Advances in Supramolecular Chemistry. Volume 6

ADVANCES IN SUPRAMOLECULAR CHEMISTRY Volume 6 ,, 2000 This Page Intentionally Left Blank ADVANCES IN SUPRAMOLECU...

Author: Gokel G.W.

28 downloads 1142 Views 24MB 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

ADVANCES IN SUPRAMOLECULAR CHEMISTRY

Volume 6

,,

2000

This Page Intentionally Left Blank

ADVANCES IN SUPRAMOLECULAR CHEMISTRY Editor: GEORGE W. GOKEL

Department of Molecular Biology and Pharmacology Washington University School of Medicine St. Louis, Missouri

VOLUME 6

9

2000

JAI PRESS INC.

Stamford, Connecticut

Copyright 92000 by JAI PRESSINC. 100 Prospect Street Stamford, Connecticut 06904-0811 All rights reserved. No part of this publication may be reproduced, stored on a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, filming, recording, or otherwise without prior permission in writing from the publisher. ISBN: 0-7623-0557-6 ISSN: 1068-7459 Manufactured in the United States of America

CONTENTS

LIST OF CONTRIBUTORS

ix

PREFACE

George W. Gokel

MOLECULAR SELF-ASSEMBLIES THROUGH COORDINATION: MACROCYCLES, CATENANES, CAGES, AND TUBES

Kumar Biradha and Makoto Fujita

CHIRAL SELF-ASSEMBLED STRUCTURES FROM BIOMOLECULES AND SYNTHETIC ANALOGUES

Martinus C. Feiters and Roeland J. M. Nolte

SPHERICAL MOLECULAR CONTAINERS: FROM DISCOVERY TO DESIGN

Leonard R. MacGillivray and Jerry L. Atwood

41

157

SYNTHETIC PEPTIDE RECEPTORS: NONCOVALENT INTERACTIONS INVOLVING PEPTIDES

Hans-J6rg Schneider, Frank Eblinger, and Mallena Sirish

RATIONAL DESIGN OF SYNTHETIC ENZYMES AND THEIR POTENTIAL UTILITY AS HUMAN PHARMACEUTICALS: DEVELOPMENT OF MANGANESE(II)-BASED SUPEROXIDE DISMUTASE MIMICS

Dennis P. Riley

DESIGNING ACTIVE SITES OF SYNTHETIC ARTIFICIAL ENZYMES

Junghun Suh

185

217

245

vi

CONTENTS

THE RELEVANCE OF SUPRAMOLECULAR CHEMISTRY FOR THE ORIGIN OF LIFE

Pier Luigi Luisi

INDEX

287 309

LIST OF CONTRIBUTORS

Jerry L. Atwood

Department of Chemistry University of Missouri-Columbia Columbia, Missouri

Kurnar Biradha

School of Engineering Nagoya University Nagoya, Japan

Frank Ebfinger

Fachrichtung Organische Chemie Universit~t des Saarlandes Saarbr0cken, Germany

Martinus C. Feiters

Department of Organic Chemistry University of Nijmegen Nijmegen, The Netherlands

Makoto Fujita

School of Engineering Nagoya University Nagoya, Japan

Pier Luigi Luisi

Institute of Polymers Zurich, Switzerland

Leonard R. MacGillivray

Steacie Institute for Molecular Sciences National Research Council of Canada Ottawa, Ontario, Canada

Roeland J.M. Nolte

Department of Organic Chemistry University of Nijmegen Nijmegen, The Netherlands

Dennis P. Riley

MetaPhore Pharmaceuticals, Inc. St. Louis, Missouri vii

viii

LIST OF CONTRIBUTORS

Hans-JC~rgSchneider

Fachrichtung Organische Chemie Universit~t des Saarlandes SaarbrCicken, Germany

Mallena Sirish

Fachrichtung Organische Chemie Universit~t des Saarlandes SaarbrCicken, Germany

Junghun Suh

Department of Chemistry and Center for Molecular Catalysis Seoul National University Seoul, Korea

PREFACE It has long been the goal of Advances in Supramolecular Chemistry to present a broad range of supramolecular science rather than to organize a focused volume. Monographs dedicated to a single subject have great and obvious value but they are inherently narrow. As the field of supramolecular chemistry has grown, a number of useful, focused monographs have appeared. Those desiring an up-to-date assessment of a particular area will clearly benefit from such volumes. In contrast, the intent in this series has always been to present an overview of scientific endeavors in the broad discipline called supramolecular chemistry that spans analytical, inoraganic, organic, physical, and biochemistry. In Volume 6 of the series, inorganic, organic, and bioorganic chemistry are represented in contributions from Germany, Japan, Korea, Switzerland, the United States, and The Netherlands. In the first chapter, B iradha and Fujita describe their pioneering work in self-assembled structures organized by the use of transition metals. Feiters and Nolte then describe their extensive studies of self-assembled structures formed from various biomolecules. MacGillivray and Atwood describe the formation of spherical molecular containers and their understanding of such structures based on Platonic and Archimedean solids. Schneider, Eblinger, and Sirish describe the fascinating family of synthetic peptide receptors and the interactions that can be explored using these host molecules. In the next chapter, Riley describes a mixture of computational chemistry, drug design, and synthetic organic and inorganic chemistry in the development of superoxide dismutase mimics. Suh

x

PREFACE

discusses the bioorganic and supramolecular principles required for the design of synthetic artificial enzymes. In the final chapter, Luisi discusses supramolecular self-assembly and its possible role in the origin of life. It is hoped that this broad, international view of supramolecular chemistry and the many directions it leads will be of interest to those already in the field. It is also hoped that those outside the field may see extensions of their own work that will bring them into it. G. W. Gokel Editor

MOLECULAR SELF-ASSEMBLIES THROUGH COORDINATION: MACROCYCLES, CATENANES, CAGES, AND TUBES

Kumar Biradha and Makoto Fujita

1. 2.

3.

4.

5.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Self-Assembling Tetranuclear Pd(II) Macrocycle . . . . . . . . . . . . . . . . . 2.1. Pd(II)-(4,4'-bipy) Squar e Complexes . . . . . . . . . . . . . . . . . . . 2.2. Stability of Square Complexes . . . . . . . . . . . . . . . . . . . . . . . 2.3. Accumulation of Pch Square Complexes . . . . . . . . . . . . . . . . . . Self-Assembling Macrocycles . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Structure and Molecular Recognition . . . . . . . . . . . . . . . . . . . 3.2. M6I.~ Macrotricyclic Complexes . . . . . . . . . . . . . . . . . . . . . . Self-Assembling [2]Catenane . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Catenane Formation from Preformed Rings: Molecular Magic Rings 4.2. Irreversible Formation of a [2]Catenane from Two Preformed Rings . 4.3. Self-Assembly of [2]Catenanes from Rectangular Molecular Boxes . 4.4. A Three-Dimensionally Interlocked Catenane . . . . . . . . . . . . . . . Self-Assembly of MaL2 Complexes . . . . . . . . . . . . . . . . . . . . . . . . 5.1. Guest-Induced Assembly of an M3L2 Complex . . . . . . . . . . . . . . 5.2. Guest-Selected Formation of Cages from a Dynamic Receptor Library

Advances in Supramolecular Chemistry Volume 6, pages 1-39. Copyright 9 2000 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 0-7623-0557-6

. . . . . .

. .

