lewis X (Figure 24.2c) [48]. The use of the exciton chirality method for a supramolecular approach for sugar analysis has been successful in amplifying CD signals, but the interpretation of the data still requires comprehensive understanding of the interaction mode of the sugar and chromophoric reagent. A large number of coordinating reagents have been reported, but few data have been organized for their practical use for structural analysis of carbohydrates, and most of their applications have been limited to very small sugars. Further systematic studies must be conducted to fully utilize the versatility of a supramolecular approach.

24.5.2. Underivatized Sugars and Glycoconjugates UV–vis CD is applicable to a subset of sugars that possess chromophores without chemical derivatization, since it shows Cotton effects in the UV–vis region typically originating from the n → π ∗ transition of carboxylic acid and acetamide. The interpretation of UV–vis CD data may be more straightforward because the origin of each absorption can be assigned more intuitively. The CD signal from the chromophores can reﬂect not only stereochemical information of the asymmetric carbon to which they are attached but also information on the global sugar unit. An interesting example showing the usefulness of UV–vis CD in conﬁgurational analysis of sugar is presented in Figure 24.9 [49]. This study, through the measurement of a series of uronic acid monosaccharides and polysaccharides, found the following spectra–structure relationships: (a) The n → π ∗ band centered at around 210 nm is positive for d-sugars and negative for l-sugars, (b) the signal becomes bisignate for C4 equatorial and monosignate for C4 axial, and (c) the CD spectral shape of polysaccharides is close to that predicted from a linear combination of constituent monomer spectra. This and other works [50] observed that the CD curves of uronic acids change with pH. Another interesting and important application of UV CD is the study of sialic acids, a family represented by Neu5Ac (see Figure 24.2), which exists on the nonreducing end of cell-surface glycans and governs a range of intercellular interactions. In spite of its biological importance, no convenient and universal method for determining its anomeric conﬁguration has been established, with the conventional NMR assignment method being inapplicable due to the lack of an anomeric proton. As shown in Figure 24.10, the α and β forms of sialic acid methyl esters can be distinguished by the sign of a CD band at around 220 nm ascribed to n → π ∗ transition of the carboxyl group: a negative band for α and positive for β (Figure 24.10). Through the measurement of CD spectra of other sialic acid derivatives, the positive Cotton effect at 193 nm was attributed to the acetamide group at C5 [51]. Application of chiroptical spectroscopy to furanosides

807

C A R B O H Y D R AT E S A N D G LY C O C O N J U G AT E S

methyl α-D-galacturonoside (Gal) HO

3.0

COONa

3.0

2.0

O HO

Pectin OMe

Gal

methyl β-D-mannuronoside (M) NaOOC HO HO

OH O OMe

[θ] × 10−3

1.0

2.0

M 200 220 240 260

Poly M

1.0 λ/nm

−1.0

methyl α-L-guluronoside (M)

[θ] × 10−3

HO

−2.0

−1.0

−3.0

−2.0

200 220 240 260 λ/nm Alginate

OMe OH NaOOC

P T

O OH

OH

−4.0 −5.0

G

−3.0

Poly G

−4.0

Figure 24.9. UV-CD spectra of uronic acid monosaccharides (left) and polysaccharides (right). Left: Structures and UV-CD spectra measured as the sodium salt form at pH 7 of methyl α-Dgalacturonoside (Gal), β-D-mannuronoside (M) and α-L-guluronoside (G). Right: UV-CD spectra of a low methoxypectin (Pectin), polymannuronate (Poly M), an alginate (Alginate), and polyguluronate (Poly G). The spectra of alginate exhibit various trough depths (T) and peak heights (P), depending on the composition of sugar units.

(ﬁve-membered ring sugars) has to be explored more, since the 1 H-NMR technique is also not effective for characterization of their anomeric conﬁguration. Obvious examples of furanosides for the application of UV CD are nucleosides. UV-CD measurement of the α and β anomers of four different d-pentofuranosides of adenine showed that the α and β anomers exhibit respectively positive and negative broad bands at around 260 nm originating from n → π ∗ and π → π ∗ transitions of the purine [52]. In analogy to optical rotation and VUV-CD spectroscopy, the UV–vis CD technique has been used to study the higher-order structure of polysaccharides bearing chromophores. For example, ion-induced structural changes of polysaccharides such as alginate and pectin derivatives [53] have been studied by UV-CD. UV-CD measurement of poly(l-guluronate) with various concentrations of calcium ions indicated that this polysaccharide forms a dimer in the presence of 0.5 equivalent of Ca2+ ions per carboxylate and that each strand begins to aggregate when excess Ca2+ ions are present [54]. Care must be taken before correlating observed spectral changes with higher-order conformation of a polysaccharide because even a monosaccharide can change its chiroptical property depending on pH and counterion. UV–vis CD studies of underivatized sugars with an acetamide or carboxylic acid group usually deal with weak signals, and available CD bands are limited to below 240 nm. This transparency at longer wavelengths is advantageous when the primary

808

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

10

22°C methyl Neu5Ac methyl ester

8 HO

6

AcHN

[q] × 10−3

COOMe O

HO HO

4 2 HO

0 −2

OH

190

200

210

220

230

240

nm

OH

AcHN HO HO

α

OMe

OMe O

β

COOMe

Figure 24.10. ECD spectra of methyl α-glycoside (dotted line) and methyl β-glycoside (solid line) of Neu5Ac methyl esters.

interest is the structure of the aglycan of glycoconjugate: Structural information on the chromophoric aglycan part is reﬂected in the UV–vis CD curve with little or no interference from the sugar part. For example, native insulin and sialo-oligosaccharidemodiﬁed insulins exhibited practically the same UV-CD curves, clearly indicating that the sugar part interfered neither with the observed CD curve nor the conformation of insulin (Figure 24.11a) [55]. In contrast, comparison of the CD spectra of synthetic antifreeze glycopeptides 5 and the aglycan polypeptide 6 revealed dramatic polypeptide conformational changes upon glycosylation. As shown in Figure 24.11b, 5 exhibited CD spectral patterns typical for polyproline type II helix, seemingly without major CD contribution from the sugar, whereas the CD spectrum for 6 suggested a random coil structure. This experiment, in association with CD studies of the analogues and an NMR experiment, suggested the involvement of the sugar acetamide group in the formation of an ordered polypeptide structure. Furthermore, measurements of antifreeze glycopeptides with various lengths (n = 1–7) showed that the formation of a polyproline type II helix requires at least 2 repeats of the glycosylated sequence [56]. The CD contribution of the sugar part is so subtle that little attention has been paid to it when attempting to interpret a UV-CD spectrum of a glycoconjugate. For instance, theoretical simulation of a glycoconjugate CD curve may be conducted for an in silico model compound in which the sugar part is substituted by a simpler group such as a methyl group. Recently, Ferreira and co-workers carried out TDDFT calculation of the UV-CD spectra of ﬂavonone glycosides 8 and 9 taking into account the sugar part. Several conformers that differ only in the dihedral angles φ and ψ were found to stably exist for both 8 and 9, but, as expected, the differences of the rotation around the glycosidic bond had no signiﬁcant effect on the spectra between 400 and 200 nm. Their Boltzmann-averaged theoretical CD curves showed moderate agreement with the observed data [57]. To our knowledge, this is the ﬁrst TDDFT-ECD calculation of a sugar-containing molecule.

24.6. VIBRATIONAL CIRCULAR DICHROISM The VCD technique is advantageous over other spectroscopies for carbohydrate structural studies in terms of applicability to all carbohydrates without any derivatization and the availability of plenty of signals in a wider spectral region. However, only about 30 papers on VCD applications to sugars have been reported so far (see Figure 24.12). This is partly because of the weak VCD intensity of sugars compared to other classes of molecules, with

809

C A R B O H Y D R AT E S A N D G LY C O C O N J U G AT E S

(a)

(b)

Molar ellipticity

10000 5000

0

0

−20000

−5000

−15000 200

210

220 230 240 250 λ/nm tri(Neu5Ac-Gal-GlcNAc)-insulin di(Neu5Ac-Gal-GlcNAc)-insulin insulin (wt)

d a natural AFGP (AFGP-8)

c

b synthetic AFGP (5)

−80000

180

H

Me

HO 5: R = HO

H N H

d b-O-linked derivative 7

200 220 λ/nm

O

H N

H

O Me OH HO O O OH

c polypeptide 6

−100000 a

(c)

RO Me O

H N

c

d

θ/(glyco)- −40000 peptide unit b −60000

−10000

n

OH O

AcHN

6: R = H HO 7: R = HO

OH HO O O OH

240

OH O

AcHN

O OMe

OH HO HO

b a

20000

O

O

OMe

O

Me O

O

Me O

O

O

HO HO

8 Me

O

HO

9 Me

OH

O

OH

O

Figure 24.11. Applications of UV–vis CD spectroscopy to glycoconjugates. (a) UV-CD spectra of wild-type insulin and sialo-oligosaccharide-conjugated insulins. The curves for tri- and di- (Neu5AcGal-GlcNAc)insulins are virtually superimposable. (b) UV-CD spectra of natural and synthetic AFGPs and their derivatives. (c) Structures of ﬂavonone glycosides that were investigated by TDDFT-ECD calculation.

the sensitivity of VCD spectroscopy being intrinsically low. In order to obtain a reliable VCD spectrum, 1–10 mg of a sugar sample must be dissolved in 5–200 μL of a solvent, usually water, deuterated water, or DMSO-d6 [58], which hinders its application to poorly soluble sugars. The solubility of even some of the most abundant polysaccharides, such as amylose and cellulose, may not be satisfactory for VCD measurement. In addition to the low intensity, broadened VCD bands observed for simple sugars (see Figure 24.13) have hampered reliable spectral interpretation until recently. This situation is in contrast to nucleic acids and polypeptides, whose secondary structures have been analyzed by strong and relatively sharp amide VCD patterns. In recent years, much progress has been made in the ﬁeld of sugar VCD, including the discoveries of spectra–structure relationships and applications to a wider range of carbohydrates and glycoconjugates, as shown in Figure 24.12. Figure 24.12 summarizes all of the VCD works on carbohydrates to the best of our knowledge in chronological order, highlighting some important reports. The ﬁrst VCD study on sugars was conducted for methyl α- and β-d-glucopyranosides in the C–H stretching region in the late 1970s [59]. Only three other studies have been conducted in the C–H region, and its application has been limited to monosaccharides and disaccharides [60–62]. However, this should not minimize the value of C–H VCD measurements. In this region, only a small number (almost equal to the number of C–H covalent bonds) of vibrational modes can be observed among 3N − 6 fundamental vibrational transitions, where N is the number of atoms in the molecule (see Figure 24.4); thus, assignment of each absorption can be much simpler than that in the mid-IR region. Nonetheless, the majority of carbohydrate VCD studies have focused on the mid-IR region due to the feasibility of experiments and the availability of numerous signals originating from, for example, C–C, C–O stretching and deformation. Since the ﬁrst report in 1984 [63],

810

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

The C-H region

The mid-IR region

1972

The first IR-CD measurement for liquid crystalline sample The first measurement of crystal and small chiral molecules The first measurement of sugar in the C-H region (ref 59) ref 60

1973–1974 1978 1980 1984 1986 1988 1993 1994 1996–1998

The first measurement of small chiral molecules The first measurement of sugar in the mid-IR region (ref 63) ref 64 ref 85 The first measurement of cyclodextrin (ref 79) Glycoside band (ref 70)

Commercialization of VCD instruments and development of DFT for VCD calculation

Vibrational chirality Probe (using methoxy) (ref 61) O

R

+ ν36 CH3 (∼2840 cm−1)

OMe S

−

ref 58: (a review article dedicated for both C-H and mid-IR sugar VCD)

ref 62

O

1999 2000 2002 2003 2004

ref 65, 66, 67

2005

The first measurement of polysaccharide (ref 75)

2006

ref 3

2007

axial

ref 80, 74 ref 81

ref 82, 83

ref 76, 77

2009

ref 78, 84

2010

ref 89

equatorial

δ C1-H1 ν C1-O1 (∼1145 cm4)

The film measurement of sugars (ref 72)

The beginning of extensive VCD studies on glycoconjugates (ref 86, 87) Accurate VCD prediction of sugars and glycoconjugates (ref 88, 4) Aromatic glycosides (ref 5)

2008

O

−

ref 68 ref 69, 71

O

O

O

(ref 4) O

O

Sialic acids (ref 73) OH

COOMe

HO O AcHN HO

α

OR

ν O1-CAr (∼1210 cm−1)

−

OPh

β

O

4C 1

HO

β

O OPh

α

axial

−

O

axial 1C 4

+

ν C1=O (1750–1720 cm−1)

Figure 24.12. History of carbohydrate VCD research, classiﬁed according to the C–H stretching region (left side) and the mid-IR region (right side). This ﬁgure covers all of the published sugar VCD works to the best of the authors’ knowledge, excluding VCD studies on nucleic acids and review articles that brieﬂy mention sugar VCD. ν = stretch, δ = deformation, as = antisymmetric.

mid-IR VCD spectroscopy has been applied to various classes of carbohydrates including monosaccharides [3–5, 63–73], disaccharides [66, 70, 72–74], oligosaccharides [70], polysaccharides [75–78], and cyclodextrins and their complexes [66, 72, 79–84]. Mid-IR VCD study has been extended to glycoconjugates such as glycopeptides [4], glycoproteins [85, 86], glycolipids [87], and glycosylated natural products [88, 89]. The high sensitivity of the VCD technique to even subtle structural differences in sugars has been recognized since earlier studies. Figure 24.13 shows examples of typical VCD and IR spectra of representative monosaccharides in the ﬁngerprint region. As shown in this ﬁgure, the inversion of a single stereocenter can dramatically alter the entire VCD spectrum, and thus the VCD technique can distinguish the type of sugar unit [65, 69, 70]. This trend is the same for measurements in the C–H region. Because these sugar samples are mixtures of anomeric isomers, these data are admixtures of the VCD spectra derived from both anomers. Until recently, the importance of separating each anomer had been merely recognized. In practice, the difference in anomeric conﬁguration can give rise to completely different VCD spectra (Figure 24.14a), and hence it is desirable to measure α and β anomers separately, especially in the case of small saccharides. More importantly, the measurement of a series of individual anomers led to the discovery of the ﬁrst spectra–structure relationship in sugar VCD. Namely, sugars with an axial glycosidic linkage (α-conﬁguration for d-sugars and β for l-sugars) exhibit a characteristic strong, sharp VCD peak at around 1145 cm−1 , while the VCD features of the corresponding equatorial anomers are silent in this region, regardless of the sugar type (Figure 24.14).

811

C A R B O H Y D R AT E S A N D G LY C O C O N J U G AT E S

(a)

HO HO

OH O HO

0.01 Δε 0 −0.01

(b)

HO

OH O

HO

OH

HO

(c) OH

HO HO HO

OH O

OH

OH

0.01 0.01 0.01 Δε 0 Δε 0 Δε 0 −0.01 −0.01 −0.01 L-Glc L-Gal 100 100 100 100 50 50 50 50 0 ε 0 ε 0 ε 0 ε 1500 1400 1300 1200 1100 1500 1400 1300 1200 1100 1500 1400 1300 1200 1100 1500 1400 1300 1200 1100

HO HO

(f)

O HO

0.01 Δε 0 −0.01

HO

HO HO

HO

(e)

(d)

OH O

OH

(g)

O HO HO

OH

HO HO

OH O

(h)

O

HO

OH

OH HO HO

0.01 0.01 0.01 Δε 0 Δε 0 Δε 0 −0.01 −0.01 −0.01 100 100 100 100 50 50 50 50 0 ε 0 ε 0 ε 0 ε 1500 1400 1300 1200 1100 1500 1400 1300 1200 1100 1500 1400 1300 1200 1100 1500 1400 1300 1200 1100

Figure 24.13. VCD (solid) and IR (dotted) spectra of monosaccharides in DMSO-d6 solution: (a) D- and L-glucose, (b) D- and L-galactose, (c) D-mannose, (d) D-gulose, (e) D-xylose, (f) L-arabinose, (g) D-lyxose, (h) D-ribose. All monosaccharides were measured as anomeric mixtures.

This absorption band, named the glycoside band, can be observed not only for methyl glycosides but also for other linkage types such as methylene (e.g., isomaltose), methine (e.g., trehalose, maltose, etc.), and even hydrogen [70]. These ﬁndings resulted in the development of a methodology to analyze the sugar structure of oligosaccharides and glycoconjugates without the need of a VCD database. The intense and sharp features of a glycoside band enabled VCD applications to simultaneous analysis of the sugar and aglycan parts of a glycopeptide (see Figure 24.16): quantiﬁcation of the α:β ratio in a given sample and real-time monitoring of degradation of maltohexaose by amyloglucosidase, the ﬁrst reported enzymatic reaction monitored by VCD [70]. DFT calculations (vide infra) and isotope labeling experiments characterized the vibrational mode of the glycoside band as C1–H1 deformation coupled with the C1–O1 stretching vibration, without any major involvement of other parts of the molecule (Figure 24.14f), thus corroborating its broad applicability [4]. This assignment implies that the glycoside band may not be observed in sugars connected with carbonyl or aromatic groups because these can perturb the coupling of C1–O1 stretching and C1–H1 deformation by altering the bond strength of the C1–O1 linkage. In the case of axial aromatic glycosides, on the other hand, a negative VCD peak originating from an aromatic C–O stretching vibration can be observed at around 1250 cm−1 , which was also conﬁrmed by DFT calculation [5]. The usefulness of VCD spectroscopy was also exempliﬁed in a structural study of sialic acids (cf. Figure 24.9). Although sialic acid does not exhibit a glycoside band due to the lack of an anomeric proton, the carboxylic acid group adjacent to the anomeric carbon gives rise to characteristic C=O stretching VCD bands, depending on the anomeric chirality (see Figure 24.12). Since carboxyls are rarely found in other common sugars, the C=O stretching VCD should predominantly reﬂect the anomeric conﬁguration of the sialic acid even for a more complex oligosaccharide bearing other sugars [73]. An intriguing issue is whether the stereochemical information on any desired chiral center can be singled out. This may not be realized unless VCD signals representing the stereochemistry of interest are signiﬁcantly strong, as in the case of a glycoside band, or are isolated from other vibrations, as in the case of sialic acid. A VCD study of a

812

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

(a)

Δe 0.01 0 −0.01

Axial glycoside

HO HO 0.02 (b) 0 −0.02

HO HO

(c)

0.01 0 −0.01 HO HO

(d)

0.02 0 −0.02

OH O HOOMe

OH O HOOMe

Δe 0.01 0 −0.01

HO HO 0.02 0 −0.02

HO

0.01 0 −0.01

0.02 0 −0.02

0.02 (e) 0 −0.02 HO HO −1 (cm )

HOOMe

OH O

OH O

H1

HO

O

O1 OMe

HO

O 0.02 0 −0.02

1500 1400 1300 1200 1100

C1–H1 deformation

C4

HO HO

OMe

O5 OMe

C1–O1 stretching

O HO

OMe

HO

C1 HO HO

HOOMe

OH O

(f)

HO

O

HO

Equatorial glycoside

HO HO

OMe

HO

OH O

OMe

1500 1400 1300 1200 1100

Figure 24.14. (a–e) VCD spectra of axial (left column) and equatorial (right column) anomers of monosaccharide methyl pyranosides in the mid-IR region of (a) D-glucose, (b) D-galactose, (c) D-xylose, (d) L-arabinose, and (e) 2-deoxy-D-glucose. The VCD peaks at around 1145 cm−1 characteristic to axial glycoside (glycoside band) are highlighted. (f) Schematic illustration for vibrational mode of the glycoside band.