2 3 3 6 6 8 8 8 12 12 14 15 19 21 21 22

2

KUMAR BIRADHA and MAKOTO FUJITA

6. Self-Assembly of Hollow, Nano-Sized M6I_4Complexes . . . . . . . . . . . . 24 6.1. The Structure of Octahedral M6L4Cage Complexes [M = Pd(ll)] . . . . 24 6.2. Guest-TemplatedSynthesis of a Kinetically Stable M6L4 Cage [M = Pt(II)] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 6.3. Molecular Recognition of Large Guests by the M6I-4 Cage . . . . . . . 27 6.4. Formation of Hydrophobic Dimers in the M6I..4 Cage . . . . . . . . . . 27 6.5. Catalysis and Acceleration of Chemical Reactions in the M6L4Cage . . 29 7. A Coordination Nanotube . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 8. A Coordination Capsule Assembled from 24 Components . . . . . . . . . . . 33 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 References and Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

1. INTRODUCTION The aim of supramolecular chemistry is to devise wholly synthetic systems that function like their natural counterparts, by accepting, storing, processing, transferring, and disseminating information at molecular level. 1-3 Such systems can be constructed by a self-assembly process which refers to the spontaneous generation of well-defined structures from component molecules under a well-defined set of conditions. Even the most complex architectures can be easily and quantitatively synthesized using a self-assembly process. Often, the template effects play an important role in the self-assembly process and lead to the formation of complex systems with remarkable selectivities and also with very high yields. 4 That is, the template effects induce the self-assembly process both kinetically and thermodynamically. The key to a self-assembly process is the identification of weak and directional noncovalent interactions such as hydrogen bonds and metal-ligand coordination bonds. Earlier studies on self-assemblies focused on hydrogen bonds because of their presence and importance in biological systems. 5 Researchers have just begun to realize the potentialities of a coordination bond 6-8 in self-assembling various architectures such as helices, 6d'9'1~grids, II boxes, 12 rods, 13 tubes, 14 interlocked compounds, 6a'Sa'c'15 and so on. This chapter focuses on the self-assemblies of transition metals and various pyridine-based ligands. Generally, self-assemblies using coordination bonds have been designed using naked metal ions such as tetrahedral Cu(I), octahedral Fe(II), Co(II), Ni(II), and square-planar Pd(II). However, with such naked metals ions, it is often difficult to control the number and direction of the coordination of the ligands. If a metal ion is protected appropriately, the number of coordination sites will be reduced and hence the direction of the coordination of the ligands can be controlled easily. Furthermore, the protection of the metal gives rise to the design of simple self-assembly from nonsophisticated monodentate ligands (e.g. 4-pyridyl group). Accordingly, our strategy for designing self-assembled architectures is to use the cis-protected Pd(II) l a and/or Pt(II) complexes l b because the direction of the coordination with ligands involves right angles. The self-assembly takes place

Molecular-Coordinated Self-Assemblies

3

N - - M-'---ONO2

M2

J

ot~:k

a: M = Pd; b: M = Pt

without losing the Pd(en)/Pt(en) framework because of the dissociation of en ligand from the metal is negligible under ordinary conditions, whereas monodentate ligands undergo rapid dissociation. Based on such an idea, novel cis-protected Pd(II) and Pt(II) building blocks la and lb were designed. These metal building blocks have shown a remarkable ability in inducing a self-assembly process. This review focuses on our efforts in designing self-assemblies of a variety of discrete architectures using la or lb and several pyridine-based ligands. The designed architectures include squares, macrocycles, catenanes, cages, and tubes. Most of these self-assemblies were shown to be stable both in solution and solid and are characterized by proton nuclear magnetic resonance (1H NMR), electron spray mass spectroscopy (ESI-MS), 16 and X-ray crystallography. 0

SELF-ASSEMBLING TETRANUCLEAR Pd(ll) MACROCYCLE 2.1.

Pd(ll)-(4,4'-bipy) Square Complexes

The linear bifunctional and exodentate nature of 4,4"-bipyridine, 2, and 90 ~ coordination angle of la were used in combination to design a square complex, 3a. ]7 It was synthesized in quantitative yield by simply mixing la and 2 in 1:1 ratio in water (Scheme 1). This compound is the first example of molecular squares which have attracted considerable current interest. 18 The solution structure was confirmed using NMR and ESI-MS, whereas the solid-state structure was determined by X-ray crystallography. 19 The crystal structure of 3a revealed an almost perfect square structure with facial conformation of pyridine nuclei (Figure 1). The square cavity in 3a is of dimension ca. 8 x 8/~ and offers encapsulation of variety of aromatic guest molecules because of its hydrophobic nature. Thus, we found the ability of 3a to recognize several neutral aromatic molecules from water (Table 1).2~For example, 1,3,5-trimethoxybenzene was recognized with K a value of 7.5 x 102. The square structure 3a was expanded further by simply inserting spacers such as p-phenylene, - C ~ C - or - C - - C - units between the pyridine units of 2. Interestingly in expanded structures, the molecular square 4 was in equilibrium with the

4

KUMAR BIRADHA and MAKOTO FUJITA 8+

C

NH2

I

N - M----ONO2 H= ON~

I

N -"- M "H2 ~

H2

~

N' ~ ~

N -/-

~NH2

//'-I--H

/~'~N

(3 "1

2

H2

---M--- N

H2N~.~

2

(NO3")s

a: M = Pd; b: M = Pt Scheme 1.

trinuclear triangle, 5, depending on the concentration of the components (Scheme 2). 19Hong and coworkers reported recently that this equilibrium can be controlled by guest-induced molecular recognition. 21That is, the small guest molecules shifted the equilibrium towards molecular triangle 5b, whereas the large guest molecules shifted the equilibrium towards molecular square 4b.

Figure 1. The molecular square in the crystal structure of 3a.

Molecular-Coordinated Self-Assemblies

5

Table 1. Association Constants for the Various Self-Assembled Macrocycles and Guests

Host

Association Constant (L tool- 1)

Guest

3a

N-(2- naphth yl)acetam ide

1800 750 330 580 30 10 n.c. 260 550 20 420 130 200 201 2500 2580 560 80 30

1,3,5-trimethoxybenzene

p-dimethoxybenzene rn-dimethoxybenzene o-dimethoxybenzene p-bis(methoxymethyl)benzene 1,4-dimethoxycyclohexane p-dimethoxybenzene m-dimethoxybenzene o-dimethoxybenzene 1,3,5-trimethoxybenzene p-dimethoxybenzene p-dimethoxybenzene p-dicyanobenzene

3b

4b

13 14

Ref.

1,3,5-trimethoxybenzene

p-dimethoxybenzene p-bis(methoxymethyl)benzene p-dicyanobenzene p-dinitrobenzene

25 17 20 20 20 20 20 20 20 20 a a 37 37 25 25 25 25 25

Note:. aUnpublished results.

N -- Pd-'N

H~ ~

~

/)"~X--'-(x

~

x

I

x

4

H2N~ /NH2 Pd

_N- Pd'- N

H~

/\

4---

"(NO3)8

4

x

~......-N

H

2

~

x

~ 5

a: X =-C C-; b: X =-CH=CH-; c: X =-C C-C C-; d: X =-C6H4 -

Scheme 2.

H2N-.--J =(N03)6

6

KUMAR BIRADHA and MAKOTO FUJITA

2.2. Stability of Square Complexes The self-assembly of Pt(II) analogue lb with 2 was very slow due to the inactivity of the Pt(II)-pyridine bond. Thus, upon treatment of lb with 2, a kinetically distributed oligomeric mixture was initially formed. 22 However, the mixture gradually turned into the thermodynamically most stable molecular square 3b after heating the solution for a few weeks at 100 ~ The use of bis-nitrate salt of 2, instead 2 itself, dramatically increased the reaction rate as well as the yield (79-81%). Similar improvements in reaction rate and yield were observed by the addition of NaNO 3 to the reaction. A significant difference in stability was found between 3a and 3b. The addition of la to 3a in D20 resulted in dissociation of 3a to give a mixture of 3a (ca. 50%) and two acyclic components which have l a and 2 in 1:2 and 2:3 ratios. In a striking contrast, 3b remained intact even upon the addition of lb, as its structure had been locked by the inert Pt-Py bond. The square complex 3b was also found to show inclusion properties similar to 3a (Table 1).