series of methyl glycosides in the C–H stretching region indicated that such information can be extracted into an isolated region by introducing a functional group that acts as a probe. The absorption of a methyl functional group overlaps with that of sugars in the mid-IR region, but it exhibits an isolated symmetric methyl stretching absorption at around 2840 cm−1 , the sign of which solely represents the chirality of C1: positive for R-conﬁguration and negative for S -conﬁguration, regardless of axial or equatorial orientation (Figure 24.15) [61]. Since such a vibrational chirality probe is applicable to disaccharides (unpublished), it will be effective for a more complicated system. Meanwhile, it has been shown that introduction of functional groups such as acetate [67], benzoate (unpublished), and 3,5-dimethylphenylcarbamate [76, 77] can amplify the intensity of VCD signals and can generate VCD spectra characteristic of the sugar structure. The use of chemical derivatization may lead to a new analytical methodology for sugar structure by using VCD. The high structural sensitivity of VCD spectroscopy enables distinction of glycosidic linkage positions. There have been few applications of chiroptical spectroscopy to analysis of linking patterns, except for a few studies using VUVCD [22] and ROA [90], partly because of the poor availability of a set of disaccharides and oligosaccharides suitable for such experiments. Recent VCD studies for the ﬁrst time examined all of the 11 d-glucobioses in the mid-IR region [70] as well as in the C–H region [62]. The difference in linking pattern had minor effects on VCD in the former, while it caused drastic VCD changes in the latter, seemingly due to inter-residue interaction such as hydrogen bonding [62]. Current methodologies for assigning a linking site by chiroptical

813

C A R B O H Y D R AT E S A N D G LY C O C O N J U G AT E S

0.005

(a)

OH O

HO HO

HO

R

3000

2900

2800

Δε 0 −0.005

200

OH

OH O

HO HO

R OMe

HO OH O

R OMe

(d) O

HO HO

HO

0

Δε 0 −0.005

R

0

0.005

Δε 0 −0.005

HO HO

R OMe

R O

HO

2800

OH

Δε 0 −0.005

200

0

0.005

Δε 0 −0.005

0

Δε 0 −0.005

ε

200

OH

100 0

0.005

Δε 0 −0.005

Δε 0 −0.005

ε

200

OMe OH

0.005

0

Δε 0 −0.005

ε

200

3000

2900

2800

−1 Wavenumber (cm )

0

O

ε

OMe

HO HO

S

HO 100 0

0.005

Δε 0 −0.005

ε

200

OH O

HO HO

OMe S

O

HO HO

OMe

100 0.005

0

Δε 0 −0.005

HO

ε

S

HO 200 100

0.005

0

ε

O OMe

HO HO

S OMe

Δε 0 −0.005

200

100

OH

OH

200

100 0

OMe S

HO

ε

200

100 0.005

O

HO HO

100

100

(f) Me

ε

200

OMe

OH O

ε

200

0.005

(e)

2900

100

(c) HO HO

3000

100

OMe 0.005

(b)

0.005

100

ε

3000

2900

2800

0

ε

O

Me

S OH

HO

OH

−1 Wavenumber (cm )

Figure 24.15. VCD (solid line) and IR (dashed line) spectra of monosaccharide methyl pyranosides bearing C1R conﬁguration (right column) and C1S conﬁguration (right column) in the C–H stretching region of (a) D-glucose, (b) D-galactose, (c) D-mannose, (d) D-xylose, (e) L-arabinose, and (f) L-fucose. The sign of the VCD signal originating from the symmetric methyl stretching shown by arrows is characteristic of the stereochemistry of C1, regardless of the sugar conformation or axial/equatorial orientation.

techniques require a spectral database, but future development of theoretical calculation and expansion of the database should reveal a spectra–structure relationship that can determine linking site. Some of the spectra–structure relationships found in smaller saccharides have been used for studies of more complex carbohydrates and glycoconjugates. Most of the mid-IR VCD signals of common sugars are conﬁned to the region below 1500 cm−1 , whereas other biomolecules have been studied by amide VCD patterns. This feature gives VCD spectroscopy a high potential as a versatile analytical tool for both glycan and aglycan parts of glycoconjugates. Based on this idea, a simultaneous structural analysis of sugar and polypeptide structures was conducted for antifreeze glycopeptides 5. As shown in Figure 24.16, a negative glycoside band observed at around 1150 cm−1 testiﬁed to the axial stereochemistry of the sugar anomeric centers, while a negative-positive couplet in the amide I region suggested a polyproline type II helical structure of the polypeptide backbone [4]. Polypeptide and sugar structures were also examined in a study on a glycoprotein [86]. Alternatively, by focusing only on amide VCD signals, protein structural

814

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

0.1 0.05

VCD

0 −0.05

A

−0.1

0.15 0.1

ΔA (× 104) IR 1800

1700

1600

1500

1400

(cm−1)

1300

1200

1100

0.05 0

Figure 24.16. VCD and IR spectra of antifreeze glycoprotein 5. The amide VCD signals suggested a polyproline type II helical structure of the polypeptide part, while the glycoside band accurately predicted the axial anomeric conﬁguration of the sugar.

changes upon complexation with a polysaccharide can be selectively monitored. A recent study on the interaction of concanavalin A and glycosaminoglycans revealed that glycosaminoglycans, which showed essentially featureless VCD spectra in the amide I region, dramatically changed the amide VCD patterns of concanavalin A by forming a complex [76]. As discussed earlier, some polysaccharides and glycoconjugates have poor solubility in solvents that are compatible with VCD measurements. Polavarapu has developed a ﬁlm technique that can extend the applicability of VCD spectroscopy to insoluble sugars [72, 75]. A structural study for α1 -acid glycoprotein was conducted in D2 O solutions, as well as in ﬁlm samples [86]. The conformations of tris(3,5-dimethylphenylcarbamate) derivatives of amylose and cellulose were also studied both in a ﬁlm and in a solution, which indicated a solvent-induced structural change [77, 78]. The feasibility of theoretical simulation of vibrational chiroptical properties compared to that for electronic transitions has been an unrecognized advantage in VCD application to sugars. Although VCD calculation of carbohydrates has been highly challenging for the reasons discussed in the ﬁrst part of this chapter, recent studies have demonstrated its usefulness for reliable interpretation of sugar VCD. In a study to identify the vibrational mode of the glycoside band, the difﬁculties in VCD calculation were circumvented by using structurally simple, hydrophobic model compounds 10a and 10b. Compounds 10a and 10b maintain the minimum features of axial and equatorial glycosides with their ring conformation ﬁxed in 4 C1 . The observed VCD and IR spectra of these compounds showed good agreement with the calculated, as shown in Figure 24.17. Furthermore, both the theoretical and experimental spectra yielded a sharp negative peak for 10a and showed a relatively ﬂat feature for 10b in the corresponding region. Analysis of the calculated results bolstered the assignment of vibrational transition of the glycoside band [4]. Further use of DFT calculation should accelerate the ﬁnding and establishment of other spectra–structure relationships. DFT calculations were also successful in predicting the pyranose conformation and orientation of hydroxyl groups of simpler sugars [4, 5]. Moreover, the absolute conﬁguration and conformation of a few glycosylated natural and synthetic compounds have recently been studied utilizing VCD calculation [5, 6, 89]. Simulation of the VCD spectrum of a bona ﬁde sugar in a coordinating solvent is highly challenging. However, the structures of small molecules, such as structural isomerism [91], diastereoisomerism [92], and chirality [93], can be determined by DFT calculation of their VCD spectra; therefore, the success in computation of sugar VCD should open a completely new way of analyzing the structures of carbohydrates and glycoconjugates—that is, sugar type, anomeric conﬁguration, linking position, conformation, and aglycan structure.

815

C A R B O H Y D R AT E S A N D G LY C O C O N J U G AT E S

10a

10b O

O

O O

O

O

OMe

OMe

0.05

observed 10a

VCD Δε

0 −0.05

Δε

0.05

calculated 10a

0

0.05

observed 10b

VCD Δε

0 −0.05

Δε

0.05

calculated 10b

0

−0.05

−0.05 500 250

IR

500 250

observed 10a

0

0 ε

IR observed 10b

calculated 10a 1500 1400 1300 1200 1100 1000 900

Wavenumber (cm−1)

500 0 ε

ε

calculated 10b 1500 1400 1300 1200 1100 1000 900

500 0 ε

Wavenumber (cm−1)

Figure 24.17. DFT calculation of sugar model compounds. Structures and the most stable conformers of each compound are shown above the spectra. Compound 10a exhibited a negative sharp peak, as highlighted, while 10b showed a rather featureless spectrum in the corresponding region. These spectra were measured in CDCl3 .

24.7. CONCLUDING REMARKS The application of chiroptical spectroscopies to carbohydrates has been limited, despite their usefulness for structural analysis of sugars. As discussed in this chapter, chiroptical techniques can solve various levels of structural complexities posed by carbohydrates and glycoconjugates, including sugar type, anomeric conﬁguration, linkage position, conformation, and aglycan structure. More importantly, some analytical methods using optical rotation and circular dichroism can provide stereochemical information that is difﬁcult to obtain by using other techniques such as NMR. Still, there is a long list of problems regarding sugar structure that remain to be examined: for example, the establishment of a generalized analytical method for the structure of furanose, anomeric conﬁguration of sialic acid, anomeric conﬁguration of 2-equatorial- and 2-deoxy-pyranose, chirality of an exocyclic asymmetric center, higher-order structure of polysaccharides, and structure of the entire glycoconjugate. Future chiroptical studies on sugars using a further systematic approach and theoretical calculation should lead to the development of a practical chiroptical methodology for sugar structure.

REFERENCES 1. A. Varki, N. Sharon, Essentials of Glycobiology, 2nd ed., A. Varki, R. D. Cummings, J. D. Esko, H. H. Freeze, P. Stanley, C. Bertozzi, G. W. Hart, and M. E. Etzler, eds., Cold Spring Harbor Laboratory Press, New York 2008, pp 75–88. 2. R. V. Stick, Carbohydrates: The Sweet Molecules of Life, Academic Press, London, 2001. 3. N. Miura, T. Taniguchi, K. Monde, S.-I. Nishimura, Chem. Phys. Lett. 2006, 419 , 326–332. 4. T. Taniguchi, K. Monde, Chem. Asian J . 2007, 2 , 1258–1266. 5. T. Taniguchi, I. Tone, K. Monde, Chirality 2008, 20 , 446–453.

816

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

6. O. Michalski, W. Kisiel, K. Michalska, V. Setnicka, M. Urbanova, J. Mol. Struct. 2007, 871 , 67–72. 7. C. S. Hudson, J. Am. Chem. Soc. 1909, 31 , 66–86. 8. D. H. Whiffen, Chem. Ind . 1956, 964–968. 9. J. H. Brewster, J. Am. Chem. Soc. 1959, 81 , 5483–5493. 10. S. Yamana, Bull. Chem. Soc. Jpn. 1958, 31 , 558–564. 11. R. U. Lemieux, J. C. Marin, Carbohydr. Res. 1970, 13 , 139–161. 12. D. A. Rees, W. E. Scott, F. B. Williamson, Nature 1970, 227 , 390–392. 13. I. Listowsky, G. Avigad, S. Englard, J. Am. Chem. Soc. 1965, 87 , 1765–1771. 14. S. Beychok, E. A. Kabat, Biochemistry 1965, 4 , 2565–2574. 15. D. A. Rees, I. W. Steele, F. B. Williamson, J. Polym. Sci., Part C 1969, 28 , 261–276. 16. I. T. Norton, D. M. Goodall, E. R. Morris, D. A. Rees, J. Chem. Soc. Chem. Commun. 1980, 545–547. 17. P. V. Bulpin, E. J. Welsh, E. R. Morris, Starch 1982, 34 , 335–339. 18. I. Listowsky, S. Inglard, Biochem. Biophys. Res. Commun. 1968, 30 , 329–332. 19. R. G. Nelson, W. C. Johnson, Jr., J. Am. Chem. Soc. 1972, 94 , 3343–3345. 20. R. G. Nelson, W. C. Johnson, Jr., J. Am. Chem. Soc. 1976, 98 , 4290–4295. 21. R. G. Nelson, W. C. Johnson, Jr., J. Am. Chem. Soc. 1976, 98 , 4296–4301. 22. K. Matsuo, K. Gekko, Carbohydr. Res. 2004, 339 , 591–597. 23. E. R. Arndt, E. S. Stevens, J. Am. Chem. Soc. 1993, 115 , 7849–7853. 24. K. Matsuo, H. Namatame, M. Taniguchi, K. Gekko, Biosci. Biotechnol. Biochem. 2009, 73 , 557–561. 25. D. G. Lewis, W. C. Johnson, Jr., Biopolymers 1978, 17 , 1439–1449. 26. J. N. Liang, E. S. Stevens, E. R. Morris, D. A. Rees, Biopolymers 1979, 18 , 327–333. 27. D. G. Cziner, E. S. Stevens, E. R. Morris, D. A. Rees, J. Am. Chem. Soc. 1986, 108 , 3790–3795. 28. H.-W. Liu, K. Nakanishi, J. Am. Chem. Soc. 1981, 103 , 5591–5593. 29. W. T. Wiesler, J. T. V´azquez, K. Nakanishi, J. Am. Chem. Soc. 1987, 109 , 5586–5592. 30. M. Ojika, H. V. Meyer, M. Chang, K. Nakanishi, J. Am. Chem. Soc. 1989, 111 , 8944–8946. 31. W. T. Wiesler, N. Berova, M. Ojika, H. V. Meyers, M. Chang, P. Zhou, L.-C. Lo, M. Niwa, R. Takeda, K. Nakanishi, Helv. Chim. Acta 1990, 73 , 509–551. 32. M. D. Lee, T. S. Dunne, C. C. Chang, M. M. Siegel, G. O. Morton, G. A. Ellestad, W. J. McGahren, D. B. Borders, J. Am. Chem. Soc. 1992, 114 , 985–997. 33. M. Ushijima, N. Komoto, Y. Sugizono, I. Mizuno, M. Sumihiro, M. Ichikawa, M. Hayama, N. Kawahara, T. Nakane, O. Shirota, S. Sekita, M. Kuroyanagi, Chem. Pharm. Bull . 2008, 56 , 308–314. 34. K. Kumaralal, B. C. Allen, Anal. Biochem. 1989, 179 , 209–215. 35. T. Anderson, N. Berova, K. Nakanishi, G. T. Carter, J. Am. Chem. Soc. 2000, 2 , 919–922. 36. H. Bittiger, G. Keilich, Biopolymers 1969, 7 , 539–556. 37. V. S. R. Rao, P. K. Qasba, P. V. Balaji, R. Chandrasekaran, Conformation of Carbohydrates, Harwood Academic Publishers, Netherlands, 1998. 38. J. Shen, T. Ikai, Y. Okamoto, J. Chromatogr. A 2010, 1217 , 1041–1047. 39. J. I. Padr´on, E. Q. Morales, J. T. V´azquez, J. Org. Chem. 1998, 63 , 8247–8252. 40. A. Ro¨en, J. I. Padr´on, C. Mayato, J. T. V´azquez, J. Org. Chem. 2008, 73 , 3351–3363. 41. A. Mukhopadhyay, E. Kolehmainen, C. P. Rao, Carbohydr. Res. 2000, 328 , 103–113. 42. M. Inouye, M. Waki, H. Abe, J. Am. Chem. Soc. 2004, 126 , 2022–2027. 43. S. Striegler, M. Dittel, J. Am. Chem. Soc. 2003, 125 , 11518–11524.

C A R B O H Y D R AT E S A N D G LY C O C O N J U G AT E S

44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80.

S. Striegler, M. G. Gichinga, Chem. Commun. 2008, 5930–5932. Y. Ferrand, M. P. Crump, A. P. Davis, Science, 2007, 318 , 619–622. K. Tsukagoshi, S. Shinkai, J. Org. Chem. 1991, 56 , 4089–4091. Y. Shiomi, M. Saisho, K. Tsukagoshi, S. Shinkai, J. Chem. Soc. Perkin Trans. I 1993, 1 , 2111–2117. O. Hirata, Y. Kubo, M. Takeuchi, S. Shinkai, Tetrahedron 2004, 60 , 11211–11218. E. R. Morris, D. A. Rees, G. R. Sanderson, D. Thom, J. Chem. Soc. Perkin Trans. II 1975, 1418–1425. L. A. Bufﬁngton, E. S. Pysh, B. Chakrabarti, E. A. Balazs, J. Am. Chem. Soc. 1977, 99 , 1730–1734. H. Ogura, K. Furuhata, Tetrahedron Lett. 1981, 22 , 4265–4268. J. S. Ingwall, J. Am. Chem. Soc. 1972, 94 , 5487–5495. D. Thom, G. T. Grant, E. R. Morris, D. A. Rees, Carbohydr. Res. 1982, 100 , 29–42. E. D. Morris, D. A. Rees, D. Thom, Carbohydr. Res. 1978, 66 , 145–154. M. Sato, T. Furuike, R. Sadamoto, N. Fujitani, T. Nakahara, K. Niikura, K. Monde, H. Kondo, S.-I. Nishimura, J. Am. Chem. Soc. 2004, 126 , 14013–14022. Y. Tachibana, G. L. Fletcher, N. Fujitani, S. Tsuda, K. Monde, S.-I. Nishimura, Angew. Chem. Int. Ed . 2004, 43 , 856–862. Y. Ding, X.-C. Li, D. Ferreira, J. Nat. Prod . 2009, 72 , 327–335. T. Taniguchi, K. Monde, Trends Glycosci. Glycotech. 2007, 19 , 147–164. C. Marcott, H. A. Havel, J. Overend, A. Moscowitz, J. Am. Chem. Soc. 1978, 100 , 7088–7089. M. G. Paterlini, T. B. Freedman, L. A. Naﬁe, J. Am. Chem. Soc. 1986, 108 , 1389–1397. T. Taniguchi, K. Monde, N. Miura, S.-I. Nishimura, Tetrahedron Lett. 2004, 45 , 8451–8453. T. Taniguchi, K. Monde, Org. Biomol. Chem. 2007, 5 , 1104–1110. D. M. Back, P. L. Polavarapu, Carbohydr. Res. 1984, 133 , 163–167. C. M. Tummalapalli, D. M. Back, P. L. Polavarapu, J. Chem. Soc. Faraday Trans. I 1988, 84 , 2585–2594. P. K. Bose, P. L. Polavarapu, Carbohydr. Res. 1999, 319 , 172–183. P. K. Bose, P. L. Polavarapu, J. Am. Chem. Soc. 1999, 121 , 6094–6095. P. K. Bose, P. L. Polavarapu, Carbohydr. Res. 1999, 322 , 135–141. M. W. Ellzy, J. O. Jensen, H. F. Hameka, J. G. Kay, Spectrochim. Acta A 2003, 59 , 2619–2633. T. Taniguchi, N. Miura, S.-I. Nishimura, K. Monde, Mol. Nutr. Food Res. 2004, 48 , 246–254. K. Monde, T. Taniguchi, N. Miura, S.-I. Nishimura, J. Am. Chem. Soc. 2004, 126 , 9496–9497. A. Synytsya, M. Urbanov´a, V. Setnicka, M. Tkadlecov´a, J. Havl´ıcek, I. Raich, P. Matejka, A. Synytsya, A. J. Cop´ıcov´a, K. Volka, Carbohydr. Res. 2004, 339 , 2391–2405. A. G. Petrovic, P. K. Bose, P. L. Polavarapu, Carbohydr. Res. 2004, 339 , 2713–2720. A. Nakahashi, T. Tanigcuhi, N. Miura, K. Monde, Org. Lett. 2007, 9 , 4741–4744. P. L. Polavarapu, C. Zhao, Fresenius J. Anal. Chem. 2000, 366 , 727–734. G. Shanmugam, P. L. Polavarapu, Appl. Spectrosc. 2005, 59 , 673–681. T. R. Rudd, R. J. Nichols, E. A. Yates, J. Am. Chem. Soc. 2008, 130 , 2138–2139. S. Ma, S. Shen, H. Lee, N. Yee, C. Senanayake, L. A. Naﬁe, N. Grinberg, Tetrahedron Asymm. 2008, 19 , 2111–2114. S. Ma, S. Shen, H. Lee, M. Eriksson, X. Zeng, J. Xu, K. Fandrick, N. Yee, C. Senanayake, N. Grinberg, J. Chromotogr. A 2009, 1216 , 3784–3793. G.-C. Chen, P. L. Polavarapu, S. Weibel, Appl. Spectrosc. 1994, 48 , 1218–1223. P. K. Bose, P. L. Polavarapu, Carbohydr. Res. 2000, 323 , 63–72.

817

818

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93.

V. Setnicka, M. Urbanov´a, V. Kr´al, K. Volka, Spectrochim. Acta A 2002, 58 , 2983–2989. P. Zhang, P. L. Polavarapu, J. Phys. Chem. A 2007, 111 , 858–871. I. Goncharova, M. Urbanov´a, Tetrahedron Asymm. 2007, 18 , 2061–2068. I. Goncharova, M. Urbanov´a, Anal. Biochem. 2009, 392 , 28–36. M. Urbanova, P. Pancoska, T. A. Keiderling, Biochim. Biophys. Acta 1993, 1203 , 290–294. G. Shanmugam, P. L. Polavarapu, Proteins 2006, 63 , 768–776. K. Monde, N. Miura, M. Hashimoto, T. Taniguchi, T. Inabe, J. Am. Chem. Soc. 2006, 128 , 6000–6001. O. Michalski, W. Kisiel, K. Michalska, V. Setnicka, M. Urbanova, J. Mol. Struct. 2007, 871 , 67–72. G. Yang, H. Tran, E. Fan, W. Shi, T. L. Lowary, Y. Xu, Chirality 2010, 22 , 734–743. A. F. Bell, L. Hecht, L. D. Barron, J. Am. Chem. Soc. 1994, 116 , 5155–5161. H. M. Min, M. Aye, T. Taniguchi, N. Miura, K. Monde, K. Ohzawa, T. Nikai, M. Niwa, Y. Takaya, Tetrahedron Lett. 2007, 48 , 6155–6158. K. Monde, T. Taniguchi, N. Miura, C. S. Vairappan, M. Suzuki, Tetrahedron Lett. 2006, 47 , 4389–4392. L. A. Naﬁe, Nat. Prod. Commun. 2008, 3 , 451–466.

25 ELECTRONIC CIRCULAR DICHROISM IN DRUG DISCOVERY Carlo Bertucci and Marco Pistolozzi

Determination of the secondary structure of peptides and proteins in solution is one of the main applications of electronic circular dichroism (CD)1 . In particular, CD has been successfully used since the late 1960s to investigate the relationship between structure and function. Different secondary structures show speciﬁc CD spectra (Figure 25.1). The observed CD spectrum of a protein can be analyzed to obtain the percentages of the secondary structure components by correlation with a set of CD spectra with known structure, as determined by X-ray analysis of protein crystals. Several empirical algorithms have been developed, and all these methods assume linear independence and add different secondary structure contributions to produce the observed spectrum [2–11]. The data obtained are reasonably accurate, but the reference database is the critical variable affecting the results of the analysis. However, the most important application of CD in protein analysis is currently the monitoring of a change in secondary structure caused by its surroundings or its interaction with other biomolecules. Investigation of this aspect is fundamental since protein activity is related to its secondary and tertiary structure, and conformational transitions of proteins are often the switch for physiological or pathological processes [12]. The possibility of monitoring changes in the CD spectra due to conformational protein transitions offers a powerful tool to follow the dynamic behavior of protein structure. This is particularly relevant in drug discovery and development areas because CD can yield fundamental information on the biochemical pathways, thereby serving to select and validate targets, and then pursue the rational design of new drugs. Complex

1

CD is the abbreviation used in this chapter for electronic circular dichroism.

Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

819

820

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

80000

60000

[θ] (deg cm2 dmol –1)

40000

20000

0

Figure 25.1. Far UV CD spectra associated

–20000

with various types of secondary structure. Solid line, α-helix; long dashed line, antiparallel β-sheet; dotted line, type I

–40000 160

180

200 220 Wavelength (nm)

240

260

β-turn; cross dashed line, extended 310 -helix or poly (Pro) II helix; short dashed line, irregular structure. (Reproduced with permission from reference 1.)

networks of interactions between proteins likely regulate biological processes. Life science is now oriented to studying the full complement of proteins involved in speciﬁc biological processes, rather than describing each individual protein. Modulation of the protein–protein interaction and protein conformation is now recognized as a valuable strategy in drug discovery. Conformational changes quite often trigger the process so that monitoring and modulation contribute to a mechanistic understanding of disease pathophysiology, as well as to the identiﬁcation of treatable targets. Furthermore, CD is becoming very useful for the structural characterization of biotechnological drugs, namely recombinant therapeutic proteins. Information on the secondary structure variation and stability of these proteins can be easily and reliably obtained. This stereochemical characterization is essential for the quality control of peptide and protein therapeutics, and it contributes to the correct formulation of these drugs to guarantee stability and in vivo half-life. All this has revived the use of CD in structural biology studies, with a special focus on monitoring the interaction between proteins and the modulation of this process when the complex formation is essential for physiological and pathological processes. Information on molecular recognition processes, both drug–protein and protein–protein interactions, can be obtained by several techniques, including X-ray crystallography, mass spectrometry, ﬂuorescence and NMR spectroscopy, and surface plasmon resonance-based optical biosensors. The use of independent techniques allows a better characterization of the structure and a more accurate picture of the dynamics of the recognition process. CD can yield deeper insights into the stereochemistry of the protein-bound

E L E C T R O N I C C I R C U L A R D I C H R O I S M I N D R U G D I S C O V E RY

drug and on protein folding, thus giving speciﬁc information on the binding mechanism. Spectroscopic investigation of drug–protein complexes—and, in particular, CD investigation—is invaluable not only to obtain a complete picture of the binding process and mechanism, but also for the quantitative characterization of the binding process at equilibrium (i.e., binding parameters) without interfering with the equilibrium process. Thus many applications of CD are relevant in the area of drug discovery and development. Among these, determination of a drug’s absolute conﬁguration, essential for a reliable study on the relationship between stereochemistry and pharmacological activity, and the binding of drugs to serum proteins, as fundamental parameters for the evaluation of drug bioavailability, are widely discussed in other chapters of this book. The use of CD for secondary structure analysis of proteins is also discussed in several chapters of this book. Here the use of CD spectroscopy is discussed from the standpoint of gaining information on ligand–protein and protein–protein interactions, and monitoring and characterizing protein conformational transition of physiological or pathological interest.

25.1. PROTEIN–LIGAND INTERACTIONS 25.1.1. Monitoring Ligand Binding to Proteins Circular dichroism can be successfully used to detect ligand–protein interactions and to measure the thermodynamic equilibrium constants associated with the interaction of the two molecules (often referred to as binding constants or afﬁnity constants). Thus CD can be an invaluable aid in drug discovery, allowing the characterization of new targets and the discovery and characterization of new drugs. Obviously, to use this technique, the interaction of the molecules must generate a change in the circular dichroism signal. In other words, an induced circular dichroism (ICD) must arise from the binding of the two molecules. In modern drug discovery, at least one of these molecules is usually a protein or a peptide, while the other can be a small molecule, a peptide, or a protein. Experimentally, measurement of the binding constants involves monitoring the ICD signal while titrating one molecule with increasing concentrations of the other. The change in the ICD signal reﬂects the association of the two molecules to make a complex and the maximum intensity is reached when the whole titrand (the molecule being titrated) is bound to the titrant (the molecule added in increasing concentrations). The titration method was widely treated by Drake, Siligardi, and co-workers [13]. Another approach to measure binding constants is to evaluate the effect of dilution on the ICD signal of a solution of the complex at ﬁxed stoichiometry [14]. In this case the change in the CD signal represents the dissociation of the complex; hence, the more sensitive the complex is to dilution, the weaker the binding [15]. According to the traditional biochemical point of view, when two molecules bind each other, the larger of the two is normally called the host, while the smaller is called the guest or ligand . Although this host –guest model is very common in literature, the choice of “who binds who” is only a point of view and is mathematically arbitrary. In CD experiments, the choice of which molecule is to be treated as a host and which is to be treated as a guest usually depends on spectroscopic limitations. In this chapter the molecule being monitored will be referred to as the guest. Depending on the nature of the interaction and the type of perturbation involved, the ICD can occur at different wavelengths. Usually, while studying protein–protein or

821

822

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

protein–drug interactions, three different regions of the UV–vis spectrum are considered: the far-UV (180–240 nm), the near-UV (240–320 nm), and the spectral region ranging from 320 nm to the end of the visible spectrum (∼780 nm). If the binding involves a change in the secondary structure of a peptide or a protein, the ICD occurs in the far-UV, where the amide backbone absorbs light. Special designed instruments [16] and synchrotron radiation CD spectrometers [17–19] measure CD spectra in aqueous solution down to 168 nm. The extended wavelength range gives further information, thus improving the reliability of the protein secondary structure prediction [20–27]. If the binding causes a perturbation of aromatic amino acids, either directly (the binding involves the interaction of the guest with the aromatic side chains of the host) or because of small rearrangements of protein tertiary structure triggered by the binding, the ICD occurs in the near-UV and reﬂects changes in the environment surrounding the aromatic and disulﬁde chromophores. If the binding involves the stabilization of a particular conformation of other chromophores (this is common when the binding of small molecules to proteins is studied), an ICD occurs at the wavelengths absorbed by those chromophores. The change in the secondary structure of a protein or a peptide triggered by the binding of ligands is easily detected by circular dichroism. If the ligand is a peptide or another protein, the change in secondary structure can involve both ligand and host simultaneously, driven by a mutual adaptation. Since the CD signal results from the linear combination of the contributions of the single structures, if no change in the secondary structure occurs, the arithmetic sum of the CD spectra of the two molecules is equal to the CD spectrum of a mixture of the two ligands, recorded in the same conditions (i.e., concentration, pH, temperature, solvent, etc.). Conversely, when the interaction of the two ligands causes a change in the secondary structure, the CD spectrum of the complex differs from the linear combination of the CD spectra of the two ligands when alone in solution. The difference between the observed spectrum and the calculated one reﬂects the secondary structure induced by the interaction. The usefulness of far-UV CD to monitor conformational shifts upon binding of biomolecules is well known. Staphylococcal enterotoxin B (SEB) is an exotoxin produced by Staphylococcus aureus and is commonly associated with food poisoning. The interaction of SEB with peptides selected among a phage-displayed peptide library was demonstrated by the induction of ordered structure monitored by CD [28]. In this case the SEB-binding peptides showed a nonordered conformation when free in solution. The addition of SEB caused a change in overall peptide structure, resulting in a more ordered secondary structure (Figure 25.2). HLP-2 and HLP-6 are two synthetic peptides designed on the basis of a helical region on helix 1 of human lactoferrin HLP-2 exhibiting bactericidal activity against Escherichia coli serotype O111. Far-UV CD studies have helped to elucidate the structure and organization of peptides when interacting with E. coli lipopolysaccharide (LPS). In the presence of LPS, HLP-2 and HLP-6 were found to bind and adopt a β-strand conformation rather than an α-helix. Furthermore, this assay was used to show that there is a time-dependent association of peptide that results in an ordered formation of peptide aggregates. The rate of interpeptide association was far greater in HLP-2 LPS than in HLP-6 LPS, which was consistent with the lag phase observed on the killing curves. On the basis of these results, it was proposed that HLP-2 folds and self-assembles at the outer membrane surface before exerting its activity [29].

823

E L E C T R O N I C C I R C U L A R D I C H R O I S M I N D R U G D I S C O V E RY

(a)

(b)

1

15

0

SEB

10

SEB-peptide 3 complex

–1 5

–3

[θ]m (mdeg)

[θ]m (mdeg)

–2 Peptide 1

–4

Peptide 2

–5

Peptide 3

–6

Peptide 5

–5

–10 –15

–7 –8 195

Peptide 3

0

205

215

225

235

245

–20 195

205

Wavelength (nm)

215

225

235

245

Wavelength (nm)

(c) 20 15

Peptide 1 Peptide 2

[θ]m (mdeg)

10

Peptide 3 Peptide 5

5 0 –5

–10 195

205

215

225

235

245

Wavelength (nm)

Figure 25.2. CD spectra of free and bound peptides and Staphylococcal enterotoxin B (SEB). (a) Spectra of free peptides (50 μM) in 50 mM PBS at pH 7.4. (b) Spectra of free peptide3 (25 μM), free SEB (25 μM), and 25 μM SEB incubated with 25 μM peptide3 in 50 mM PBS at pH 7.4. (c) Spectra of bound peptides obtained by subtracting the SEB spectrum from that of SEB–peptide complex. (Reproduced with permission from reference 28.)

An interesting example of the application of CD spectroscopy in ligand–protein binding studies is the demonstration of the antitumor drug paclitaxel binding to Bcl-2, an anti-apoptosis protein [30]. The secondary structure change occurring when the drug binds to the protein (Figure 25.3) was proved to involve 10–12 amino acids in a loop region by analyzing a series of small peptides, thus selecting this unfolded loop as a potential target for a rational drug design of nontaxoid compounds. A random library of phage-displayed peptides was screened for binding to a biotinylated derivative of paclitaxel. A subset of the peptides was identiﬁed exhibiting signiﬁcant similarity to a nonconserved region of the anti-apoptotic human protein Bcl-2. ELISA assays conﬁrmed the binding of paclitaxel to Bcl-2, and CD analysis showed a conformational change in the protein upon drug binding [30]. Subsequently [31], this interaction was further characterized by optical biosensor surface plasmon resonance (SPR) measurements, and the role of paclitaxel in modulating the Bcl-2 activity demonstrated. It is worth stressing that the absence of differences between the linear combination of the CD spectra of the ligands and the CD spectrum of the mixture does not demonstrate that the binding does not occur, but that the binding does not induce a major change

824

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

20

Ellipticity

12

4

–4

–12

Figure 25.3. Circular dichroism spectrum of

200

220

240

Wavelength (nm)

260

human Bcl-2/GST fusion protein with (continuous line) and without (broken line) paclitaxel. (Reproduced with permission from reference 30.)

in the overall secondary structure. As an example, the observed CD spectrum of the mixture of the heat shock protein Hsp90 (a chaperone involved in many pathological and physiological pathways) and its co-chaperone Cpr6 is identical to the linear combination of the HSP90 and Cpr6 spectra alone, indicating no change in the overall secondary structure of either protein on binding [32]. Instead, the observed CD spectrum of the complex of Hsp90 and another co-chaperone, named Sti1, cannot be reproduced by linear combination of the single spectra of HSP90 and Sti1, indicating a change in the secondary structure of one or both proteins driven by the binding process (Figure 25.4). Two main mechanisms occur in the interaction phenomena: induced-ﬁt and rigid-body [33]. Induced-ﬁt interactions are characterized by a change in the structure of one or both components when they bind together. By contrast, rigid-body interactions can occur without either secondary or tertiary structural changes in either component. These two interactions can produce changes in the local environments for residues lying at the interface. Although the far-UV CD discloses the binding of proteins and peptides, in many cases the binding of ligands (either small molecules or small peptides) to proteins does not involve a major change in the secondary structure of the protein. More often, ICD signals are detectable in the near-UV region, where the aromatic amino acids, mainly tyrosines and tryptophans, absorb light. This region is more sensitive to local changes in conformation and hence often represents a more suitable monitor feature to study binding phenomena, in particular for the determination of binding constants [34–36].

25.1.2. Measurement of Binding Constant by Titration Experiments The guest should be selected in order to detect a CD signal clear of the host and any other absorbing species that, albeit not yielding any CD signal, can add noise due to strong absorption in the selected monitoring wavelength range. In the most common titration experimental setup, increasing concentrations of the host (the titrant) are added to a constant concentration of the guest (the titrand) until the entire guest is bound. To obtain reliable data, it is crucial to reach guest saturation since the ﬁnal values are needed for the nonlinear regression to be computed.

825

E L E C T R O N I C C I R C U L A R D I C H R O I S M I N D R U G D I S C O V E RY

ΔA = AL – AR (x 10–5)

0

H –20

–40

C CALC

–60

OBS

200

240

220

Wavelength (nm) (a)

ΔA = AL – AR (x 10–5)

0

S

Figure 25.4. Far-UV CD spectra for Hsp90–co-chaperone interactions. (a) Far-UV spectra for Hsp90, H; Cpr6, C; Hsp90 1 Cpr6 observed spectrum, OBS; and linear combination of the separate spectra for Hsp90 and Cpr6 in

–20 H

the same proportion as the molar ratio of Hsp90 and Cpr6 in the complex, CALC. The observed and calculated spectra for the complexes are essentially identical, indicating no

CALC –40

200

OBS

change in the overall secondary structure of either protein on binding. (b) Same as (a) but for Sti1, S. The observed spectrum for the complex cannot be reproduced by linear

240

combination of the separate spectra for Hsp90 and Sti1, indicating a net change in the overall secondary structure of one or both proteins on binding. (Reproduced with

220 Wavelength (nm) (b)

permission from reference 32.)

Considering the binding of a single guest (G) to a single host (H) at equilibrium, −→G–H (complex) G + H←− the CD, expressed in A, at any wavelength is proportional to the concentration of the bound and unbound species of the host (H) and the guest (G) components. A = AH + AG + AHG .

(25.1)

826

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Based on Lambert–Beer’s Law, we have A = ε · b · c, where ε is the molar ellipticity, b is the pathlength (cm), and c is the concentration (M). Considering a titration with a concentration at equilibrium of the host [H], guest [G], and host–guest complex [HG], the CD of each species can be described as follows: AH = εH · b · [H ],

(25.2)

AG = εG · b · [G],

(25.3)

AHG = εHG · b · [HG].

(25.4)

Substituting Eqs. (25.2)–(25.4) with Eq. (25.1), we obtain AHG = {εHG · [HG] + εH · [H ] + εG · [G]} · b.

(25.5)

At each titration step the total concentration of both the host and the guest is the sum of the respective free and bound fractions. Thus it follows that [HT ] = [H ] + [HG],

(25.6)

[GT ] = [G] + [HG].

(25.7)

Rearranging Eqs. (25.6) and (25.7), the concentrations of the free host and guest can be obtained: [H ] = [HT ] − [HG],

(25.8)

[G] = [GT ] − [HG].

(25.9)

Substituting Eqs. (25.8) and (25.9) with Eq. (25.5) leads to AHG = {εHG · [HG] + εH · ([HT ] − [HG]) + εG · ([GT ] − [HG])} · b. (25.10) Developing the equation, we obtain AHG = {εHG · [HG] + εH · [HT ] − εH · [HG] + εG · [GT ] − εG · [HG]} · b, (25.11) AHG = {(εHG − εH − εG ) · [HG] + εH · [HT ] + εG · [GT ]} · b.