2.3. Accumulation of Pd4 Square Complexes The square compounds 3a and 3b were used as building blocks to construct a one-dimensional staircase and tubular networks by bridging the corners of the squares with X - P t - X (X = C1, Br) units. 23 We anticipated these architectures in the mixed valence complexes which can be easily synthesized by simply mixing two metal complexes that have different valences. 24 Accordingly, the molecular square 3b was prepared and then mixed with [Pt(en)2Br2]C12 (6) in an aqueous solution of NaNO 3 (Scheme 3). The single crystals of 1:3 complex 7 of 3b and 6 were obtained and characterized by X-ray crystallography. The crystal structure shows one square moiety, 3b, and two moieties of 6 form a staircase like network, while the third moiety of 6 acts as a guest molecule and occupies in the square cavity of 3b (Figure 2). The Pt(II)-Br distance in 7 (3.22 and 3.37 ]k) indicates a weak interaction between Pt(II) and the Br atom. The geometry of the square in 7 is similar to that of the above-described square complex 3b. It is interesting to note that a cationic

~

~ . - u

N

N

-'1 8+

_~~~~

i H2E~rH2

(NO,j')e

3b M-

~..N,r'I'..N~

1-12er I-I2

I

6=M'

Scheme 3.

e

e

,

Molecular-Coordinated Self-Assemblies

7

Figure 2. Staircase network in the crystal structure of 7. Note that one moiety of 6, represented in space filling mode, occupied in the scare cavity of 3b.

species 6 was encapsulated by a cationic square 3b because of the perfect match of their shape and size. It indicates that an unfavorable electrostatic effect was overridden by a favorable shape factor. Similarly, the accumulation of the squares 3a or 3b were also observed via anionic complex [PTC1412-(8). The formation of complex (3a.(8)4) in 94% yield was observed when the aqueous solutions of 3 and 8 were combined at any given ratio (Scheme 4). The product was characterized by UV-vis and elemental analysis. The UV-vis shows new absorption peaks at 390 and 490 nm, which indicates the existence of an interaction between Pd(II) and Pt(IV). The elemental analysis of the complex confirms ratio of 3a and 8 is 1:4. These observations suggest that the structure of this complex could be a polytube 9 in which each corner of the square is linearly bridged by the linear Br-Pt-Br motif. It seems logical that the accumulation of the cationic squares with anionic species leads to a tubular network rather a staircase network because of favorable electronic effects.

N

N

_

--1

_

'0.

'§ lilt'

,

M,'

'.

ir

'.

if'

br

:

"

:

~..,~1

ci" c,

(NOs") 8

M=

.

I

8=M'

Scheme 4.

8

KUMAR BIRADHA and MAKOTO FUJITA 3.

SELF-ASSEMBLING M A C R O C Y C L E S

3.1. Structure and Molecular Recognition In the previous section we have shown that the rigid bifunctional ligands can form molecular squares or triangles. Here we show that the flexible bidentate ligands can form binuclear macrocycles and discuss their ability towards the inclusion of guest molecules. The macrocycles 10-18 were prepared in quantitative yields by simply mixing the corresponding ligands and la in 1:I ratio in water. 25-28 Structures 10, 12, 17, and 18 were fully characterized by X-ray crystallography (Figure 3) and the other structures were characterized by fast atom bombardment mass spectrometry (FABMS). The guest inclusion properties of these macrocycles were studied using IH NMR. It is interesting to note that some of them showed a unique ability for molecular recognition. For example, macrocycle 14 having two tetraflurophenylene units exhibited a remarkable molecular recognition ability for electron-rich aromatic compounds. Thus, the association constants increased with increasing electron density of the guest molecules (see Table 1). The inclusion geometry of these macrocycles was suggested from the crystal structure of the related infinite Cd complex {[Cd(kt-13)2](NO3)2.(p-NH2C6HaNO2) }n, 19, in which the macrocyclic framework of 13 repeats one-dimensionally (Figure 4). 29 Efficient edge-face aromatic interactions in the cavity are probably a driving force for high efficient molecular recognition. 3.2. M6L4 Macrotricyclic Complexes Here we discuss the formation of macrotricycles by exotridentate ligands and also discuss their remarkable ability to bind dicarboxylates. Treatment of the Pd(II) complex, la, with the ligands 20 and 21 resulted in the self-assembly of macrotricyclic frameworks 22 and 23, respectively (Scheme 5). 30 The formation of 22 and 23 were characterized in solution by IH NMR and in the solid state by X-ray crystallography (Figure 5). Further, the addition of NaC10 4 to the aqueous solution of 22 or 23 resulted in the precipitation of corresponding CIO 4 salts in 85% yield. These frameworks were held together by 10 molecular components (six metal ions and four ligands) and the most significant feature is that they have nanometer dimensions in spite of the small size of the molecular components. The X-ray structures show 22 and 23 have molecular dimensions of ca. 30 x 23 x 22/~ and 27 x 24 x 14/~, respectively, and have the same topologies. Although they have same topologies, their cavity shapes differ significantly: 22 has an elliptical cavity, while 23 has a bowl shape cavity. The proposed pathway leading to formation of 22 or 23 involves a two-step self-assembly process; that is via the formation of intermediates, 24 and 25 for 22 and 23, respectively. The first step is the formation of 24 and 25, whereas the second step is self-assembly of 24 and 25. This pathway was supported by the isolation of

9,,,,..

E tt~ tt~

"X2

6 0

Z

z~

z=?

/gu N Z

L_/

=?z

N.=/ L_/

=~z"~"z =?

o

II

II

X

\

L_/

/

:z:m

Z

v~~.

/

mm

Z

e~

x~,,

., r - - x Z ~ ., mZ

Z L_/

~z

~,V#I

zs

..

z a~

zm

L__I

:~ z

/ \

../--~

mz

< F-.

9 9 <

v

< I <

<

0

0

z~ *"

~E

~

0

~

m

f-

~

t-

0

E ~ t_ ~ 0 ..1~ m

0

E ~ t~ ,2.,

t_

~

~"

o

o..

a;

o

E

0

e-J

0

0

.~

0

U

.0~

cZ

r-

~~,~

m ~

c-

0

u~

0

~.e~ ~--

~ -

Molecular-Coordinated Self-Assemblies

11

a) H2~--~l'h

-']12+

H2

F~x LON02 la

+

D20

4 3")12 20

22

b) H2

H2

H2

H2

ONO,

12+

~NO, la

_

3~12

21

23

Scheme 5.

the intermediates 24 and 25 which have the partial framework of macrocycles 22 and 23, respectively. The guest inclusion studies by 1H NMR indicate that 22 and 23 can effectively bind dicarboxylates dianions. For example, 22/23 encapsulated 1,4-phenylenediacetic acid and sodiumtherphthalate. Interestingly, monocarboxylate, such as pmethoxyphenylacetate, did not complex with 22 but complexed effectively with 23. This indicates that for host 22 the complexation with dicarboxylates could be taking place by a two-point electrostatic attraction between negative (CO0-) and positive

12

KUMAR BIRADHAand MAKOTO FUJITA

Figure 5. Space-filling representation of macrotricycles in the crystal structures of (a) 22 and (b) 23. Please note the difference in the shapes of 22 and 23.

i"\t'k \/J 24: Pd = ( e n ) P d

25: Od = ( e n ) P d

charges (pd2+), while in host 23 the binding could be taking place due to its bowl shape.

4.

SELF-ASSEMBLING [2]CATENANE

Catenanes have attracted considerable current interest particularly owing to their recently explored potential as molecular-scale devices. 34"36 The incorporation of a coordination bond into catenane frameworks makes it possible to realize a quick molecular motion, which is somewhat similar to that of well-known magic rings: under thermodynamic conditions a catenane framework rapidly arises from two preformed molecular tings. In the following sections, we describe the self-assembly, chemical manipulation, and characterization of several self-assembled catenanes.