(25.12)

Both εHG and [HG] are unknown, but [HG] can be obtained from the association constant, in fact: KA =

[HG] . [H ] · [G]

(25.13)

Thus, [HG] is [HG] = KA · [H ] · [G].

(25.14)

Substituting Eqs. (25.8) and (25.9) with Eq. (25.14) leads to [HG] = KA · ([HT ] − [HG]) · ([GT ] − [HG]).

(25.15)

827

E L E C T R O N I C C I R C U L A R D I C H R O I S M I N D R U G D I S C O V E RY

Developing the equation and rearranging the following second-degree equation can be derived: KA · [HG]2 − [HG] · (KA · [HT ] + KA · [GT ] + 1) + KA · [HT ] · [GT ] = 0.

(25.16)

Solving the equation for [HG], we obtain (KA · [HT ]+KA ·[GT ]+1) ± (KA ·[HT ]+KA ·[GT ]+1)2 −4KA2 ·[HT ]·[GT ] [HG] = . 2·KA (25.17) Substituting Eq. (25.17) with (25.12) leads to the ﬁnal equation: AHG = (εHG − εH − εG ) (KA ·[HT ]+KA ·[GT ]+1) ± (KA ·[HT ]+KA ·[GT ]+1)2 −4KA2 ·[HT ]·[GT ] × 2·KA +εH ·[HT ]+εG ·[GT ] · b. (25.18) The mathematical solution of Eq. (25.18) is complex, requiring computational methods. The two solutions can be calculated by a data analysis computer program implemented with the Marquardt–Levemberg nonlinear least square algorithm, using Eq. (25.18) as nonlinear ﬁtting equation. The KA results from one of the two solutions to Eq. (25.18) best ﬁtting the experimental data plotted as the titration CD data (expressed in A units) at single wavelength versus the molar concentration of the host ([HT ]). Equation (25.18) can be simpliﬁed if the guest has no CD (εG = 0), either because it has no absorbing chromophores in the UV region studied [37] or because it is nonchiral [38], or if the ICD of the guest occur at wavelengths higher than 300–320 nm where the protein has no CD (εH = 0). The titration method has been successfully employed to characterize many drug–protein and protein–protein interactions. As an example, the interaction between Hsp90 and its co-chaperone p50cdc37 was characterized by titration experiments. The titration of Hsp90 with increasing concentrations of p50cdc37 resulted in a saturation of the near-UV CD signal at a molar ratio close to 1:1. The change in the ICD shown upon further addition of p50cdc37 is consistent with the homodimerization of p50cdc37 , or with the existence of a second p50cdc37 binding site on Hsp90. The following self-titration experiment, where increasing concentrations of p50cdc37 were monitored, showed that the spectral changes in the near-UV of the different CD spectra are consistent with those observed when p50cdc37 is in excess over Hsp90 and are consistent with homodimerization. Analysis of the data obtained allowed estimation of both the binding afﬁnity constant of the Hsp90-p50cdc37 complex and the dimerization constant of p50cdc37 [39].

25.1.3. Measurement of Binding Constant by Dilution Experiments Since the intensity of the ICD is proportional to the concentration of the drug–protein complex, CD can be used to measure the binding constant of a drug to a protein in a very elegant way [12]. This approach consists of monitoring the effect of sequential dilutions

828

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

of the drug–protein complex at ﬁxed stoichiometry on the ICD signal while increasing the pathlength in proportion to the dilution (e.g., increasing the pathlength tenfold while diluting the solution of the complex tenfold) [14, 40]. The decrease in the ICD signal detected by this method is correlated with the association constant of the complex, which can be directly derived from the experimental data. Using the Lambert–Beer law, the following linear equation was derived for a 1:1 drug–protein stoichiometry: p

CD l

=

1 · ε

1 CD +√ , l K · ε

(25.19)

where p is the protein concentration, CD is the measured signal, l is the cell pathlength, K is the association constant, and ε is the differential molar extinction coefﬁcient. Measuring the ICD while changing p and l yields K and ε from the slope and the intercept of a linear ﬁt to the data. Obviously this approach requires the binding phenomenon to yield a detectable ICD. Furthermore, if the drug is chiral, its intrinsic CD signal must be much lower than the ICD to obtain reliable afﬁnity data. On the other hand, this approach can provide data on the highest-afﬁnity binding site of a drug, since usually only drug binding to that site gives an ICD [15]. This is particularly important when studying the binding of drugs to serum albumins since (at therapeutic concentrations) only the main binding site of the drug has pharmacokinetic relevance [15]. This approach is not applicable to all drug–protein systems. An example of successful application is the measurement of the binding constants of ketoprofen to human and rat serum albumins [41], which yielded binding constants in agreement with those obtained by ultraﬁltration [42].

25.2. CONFORMATIONAL TRANSITIONS OF PEPTIDES AND PROTEINS The most traditional application of CD to monitoring a conformational transition of a protein is the transition from the folded to the unfolded state, typically in the study of secondary structure stability during temperature changes. However, many systems undergo transitions involving changes from α-helical to β-sheet state. The conformational transition may be followed by oligomerization of the protein, aggregation, and precipitation [43]. A recent study of α to β or unfolded to β conformational transitions of peptides and proteins has attracted considerable attention because these events were proved to be key molecular processes in a variety of degenerative diseases such as Alzheimer’s disease, Parkinson’s disease, Huntington’s disease, prion diseases, and amyotrophic lateral sclerosis [44]. These neurodegenerative diseases are increasingly being recognized to share protein aggregation as common molecular mechanisms. CD has been extensively used to monitor amyloid peptide conformational changes [45–49]. As an example, CD was successfully used to monitor the aggregation of β –amyloid (1–40) induced by the peptide interaction with a protein, human recombinant acetylcholinesterase (AChE) [50]. The self-aggregation process of β –amyloid (1–42) peptide upon changing concentration, solvent, and temperature parameters has also been monitored [51]. CD spectra of β –amyloid (1–42) between 260 and 190 nm at different times after preparation of the solution are reported in Figure 25.5a. Initially the peptide shows an unordered structure, while a β-sheet secondary structure prevails after 8 h, as suggested by the negative CD band at 215 nm and the positive band at 195 nm (Figure 25.5a). The change in the absolute CD value at 215 nm against time has a

829

E L E C T R O N I C C I R C U L A R D I C H R O I S M I N D R U G D I S C O V E RY

10

CD[mdeg]

0

–10

–20 190

200

220 Wavelength (nm)

240

260

Ab42 (50 μM) that was incubated in

(a)

Absolute CD signal at 215 nm (mdeg)

Figure 25.5. CD kinetic study of

20

phosphate buffer (8.6 mM, pH 8.0) that contained NaCl (10 mM) at 30◦ C. (a) Overlaid CD spectra (range

15

190–260 nm) recorded at 45-min time intervals from 1 h to 12 h after the solubilization. Plot color code:

10

light gray denotes CD spectrum recorded after 1 h, black, denotes CD spectrum recorded after 12 h. The increase in the β-sheet content is indicated by the increase of a negative band at 215 nm and of a positive band at 195 nm. (b) Time course absolute CD signal at

5

0 0

2

4

6 Time (h) (b)

8

10

12

215 nm: sigmoidal trend. (Reproduced with permission from reference 51. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.)

sigmoid behavior (Figure 25.5b) with three kinetic steps: a lag phase, where the initial unordered conformation does not change signiﬁcantly, an exponential growth phase, with an increase in the β-sheet content, and a plateau phase marking the start of ﬁbril formation. CD will characterize the kinetics of the secondary structure change from the beginning of the peptide aggregation process up to ﬁbril formation. The aggregation process can be studied by many techniques like light scattering, [52], atomic force microscopy [53], turbidimetry [54, 55], X-ray diffraction [56, 57], and imaging techniques like electron microscopy [58, 59]. The peculiar use of CD is to monitor the conformation transition associated with ﬁbril formation and to screen inhibitors of the aggregation process. In addition, these CD studies yield information on the mechanism of the inhibition process, by discriminating inhibitors acting on the lag phase or on the exponential phase of the kinetics. During the lag phase the peptide maintains a nonamyloidogenic state, so that an inhibitor prolonging this phase will slow down the aggregation process. If the inhibitor acts on the exponential phase, the rate of the conformational transition to the β-sheet form and insoluble ﬁbril formation will be reduced [50, 51]. As an example,

830

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

memoquin, a multifunctional molecule under investigation for its activity in Alzheimer’s disease [60], causes a concentration-dependent decrease in the rate of the conformational transition exponential phase [51] (Figure 25.6). Conformational transitions of proteins are not necessarily related to degenerative processes, because folding/unfolding processes can occur under physiological conditions, the conformational switch quite often being trigger for protein function. As an example, calexcitin/cp20 is a low-molecular-weight GTP- and Ca2+ -binding protein, which is phosphorylated by protein kinase C during associative learning and reproduces many of the cellular effects of learning, such as the reduction of potassium currents in neurons [61]. The secondary structure of cloned squid calexcitin was determined by CD in aqueous solution where it resulted one-third in α-helix and one-ﬁfth in β-sheet. The secondary structure depends on calcium binding, which determines an increase in αhelix and a decrease in β-sheet, as estimated by CD. The switch of calexcitin secondary structure upon calcium binding, conﬁrmed by intrinsic ﬂuorescence spectroscopy and non-denaturing gel electrophoresis, was proved to be reversible and, most importantly, it occurs in a physiologically meaningful range of Ca2+ concentrations. The calciumdependent conformational equilibrium of calexcitin could serve as a molecular switch for the short-term modulation of neuronal activity following associative conditioning. The CD spectrum of calexcitin (Figure 25.7) was deconvoluted into the ﬁve standard components of the secondary structure [62, 63]. When the concentration of free calcium in solution is reduced, the proﬁle of the calexcitin CD spectrum changes signiﬁcantly (Figure 25.7) with a reduction of the α-helical content. A full titration of calexcitin secondary structure versus Ca2+ concentration was then carried out (Figure 25.8). A conformational transition between the two states with signiﬁcant differences in α and β structures was shown to occur in a relatively small range of Ca2+ concentrations. This behavior may therefore be used experimentally to follow the extent of the structural switch (Figure 25.8). The change in calexcitin CD was then ﬁtted to a one-step equilibrium equation, with an apparent thermodynamic constant of approximately 10−6 M for Ca2+ binding [61]. Another interesting CD application for monitoring the inﬂuence of the peptide environment on its secondary structure is the study of Polymixin B structure ﬂexibility related 110 100 90 80

% CD

70

Figure 25.6. Trend of the β-sheet content

60

increase over time for 50 μM β-amyloid (1–42), without (gray long dashed line) and

50

with (black lines) 10 μM [drug/peptide = 1:5] (black solid line) and 50 μM memoquin [drug/peptide = 1 : 1] (black short dashed

40 30 20

line). Reported data are the mean of two independent experiments and are expressed

10 0 0

10

20

40

30 t (h)

50

60

70

as % of maximum CD signal. (Reproduced with permission from reference 51. Copyright Wiley-VCH Verlag GmbH & Co. KGaA.)

831

E L E C T R O N I C C I R C U L A R D I C H R O I S M I N D R U G D I S C O V E RY

6

4

Figure 25.7. CD spectrum of cloned calexcitin

H

A

P

T

O

33

15

3

19

30

(2)

(1)

(1)

(1)

(full line) in 20 μL of 5 mM Tris–HCl (pH 7.5); cell pathlength is 0.01 cm; the concentration of calexcitin is 9.5 μM. The calcium concentration in the solution is estimated at 10 μM. The

(2)

Δε

2

deconvoluted secondary structure is reported in the inset, together with the standard deviation (in parentheses). H, α-helix; A, antiparallel

0

β-sheet; P, parallel β-sheet; T, turns; O, other structures. The dotted line is the spectrum of the same solution after buffering the calcium at

–2

–4 180

190

200

210

220 nm

230

240

250

260

10 nM. (Reproduced with permission from reference 61.)

2.600 2.400 2.200

α/β

2.000 1.800 a/b = 1.22+1.35*[Ca++] / ([Ca++]+Kd)

1.600 1.400 1.200 1.000 10–10 10–9

10–8

10–7

10–6

[Ca++] (M)

Kd Chisq

Value 8.6958e–07 0.039139

Error 6.7801e–08 NA

R2

0.99119

NA

10–5

10–4

10–3

Figure 25.8. Effect of Ca2+ binding on the relative composition of α-helix and β-sheet (as a sum of parallel and antiparallel structures) of calexcitin’s secondary structure. (Reproduced with permission from reference 61.)

to its high selectivity. The structural changes observed for this potent and selective antimicrobial at the interface of anionic phospholipid vesicles were consistent with those registered in aqueous buffer solution in the presence of triﬂuoroethanol to lower the polarity of the solvent [64]. Polymixin B CD spectra carried out in both conditions suggested the adoption of a more ordered structure in the presence of triﬂuoroethanol, in agreement with its hypothesized membrane-mimetic behavior. CD and NMR studies suggested that the topological ﬂexibility of Polymixin B might represent a critical element allowing different functions depending on the local polarity of the bacterial cell surface [64]. The usefulness of CD in providing information on the mechanism of the ligand–target molecular recognition process is clearly demonstrated by the study reported for lepirudin. This drug is a recombinant hirudin used as an anticoagulant when heparins determine heparin-induced thrombocytopenia as a drawback. Hirudin is a potent and speciﬁc inhibitor of thrombin, ﬁrst obtained from the salivary gland of the European leech [65]. CD yielded information on the mechanism of the recognition process by comparing the CD spectra of the single peptides and their complex. The CD spectrum of the hirudin–thrombin complex was not additive with respect to the individual spectra of thrombin and hirudin, suggesting that dissociation of the

832

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

hirudin aggregates should be the driving factor determining the conformational changes monitored during the complex formation. An interesting example of the use of CD spectroscopy for monitoring modulation of peptide–protein interactions is the study on the binding mechanism of Enfurvitide (Fuzeon®), a potent inhibitor of HIV-1 entry in its host cell, with gp41, a transmembrane envelope glycoprotein modulating the fusion between the viral and cellular membrane. In its active state the gp41 core consists of a trimer of heterodimers comprising a leucine–isoleucine zipper sequence. CD spectra suggested the inﬂuence of Enfurvitide on the leucine–isoleucine zipper-like sequence without perturbing the trimer of heterodimers conformation. This ﬁnding disclosed that Enfurvitide acts by binding gp41 in an intermediate state thereby inhibiting the glycoprotein transition to the fusion-active conformation [66]. CD is becoming more and more useful in drug discovery because of the efforts of most pharmaceutical companies to design new drugs as modulators of functional protein–protein interactions. This new system biology-oriented trend in pharmaceutical chemistry raises the need to characterize protein homo- or heterodimer functional systems to discover and validate the new targets. In particular, the protein domains important for protein complex stability must be determined for the rational design of new drugs. Drug design is much more difﬁcult when the target is a functional protein dimer or trimer with respect to the more deﬁned structure of a receptor or enzyme. The stability and function of these protein oligomers are often regulated by their conformation; thus, once again, CD can be a great help in monitoring conformational changes that serve to modulate physiological and pathological processes. As an example, CD was used to study (a) the mechanism of Herpes simplex virus 1 (HSV-1) entry into cells and (b) virus–cell fusion. This process is mediated by the HSV-1 glycoproteins, namely four glycoproteins, gD, gB, gH, gL [67]. Among these, gH is the only one exhibiting the structural–functional features typical of viral fusion glycoproteins—that is, a i.e. a candidate fusion peptide and, downstream from it, a heptad repeat (HR) segment able to form a coiled coil, named HR-1 [67]. A second functional segment, HR-2, is capable of physical interaction with HR-1. Speciﬁcally, mutational analysis of gH aimed at increasing or decreasing the ability of HR-2 to form a coiled coil resulted in increased or decreased fusion activity, respectively. A mimetic peptide with the HR-2 sequence inhibited HSV-1infection in a speciﬁc and dose-dependent manner, thus suggesting that the HR-1/HR-2 interaction is a potential target for the rational design of an inhibitor of fusion entry [68, 69]. CD spectroscopy showed that both HR-2 and HR-1 mimetic peptides adopt mainly random conformation in aqueous solution, while a decrease in peptide environmental polarity determines a conformational change. A signiﬁcant increase in the α-helical conformation content was observed, in particular for the HR-1 peptide (Figure 25.9). The least polar environment was obtained by adding triﬂuoroethanol up to 15% to the buffer solution. A signiﬁcant conformational change was observed for HR1-25 upon adding the organic solvent. The CD spectra indicate an almost unordered structure in aqueous buffer, whereas a prevailing α-helix structure is observable at 15% TFE concentration, with the typical negative bands at 208 nm and 222 nm [69]. The HR2-25 peptide underwent minor conformational changes during titration, and even at 15% TFE the spectrum revealed a mostly unordered structure. These results on the propensity of the isolated peptides to assume ordered structure ﬁt the bioinformatic study on the N-terminal domain of the protein. Difference CD spectra methodologies have been used to study the structural effect of the interaction between the two peptides. A non-negligible signal (Figure 25.9}

833

E L E C T R O N I C C I R C U L A R D I C H R O I S M I N D R U G D I S C O V E RY

HR1-25

40

Millidegrees

20 10

Millidegrees

30

HR2-25

10

0% TFE 2.5% TFE 5% TFE 7.5% TFE 10% TFE 12.5% TFE 15% TFE

0

0 0% TFE 2.5% TFE 5% TFE 7.5% TFE 10% TFE 12.5% TFE 15% TFE

–10

–10 –20

–20

–30 200

210

220 230 240 Wavelength (nm)

250

200

260

210

220

230

240

1:1 Mixture 10

0% TFE 2.5% TFE 5% TFE 7.5% TFE 10% TFE 12.5% TFE 15% TFE

20 5

0 0% TFE 2.5% TFE 5% TFE 7.5% TFE 10% TFE 12.5% TFE 15% TFE

–10 –20 –30 –40 220

230

Wavelength (nm)

240

250

260

Millidegrees

Millidegrees

10

210

260

Difference Spectra

30

200

250

Wavelength (nm)

0

–5

–10 200

210

220 230 240 Wavelength (nm)

250

260

Figure 25.9. CD spectra of HR1–C25, HR2–C25, and their 1:1 mixture: [peptide] 50 μM, PBS, pH 7.4, TFE as the co-solvent (0%, 2.5%, 5%, 7.5%, 10%, 12.5%, 15%). Difference CD spectra as a function of TFE concentration (0%, 2.5%, 5%, 7.5%, 10%, 12.5%, 15%) is also depicted. The difference CD spectra are calculated subtracting the two individual peptide spectra from those of the mixture. (Reproduced with permission from reference 69.) (See insert for color representation of the ﬁgure.)

was observed (about 12% signal value of the mixture at 15% TFE), even if weaker than usually detected in analogous models [70, 71]. Superimposition of the spectra recorded at increasing concentrations of TFE shows a clean conserved isodichroic point at 203 nm for each peptide and for the mixture [Figure 25.9], suggesting a two-state transition from an unordered conformation to an α-helical conformation. Thus HR1 and HR-2 mimetic peptides formed a stable complex, as also revealed by thermal denaturation experiments on the isolated peptides and their mixture (Figure 25.10). The melting temperature of the HR-1/HR-2 1/1 complex was close to that of the most stable peptide [69].