4.1. Catenane Formation from Preformed Rings: Molecular Magic Rings We found an unprecedented formation of [2]catenane under an equilibrium in which the catenane framework arises from two preformed molecular rings. 37 The catenane 27 and its component ring 13 self-assembled from l a and 1,4-bis(4pyridylmethyl)benzene (26) in aqueous solution (Scheme 6). Spectroscopic studies

Molecular-Coordinated Self-Assemblies

13

H2 \;ONC~

~N/

H2

~

\ONO2 la

' ~ ' H20

N 26

H~, /NH2 Pd\

H-zN,~ /NI-~

/N Pd

1"I2N\ ?1"12

"(NOs)I

/

27

Pd \

13

1 mM b

<1

9

>99

2 mM

11

9

89

5 mM

38

9

62

10 m M

59

9

41

20 m M

75

9

25

50 m M

91

9

9

aMeasured in D20 at room temperature, bNet concentration of

Pd(ll).

Scheme 6.

confirmed the existence of the equilibrium between 27 and monomeric ring 13 in solution. At lower concentrations (<2 mM) the equilibrium lies in favor of 13, whereas at higher concentrations (>50 mM) 27 is the overwhelmingly dominant species. The convenient mechanism for the interconversion of 13 to 27 could be involved with the breaking and reassembling of one of the rings at Pd-N linkage in order to thread the other ring (mechanism A in Scheme 7). 38 The studies on equilibrium

14

KUMAR BIRADHA and MAKOTO FUJITA

Q A Scheme 7.

using one of the NMR techniques, namely truncated driven nuclear Overhauser effect (TOE), disprove this mechanism and suggest two sequential ligand exchanges between two molecular rings concomitant with a twisting of the rings around each other (mechanism B in Scheme 7). It is interesting to note that the mechanism B is reminiscent of the MiSbius strip approach to a [2]catenane through ligand exchange. 39

4.2. Irreversible Formation of a [2]Catenane from Two Preformed Rings In the above described reaction the presence of an equilibrium means that [2]catenane 27 once formed easily dissociates into two separate rings. If the labile coordinate bond can be frozen after the catenane assembles, one can obtain a complete catenane that never dissociates into two rings. Here we describe such an irreversible formation of Pt counterpart, catenane 29, from lb and 26 (Scheme 8). 40 This was achieved by employing a concept of "molecular lock" which exploits the dual character of a Pt(II)-pyridine coordinate bond (Scheme 9a). This bond can be likened to the lock since it is irreversible (locked) under ordinary conditions, but becomes reversible ("released") in highly polar media at an elevated temperature. Since the Pt(II)-Py bond is inert (locked) under ordinary conditions, macrocycle 28 is not in equilibrium with any other structures. However, heating at 100 ~ in the presence of NaNO 3 makes the Pt(II)-pyridine bond labile (released) and two rings of 28 slide into a catenated dimer 29 (Scheme 9b). After 29 assembled, its structure is locked by removing salt and cooling to room temperature to give a catenane which never dissociates into its component tings. The structure of 29 was confirmed by X-ray crystallography (Figure 6). The crystal structure shows that the catenane was stabilized by strong n-rr interactions.

Molecular-Coordinated Self-Assemblies

H2

H=

joNo,

~

lb

15

N

, ~ H=O

26

A 2

~'

irreversibly

,~

HzN _/NI-I= e(NO=)4

29

28 Scheme 8.

4.3. Self-Assembly of [2]Catenanes from Rectangular Molecular Boxes Here we show that the rectangular molecular boxes composed of transition metals and organic ligands are predictably and quantitatively catenated if the boxes involve parallel aromatic rings with an interplanar separation of 3.5/~. The catenane 31 self-assembled quantitatively when the Pd complex la is treated with ligand 30 (Scheme 10).41 The stability of 31 in solution is remarkable as the dissociation of the catenane into its component rings was not observed even at low concentration or in less polar media (D20:CD3OD 1:1). The structure of 31 was determined using X-ray crystallography as well as careful examination of NMR and mass spectra. The crystal structure reveals that the catenane was stabilized by an efficient stacking of four aromatic rings and leaves no empty space within the catenane (Figure 7). The solution structure was characterized in terms of topological chirality: 42 i.e. the clockwise and anticlockwise interlocking of the second ring onto the first ring gives rise to the enantiomers of 31. The rectangular box was expanded further by the insertion of additional phenylene ring to 30. Thus we observed relatively flexible catenane 33 from la and 32 (Scheme 11). Again the interplanar separation of ca. 3.5/~ seems to become an essential factor for the stabilization of catenane 33.

16

KUMAR BIRADHAand MAKOTO FUJITA

(a) .~+1.,,~__ irrever$iblt (locked)

_.-.+salt'heating~ "~"+--/--salt, cooling

~"-~--"--~ " +/--'!

reversible (releosezi) I

(b)

2

k

- - J

A

D

release[( heating +salt' )

locking I ( cooling "salt, ) self-assembly

2

B

,p

C Scheme 9.

Figure

6. Space-fillingrepresentationof catenane in the crystal structureof 29.

Molecular-Coordinated Self-Assemblies

N\ fON02 [~NTd~oNO2

17

H2

~____~~..Ns------- Pd--NH2

H2

la +

H20

H21NJ

N--"

30

(NO3")S

31

Scheme 10.

Interestingly, the self-assembly of catenanes was also observed from a three-component system. Three components, la, 2, and 34, self-assembled in water in a 2:1:1 stoichiometry to give catenane 35 in very high quantitative yield of 94% (Scheme 12). The structure of 35 was conformed by X-ray crystallography (Figure 8), ESI-MS, and ~H NMR. It is noteworthy that the thermodynamic stability of 35 overcomes the combination problem which arises in the self-assembly of larger sets of components. At least, the formation of the three component macrocycles 3a, 36, and 37 is possible because their thermodynamic stability is comparable with that of 35. Further, the formation of more flexible and expanded catenane 40 was also observed when the components la, 38, and 39 were combined in water in 2:1:1 ratio (Scheme 13). The structure of 34 was deduced from an ESI-MS study.

Figure 7. Space-filling representation of catenane in the crystal structure of 31.

18

KUMAR BIRADHA and MAKOTO FUJITA

~N~\ON, ,~

C ....~"~

~

--N

,~N ......~ .....~,"~:~..~;---~.

32

[. ,'1

33

Scheme 11. 1t2

r,,.N\

,

/ONO2

LN/%N~ H2

NI-I2 I| ~

~ " - ~ , , ~ -~P d - - N H

\ ~ P d ~ ~

la +

2

H20 34 9

H~.---p. ~rL~..-, I <, / ~ " ' - - - - P ~ - N H o H2N~,/

(NO3")8

35

Scheme 12.

~

H21I

" NI'I 2

N~Pd~N.

s)-'--@

~

. N---Pd---N

N""~

~/'N'J

"(NO3)4

37 36

"(N03)4

Molecular-Coordinated Self-Assemblies

19

H2

N,,/oNo2

IN

No2

H2

Nil= .-

~

H2.N'~ I/ N...--p.d--NI-I=

la

H20

38 +

H2N.~

(NO3)8

40 39

Scheme 13.

4.4. A Three-Dimensionally Interlocked Catenane A three-dimensional catenane was designed using retrosynthetic analogy 43 as shown in Scheme 14. This analogy suggests that the components 1, 41, and 42 can form a cage 43 that can self-assemble into the three-dimensionally interlocked catenane 44. 44 The components 41 and 42 could be termed as floor and ceiling of cage 43. The distance between floor and ceiling has been adjusted such that another copy of cage 43 would fit in the cavity. That is, in cage 43 the interplanar surface-to-surface distance between the floor and ceiling is ca. 3.5 ~, an ideal distance for aromatic n - n interactions. Thus, three components 1, 41, and 42 were combined in a 3:1:1 ratio in D20. The formation of 3D-interlocked catenane 44 in

Figure 8. Space-filling representation of catenane in the crystal structure of 35.

20

KUMAR BIRADHA and MAKOTO FUJITA

.

(

III

H2

N~M,.ONOz

I~N"

"ONO2

,,,~"

H21

NII~M

9

N~r

§

< H.N~ 2 ~ N ~ N ~ ~ ~ " INH' ~ N ' I "NH= (NO=')l HaNJ

N

41

42

~ ' ~ i H2

43

H,N"%

.T_..,~W.Jd "N/I"I2

Scheme 14.

solution was characterized using 1H NMR. Further, the determination of crystal structure of 44b provided a reliable evidence for the catenane formation. The crystal structure shows that 44b consists of two identical cage frameworks interlocked with each other (Figure 9) and also shows the presence of an efficient quadruple stacking of aromatic rings. This type of stacking makes the interlocked structure the most stable of any possible structures formed by the complexation of I with 41 and/or 42. A striking observation is that two preformed Pd(II) cage compounds of 41 (M6L4 type) and of 42 (M3L 2 type), which will be described in the following sections, are reorganized into the 3D-interlocked complex 44a in a high yield when they are

Molecular-Coordinated Self-Assemblies

21

Figure 9. Illustrations of the crystal structure 44b. (a) A top and (b) a side view of the three-dimensional catenane 44b.

mixed in aqueous solution. The NMR of the solution initially showed that the cage of component 41 existed intact, whereas cage of component 42 was considerably converted into oligomeric components. After the solution was allowed to stand at room temperature for 1 day all components were completely reorganized into catenane 44. At an elevated temperature, 80 ~ the reorganization was significantly accelerated and completed within 10 min. These results clearly demonstrate that 44 is a thermodynamic product and is the most stable among all possible structures which can be assembled by numerous combination of components 1, 41, and 42.

5. SELF-ASSEMBLY OF M3L 2 COMPLEXES In this section two types of self-assemblies of M3L 2 complexes will be described. In one case the self-assembly was occurred only in the presence of a suitable guest molecule (guest-induced fit) and in the other case the type of self-assembly was selected by the guest from a receptor's library (guest-selected cage). 5.1.

Guest-InducedAssemblyof an

M3L z Complex

A tridentate ligand 1,3,5-tris(4-pyridylmethylbenzene) (45) was treated with

cis-protected Pd(II) complex la and sodium 4-methoxyphenylacetate in water (Scheme 15). 45 The 1H NMR spectra revealed a high symmetry of the product 46 and the expected 2:3 ratio of the components la and 45. While the ESI-MS studies confirmed the presence of one guest per cage. Further, the aromatic and methoxy protons of the guest exhibit significant upfield shifts (ca. 3 ppm) whereas the shifts for CH2COO- were negligible. This indicates that the hydrophobic portion of the guest stays inside the cage-like cavity while the hydrophilic portion protrudes

22

KUMAR BIRADHA and MAKOTO FUJITA

.2

3(la)

"218+

4H2 ' 45

~

N 46

L~N~J

H2 ~[~3")12

Scheme 15.

outside. In the absence of guest molecules an intractable mixture of oligomeric products was obtained. The relative abilities of guest molecules to induce self-assembly 46 were estimated from NMR yields of the 1:1 host-guest complexes (Scheme 16). These results demonstrate that the bulky substituents, whose sizes are comparable with the cavity size, are effectively induced the self-assembly 46. Thus, the guest-induced assembly 46 provides a new model for induced-fit molecular recognition because the recognition site of a receptor was organized only in the presence of specific guests. A macrocyclic complex showing similar characteristics was reported recently. 46

5.2. Guest-Selected Formation of Cages from a Dynamic Receptor Library Again, a tridentate component, but without threefold symmetry, 47 was selected as a ligand because it can be in equilibrium with two M3L2 cage complexes and also with several oligomers (Scheme 17).47 The reaction of la and 47 in 3:2 ratio resulted in a mixture of oligomeric compounds. However, addition of 1,3,5-benzenetricarboxylic acid selectively turns the mixture into the asymmetric cage complex 48. 1H NMR spectra revealed desymmetry for both the host and the guest frameworks. The formation of a host-guest complex in 1:1 ratio was also supported by ESI-MS. Several aromatic guests including water immiscible compounds, such as benzene and xylene, were also found to be effective in inducing the formation of 48. Interestingly the spherical molecules such as CBr 4 or CBrCI 3 selectively induced the formation of symmetrical cage 49 in 85 % yield. The NMR spectra show that the ligand framework maintains its symmetry in good accordance with the symmetrical structure 49. Encapsulated CBrCI 3 was observed by the 13C NMR with a remarkable upfield shift (ca. 2.9 ppm). The origin for this selectivity can be

23

Molecular-Coordinated Self-Assemblies

-2.10

" - " -1.71CH3 -2.4 -0.92 94%

~ u -0.70 -2.76-1.33 82%

CH3.-~CHz-CH2-COONa

-1.94

-2.5 -2.0

-1.40 -1.1 3

~"~CH2COONa ~-..Y-1.02

" " - -1.69 -2.2

84%

87%

(~. -1.2 COONa

-1.6 (a,e)t~

'~---CHzCOONa

/

ca-2--3 92%

76%

-0.7

~ I-3.1 ~ " ~ I ~-2.1 ICH2

""T

il

60%

cOONa

"1 -0.sa

"2"~-1.2 -3.1 -2.2

-0.80

-I .25

~ " -1.45 -0.85

81%

CH2COONa r CHzCOONa

cH:OON,

not complexed

76%

36%

CH3COONa not complexed

'The NMR yield of 46 is determinedat [45]o= 2 raM, [la]o = 3 mM, and [guest] = I_~ raM. The negative values present upfield shift (Dd in ppm) in IH NMR measuredat [45]0 = 6 raM, [la]o = 9 raM, and [guest] = 1.5 raM: at thisconditions, Dd values of the guests (except

C6H4(COONa)2) were saturated, bAxialand equatorial protons are donated by a and e,

respectively. CDdvaluescouldnotbeanalyzeddueto overlapwith uncharacterizedsignals. Scheme 16.

attributed to the differences in shapes of the cavities of 48 and 49. The molecularmodeling calculations suggest that the 48 has a fiat cavity while 49 has a spherical cavity. Therefore, the fiat guests selected host 48 and spherical guests selected the host 49. The interconversion between two cages was also observed when 48-pxylene was treated with the excess amount of CBrCI 3. Within a day the guest was exchanged and the host 48 was slowly converted to 49 to form the more symmetrical 49.CBrC13 complex.

24

KUMAR BIRADHA and MAKOTO FUJITA

Scheme 17.

6. SELF-ASSEMBLYOF HOLLOW, NANO-SIZED M6L4 COMPLEXES Precise control of nanometer-scale structures has become one of the most important subjects in both physics and chemistry. 48-5~Of many nano-scale materials, ultrafine particles are of intense interest as they are expected to show unique properties which fine particles have never possessed. Usually ultrafine particles are prepared by grinding an already fine particle. We discuss here the synthesis of such molecularbased ultrafine particles through supramolecular self-assembly of 10 simple molecular components. Further, their stability and inclusion and catalytic properties will be described.

6.1. The Structure of Octahedral M6L4 Cage Complexes [M = Pd(ll)] A nano-sized three-dimensional supramolecule 50a self-assembled quantitatively from six molecules of la and four molecules of tridentate ligand 41 (Scheme 18). 51 The complex 50a is a thermodynamically stable product because the formation of the product is not affected by the presence of excess of la. The structure of the clatharate complex of 50a with the adamantane carboxylate ion was determined

25

Molecular-Coordinated Self-Assemblies M

I 12+

6(1) D20

N

N9

/ ~

"(NO3"~12

41 a: M = Pd, b: M = Pt

Scheme 18.

by X-ray crystallography (Figure 10), which showed that four guest molecules are tightly encapsulated inside the nano-sized cavity of 50a. The inclusion geometry of the guest in the cavity is interesting as the hydrophobic and hydrophilic groups (COO-) are located inside and outside of the cavity, respectively (Figure 10b). The 1H NMR study suggested that the same host-guest aggregate (50a).(45) 4 was also organized even in aqueous media. The nanocage 50a expanded further by using the larger ligands 41a and 41b instead of 41. The molecular sizes of these molecules amount to several nanometers

Ar

NAN

Ar~', N/~J~.Ar 41a: Ar =

41b: Ar =

(3r-o

26

KUMAR BIRADHA and MAKOTO FUJITA

Figure 10. The nanosized aggregate in the crystal structure of 50a.(adamantane carboxylate)4. (a) The host framework. (b) A space-filling representation showing the small entrance of nanosized cavity.

and studies using a method of common laser light scattering indicated their ultrafine particle characteristics.