25.3. CD IN THE ANALYSIS OF RECOMBINANT PROTEIN THERAPEUTICS The usefulness of CD spectroscopy in the development of therapeutic peptides is widely documented. Since the end of the 1960s, when the peptides market emerged, CD has been used to study molecules of great pharmacological interest like gramicidin antibiotic [72] and oxytocin [73]. Nowadays the use of proteins or peptides as pharmaceutical drugs is increasing for their efﬁcacy in the treatment of many diseases. However, the

834

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

HR1-25

Mixture

HR2-25

(b)

ƒ folded

d (θ) / dT

(a)

1.0

0.5

0.0 0

15

30 45 60 Temperature, °C

75

90

10 20 30 40 50 60 70 80 90 Temperature, °C

Figure 25.10. Melting temperature in PBS 10% TFE of HR1-25, HR2-25 and their mixture. (a) First-order derivatives of the unfolding curves of the two individual peptides and the 1:1 mixture; the maximum of the derivative allow the determination of Tm: 42◦ C for HR1–25, and 37◦ C for the mixture. HR2–25 is not affected by thermal denaturation. (b) Unfolding curves of HR1–25 expressed as folded fraction. (Reproduced with permission from reference 69.)

activity of these drugs can be compromised by their instability. In particular, structural differences in protein therapeutics may occur because of oligomerization, changes in tertiary conformation, in vivo modiﬁcation, or degradation of primary structure [74]. Thus novel delivery systems for peptide and protein therapeutics have been developed to improve their pharmacodynamics and pharmacokinetics [34, 75, 76]. Among the recombinant protein therapeutics with enzymatic or regulatory activity, recombinant insulin is surely the most famous, and it is widely used to treat diabetes mellitus. Several formulations of insulin are on the market, and signiﬁcant differences in drug activity have been documented. The stereochemical stability of insulin has been extensively investigated to optimise its bioavailability and half-life, developing different formulations and routes of administration. The CD of insulin in solution is typical of a predominant α-helix structure, with two negative bands centred at 222 and 208 nm, and a positive band at 190 nm [77]. Several physicochemical tests were performed to evaluate product quality, and establish the correlation between structure and activity [78]. Recombinant insulin has been characterized for its stereochemical stability in different formulations [79–81] or under physical stress, for example freeze-drying treatment to obtain the drug in a lyophilised form [82]. Furthermore, determination of insulin’s secondary structure and monitoring of its stability are becoming essential for the quality control assay of insulin in formulations. In particular, the problem of aggregation phenomena is critical for insulin and many other protein therapeutics because it represents a severe limitation for drug stability and disposition. The insulin formulations Humulin®, Regular Iletin I® and Regular Iletin II® were studied by CD to evaluate the dependence of conformation and aggregation state, production, storage and delivery phases of the drug. The ﬁbrillation process is characterized by a transformation of α-helix into β-sheet, and the consequent aggregation of insulin. The three steps of the peptide ﬁbrillation process (nucleation, elongation and precipitation) and the experimental conditions (concentration, temperature, pH, ionic strength) inﬂuencing these changes are known [83–85]. Recently [86], a multipathway process was proposed on the basis of CD and atomic force

835

E L E C T R O N I C C I R C U L A R D I C H R O I S M I N D R U G D I S C O V E RY

40°C

q [10–3 deg]

35°C 200

200

150

100

100

0

50

–100

0

–200 350 400 450 500 550

350 400 450 500 550 50°C

q [10–3 deg]

45°C 200

0

100

–50

Figure 25.11. Stochastic formation of chiral variants of insulin ﬁbrils. ICD of ThT added to insulin ﬁbrils formed through a 48-h-long agitation of the

–100

0

–150

native protein at 1400 rpm and at different temperatures. The multiple spectra in each panel correspond to 10

–100 –200 –200

–250 350 400 450 500 550 Wavelength (nm)

350 400 450 500 550 Wavelength (nm)

samples vortexed simultaneously in each series. (Reproduced with permission from reference 87.)

microscopy analysis under sample agitation. Interestingly, monitoring the induced CD spectra of thioﬂavin T (ThT), an amyloid binding dye (Figure 25.11) [87], disclosed conformational variants of insulin with a twisted superstructure. The chiral bifurcation of the aggregating insulin occurs with a prevailing chiral sense of the higher-order structures. The temperature-induced reversal of the insulin aggregate chiral sense was reﬂected by the sign inversion of the induced CD spectra of thioﬂavin T [85]. Many studies were then carried out to inhibit the aggregation phenomenon and ﬁbril precipitation. As an example, CD proved able to monitor the effect of agitation on the secondary structure of insulin in solution and in the liquid crystalline cubic phase gel of glyceryl monooleate, used to enhance the chemical stability. Analysis of the CD and absorption spectra of insulin demonstrated a complete loss of the native conformation in solution, while insulin did not aggregate in the gel phase, thus being preserved from precipitation [88]. Several excipients were then used to avoid the ﬁbrillation process. Among these, the α-crystallin protective effect against aggregation was particularly efﬁcient. A CD study performed in a relative wide range of temperatures demonstrated the anti-aggregating effect of α-crystallin applying stirring to stress the formulation [Figure 25.12]. This effect has been related to the stabilization of protein intermediates in a crucial secondary structure [89]. Engineering of the host system to coexpress a chaperone with the therapeutic proteins has also been proposed [75]. CD was also successfully used in the development of an alginate/chitosan-based formulation with the insulin complexed to a cationic β-cyclodextrin polymer. Insulin was relatively stable in simulated gastric ﬂuid, making the formulation promising for the development of an oral delivery system to avoid the drawbacks of parental insulin administration [90].

836

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

Molar ellipticity (Degrees.cm2.dmol–1.residue–1)

5000

0

(a)

(b)

–100 0

–200 –300

–5000

–400 –10000

–500

–15000 7500

–600 400

(c)

(d)

300

5000

200

2500

100 0

0

–2500 –5000 190

–100 200

210

220

230

–200 240 250 Wavelength (nm)

260

270

280

290

300

Figure 25.12. The top row shows the circular dichroism spectra of the insulin formulations (0.58 mg mL−1 ) in the far-UV (a) and near-UV (b) regions. The spectra of the insulin formulation without excipient are shown as a solid line (unstressed) and dotted line (stressed by applying stirring). The spectra of the α-crystallin containing insulin formulation are shown as a dashed line (unstressed) and a dash–dotted line (stressed by applying stirring). Spectra a and b are normalized to insulin concentration. In the bottom row, spectra of α-crystallin in the absence of insulin are shown in the far-UV (c) and near-UV (d) region. Unstressed (solid) and stressed (dotted line) α-crystallin (0.2 mg mL−1 ) samples are shown. Spectra c and d are normalized to α-crystallin concentration. (Reproduced with permission from reference 89.)

The nature of excipients in formulations can affect protein secondary structure, but signiﬁcant conformational changes can be due also to different manufacturing processes. Thus the characterization of protein therapeutics deserves a CD investigation to compare the stereochemistry of the drug and its stability in products from different companies. As an example, this type of analysis was applied to Epogen® and Eprex®, two formulations of recombinant human erythropoietin alfa, used for the treatment of anemia. This hormone is responsible for regulating red blood cell production [91]. Erythropoietin also has a history of use as a performance-enhancing drug in endurance sports (blood doping). The analysis performed by CD on Epogen® and Eprex® disclosed different behavior of the same protein produced by different manufactures. In fact slight differences between the two proteins were observed using far-UV CD (Figure 25.13), suggesting a difference in the α-helix content. Furthermore, while the erythropoietin alfa from Epogen® showed a complete reversibility of the conformation upon thermal denaturation analysis, the protein from Eprex® did not recover its native conformation under extreme conditions (Figure 25.14). This example highlights that proteins with the same amino acid sequence fold in almost the same manner but may present differences in biophysical characteristics that might be related to the manufacturing process [92]. CD is also widely used to analyze proteins from different species or isoforms. Signiﬁcant differences in the propensity to assume a α-helix conformation upon changing the polarity of the solvent were observed

837

E L E C T R O N I C C I R C U L A R D I C H R O I S M I N D R U G D I S C O V E RY

Mean Residue Ellipticity (10–3 deg cm2 dmol–1)

5 0 –5 –10 –15

Figure 25.13. Far-UV CD spectra of Epogen bulk

–20 –25 200

(solid), puriﬁed Epogen (dashed), and puriﬁed Eprex (dotted–dashed). The CD spectra of Epogen bulk and

210

220

230

240

250

260

Wavelength (nm)

puriﬁed Epogen resulted perfectly superimposable. (Reproduced with permission from reference 92.)

5

Mean Residue Ellipticity (10–3 deg cm2 dmol–1)

0 –5

Figure 25.14. Overlays of Epogen bulk (solid), puriﬁed Epogen (long dashes), and puriﬁed Eprex

–10

(dot and short dashes) at 25◦ C before thermal unfolding and Epogen bulk (double dots and short dashes), puriﬁed Epogen (short dashes), and puriﬁed

–15

Eprex (dots) at 25◦ C after thermal unfolding. The CD spectra of Epogen bulk and puriﬁed Epogen either before and after thermal unfolding resulted perfectly

–20 –25 200

210

220

230

240

Wavelength (nm)

250

260

superimposable. (Reproduced with permission from reference 92.)

for human and salmon calcitonin. This hormone is involved in calcium–phosphorus metabolism and is used for diseases like osteoporosis. The two calcitonins showed a high α-helix content at low dielectric constant solvents. The propensity to assume an α-helical structure is more evident in human than in salmon calcitonin, as determined by analysis of CD spectra in the 260 to 185 nm spectral range [93]. Understanding these differences can be useful in determining the protein’s structure–function relationship. In addition, recombinant human calcitonin was studied by CD to monitor the kinetics of the ﬁbrillation process [93]. Another interesting application of CD in the area of recombinant protein therapeutics is the structural analysis of recombinant human interferon α-2b and evaluation of its stability in terms of half-life of the administered drug [94]. Interferons are cytokines produced by eukaryotic cells, and they show several biological activities, like antiviral effect, immunomodulatory properties, and cell growth inhibition. Thus interferons have been used for a wide range of infectious and proliferative disorders. In particular, recombinant α-interferons were approved for the treatment of hairy-cell leukemia, AIDS-related Kaposi’s sarcoma, and hepatitis C. The problem with both interferon α-2a (Roferon) and interferon α-2b (Intron® A, the recombinant human version) is their short half-life. Patients need three injections per week to maintain the serum concentration.

838

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

1

0

0

–1

0.0

ΔA x 104

1.5

1

ΔA x 104

Δε (mol–1 cm–1)

The strategy adopted to increase the serum half-life of Intron® was the attachment of a monomethoxy poly-(ethylene glycol) polymer to the protein (PEG Intron) [95, 96]. A check on the CD spectra of nonpegylated and pegylated interferon α2b samples in the near-UV and far-UV spectral regions did not disclose any signiﬁcant difference between the two forms [95]. Thus the pegylation of interferon α-2b does not affect its tertiary and secondary structure. Another interesting application of CD in medicinal chemistry concerns the stereochemical characterization of protein therapeutics with special targeting activity, like therapeutic antibodies. As an example, the therapeutic antibody MMA 383 was extensively investigated by CD. Information on the relationship between structural changes induced by freeze-drying process, and the loss of in vivo immunogenic properties of the antibody was obtained [97]. MMA 383 has been designed as an immunogenic surrogate for the Lewis Y carbohydrate, a potential cancer antigen because of its signiﬁcant expression on the surface of most cancer cells of epithelial origin. The speciﬁcity is related to the limited expression of the antigen in normal tissues. CD spectra of lyophilized and nonlyophilized MMA 383 antibody were analyzed in the near-UV spectral region at 50◦ C and at 11◦ C. The spectra at 50◦ C were almost superimposable, whereas the spectra at 11◦ C showed signiﬁcant differences between the lyophilized and nonlyophilized samples. The most prominent change was observed at about 240 nm, where the disulﬁde electronic transition occurs (Figure 25.15). Instead, the CD spectra for both the lyophilized and nonlyophilized antibody forms at shorter wavelengths, associated with the protein backbone, were not inﬂuenced by temperature, suggesting that the secondary structure is maintained. Thus the loss of antibody biological activity is not related to a change in its secondary structure, but the conformational change at the disulﬁde might be responsible for the reduction of MMA 383 ﬂexibility upon lyophilization, as conﬁrmed by electron microscopy [97]. A signiﬁcant change in the CD spectrum in the 230 to 250-nm spectral range may then be diagnostic for loss of biological activity.

–1

–2

–2

–1.5

200

220

240

Wavelength (nm) (a)

–3

240 260 280 300 Wavelength (nm) (b)

–3

240 260 280 300 Wavelength (nm) (c)

Figure 25.15. (a) CD spectrum of MMA 383 antibodies at 24◦ C; antibody concentrations were 0.1 mg mL−1 , and 0.02-cm-pathlength cuvettes were used. No differences were observed between the CD spectra of lyophilized and nonlyophilized antibodies. (b, c) Temperature-induced changes in the near-UV CD spectra: (b) Nonlyophilized and (c) reconstituted, lyophilized antibodies. The CD spectra were measured at 24◦ C (black lines) and 11◦ C (gray lines); 1-cm cuvette, 1-mg mL−1 samples. In the 236-nm region the temperature-dependent changes were more pronounced for the nonlyophilized antibodies. The data in (a) are presented in terms of the molar differential extinction coefﬁcient. The raw data are shown in (b) and (c). (Reproduced with permission from reference 97.)

E L E C T R O N I C C I R C U L A R D I C H R O I S M I N D R U G D I S C O V E RY

25.4. CONCLUSIONS AND PERSPECTIVES CD is a powerful technique that enables structural components to be monitored in solution before and after the formation of drug/protein and protein/protein complexes. Thus CD is currently widely used in all the phases of drug discovery and development. The absolute conﬁguration of enantiomeric forms of chiral drugs can be reliably determined by means of the available programs for ab initio calculations. In addition, analysis of the induced CD spectra observed in the drug binding to a target protein yields information on the stereochemistry of the bound drug and allows the binding parameters to be determined. Furthermore, CD is particularly useful in monitoring protein–protein interactions, which represent potential new targets in modern medicinal chemistry. Lastly, monitoring conformational transitions of peptides and proteins is fundamental when they trigger oligomerization and/or aggregation, these processes often being the key step in physiological and pathological events. Major perspectives in the use of circular dichroism are in the new area of interactome to characterize molecular recognition phenomena and large protein complexes by investigating the dynamics of the folding/unfolding and conformational switch processes. A wider use of this technique is also expected for monitoring the structural stability and aggregation of recombinant proteins in the different phases from drug discovery to therapy. This will serve to optimize drug efﬁcacy by developing the correct formulation and by choosing the best route of administration for each of the protein therapeutics. A new emphasis on the use of the CD will arise from the simultaneous detection of CD, linear CD, absorption, turbidity, light scattering, and ﬂuorescence [98].

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

S. M. Kelly, T. J. Jess, N. C. Price, Biochim. Biophys. Acta 2005, 1751 , 119–139. C. T. Chang, C. S. Wu, J. T. Yang, Anal. Biochem. 1978, 91 , 13–31. S. W. Provencher, J. Gl¨ockner, Biochemistry 1981, 20 , 33–37. L. A. Compton, W. C. Johnson Jr., Anal. Biochem. 1986, 155 , 155–167. B. A. Wallace, C. L. Teeters, Biochemistry 1987, 26 , 65–70. G. B¨ohm, R. Muhr, R. Jaenicke, Protein Eng. 1992, 5 , 191–195. A. Perczel, K. Park, G. D. Fasman, Anal. Biochem. 1992, 203 , 83–93. M. A. Andrade, P. Chac´on, J. J. Merelo, F. Mor´an, Protein Eng. 1993, 6 , 383–390. N. Sreerama, R. W. Woody, Anal. Biochem. 1993, 209 , 32–44. P. Pancoska, V. Janota, T. A. Keiderling, Anal. Biochem. 1999, 267 , 72–83. J. G. Lees, A. J. Miles, R. W. Janes, B. A. Wallace, BMC Bioinform. 2006, 7 , 507–516. C. Bertucci, M. Pistolozzi, A. De Simone, Anal. Bioanal. Chem. 2010, 398 , 155–166. D. J. Freeman, G. Pattenden, A. F. Drake, G. Siligardi, J. Chem. Soc. Perkin Trans. 2 1998, 129–135. G. A. Ascoli, C. Bertucci, P. Salvadori, J. Pharm. Sci . 1995, 84 , 737–741. G. A. Ascoli, E. Domenici, C. Bertucci, Chirality 2006, 18 , 667–679. W. C. Johnson Jr., Rev. Sci. Instrum. 1971, 42 , 1283–1286. P. A. Snyder, E. M. Rowe, Nucl. Instrum. Methods 1980, 172 , 345–349. J. C. Sutherland, A. Emrick, L. L. France, D. C. Monteleone, J. Trunk, Biotechniques 1992, 13 , 588–590. A. J. Miles, B. A. Wallace, Chem. Soc. Rev . 2006, 35 , 39–51.

839

840

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31.

32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51.

W. C. Johnson Jr., Photochem. Photobiol . 1986, 44 , 307–313. B. A. Wallace, Nat. Struct. Biol . 2000, 7 , 708–709. B. A. Wallace, R. W. Janes, Curr. Opin. Chem. Biol . 2001, 5 , 567–571. K. Matsuo, Y. Sakurada, R. Yonehara, M. Kataoka, K. Gekko, Biophys. J . 2007, 92 , 4088–4096. K. Matsuo, H. Watanabe, K. Gekko, Proteins 2008, 73 , 104–112. K. Matsuo, H. Namatame, M. Taniguchi, K. Gekko, Biochemistry 2009, 48 , 9103–9111. M. R. Hicks, T. R. Dafforn, A. Damianoglou, P. Wormell, A. Rodger, S. V. Hoffmann, Analyst 2009, 134 , 1623–1628. I. W. Hamley, V. Castelletto, C. Moulton, D. Myatt, G. Siligardi, C. L. P. Oliveira, J. S. Pedersen, I. Abutbul, D., Danino, Macromol. Biosci . 2010, 10 , 40–48. F. C. Dudak, E. A. Soykut, M. E. Oˇguz, F. Yas¸ar, I. H. Boyaci, J. Mol. Recognit. 2010, 23 , 369–378. D. S. Chapple, R. Hussain, C. L. Joannou, R. E. W. Hancock, E. Odell, R. W. Evans, G Siligardi, Antimicrob. Agents Chemother. 2004, 48 , 2190–2198. D. J. Rodi, R. W. Janes, H. J. Sanganee, R. A. Holton, B. A Wallace and L. Makowski, J. Mol. Biol . 1999, 285 , 197–203. C. Ferlini, L. Cicchillitti, G. Raspaglio, S. Bartollino, S. Cimitan, C. Bertucci, S. Mozzetti, D. Gallo, M. Persico, C. Fattorusso, G. Campiani, G. Scambia, Cancer Res. 2009, 69 , 6906–6914. C. Prodromou, G. Siligardi, R. O’Brien, D. N. Woolfson, L. Regan, B. Panaretou, J. E. Ladbury, P. W. Piper, L. H. Pearl, EMBO J . 1999, 18 , 754–762. B. A. Wallace, R. W. Janes, Biochem. Soc. Trans. 2010, 38 , 861–873. A. J. Chirino, A. Mire-Sluis, Nat. Biotechnol . 2004, 22 , 1383–1391. L. Whitmore, B. A. Wallace, Biopolymers 2008, 89 , 392–400. K. Nakanishi, N. Berova, R. W. Woody, Circular dichroism, principles and applications, VCH Publishers. New York, 1994. D. A. Renzoni, D. J. R. Pugh, G. Siligardi, P. Das, C. J. Morton, C. Rossi, M. D. Waterﬁeld, I. D. Campbell, J. E. Ladbury, Biochemistry 1996, 35 , 15646–15653. G. Siligardi, R. Hussain, Enantiomer 1998, 3 , 77–87. G. Siligardi, B. Panaretou, P. Meyer, S. Singh, D. N. Woolfson, P. W. Piper, L. H. Pearl, C. Prodromou, J. Biol. Chem. 2002, 277, 20151–20159. C. Bertucci, E. Domenici, P. Salvadori, Chirality 1990, 2 , 167–174. M. Pistolozzi, C. Bertucci, Chirality 2008, 20 , 552–558. K. Kaneko, H. Fukuda, V. T. Chuang, K. Yamasaki, K. Kawahara, H. Nakayama, A. Suenaga, T. Maruyama, M. Otagiri, Drug Metab. Dispos. 2008, 36 , 81–86. N. J. Greenﬁeld, Trends Anal. Chem. 1999, 18 , 236–244. C. A. Ross, M. A. Poirier, Nat. Med . 2004, 10 Suppl ., S10–S17. C. J. Barrow, M. G. Zagorski, Science 1991, 253 , 179–182. C. J. Barrow, A. Yasuda, P. T. Kenny, M. G. Zagorski, J. Mol. Biol . 1992, 225 , 1075–1093. L. Otvos Jr., G. I. Szendrei, V. M. Lee, H. H. Mantsch, Eur. J. Biochem. 1993, 211 , 249–257. S. Tomaselli, V. Esposito, P. Vangone, N. A. van Nuland, A. M., Bonvin, R. Guerrini, T. Tancredi, P. A. Temussi, D. Picone, ChemBioChem. 2006, 7 , 257–267. Y. Fezoui, D. B. Teplow, J. Biol. Chem. 2002, 277 , 36948–36954. M. Bartolini, C. Bertucci, V. Cavrini, V. Andrisano, Biochem. Pharmacol . 2003, 65 , 407–416. M. Bartolini, C. Bertucci, M. L. Bolognesi, A. Cavalli, C. Melchiorre, V. Andrisano, ChemBioChem. 2007, 8 , 2152–2161.