6.2. Guest-Templated Synthesis of a Kinetically Stable M6L4 Cage [M = Pt(ll)] Similar to all the Pd(II) self-assemblies described so far, 50a is also a result of thermodynamical equilibration and it is not stable to maintain its framework under extreme conditions (e.g. acidic, basic, or nucleophilic conditions). In order to prepare a kinetically stable M6L4 complex the Pt(II) complex l b was used instead of la. 52 In contrast to Pd cage, the formation of Pt cage 50b was quite slow to give it in a reasonable yield. However, the addition of guest molecule, adamantanecarboxylate, dramatically improved the reaction rate as well as the yield. The host-

Molecular-Coordinated Self-Assemblies

27

guest ratio and the guest inclusion geometry were found to be similar to those of Pd structure 50a. Usually, a receptor framework organized by guest "induced-fit" is lost when the guest is removed. In contrast to this, the Pt cage 50b did not lose its cage structure even after removal of the guest because of the locking of Pt(II)-Py bond after the self-assembly event. As anticipated, the Pt(II) complex was very stable and did not decompose even in the presence of an acid (HNO3), a base (K2CO3) , or a nucleophile (NEt3) due to the inertness of a Pt(II)-pyridine coordinate bond.

6.3. Molecular Recognition of Large Guests by the M6L4 Cage Generally, the cage compounds prepared by covalent synthesis can encapsulate only one or two small molecules because the construction of the larger frameworks by conventional covalent synthesis is very difficult. The recent advances in noncovalent synthesis made it easy to create such larger frameworks with very large cavities. The above described cage compounds, 50a and 50b, have very large cavity with a diameter of 2 nm and exhibited a remarkable ability to encapsulate large, neutral molecules. Four molecules of either O-carborane 51 or four molecules of adamantane 52 were found to include in the cavity of M6L4 cage (Scheme 19).53 The encapsulation of O-carborane was faster than that of 52 because of the slow transfer of adamantane from hexane to the water phase. Interestingly, the adamantane 52 was found to transfer into the aqueous phase even in a solid-liquid two-phase system. Similarly, 50 also included 2-adamantanol and 1-adamantanol and aromatic compounds such as 1,3,5-trimethoxybenzene, anisole, and toluene. Notably, the complexation is faster with smaller guest molecules and slower with larger guest molecules. For example the tri-tert-butylbenzene, which is slightly larger than the portal of 50, was encapsulated very slowly.

6.4. Formation of Hydrophobic Dimers in the M6L4 Cage The cage compound 50 exhibited a remarkable ability to encapsulate Cshaped molecules such as cis-azobenzene 53 and cis-stilbene 54 derivatives. 54

Scheme 19.

28

KUMAR BIRADHA and MAKOTO FUJITA H

H

B ~.---~-~-B

/

H

\

H

H

52

51

These guest molecules enclatharated in the cavity via the formation of a hydrophobically interacted dimer with a topology reminiscent of Rebek's hydrogen-bonded tennis ball 55 (Scheme 20). The selective enclatharation of the cis-isomer was observed when cis-trans mixtures of either 53 or 54 in hexane are stirred with the D20 solution of 50. The NMR spectra confirm the encapsulation of dimers of cis-isomers in the cavity. The cis-isomers were significantly stabilized in the cavity and not isomerized to a trans-isomer even after a few weeks under visible light at room temperature. The molecular-modeling calculations suggest that the dimer of 53 or 54 is a perfect fit for cavity of 50. Dimerization of guests prior to enclatharation is unlikely because the dimension of spherical dimer (ca. 11 /~, in diameter) is larger than that of the window of 50 (ca. 7/~ diameter). Therefore, two guest molecules should be subsequently but not simultaneously enclatharated in the cavity and turned in situ into the stable hydrophobic dimer.

Scheme 20.

Molecular-Coordinated Self-Assemblies

29

6.5. Catalysis and Acceleration of Chemical Reactions in the M6L4 Cage The reactivity and catalysis represent one of the most important features of the functional properties of supramolecular systems, la'56 The presence of a large cavity in S0a motivated us to test its ability to catalyze the oxidation of styrenes 55 and isomerization of allylbenzenes 56. 57 When la and 41 were mixed in D20 in 4:2 ratio, formation of only 50a was observed and excess l a was remained in the solution. Our strategy is to use this excess of la as a mediator between organic and aqueous phases; that is to use it to cyclically and continuously transfer the substrate into aqueous phase that contains 50a and then the formed product into the organic phase (Scheme 21). It was observed that the 50a can accommodate three molecules of styrene 55a or two molecules of allylbenzene 56a in its cavity. The oxidation of styrene at 80 ~ in the aqueous solution of either l a or 50a gave acetophenone 57a only in 4% yield. Importantly, the presence of both l a and 50a in the aqueous solution increased the yield of the reaction up to 86% (Scheme 22a). Similarly, the isomerization of allylbenzenes catalyzed by the presence of la and 50a in aqueous solution to give 13-methylstyrene 58a in 50% yield, whereas the reaction did not occur in the absence of either l a or 50a (Scheme 22b). The presence of trimethoxybenzene in the reaction media inhibited these reactions because the cavity of 50a was strongly occupied by it. The yields of these reactions reveal that as the size and electron deficiency of the substrate increases the yield of the reaction decreases. These are the good examples of organic reactions in solutions but with no organic solvents.

Scheme 21.

30

KUMAR BIRADHA and MAKOTO FUJITA

a)

X3

X3 la and 50a in D20 L

X

1 day, 80 ~

X

Xl 55

X1 57

a: Xl = X2 = X3 = H; yield = 86% b: X~ = CH3; X2 = X3 = H; yield = 70% c: X1 = OCH3; X2 = X3 = H; yield = 57% d: X1 = NO2; X2 = X3 = H; yield = 15% e: X l = H; X2-X3 = -CH---CH-CH=CH-; yield = 12%

b) X3 a, and,o inD2O X2,,,,,"

~ I

1 day 800C

T

X1

Xl 58

56 a: X 1 -- X 2 = X3 = H; yield = 50% b: X l = OCH3; X 2 = X 3 = H; yield = 58% c: X 1 = H; X2-X3 = -O-CH2-O-; yield = 27%

Scheme 22.

7. A C O O R D I N A T I O N NANOTUBE M o l e c u l a r - b a s e d tubular structures have attracted c o n s i d e r a b l e current interest b e c a u s e of their potential abilities for selective inclusion and transportation of ions and m o l e c u l e s and catalysis of specific c h e m i c al transformations. 58 H e r e we describe the design of coordination nanotubes using the oligo(3,5-pyridine)s 59 and

cis-protected

Pd(II) c o m p l e x l a . 59 In case of pentakis(3,5-pyridine) ligand 59a, a

Molecular-Coordinated Self-Assemblies N

31 N

N

oee

Oee

59 ~

. ONa

61 coordination nanotube 60a is expected from four molecules of 59a and 10 molecules of la (Scheme 23). 46 Similarly, four of 59b and eight of la, and four of 59e and six of la, were anticipated to form coordination nanotubes 60b and 60c, respectively. However, the formation of coordination nanotubes were observed only

-"720+ N <:=

H2 [~N ~pd/ONO ' N" "ONO 2 142

,::>

N ,:C>

[

=

la

I ,F(C '.~ ,~ ~' ~,

,, >

I "~"-~.