E L E C T R O N I C C I R C U L A R D I C H R O I S M I N D R U G D I S C O V E RY

52. A. Lomakin, D. S. Chung, G. B. Benedek, D. A. Kirschner, D. B. Teplow, Proc. Natl. Acad. Sci. USA 1996, 93 , 1125–1129. 53. J. D. Harper, S. S. Wong, C. M. Lieber, P. T. Lansbury, Chem. Biol . 1997, 4 , 119–125. 54. A. S. Pai, I. Rubinstein, H. Ony¨uksel, Peptides 2006, 27 , 2858–2866. 55. S. S. Wang, Y. T. Chen, S. W. Chou, Biochim. Biophys. Acta 2005, 1741 , 307–313. 56. H. Inouye, P. E. Fraser, D. A. Kirschner, Biophys. J . 1993, 64 , 502–519. 57. P. E. Fraser, D. R. McLachlan, W. K. Surewicz, C. A. Mizzen, A. D. Snow, J. T. Nguyen, D. A. Kirschner, J. Mol. Biol . 1994, 244 , 64–73. 58. D. B. Teplow, Amyloid 1998, 5 , 121–142. 59. M. Sunde, C. Blake, Adv. Protein Chem. 1997, 50 , 123–159. 60. A. Cavalli, M. L. Bolognesi, S. Capsoni, V. Andrisano, M. Bartolini, E. Margotti, A. Cattaneo, M. Recanatini, C. Melchiorre, Angew. Chem. Int. Ed. Engl . 2007, 46 , 3689–3692. 61. G. A. Ascoli, K. X. Luu, J. L. Olds, T. J. Nelson, P. A. Gusev, C. Bertucci, E. Bramanti, A. Raffaelli, P. Salvadori, D. L. Alkon, J. Biol. Chem. 1997, 272 , 24771–24779. 62. J. P. Hennessey, W. C. Johnson Jr., Biochemistry 1981, 20 , 1085–1094. 63. W. C. Johnson Jr., Methods Enzymol . 1992, 210 , 426–447. 64. M. D. Bruch, Y. Cajal, J. T. Koh, M. K. Jain, J. Am. Chem. Soc. 1999, 121 , 11993–12004. 65. S. Konno, J. W. Fenton 2nd, G. B. Villanueva, Arch. Biochem. Biophys. 1988, 267 , 158–166. 66. Y. Kliger, Y. Shai, J. Mol. Biol . 2000, 295 , 163–168. 67. G. Campadelli-Fiume, M. Amasio, E. Avitabile, A. Cerretani, C. Forghieri, T. Gianni, L. Menotti, Rev. Med. Virol . 2007, 17 , 313–326. 68. T. Gianni, A. Piccoli, C. Bertucci, G. Campadelli-Fiume, J. Virol . 2006, 80 , 2216–2224. 69. B. Sanavio, A. Piccoli, T. Gianni, C. Bertucci, Biochim. Biophys. Acta 2007, 1774 , 781–791. 70. S. Liu, G. Xiao, Y. Chen, Y. He, J. Niu, C. R. Escalante, H. Xiong, J. Farmar, A. K. Debnath, P. Tien, S. Jiang, Lancet 2004, 363 , 938–947. 71. B. Tripet, M. W. Howard, M. Jobling, R. K. Holmes, K. V. Holmes, R. S. Hodges, J. Biol. Chem. 2004, 279 , 20836–20849. 72. F. Quadrifoglio, D. W. Urry, Biochem. Biophys. Res. Commun. 1967, 29 , 785–791. 73. S. Beychok, E. Breslov, J. Biol. Chem. 1968, 243 , 151–154. 74. C. Bertucci, M. Pistolozzi, A. De Simone, Curr. Pharm. Biotechnol . 2011, 12 , 1508–1516. 75. B. Leader, Q. J. Baca, D. E. Golan, Nat. Rev. Drug. Discov. 2008, 7 , 21–39. 76. I. M. Tomlinson, Nat. Biotechnol. 2004, 22 , 521–522. 77. S. G. Melberg, W. C. Johnson, Jr., Proteins 1990, 8 , 280–286. 78. I. S. Johnson, Nat. Rev. Drug Discov . 2003, 2 , 747–751. 79. B. Sarmento, D. C. Ferreira, L. Jorgensen, M. van de Weert, Eur. J. Pharm. Biopharm. 2007, 65 , 10–17. 80. E.-S. Khafagy, M. Morishita, Y. Onuki, K. Takayama, Adv. Drug Deliv. Rev . 2007, 59 , 1521–1546. 81. M. Amidi, H. C. Pellikaan, A. H. de Boer, D. J. A. Crommelin, W. E. Hennink, W. Jiskoot, Eur. J. Pharm. Biopharm. 2008, 68 , 191–200. 82. Y. Zhang, Y. Deng, X. Wang, J. Xu, Z. Li Int. J. Pharm. 2009, 371 , 71–81 83. L. Nielsen, S. Frokjaer, J. Brange, V. N. Uversky, A. L. Fink, Biochemistry 2001, 40 , 8397–8409. 84. L. Nielsen, R. Khurana, A. Coats, S. Frokjaer, J. Brange, S. Vyas, V. N. Uversky, A. L. Fink, Biochemistry 2001, 40 , 6036–6046. 85. R. Liu, M. He, R. Su, Y. Yu, W. Qi, Z. He, Biochem. Biophys. Res. Commun. 2010, 391 , 862–867. 86. A. Loksztejn, W. Dzwolak, J. Mol. Biol . 2010, 395 , 643–655.

841

842

C O M P R E H E N S I V E C H I R O P T I C A L S P E C T R O S C O P Y, V O L U M E 2

87. A. Loksztejn, W. Dzwolak, J. Mol. Biol . 2008, 379 , 9–16. 88. Y. Sadhale, J. C. Shah, Int. J. Pharmaceut. 1999, 191 , 51–64. 89. T. Rasmussen, R. Tantipolphan, M. van de Weert, W. Jiskoot, Pharm. Res. 2010, 27 , 1337–1347. 90. N. Zhang, J. Li, W. Jiang, C. Ren, J. Li, J. Xin, K. Li, Int. J. Pharm. 2010, 393 , 212–218. 91. S. E. Graber, S. B. Krantz, Annu. Rev. Med . 1978, 29 , 51–66. 92. S. Deechongkit, K. H. Aoki, S. S. Park, B. A. Kerwin, J. Pharm. Sci . 2006, 95 , 1931–1943. 93. T. Arvinte, A. F. Drake, J. Biol. Chem. 1993, 268 , 6408–6414. 94. M. Cindri´c, N. Gali´c, M. Vuleti´c, M. Klari´c, V. Drevenkar, J. Pharm. Biomed. Anal . 2006, 40 , 781–787. 95. M. Grace, S. Youngster, G. Gitlin, W. Sydor, L. Xie, L. Westreich, S. Jacobs, D. Brassard, J. Bausch, R. Bordens, J. Interferon Cytokine Res. 2001, 21 , 1103–1115. 96. Y. S. Wang, S. Youngster, M. Grace, J. Bausch, R. Bordens, D. F. Wyss, Adv. Drug Deliv. Rev . 2002, 54 , 547–570. 97. N. Taschner, S. A. M¨uller, V. R. Alumella, K. N. Goldie, A. F. Drake, U. Aebi, T. Arvinte J. Mol. Biol . 2001, 310 , 169–179. 98. T. Arvinte, T. T. T. Bui, A. A. Dahab, B. Demeule, A. F. Drake, D. Elhag, P. King, Anal. Biochem., 2004, 332 , 46–57.

INDEX

ab initio methods, 427, 432–434, 437, 443, 454, 456, 476, 762–763, 765, 774, 776 abscisic acid, 134 absolute conﬁguration, of (+)-1,8a-dihydro-3,8dimethylazulene, 169–170 acenocoumarol, 670 acetonitrile, 360 acetoxyﬂavans, 98 1-acetoxymethyl-2,3,4,4-tetramethylcyclopentane, 413 acetylenealcohols AC assignment, 157–158 acetylmajapolene B, 407 achiral chromophore, chirality sensing with, 318–319 through coordination, 321–324 through hydrogen-bonding interaction, 319–321 through noncovalent and covalent interactions, 324–326 actinomycin D, 647–649 acyclic 1,2-glycols and polyols, 140–142 Aeromonas hydrophila, 619 A-form DNA, 636 and B–A transition, 577–578 africanane, 406 Aizoon canariense, 229 alkaloids, 400–401 alkylchroman-4-ones, 108 Allylic Axial Chirality Rule, 43 allylic benzoate exciton method, for allylic alcohol AC determination, 146–147 α, β-didehydro -α-amino acid-based polypeptides, 556–562 α, β-unsaturated ketones, 41 α-helix, 477–478, 488, 500–504, 768–770, 773 α1-acid glycoprotein (AGP), 488, 690–692, 784 binding examples to genetic variants, 693–700 stereoselective binding to, 692–693 α-pinene, 397 amide bonds, 522–523 amide chromophore, 476 bi- and polycyclic β-lactam ring systems, 56–57 penicillins and cephalosporins, 57–67 amide complexes anion-controlled switching, 260–263 aminoalcohols, 321 2-aminopurine substitutions, 624 amphiphilic thin ﬁlms, 312

amyloid ﬁbril formation, 780–782 amyloid peptide and ginkgolides, nature of interactions between, 414 ancistrocladium B, 365–366 HPLC-CD and quantum chemical CD calculations, 367–369 HPLC separation conditions, optimizing, 366 information obtained by HPLC-MS and HPLC-NMR measurements, 367 rotation barrier estimation, 369–370 Ancistrocladus sp.,363 Anemia tomentosa var. anthriscifolia, 409 7β –angeloyloxy-8α-isovaleroyloxylongipin-2-en-1-one, 410 anomalous dispersion, 5–6 anthracene chromophores, 127 antitumor antibiotic AT2433-A1 AC, with secondary amino group, 159 Apiospora montagnei, 397 Applequist’s method, 489 Arago, Dominique-Franc¸ois Jean, 4 Armstrong, H. E., 27 Arndtsen, Adam, 4 Artimisia maritima, 229 Ascochyta sp., 105, 229 Aspergillus niger, 397 atropisomerism, 153, 465–466 guanine derivative-platinum(II) complexes, 466–468 of natural products, 174–176 nitrito-κ-O complexes, 466 axially chiral natural products, 416–417 axial orientation, 78, 83, 98 B/A conformation of (dC)n • (dG)n DNA sequences, 578 basic/helix–loop–helix/leucine zipper transcription factors, 624–625 B–A transition, 577–578 benzamide, phthalimide, and 2,3-naphthalenedicarboximide chromophores, 131–132 benzene chromophore, 433 chroman-4-one chromophore, 108–109 derivatives, 73–76 dihydroisocoumarin chromophore, 104–108 with fused heterocyclic ring, 81–83

Comprehensive Chiroptical Spectroscopy, Volume 2: Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules, First Edition. Edited by N. Berova, P. L. Polavarapu, K. Nakanishi, and R. W. Woody. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

844

benzene chromophore (Continued ) benzodioxane chromophore, 83–85 chroman chromophore, 96–100 2,3-dihydrobenzo[b]furan derivatives, 90–96 isochroman chromophore, 85–90 tetralins with achiral substituents, 79–81 and tetrahydroisoquinolines, 77 tetralone derivatives, 101–104 benzoates, 149 benzodiazepines, 678–685 benzodioxane chromophore, 83–85 berberine CD analyses, binding of DNA quadruplex, 658–660 β-amino acids, 562 β,β -bisporphyrins, 378–379 β-cyclodextrins, 343 β-diketone, 675–676 β-hairpins, 515 β-peptides, 562–569 β-pinene, 399 β-sheet, 478, 488, 512–516, 730–731, 738, 770–773 β-turns, 478, 516–518, 770, 772 B-form DNA, 636 hairpin, 576 bi[10]paracyclophane, 379–381 bicyclo[3.3.1]nonanediones, 431 Bijvoet, J. M., 4, 116 bilirubin, 667–670 binaphthoquinones AC, axially chiral, 152–153 biological polymers, dynamic stereochemistry in, 256–257 biomolecule binding, 339–342 biopolymers. See also vibrational circular dichroism dynamic stereochemistry, monitored by optical rotation and optical rotatory dispersion, 255–256 Biot, Jean-Baptiste, 4, 5, 7 BioTools, 713–714, 741 1,1’-biphenanthryl compounds anomalous CD CEs, 159–161 biphenylboronates, 437 4-biphenylboronic acid, 326 bis-porphyrin tweezers, 440 Blennoria sp., 230 Boltzmann, L., 26 Born, Max, 22 boronic acids, 325 bovine α-lactalbumin (bLA), 738 Brace, DeWitt Bristol, 12 brassicanal C, 411 Brewster’s benzoate rule, 146 Brillouin, M., 29 Bruhat, Georges, 13, 18, 20–21, 25, 27 Bruker, 713 Butia capitata, 375 butylcyanurate (BuCYA), 344 B–Z transition, 579 calicheamycin, 643–645 calycanthine, 117 Calypogeia granulata Inoue, 169 Cambridge Structural Database (CSD), 226

INDEX

camphor, 397 camphor–CDCl3 complex, 399 carbacephams, 60–61 carbohydrates, 782–784 classes, 798 optical rotation and, 800–802 UV-VIS CD derivatized sugars, 804–806 vibrational circular dichroism and, 808–815 VUVCD and, 802–804 carbonyl chromophore, 430–432 carboxylic acids, 395 garcinia acid and hibiscus acid, 395–396 hexylitaconic acid and methyl esters, 397 and oleﬁn compounds, 132 Carpobrotus edulis, 230 carprofen, 685 Carthamus oxyacantha, 236 carvone, 398 cellular uptake and anticancer cc-1065 prodrug analogues target, in live tumor cells, 642–643 cephalochromin, 417 cephalocyclidin A AC, 148 cephalosporins, 57–67 cephams, 59, 61 Chaetomium globosum, 228 chemical and mechanical systems, 308–314 chiral chromophores, 39 amide chromophore in bi- and polycyclic β-lactam ring systems, 56–57 penicillins and cephalosporins, 57–67 dichalcogenides, 51–55 dienes and trans-enones, 39–51 chiral stationary phase (CSP), 358 chiral troponoid spiro compounds, circular dichroism and absolute stereochemistry of, 172–173 Chironomus thummi thummi, 484 chlorpromazine, 694 Christiansen, Christian, 5 chroman-4-one chromophore, 108–109 chroman chromophore, 96–100 chromomycin, 645–647 chromophoric molecules more than two chromophores, 438–439 presence of only one, 430–433 benzene chromophore, 433 presence of two, 433–434 identical chromophores, 434–435 two different chromophores, 435–438 ciguatoxin ACs and related compounds, 147 cinchonidine, 400 cinnamate and β-naphthoate, 130 circular dichroism curves, 7–9 nomenclature, 22, 24 circular intensity difference (CID), 761 circularly polarized ﬂuorescence excitation (CPE), 252 dynamic stereochemistry monitored by, 278–279 circular polarized luminescence (CPL), 252, 278–279

INDEX

dynamic stereochemistry monitored by, 280 clavams, 62–63 cleavamine, 134 clerodin, 127–128 collagen, 526, 529 Condon, E. U., 4, 116 CONFLEX program, 393 conformational studies, by ORD, 255 Coniothyrium sp., 229 conocarpan, 92 cortistatin A AC, 155 Cotton, Aim´e Auguste, 6, 29–31, 30, 115 Cotton effect, 3–4, 42, 45–46, 50, 55, 73, 115, 122, 167, 183, 255, 270, 310–311, 372, 528, 535, 646, 657, 667 advances in instrumentation and theory and, 20–22 biographical notices, 24–31 change from Cotton’s phenomenon to, 23 circular dichroism and optical rotatory dispersion curves, 7–9 discovery of, 6–7 reaction of learned world to, 9–12 early chiroptical studies and, 4–6 in inorganic chemistry, 14–16 nomenclature, 22–24 in organic chemistry, 16–20 tartrates, 12–14 coupled oscillator theory, 116–117 couplet, 463 cowpea mosaic virus (CPMV), 787 crocetin, 688 cruciferous phytoalexins, 410–411 Cryptsporiopsis sp., 229 CU complex, chiral redox-controlled molecular ﬂipper, based on, 269–271 Curcuma longa L., 675 curcumin, 675–678 Curvularia sp., 229 3-cyanomethyl-3-hydroxyoxindole, 410 cyclodextrin, 783–784 binding with, 336–339 cyclosporins, 415–416 cytochrome c (cyt c), 738 cytosine quadruplexes, 581–582 Daphniphyllum alkaloids, 231 Daphniphyllum macropodum, 231 Darmois, Eug`ene, 19, 20 daunorubicin binding to DNA, enantio-selective, 649 Davydov, A. S., 116 d –d transitions of tris- and bis-bidentate transition metal complexes, 452–456 theoretical studies on structure-spectra relationships, 456–457 δ-tocopherol, 98 dendrimers, 329 dendrons, 329 Dendrooctoonus frontalis, 413 density functional theoretical (DFT) method, 218, 392, 718, 723

845

11-deoxydiaporthein A AC, by exciton allylic benzoate method, 155 deprotonation, 309 deramciclane, 694 Descamps, R., 20 Deussen, E., 19 DeVoe’s method, 221–222, 237, 238, 239, 434, 437 3α,6β-diacetoxytropane, 402 7,9-diacetyloxylongipin-2-en-1-one, 410 7β,8α-diacetyloxylongipin-2-en-1-one, 410 diamines, 321 diazepam, 685, 696 diazobenzene-ferrocene systems, 283 dibenzoate chirality rule, 117 dichalcogenides, 51–55 dicoumarol, 694, 696 dicurcuphenol B, 416 dicurcuphenol C, 416 Diene Helicity Rule (DHR), 42–43 and model, 50 dienes, 39–51 1,3-diethynylallene compounds ACs, 158 diglycosyl disulﬁdes, 243 2,3-dihydrobenzo[b]furan derivatives, 90–96 substituted, 92, 97 unsubstituted, 90–91 dihydro-β-agarofuran sequiterpene AC, 149 dihydroisocoumarin chromophore, 104–108 dinemasone B AC, 156 dinuclear complexes, 457–458 dinuclear ruthenium-based switches, redox-controlled monitored by electronic near-IR CD, 271–273 diols, 437 dioncophylline, 240 dioxibrassinin, 410 diphenyl dichalcogenides, 243 diselenides, 243 distamycin, 640–642 dithienylethenes (DET), 300, 301 Djerassi, Carl, 20, 22, 23 DNA strands, rich in guanine and adenine, 582 Drude, Paul, 10 drug and natural product binding, to nucleic acids, 635–636 CD analyses of berberine binding of DNA quadruplex, 658–660 CD analyses of cationic porphyrin binding selectivity to DNA quadruplex, 656–658 diagnostic electronic CD spectral changes, 637–640 DNA minor groove binding natural products actinomycin D, 647–649 calicheamcin and esperamicin, 643–645 cellular uptake and anticancer cc-1065 prodrug analogues target in live tumor cells, 642–643 chromomycin, 645–647 enantio-selective daunorubicin binding to DNA, 649 netropsin and distamycin, 640–642 porphyrin selective binding to B -and Z-form DNA, 652 cationic zinc porphyrin as Z-DNA probe, 652

846

drug and natural product binding, to nucleic acids (Continued) spermine-induced Z-form DNA recognition by tetraanionic nickel porphyrin, 654–655 drug binding detection, by circular dichroism, 666 drug discovery and electronic circular dichroism, 819–821 CD in analysis of recombinant protein therapeutics, 833–838 protein–ligand interactions binding constant measurement by dilution experiments, 827–828 binding constant measurement by titration experiments, 824–827 conformational transitions of peptides and proteins, 828–833 ligand binding monitoring to proteins, 821–824 dye interactions with PNA–XNA duplexes, 606–607 cyanine dye aggregation, 608–609 cyanine dyes in PNA–PNA duplexes, 609–610 forced intercalation, 610 dynamic systems, 289–290 chemical and mechanical systems, 308–314 photoactive systems, 294–307 thermal systems, 291–294 E -isomer, 413 electric transition dipole moments (ETDMs), 117–118, 121, 129, 145, 212 electric transition moment, 124 electronic circular dichroism (ECD), 423. See also individual entries metal-based switches monitored by, 256 dynamic stereochemistry in biological polymers, 256–257 dynamic stereochemistry in synthetic polymers, 257–258 electron-switched supramolecular chiral polythiophene aggregates, 258 environment-induced switches, 258–265 photochemically triggered chiral metal switches, 273–277 redox-triggered switches, 265–273 electron-switched supramolecular chiral polythiophene aggregates, 258 enantiomers, 392 enantiopure compounds, 137–138 endo-borneol and derivatives, 398–399 endoperoxides, 407–408 enediynes, 643 Enfurvitide (Fuzeon®), 832 enones, 39–51 environment-induced switches, 258–259 amide complexes anion-controlled switching, 260–263 guest-controlled tripodal ligand chirality, 259 Pfeiffer effect practical application, for analyzing chiral diamines, 260 solvent-controlled switching of metal complexes, 263–265 zinc porphyrin tweezers, 259 8-epiisolippidiol-3-O-β-d-glucopyranoside (3β,8β,dihydroxy-1α, 4β, 5α, 6β, 7α, 11β-guai-10(14)-ene-6,12-olide-3-O-β-d-glucopyranoside), 406 9-epi-presilphiperfolan-1-ol, 409 Epogen®, 836 Eprex®, 836

INDEX

equatorial orientation, 79, 83, 87, 92–93, 97–98, 104–105, 108 eremophilanoids, 408–409 eremophilanolide, 409 Escherichia coli, 623, 822 esperamicin, 643–645 ethylenediaminetetraacetate (EDTA), 255 eudesamanolides, 409 exciton chirality method, 319 applications acetylenealcohols AC assignment, 157–158 to acyclic 1,2-glycols and polyols, 140–142 allylic benzoate exciton method for allylic alcohol AC determination, 146–147 axially chiral binaphthoquinones AC, 152–153 cephalocyclidin A AC, 148 ciguatoxin ACs and related compounds, 147 cortistatin A AC, 155 11-deoxydiaporthein A AC by exciton allylic benzoate method, 155 1,3-diethynylallene compounds ACs, 158 dihydro-β-agarofuran sequiterpene AC, 149 dinemasone B AC, 156 exciton coupling between polyacene and related chromophores, 137–140 gymnocin-B AC, 151–152 kolokoside A aglycone, 156 leucettamol A AC, 153–155 nondegenerate system, 143–145 oligonaphthalenes AC, 156–157 phomopsidin AC, 150 phorboxazole AC, 151 pinellic acid AC, 150–151 pre-anthraquinones AC, 153 spiroleptosphol AC, 153 spiroxin A AC strevertenes AC, 149–150 trans-acenaphthene-1,2-diol AC, 155–156 urothion AC, 148 UV λmax chromophore separation versus exciton CD, 145–146 CD spectra and CE rotational strength, 122 consistency with X-ray Bijvoet method, 127–129 deﬁnition, 120 examples, 129 benzamide, phthalimide, and 2,3-naphthalenedicarboximide chromophores, 131–132 carboxylic acids and oleﬁn compounds, 132 cinnamate and β-naphthoate, 130 para-substituted benzoate chromophores for glycols, 129–130 red-shifted chromophores, 132–133 tetraphenyl-porphyrin-carboxylic acid (TPP-COOH), 130–131 historical review, 115–117 molecular exciton theory of binary system with two chromophores, 123–125 N -mer and dimer and, 126–127 porphyrin tweezers method and, 135–137

INDEX

preexisting chromophores in natural products and, 133–135 substituted benzene and polyacene chromophores, 134 principle of, 117–118, 119 nonempirical nature, 118–122 theoretical analysis and caution antitumor antibiotic AT2433-A1 AC with secondary amino group, 159 1,1 -biphenanthryl compounds anomalous CD CEs, 159–161 theoretical simulations, 127 extrinsic chromophores, 485 Eyring, H., 4, 116 Fagonia cretica, 233 fatty acids, 688–690 Fehling’s solution, 7 fenchone, 398 Fischer, E., 115 ﬂavanones, 108 ﬂuorescence-detected circular dichroism (FDCD), 252 dynamic stereochemistry monitored by, 278–279 F¨orster resonance energy transfer (FRET), 515 4f –4f transitions, ECD in, 459–462 Fourier Transform and dispersive instruments, 711–716 frontalin (1,5-dimethyl-6,8-dioxabicyclo[3.2.1]octane), 413 furanocoumarins, 412 furanones, 411–412 furochromones, 408 fused heterocyclic ring, benzene chromophore with, 81–83 benzodioxane chromophore, 83–85 chroman chromophore, 96–100 2,3-dihydrobenzo[b]furan derivatives, 90–96 isochroman chromophore, 85–90 fusidilactone B, 233 Fusidium sp., 233 γ –cyclodextrin, 338 γ-turns, 518–522 garcinia acid, 393–396 Garcinia cambogia, 395 gauche effect, 52, 142 Gaussian distribution, 169 ginkgolides, 414–415 globosuxanthone, 228 glycoconjugates. See also carbohydrates biological and structural basis of, 796–800 underivatized sugars and, 806–808 glycoproteins, 784 Gnidia involucrate, 377 gossypol, 416 G-quadruplexes, 656–657, 659–660 Gracilaria folifera, 229 green ﬂuorescent protein (GFP), 488 Grossmann, H., 13 guanine derivative-platinum(II) complexes, 466–468 guanine quadruplexes, 579–581 guest-controlled tripodal ligand chirality, 259 gymnocin-B AC, 151–152

847

Habropetalum dawei, 381 H-aggregates, 309 halenaquinol family marine natural products, absolute stereochemistry of, 172–176 halogenated sesquiterpenes, 407 Hantzsch, Arthur Rudolf, 14, 16, 27 Harada, N., 117 helix-coil equilibria, 501 helix-coil transition, 576 helix reversal (HRP), 293 heme proteins, 489–490 1,1,1,3,3,3-hexaﬂuoroisopropanol (HFIP), 500–501 hexylitaconic acid and methyl esters, 397 hibiscus acid, 395–396 high-performance liquid chromatography (HPLC), 356 CD spectra interpretation, 362–365 chromatographic system choice, 357–360 HPLC-CD absolute axial conﬁguration of knipholone and knipholone anthrone, 370–377 ancistrocladium B, 365–370 β,β -bisporphyrins, 378–379 bi[10]paracyclophane, 379–381 detection wavelength choice, 360–361 device, 356–357, 357 phylline, 381–382 quantitative analysis and EE determination using, 361–362 software recommendations, 382–383 UV shift application, 365 hinokiresinol, 413 hirudin, 831 homeodomain transcription factors, 628 Htelo DNA, 657 human serum albumin (HSA), 666–667 benzodiazepines, 678–685 bilirubin, 667–670 curcumin, 675–678 fatty acid binding, 688–690 nonsteroidal anti-inﬂammatory drugs (NSAIDs), 685–688 quercetin, 672–675 warfarin, 670–672 Humulin®, 834 1-hydroxy-15-acetoxyeudesm-11(13)-en-6,12-olide, 409 6-hydroxyeuryopsin and acetyl derivative, 408 hydroxyﬂavanones, 108 6β-hydroxyhyoscyamine, 401–402 HYPERCHEM program, 393 ibuprofen, 668, 670, 672, 677, 685–686, 688 imatinib mesylate, 697 immunoglobulin (IMUN), 737 i-motifs, 603–604 incident circular polarization (ICP) ROA, 759, 760, 762–764 induced circular dichroism (ICD), 258, 307, 635, 637, 640, 643, 821 induced rotatory contribution, 75, 79 inorganic stereochemistry, electronic circular dichroism applications to, 451–452

848

inorganic stereochemistry, electronic circular dichroism applications to (Continued) atropisomerism and ECD, 465–466 guanine derivative-platinum(II) complexes, 466–468 nitrito-κ-O complexes, 466 d –d transitions of tris- and bis-bidentate transition metal complexes, 452–456 theoretical studies on structure-spectra relationships, 456–457 ECD in 4f –4f transitions, 459–462 exciton ECD in intraligand transitions β-diketonato lanthanide(III) complex, 464–465 tris- and bis-chelate complexes, 462–463 polynuclear complexes with conﬁgurational chirality dinuclear complexes, 457–458 tetranuclear complexes of hexol type, 458–459 insulin, 738, 834 intercalators, 638 interferons, 836–837 intraligand transitions, exciton ECD in β-diketonato lanthanide(III) complex, 464–465 tris- and bis-chelate complexes, 462–463 intrinsically disordered proteins (IDPs), 486–487 iridoids, 403–404 iron translocation, in triple-stranded helical complexes, 265–266 isochroman chromophore, 85–90 isoepitaondiol diacetate, 404–405 isoﬂavans, 98–99, 100 iso-schizozygane alkaloids, 401 Jaeger, Franciscus Mauritius, 15, 28 J-aggregates, 309, 313, 334 Jasco, 713 Kauzmann, W., 4, 116 Kirkwood, J. G., 4, 116 klaivanolide, 412–413 Klecki, L´eon, 26 knipholone and knipholone anthrone, absolute axial conﬁguration of HPLC-CD instructive examples, 376–377 by quantum mechanical calculations, 374 remeasuring CD spectra, 370–373 stereochemical correlations, by coelution experiments, 374–376 Kodaka rule, 336 kolokoside A aglycone, 156 Krameria cystisoides, 92 Kramers–Kronig (KK) transforms, 424 Kuhn, W., 9, 11–12, 16, 20, 21, 29, 116 Kundt, August, 5 l-Ala-rich peptides, 531–532 Lambert–Beer law, 827, 828 laminarin, 783 Landolt, Hans, 5, 10 Larix decidua, 234, 242 Laurencia sp., 407

INDEX

Le Bel, J. A., 115 lepirudin, 831 Leroux, F.-P., 5 leucettamol A AC, 153–155 l-histidine chelation, 255 Lifschitz, Israel, 15–16, 18, 22, 27–29, 28 light-powered molecular motors discovery and development of, 178 absolute conﬁguration internal reference in X-ray analysis, 185–191 enantiopure chiral oleﬁns and CD spectra, 179 unexpected thermal racemization of cis-oleﬁn 13 unique photo-and thermochemical behavior of chiral dimethyl oleﬁn, 191–194 with high rotation speed, 194–196 continuous rotation, 204 motor rotation dynamics, 202–203 synthesis, CD spectra, X-ray structure, and absolute stereochemistry, 198, 205 unstable motor rotation isomers and CD spectra, 198 limonene, 399 limonene oxide, 398 linear momentum, 124 Lippia integrifolia, 406 lippifoliane, 406 Lippmann, Gabriel, 10, 30 Lobophytum crassum, 229 Loeb, A., 13 longipinane derivatives, 410 Lowry, T. M., 4, 9, 18, 22, 26–27, 31 Lycium intricatum, 229 MACROMODEL program, 393 macromolecule chiral conformation of, 326–327 chirality induction in achiral polymer, 330–332 chiral memory in polymer, 333–334 chiral molecular recognition with polymer, 334–335 inherent, 327–330 macropodumine B, 231 macropodumine C, 231 magnetic moment operator, 124 magnetic transition dipole moment (MTDM), 211 magnetic transition moment, 125 majapolene B, 407 majority-rules effect, 291–293, 328–329, 342 Mason, S. F., 117 Mathieu, J.-P., 29 McDowell, M. F., 12 Meliotus dentatus, 229 membrane proteins, 487–489 Mentha arvensis, 233 menthene, 399 menthenol, 399 Merck Molecular Force Field (MMFFs), 136 merocyanine modiﬁed poly(l-glutamate) (MEPGA), 299–300, 301 meroditerpenoids, 404–405 meso-compounds, 379 meta-ethynylpyridine polymers, 334

849

INDEX

metallo-organic compounds, 251–253 dynamic stereochemistry monitored by CPL, 280 monitored by FDCD and CPE, 278–279 monitored by VCD, 277–278 metal-based switches monitored by ECD, 256 dynamic stereochemistry in biological polymers, 256–257 dynamic stereochemistry in synthetic polymers, 257–258 electron-switched supramolecular chiral polythiophene aggregates, 258 environment-induced switches, 258–265 photochemically triggered chiral metal switches, 273–277 redox-triggered switches, 265–273 metal-binding induced switches studies, by optical rotation and ORD, 253 conformational studies by ORD, 255 dynamic stereochemistry in biopolymer monitored by optical rotation and optical rotatory dispersion, 255–256 metal-based chiroptical switches controlled by polarized light, 254 octahedral complexes of transition metals, 253 solid-state metal-based chiroptical switches photo-induced switching, 281–283 pressure-based switches, 280–281 temperature-induced dynamic stereochemistry, 281 1-methoxyspirobrassinin, 411 1-methoxyspirobrassinin methyl ether, 411 2-methylchroman-4-one, 108 5-O-methylvisamminol, 408 derivatives, 408 M -helicity, 78, 84–86, 93, 97–98 Michelson interferometer, 711 microcrystalline pellet and disc method, 224 Microdiplodia sp., 228 Microsphaeropsis sp., 229, 234, 242 Mills’ rule, 146 mismatch penalty (MMP), 293 Mitchell, S., 9, 11, 19 Mofﬁtt, W., 116 molecular absolute conﬁguration determination, 421–422 approaches, 426–430 chromophoric molecules, 430–439 transparent molecules, 439–443 electronic circular dichroism, 423 optical rotation and optical rotatory dispersion, 422–423 vibrational circular dichroism, 424–426 molecular assembly, chiral chiral hetero-aggregate, 342–344 chiral homo-aggregate, 342 chiral memory in supramolecular assembly, 344 molecular exciton theory of binary system, with two chromophores, 123–125 molecular modeling, 231 MOLSCRIPT diagrams, 766 molten globule (MG), 485 monoterpenes and derivatives, 397 montanine-type alkaloids, 403

Mouton, Henri, 31 multivariate analysis of protein ROA spectra, 777–778 Murraya alternans, 412 mutarotation, 17 myoglobin (MYO), 737 myrtenal, 399 Nakanishi, K., 117 naphthalene, 140 naphthyridine, 320 naproxen, 685 narcissus mosaic virus (NMV), 787 Natanson, L., 10–11, 21, 26 natural products, absolute conﬁguration assignment of, 217–218, 242–243 computational method choice, 221–222 conformation and conﬁguration, 218–219 solid-state ECD/computational approach origin, 219–220 solid-state ECD/TDDFT approach applicability, 225–226 application examples, 227–234 crystalline intermolecular couplings approach, 235–242 principle, 222–224 solid-state CD measurements, 224–225 natural products, preexisting chromophores in, 133–135 natural products structure determination, using VCD, 387–389 absolute conﬁguration determination, 395 applications alkaloids, 400–401 axially chiral natural products, 416–417 carboxylic acids, 395 cruciferous phytoalexins, 410–411 endoperoxides, 407–408 eremophilanoids, 408–409 eudesamanolides, 409 furanocoumarins, 412 furanones, 411–412 furochromones, 408 ginkgolides, 414–415 halogenated sesquiterpenes, 407 iridoids, 403–404 klaivanolide, 412–413 longipinane derivatives, 410 meroditerpenoids, 404–405 monoterpenes and derivatives, 397 montanine-type alkaloids, 403 norlignan, 413–414 peptides, 415–416 pheromones, 413 presilphiperfolanes, 409 sesquiterpenes, 405–406 taxol, 414 tropane alkaloids, 401–403 verticillane diterpenoids, 405 experimental VCD spectra, 389, 397 quantum chemical VCD predictions, 392 conformational analysis, 393–394 molecules with multiple chiral centers and chirality, 395 molecules with single source of chirality, 395

850

netropsin, 640–642 neutral-to-base (N–B) transition, 668, 670 Nishimoto–Mataga equation, 169 nitrito-κ-O complexes, 466 N -methylation, 548–549 Nodulisporium sp., 108 nondegenerate system, 143–145 nonlinear mapping (NLM), 778 nonsteroidal anti-inﬂammatory drugs (NSAIDs), 685–688 nopinone, 399 norlignan, 413–414 n –π ∗ transition, 476 nucleic acids, 575, 718, 784–786 A-form and B–A transition, 577–578 B/A conformation of (dC)n • (dG)n DNA sequences, 578 B-form and hairpin, 576 cytosine quadruplexes, 581–582 denatured DNA, B-form, and helix-coil transition, 576, 577 DNA strands, rich in guanine and adenine, 582 empirical RNA and DNA VCD and correlation with helical form and applications, 719–722 extensive changes in poly(dA-dT) circular dichroism, 583 guanine quadruplexes, 579–581 modeling of nucleic acid VCD, 722–723 poly(amino2 dA-dT) and poly(dG-methyl5 dC), 583 Z-forms and B–Z transition, 579 5-octadecyloxy-2-(2-thiazolylazo)phenol (TARC18), 312 octahedral complexes of transition metals, 253 octant rule, 363, 431 Oettingen, A. von, 26 oleﬁns, 132, 178 absolute stereochemistry of chiral, 182 enantiopure chiral, 179 unexpected thermal racemization of cis-, 184 unique photo-and thermochemical behavior of chiral dimethyl, 191–194 oligonaphthalenes AC, 156–157 oligopeptides, 726 Olmstead, L. B., 12 online stereochemical analysis, of chiral compounds. See High-performance liquid chromatography (HPLC) optical rotatory dispersion, 22–25 curves, 7–9 and optical rotation, metal-binding induced switches studies by, 253 conformational studies by ORD, 255 dynamic stereochemistry in biopolymers, 255–256 metal-based chiroptical switches controlled by polarized light, 254 octahedral complexes of transition metals, 253 Oseen, C. W., 22 osoplumericin, 403 Ostwald, Wilhelm, 9 oxacephams, 57, 59, 61 3-oxo-1,8-cineole, 399 pacifenol-related compounds, 407 paclitaxel (Taxol), 414, 670