I fZ"3. iii i

~

(

59a

I

- ~ I - , " ~ F ~" I

I/'~

l

N

O~')2o

60a

" • 16+ P

d

~

.

,,

~

~

~'112+

,

5), (NO~')12 H2 P d =

(NO.~'),e

BOb

Scheme 23.

"~

N~" H2

32

KUMAR BIRADHA and MAKOTO FUJITA

in the presence of a rod-like template molecule such as sodium 4,4'-biphenylenedicarboxylate (61). Nanotubes 60a, 60b, and 60c were characterized using NMR and ESI-MS. In NMR, the protons of 61 were upfield-shifted by 2.6 ppm indicating its encapsulation in the nanotube. The similar template effect was observed with two other rod-like molecules, biphenyl and p-terphenyl. Spherical and large molecules such as adamantane carboxylate failed to template the nanotubes. Interestingly, it is found that the formation of these tubes is a complete reversible process; that is that the tube dissociates into its components by the removal of guest molecule and again associates by the addition of guest of molecule. In the case of 60a, the shuttle movement of guest molecule was observed: the guest stays at a fixed position of the tube, shuttles on the NMR time scale at 60 ~ and rapidly moves or partially goes out at above 60 ~ The NMR studies of nanotube 60b revealed that it is a 1:1 mixture of structural isomers 60bl and 60b2 (Scheme 24). In isomer 60bl each ligand is placed on a C2 symmetry and only seven protons corresponding to half of 59b were observed. Whereas in isomer 6062, all 14 protons of 59b were observed as the C2 symmetry of the ligands was destroyed. The tube 60c was characterized by X-ray crystallography. The crystal structure displays tubular structure of 59c that is efficiently assembled around template 61 via strong rt-x and C - H - x interactions (Figure 11). The shape of the tube, which should ideally be square, is significantly distorted in order to have strong aromatic interactions. That is, the two faces, which are interacting with 61 via n-r~ interactions, are squeezed towards inside, while the remaining two faces, which are interacting with 61 via C - H - x interactions, are pushed outwards.

60bl

60b2 Scheme 24.

Molecular-Coordinated Self-Assemblies

33

Figure 11. Illustrations for the crystal structure of 60c.(61)2. (a) A side view and (b) a top view of the nanotube 60c. The guest molecule, 61, was present in space-filling mode.

Another interesting feature of this crystal structure is the presence of a second molecule of 61 that is encapsulated between the nanotubes.

8.

A COORDINATION CAPSULE ASSEMBLED F R O M 24 C O M P O N E N T S

Here we discuss a nanometer-sized molecular hexahedron self-assembled from 24 components. The triangle is a basic unit for the self-assembly of a polyhedron. Therefore to construct a molecular polyhedron we selected an exohexadentate ligand, 1,3,5-tris(3,5-pyrimidyl)benzene (62) as a triangular unit. This ligand is an almost coplanar triangle and is expected to give an edge-sharing polyhedron when it is self-assembled with la. Of the several possibilities, the assembly of the molecular hexahedron 63 was strongly suggested by IH NMR (Scheme 25). When ligand 62 is treated with l a in D20, we observed the predominant formation of a single component whose 1H NMR spectrum showed seven singlet-like signals in

34

KUMAR BIRADHA and MAKOTO FUJITA

a)

9 -1

>

000

b)

tetn~hedron

octahedron

he xahedron i

Scheme 2.;.

an integral ratio of 2:2:2:2:2:1:1. This observation confirms that, after complexation, the ligand 62 is placed in a less-symmetrical environment with one symmetry axis passing through a 3,5-pyrimidal (pym) ring and a core benzene ring (Scheme 26). This symmetry is in good agreement with the trigonal bipyramidal structure of the molecular hexahedron 63 in which the pym groups at the apical corners are nonequivalent to those at equatorial corners. The metal-linked dimer 64 and the trimer 65, the possible intermediates for assembly process of 63, were observed when ligand 62 was treated with la in D20 in 1:1 and 3:4 ratios respectively.

4

01

~Z

c ,-5

Z

i ,

i

z~"-z

| ,

|

Z

IZ /Q,\

ZI

....

:7

:L

r~

36

KUMAR BIRADHA and MAKOTO FUJITA a

Pd. / ~ . N .-Pd "N

.L J.

Pd-.N d~.

.Pd

I !

!

(a:b:c:d:e:f:g = 1:2:2:2:1:2:2)

Scheme 26.

Reliable evidence for the trigonal bipyramidal structure of 63 is provided by X-ray crystallography (Figure 12). The crystal structure clearly demonstrates that the assembly is a trigonal bipyramidal capsule with a chemical formula of C144H216NlosPd18, a molecular mass of 7103 Da, and a dimension of 3 x 2.5 x 2.5 nm. Each equatorial corner of the hexahedron is the assembly of four triangle units, where a (Pd(II)-pym)4 gives a small pin hole (2 x 2 ]k). Through these holes only small molecules such as water and molecular oxygen can pass, but ordinary organic molecules cannot enter or escape through it. The free volume inside the capsule, into which guests can be accommodated, is ca. 900 ]k3, implying that complex 61 can host large molecules such as buckminsterfullerene C6o.

Figure 12. Space-filling representations of the molecular hexahedron, 63, in its crystal structure. (a) A view from an equatorial direction; (b) a view from an apical direction.

Molecular-Coordinated Self-Assemblies

37

ACKNOWLEDGMENTS The authors would like to thank Prof. Kentaro Yamaguchi of the Chemical Analysis Center, Chiba University, for X-ray crystallography throughout this study. Thanks are also due to all coworkers, particularly Dr. Takahiro Kusukawa, Dr. Nobuhiro Takeda, Dr. Shuichi Hiraoka, Mr. Fumiaki Ibukuro, Mr. Masaru Aoyagi, Mr. Norifumi Fujita, Mr. Hirokazu Ito, and Mr. Kazuhiko Umemoto, who by their efforts have contributed to the success of our research.