INDEX

palmarumycin M1 , 242 papyracillic acid A, 234 para-substituted benzoate chromophores for glycols, 129–130 Pasteur, Louis, 115 Pauthenier, Marcel, 27 penams, 62–63 penicillins, 61–63 Penicillium sp., 85 peptide chromophore, 533–535 peptide nucleic acids, 587 duplexes and triplexes, with DNA and RNA, 595–596 PNA–DNA and PNA–RNA duplexes, 596–601 PNA–DNA–PNA triplexes, 601–603 PNA in quadruplexes and i-motifs, 603–604 dye interactions with PNA–XNA duplexes, 606–607 cyanine dye aggregation, 608–609 cyanine dyes in PNA–PNA duplexes, 609–610 forced intercalation, 610 single strand, and PNA–PNA duplexes, 589–595 thermal denaturation studies, 604–606 peptides, 415–416, 499 α-helix, 500–504 β-sheets, 512–516 β-turns, 516–518 γ-turns, 518–522 poly(l-pro) helices and collagen triple helix, 522–530 310 -helix, 504–512 unordered conformation, 530–536 peptide VCD studies, 723 applications, 727–732 empirical correlation with secondary structure, 724–727 theory, 732–735 peptidomimetics, electronic circular dichroism of, 545 α, β-didehydro -α-amino acid-based polypeptides, 556–562 fully extended peptide conformation (2.05 -helix), 546–548 poly-β-peptides, 562–569 poly-depsipeptides, 554–556 poly-N (alkyl)-α-amino acids, 548–550 poly-peptoids, 550–553 perillaldehyde, 398 pexiganan, 415 Pfeiffer effect, 258–259, 459 practical application, for analyzing chiral diamines, 260 Pﬂeiderer, W., 13 Z Phe compounds, 558–559 P -helicity, 78, 86, 93, 97–98, 105, 108, 176 phenprocoumon, 670 phenylacetylene, 140 phenylbutazone, 674 phenylpropanoids, 413 pheromones, 413 Pholidota chinensis, 377 Phoma sp., 229, 233, 235 phomopsidin AC, 150 Phomopsis sp., 106, 219 phomoxanthone, 220 phorboxazole AC, 151 photoactive systems, 294–307 photochemically triggered chiral metal switches, 273–274

851

INDEX

azobenzene-based molecular scissors, 274–275 chirality transfer via ternary complex, 275, 277 host-controlled guest chirality, 275 photo-induced switching, 281–283 photoswitchable chiroptical DNA complex, 306 phylline, 381–382 π -electron compounds, CD spectra of chiral extended, 167 absolute conﬁguration of chiral C60 -fullerene cis-3 bisadducts, determined by X-ray crystallography and CD spectra, 207, 209 absolute stereochemistry and CD spectra, of alleno-acetylenic macrocycle, 209–212 light-powered chiral molecular motor, with high rotation speed, 194–196 continuous rotation, 204 motor rotation dynamics, 202–203 synthesis, CD spectra, X-ray structure, and absolute stereochemistry, 198, 205 unstable motor rotation isomers and CD spectra, 198 light-powered molecular motors, discovery and development of, 178 absolute conﬁguration internal reference in X-ray analysis, 186–191 enantiopure chiral oleﬁns and CD spectra, 179 unexpected thermal racemization of cis-oleﬁn 13, 184 unique photo-and thermochemical behavior of chiral dimethyl oleﬁn, 191–196 SCF-CI-DV MO method absolute stereochemistry of chiral oleﬁns using, 182 established examples, 169 theoretical calculation of CD and UV spectra by, 168–169 pinellic acid AC, 150–151 Piper regnelli, 92, 95 Pleione yunnanensis, 95 plumericin, 403 polarizable continuum model (PCM), 392, 733 polarized light, metal-based chiroptical switches controlled by, 254 poly(allylamine hydrochloride) (PAH), 310 poly(amino2 dA-dT), 583 poly(dA-dT) circular dichroism, extensive changes in, 583 poly(dG-methyl5 dC), 583 poly(hexyl isocyanate) (PHIC), 296–297 poly(isocyanates) (PIC), 296–298, 299 poly(l-Ala), 500–501 poly(l-Lys) β-sheet, 514 poly(l-lysine), 766 poly(l-pro) helices and collagen triple helix, 522–530 poly(β-benzyl l-aspartate), 773 polyacene and related chromophores, exciton coupling between, 137–140 poly-depsipeptides, 554–556 poly-N (alkyl)-α-amino acids, 548–550 polynuclear complexes with conﬁgurational chirality dinuclear complexes, 457–458 tetranuclear complexes of hexol type, 458–459 polypeptides, unordered, 478–479, 530 poly-peptoids, 550–553 polyproline (II) helix, 774–777

polythiophene, 332 electron-switched supramolecular chiral aggregates, 258 Pope, W. J., 27 porphyrin tweezers, 135–137 redox-triggered, 271 potato virus X (PVX), 787 pre-anthraquinones AC, 153 presilphiperfolanes, 409 pressure-based switches, 280–281 principal component analysis (PCA), 778 principal component method of factor analysis (PC/FA), 740–741 Prionosciadium thapsoides, 408 prismatomerin, 403–404 progesterone, 697 property transfer method, of Bour, 734–735 propranolol, 692 Pro–Pro bonds, 523 PROTA software, 741 protein–nucleic acid interactions, circular dichroism of, 615 structural changes DNA bending, 619–621 DNA quadruplex alterations, 618–619 nucleotide level interactions, 623–624 protein aromatic residue effects, 629–630 protein secondary structure alterations, 624–629 single-stranded nucleic acid distortions, 621–623 transitions between DNA secondary structures, 616–618 proteins, electronic circular dichroism of, 475–476 applications heme proteins, 489–490 membrane proteins, 487–489 protein folding, 485–487 retinal proteins, 490–491 computer resources, 491 extrinsic chromophores, 485 peptide backbone contributions, 476 amide chromophore, 476 CD of secondary structural elements, 477–479 protein CD general aspects, 476–477 secondary structural analysis of proteins, 479–483 protein side chains, 483 far-UV CD, 484–485 near-UV CD, 483–484 protein-binding drugs, in blood, 665 protein VCD, 735–736 qualitative spectral interpretations, 736–738 secondary structure quantitative analyses, 740–743 structure and applications, 738–740 Pseudoanguillospora sp., 89, 417 pseudo-axial orientation, 98 Pseudomonas cepacia, 87 Pt-bridged cofacial diporphyrins via carbon-metal σ bonds, 271, redox-switchable pulegone (14)–CDCl3 complex, 400 pullulan, 783 quadrant rule, 74 quadrone, 405

852

quasi-axial orientation, 79 quasi-equatorial orientation, 685 quassin, 134 quercetin, 672–675 Raman optical activity (ROA), 759–760 biomolecules, 765–766 carbohydrates, 782–784 glycoproteins, 784 nucleic acids, 784–786 ROA signatures of polypeptide and protein ROA spectra, 766–782 viruses, 786–787 observables, 761–762 spectra calculation, 762–765 random coils, 727 recombination activating protein (RAG1), 628–629 redox-triggered switches, 265 chiroptical tripodal ligands, 266–269 iron translocation in triple-stranded helical complexes, 265–266 redox-controlled dinuclear ruthenium-based switches monitored by electronic near-IR CD, 271–273 redox-controlled molecular ﬂipper, based on chiral CU complex, 269–271 redox-switchable Pt-bridged cofacial diporphyrins via carbon-metal σ bonds, 271 redox-triggered porphyrin tweezers, 271 red-shifted chromophores, 132–133 Regular Iletin I®, 834 Regular Iletin II®, 834 reprotonation, 309 resonance light scattering (RLS) spectroscopy, 653 restricted multiple regression (RMR) analyses, 741 retinal proteins, 490–491 (R) isoﬂavanones, 108 Rosenfeld, L´eon, 22, 116 Rupe, Hans, 16 R-warfarin, 670–672, 684 S.ﬂabelliforme, 404 S. grahamii, 402 sargaol acetate, 404 scattered circular polarization (SCP) ROA, 759, 760, 762–764 SCF-CI-DV MO method absolute stereochemistry of chiral oleﬁns using, 182 established examples, 169 theoretical calculation of CD and UV spectra by, 168–169 Schellman, J. A., 117 schizozygane alkaloids, 400 Schizozygia caffaeoides, 400, 401 seco-drugs, 643 sector rule, 74, 76, 78, 336, 454 Senecio kleinii, 235 Senecio toluccanus, 408 sergeant-and-soldiers effect, 291–293, 328–329, 342 Servant, R., 19 sesquiterpenes, 405–406

INDEX

SHELXL, for structure reﬁnement, 223 side chains, 483, 773–774 far-UV CD, 484–485 near-UV CD, 483–484 Snatzke’s helicity rule, 78 solid-state metal-based chiroptical switches photo-induced switching, 281–283 pressure-based switches, 280–281 temperature-induced dynamic stereochemistry, 281 solvent-controlled switching of metal complexes, 263–265 Sonchus pinnatus, 402 Soret, Jacques-Louis, 5 Soret band, 130, 489 SPARTAN program, 401 spiroaromatics, 138, 139 spirobrassinin, 411 spiroleptosphol AC, 153 spiropyran modiﬁed poly(l-glutamate) (SPPGA), 299, 300 spiroxin A AC spontaneous symmetry breaking, in supramolecular system, 345–347 Staphylococcus aureus, 822 Stevia monardifolia, 410 Strempeiopsis strempelioides, 400 Streptococcus sanguis, 805 Streptomyces griseus, 645 Streptomyces plicatus, 645 Streptomyces zelensis, 642 strevertenes AC, 149–150 structural glycobiology, 782 stypotriol, 405 substituted benzene and polyacene chromophores, 134 sulﬁnpyrazone, 674 Sulfolobus acidocaldarius, 619 supramolecular complexation with chiral molecular host, 335 binding with cyclodextrin, 336–339 biomolecule binding, 339–342 supramolecular systems, electronic circular dichroism of, 317 chiral conformation of macromolecule, 326–327 chirality induction in achiral polymer, 330–332 chiral memory in polymer, 333–334 chiral molecular recognition with polymer, 334–335 inherent, 327–330 chirality sensing with achiral chromophore, 318–319 through coordination, 321–324 through hydrogen-bonding interaction, 319–321 through noncovalent and covalent interactions, 324–326 chiral molecular assembly chiral hetero-aggregate, 342–344 chiral homo-aggregate, 342 chiral memory in supramolecular assembly, 344 spontaneous symmetry breaking in supramolecular system, 345–347 supramolecular complexation with chiral molecular host, 335 binding with cyclodextrin, 336–339 biomolecule binding, 339–342 S -warfarin, 670–671, 684 synthetic polymers, dynamic stereochemistry in, 257–258

853

INDEX

tartrates, 12–14 taxol, 414 Taxus brevifolia, 414 temperature-induced dynamic stereochemistry, 281 tenoxicam, 688 tetra-anionic nickel porphyrin, 654–655 tetralins with achiral substituents, 79–81 and tetrahydroisoquinolines, 77 tetralone derivatives, 101–104 tetranuclear complexes of hexol type, 458–459 tetraphenyl-porphyrin-carboxylic acid (TPP-COOH), 130–131 thermal denaturation studies, 604–606 thermal systems, 291–294 Thermus aquaticus, 623 310 -helix, 478, 504–512, 773 tiﬂuadom, 685 time-dependent density functional theory (TDDFT) method, 55, 64, 65, 76, 105–106, 221, 229, 236, 374, 431, 456, 463, 803 tobacco rattle virus (TRV), 780 toﬁsopam, 685 trans-acenaphthene-1,2-diol AC, 155–156 trans-diaxial conﬁguration, 87, 108 trans-eusiderin, 85 transmissible spongiform encephalopathies (TSE), 782 transparent (nonchromophoric) molecules presence of no chromophores but functional groups, 439–442 and no functional groups, 442–443 triﬂuoroacetic acid (TFA), 549 1,3,3-trimethyl-2-oxabicyclo[2.2.2]octan -5-one, 399 tripodal ligands, chiroptical, 266–269 tropane alkaloids, 401 3α,6β-tropanediol, 402–403 tryptophan, 696, 774 Tschugaev, Leo Alexandrovitsch, 17, 22, 25–26 unfolded proteins, 778–780 urothion AC, 148 UV λmax chromophore separation versus exciton CD, 145–146 van’t Hoff, J. H., 115 verticilla-3E,7E-dien-12-ol

verticillane diterpenoids, 405 [6,6]-vespirene, 138, 139 vibrational circular dichroism (VCD), 424–426, 707–711. See also natural products structure determination, using VCD biopolymer VCD survey, theoretical, 717–718 dynamic stereochemistry monitored by, 277–278 experimental methods, 711 Fourier Transform and dispersive instruments, 711–716 sampling considerations for biopolymer VCD, 717 nucleic acid VCD spectra and applications, 718 empirical RNA and DNA VCD and correlation with helical form and applications, 719–722 modeling of nucleic acid VCD, 722–723 peptide VCD studies, 723 applications, 727–732 empirical correlation with secondary structure, 724–727 theory, 732–735 protein, 735–736 qualitative spectral interpretations, 736–738 secondary structure quantitative analyses, 740–743 structure and VCD applications, 738–740 vibronic contribution, 75 vindoline, 134 Violle, J., 29–30 viruses, 786–787 Volk, H., 13 warfarin, 670–672, 674, 677, 684, 689, 693 Wedeneewa, N., 9, 13, 18 Werner, Alfred, 14 Werner complexes, 14–16 X-ray Bijvoet method, 116–117, 127–129, 178 X-ray diffraction analysis, 49 Zα domains, 616 Z-form DNA, 636 and B–Z transition, 579 probe, cationic zinc porphyrin as, 652–654 spermine-induced recognition, by tetra-anionic nickel porphyrin, 654–655 zinc ﬁnger transcription factors, 625–628 zinc porphyrin tweezers, 259 Z –isomer, 413

Comprehensive Chiroptical Spectroscopy, Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules (Volume 2)

Comprehensive Chiroptical Spectroscopy, Instrumentation, Methodologies, and Theoretical Simulations (Volume 1)

Comprehensive Natural Products II: Chemistry and Biology: Volume 2

Classics in Spectroscopy Isolation and Structure Elucidation of Natural Products

Comprehensive Natural Products II: Chemistry and Biology: Volume 4

Comprehensive Natural Products II: Chemistry and Biology: Volume 5

Comprehensive Natural Products II: Chemistry and Biology: Volume 3

Comprehensive Natural Products II: Chemistry and Biology: Volume 7

Comprehensive Natural Products II: Chemistry and Biology: Volume 6

Comprehensive Natural Products II: Chemistry and Biology: Volume 1

Comprehensive Natural Products II: Chemistry and Biology: Volume 9

Comprehensive Natural Products II: Chemistry and Biology: Volume 8

Organosulfur Chemistry: vol 2 Synthetic and stereochemical aspects

Rubber Analysis ї Polymers, Compounds and Products

Comprehensive Natural Products II Vol.2 Natural Products Structural Diversity-II

Comprehensive Bioactive Natural Products, Volume 8: Quality Control & Standardization

Comprehensive Bioactive Natural Products, Volume 4: Antioxidants & Naturaceuticals

Comprehensive Bioactive Natural Products, Volume 6: Extraction, Isolation & Characterization

Comprehensive Bioactive Natural Products, Volume 1: Potential & Challenges

Comprehensive Natural Products II Vol.7 Cofactors

Studies in Natural Products Chemistry, Volume 35

Phenazine Natural Products: Biosynthesis, Synthetic Analogues, and Biological Activity

The total synthesis of natural products Volume 2

Comprehensive Natural Products II Vol.9 Modern Methods in Natural Products Chemistry

Syntheses of Natural Products

Synthetic Applications of 13-Dipolar Cycloaddition Chemistry Toward Heterocycles and Natural Products

Synthetic Applications of 1,3-Dipolar Cycloaddition Chemistry Toward Heterocycles and Natural Products

Studies in Natural Products Chemistry: Bioactive Natural Products Part L

Studies in Natural Products Chemistry, Bioactive Natural Products

Intermetallic Compounds, Volume 3, Structural Applications of

Fibonacci and Gann Applications in Financial Markets: Practical Applications of Natural and Synthetic Ratios in Technical Analysis

Comprehensive Chiroptical Spectroscopy, Applications in Stereochemical Analysis of Synthetic Compounds, Natural Products, and Biomolecules (Volume 2)

Comprehensive Chiroptical Spectroscopy, Instrumentation, Methodologies, and Theoretical Simulations (Volume 1)

Comprehensive Natural Products II: Chemistry and Biology: Volume 2

Classics in Spectroscopy Isolation and Structure Elucidation of Natural Products

Comprehensive Natural Products II: Chemistry and Biology: Volume 4

Comprehensive Natural Products II: Chemistry and Biology: Volume 5

Comprehensive Natural Products II: Chemistry and Biology: Volume 3

Comprehensive Natural Products II: Chemistry and Biology: Volume 7

Comprehensive Natural Products II: Chemistry and Biology: Volume 6

Comprehensive Natural Products II: Chemistry and Biology: Volume 1

Comprehensive Natural Products II: Chemistry and Biology: Volume 9

Comprehensive Natural Products II: Chemistry and Biology: Volume 8

Organosulfur Chemistry: vol 2 Synthetic and stereochemical aspects

Rubber Analysis ї Polymers, Compounds and Products

Comprehensive Natural Products II Vol.2 Natural Products Structural Diversity-II

Comprehensive Bioactive Natural Products, Volume 8: Quality Control & Standardization

Comprehensive Bioactive Natural Products, Volume 4: Antioxidants & Naturaceuticals

Comprehensive Bioactive Natural Products, Volume 6: Extraction, Isolation & Characterization

Comprehensive Bioactive Natural Products, Volume 1: Potential & Challenges

Comprehensive Natural Products II Vol.7 Cofactors

Studies in Natural Products Chemistry, Volume 35

Phenazine Natural Products: Biosynthesis, Synthetic Analogues, and Biological Activity

The total synthesis of natural products Volume 2

Comprehensive Natural Products II Vol.9 Modern Methods in Natural Products Chemistry

Syntheses of Natural Products

Synthetic Applications of 13-Dipolar Cycloaddition Chemistry Toward Heterocycles and Natural Products

Synthetic Applications of 1,3-Dipolar Cycloaddition Chemistry Toward Heterocycles and Natural Products

Studies in Natural Products Chemistry: Bioactive Natural Products Part L

Studies in Natural Products Chemistry, Bioactive Natural Products

Intermetallic Compounds, Volume 3, Structural Applications of

Fibonacci and Gann Applications in Financial Markets: Practical Applications of Natural and Synthetic Ratios in Technical Analysis

Recommend Documents