REFERENCES AND NOTES 1. (a) Lehn, J.-M. Angew. Chem., Int. Ed. Engl. 1988, 27, 89; (b) Lehn, J.-M. Angew. Chem., Int. Ed. Engl. 1990, 29, 1304. 2. Lehn, J.-M. Supramolecular Chemistry; VCH: Weinham, 1995. 3. Supramolecular Chemistry; Balzani, V.; DeCola, L., Eds.; Kluwer Academic: The Netherlands, 1992. 4. Sauvage, J.-P.; Hosseini, M. W. Comprehensive Supramolecular Chemistry; Lehn, J.-M., Ed.; Pergamon Press: Oxford, 1995, Vol. 9. 5. A recent review on hydrogen bonding control of molecular assembly: Hamilton, A. D. Comprehensive Supramolecular Chemistry; Lehn, J.-M., Ed.; Pergamon Press: Oxford, 1995, Vol. 9, Chap. 18. 6. Reference 4 includes excellent reviews of the metal-associated self-assembly. In particular, see the following: (a) Sauvage, J.-P.; Dietrich-Buchecker, C.; Chambron, J.-C.; (b) Sanders, J. K. M., Chap. 4; (c) Baxter, P. N. W., Chap. 5; (d) Constable, E. C., Chap~ 6; (e) Fujita, M.; Ogura, K., Chap. 7. 7. TransitionMetals in Supramolecular Chemistry; Fabbrizzi, L.; Poggi, A., Eds.; Kluwer Academic: The Netherlands, 1994. 8. (a) Fujita, M.; Ogura, K. Bull. Chem. Soc. Jpn. 1996, 69, 1471; (b) Fujita, M.; Chem. Soc. Rev. 1998, 27, 417; (c) Fujita, M. Acc. Chem. Res. 1999, 32, 53. 9. (a) Lehn, J.-M.; Rigault, A.; Siegel, J.; Harrowfield, J.; Chevrier, B. Proc. Natl. Acad. Sci. USA 1987, 84, 2565; (b) Lehn, J.-M.; Rigault, A. Angew. Chem., Int. Ed. Engl. 1988, 27, 1095; (c) Kramer, R.; Lehn, J.-M.; Rigault, A. M. Proc. Natl. Acad. Sci. USA 1993, 90, 5394; (d) Koert, U.; Harding, M. M.; Lehn, J.-M. Nature 1990, 346, 339. 10. Constable, E. C. Tetrahedron 1992, 48, 10013. 11. (a) Youinou, M.-T.; Rahmouni, N.; Fischer, J.; Osborn, J. A. Angew. Chent, Int. Ed. Engl. 1992, 31,733; (b) Baxter, P. N. W.; Lehn, J.-M.; Fischer, J.; Youinou, M.-T. Angew. Chem., Int. Ed. Engl. 1994, 33, 2284. 12. A cylindrical Cu(I) hexanuclear complex: (a) Baxter, P.; Lehn, J.-M.; DeCian, A. Angew. Chent, Int. Ed. Engl. 1993, 32, 69; (b) A related Cu(I) trinuclear complexes: Leize, E.; Dorsselaer, A. V.; Kramer, R.; Lehn, J.-M. J. Chem. Soc., Chem. Commun. 1993, 990. 13. Delaigue, X.; Hosseini, M. W.; Leize, E.; Kieffer, S.; Doreeslaer, A. V. Tetrahedron Lett. 1993, 34, 7561. 14. Fuchs, R.; Habermann, N.; Klufers, P. Angew. Chem., Int. Ed. Engl. 1993, 32, 852. 15. (a) Dietrich-Buchecker, C. O.; Sauvage, J.-P.; Kintzinger, J. P. Tetrahedron Lett. 1983, 24, 5095; (b) Dietrich-Buchecker, C. O.; Sauvage, J.-P. Chem. Rev. 1987, 87, 795; (c) Sauvage, J.-P. Acc. Chem. Res. 1990, 23, 319. 16. ESI-MS is effective for the analysis of transition metal complexes. (a) Bitsch, E; DietrichBuchecker, C. O.; Khrmiss, A.-K.; Sauvage, J.-P.; Van Dorsselaer, A. J. Am. Chem. Soc. 1991, 113, 4023; (b) Leize, E.; Van Dorsselaer, A.; Kr~imer, R.; Lehn, J.-M. J. Chem. Soc., Chem. Commun. 1993, 990.

38

KUMAR BIRADHA and MAKOTO FUJITA

17. Fujita, M.; Yazaki, J.; Ogura, K. J. Am. Chem. Soc. 1990, 112, 5645. 18. (a) Stang, J.; Cao, D. H. J. Am. Chem. Soc. 1994, 116, 4981; (b) Stang, P. J.; Olenyuk, B. Acc. Chem. Res. 1997, 30, 502; (c) Nielson, R. M.; Hupp, J. T.; Yoon, E. I. J. Am. Chem. Soc. 1995, 117, 9085; (d) Drain, C. M.; Lehn, J.-M. J. Chem. Soc., Chem. Commun. 1994, 2313; (e) Rauter, H.; Hillgeris, E. C.; Erxleben, A.; Lippert, B. J. Am. Chem. Soc. 1994, 116, 616. 19. Fujita, M.; Sasaki, O.; Mitsuhashi, T.; Fujita, T.; Yazaki, J.; Yamaguchi, K.; Ogura, K. J. Chem. Soc., Chem. Commun. 1996, 1535. 20. Fujita, M.; Yazaki, J.; Ogura, K. Tetrahedron Lett. 1991, 32, 5589. 21. Lee, S. B.; Hwang, S.; Chung, D. S.; Yun, H.; Hong, J.-I. Tetrahedron Lett. 1998, 39, 873. 22. Fujita, M.; Yazaki, J.; Ogura, K. Chem. Lett. 1991, 1031. 23. Aoyagi, M.; Biradha, K.; Fujita, M. Bull. Chem. Soc. Jpn. 1999, 72. 24. Matsumoto, N.; Yamashita, M.; Kida, S. Bull, Chem. Soc. Jpn. 1978, 51, 2334. 25. Fujita, M.; Nagao, S.; Iida, M.; Ogata, K.; Ogura, K. J. Am. Chem. Soc. 1993, 115, 1574. 26. Fujita, M.; Aoyagi, M.; Ogura, K. Inorg. Chim. Acta 1996, 246, 53. 27. Fujita, M.; Oka, H.; Yamaguchi, K.; Ogura, K. Unpublished results. 28. Fujita, M.; Kondo, T.; Yamaguchi, K.; Ogura, K. Unpublished results. 29. Fujita, M.; Kwon, Y. J.; Yamaguchi, K.; Ogura, K. Unpublished results. 30. Fujita, M.; Yu, S.-Y.; Kusukawa, T.; Funaki, H.; Ogura, K.; Yamaguchi, K.Angew. Chem., Int. Ed. Engl. 1998, 37, 2082. 31. Sauvage, J.-P. New J. Chem. 1993, 17, 617. 32. Recent developments in highly efficient catenane synthesis: (a) Stoddart, J. E; Raymo, E; Amabilino, D. B. Chap. 3 of ref 4. (b) Ashton, P. R.; Goodnow, T. T.; Kaifer, A. E.; Reddington, M. V.; Slawin, A. M. Z.; Spencer, N.; Stoddart, J. E; Vicent, C.; William, D. J. Angew. Chem., Int. Ed. Engl. 1989, 28, 1396; (c) Anelli, P. L.; Ashton, P. R.; Ballardini, R.; Balzani, V.; Delgado, M.; Fandolfi, T.; Goodnow, T. T.; Kaifer, A. E.; Philip, D.; Pietraszkiewicz, M.; Prodi, L.; Reddington, M. V.; Slawin, A. M. Z.; Spencer, N.; Stoddart, J. E; Vicent, C.; Williams, D. J. J. Am. Chem. Soc. 1992, 114, 193; (d) Amabilino, D. B.; Ashton, P. R.; Brown, C. L.; Cordova, E.; Godinez, L. A.; Goodnow, T. T.; Kaifer, A. E.; Newton, S. P.; Pietraszkiewicz, M.; Philp, D.; Raymo, E M.; Reder, A. S.; Rutland, M. T.; Slawin, A. M. Z.; Spencer, N.; Stoddart, J. E; Williams, D. J. J. Am. Chem. Soc. 1995,117,1271; (e) Armspach, D.; Ashton, P. R.; Ballardini, R.; Balzani, V.; Godi, A.; Moore, C. P.; Prodi, L.; Spencer, N.; Stoddart, J. E; Tolley, M. S.; Wear, T. J.; Williams, D. J. Chem. Eur. J. 1995, 1, 33. 35. (a) Hunter, C. A. J. Am. Chem. Soc. 1992, 114, 5303; (b) Adams, H.; Carver, F. J.; Hunter, C. A. J. Chem. Soc., Chem. Commun. 1995, 809; (c) V'ogtle, E; Meier, S.; Hoss, R. Angew. Chem-, Int. Ed. Engl. 1992, 31, 1619; (d) Ottens-Hildebrandt, S.; Nieger, M.; Rissanen, K.; Rouvinen, J.; Meier, S.; Harder, G.; V'6gtle, E J. Chem. Soc., Chem. Commun. 1995, 809. 36. (a) Gruter, G.-J.; de Kanter, E J. J.; Markies, P. R.; Nomoto, T.; Akkerman, O. S.; Bickelhaupt, E J. Am. Chem. Soc. 1993, 115, 12179; (b) Piguet, C.; Bernardinelli, G.; Williams, A. E; Bocquet, B.Angew. Chem., Int. Ed. Engl. 1995, 34, 582; (c) Mingos, D. M. P.; Yau, J.; Menzer, S.; Wdliams, D. J. Angew. Chem., Int. Ed. Engl. 1995, 34, 1894. 37. Fujita, M.; lbukuro, E; Hagihara, H.; Ogura, K. Nature 1994, 367, 720. 38. Fujita, M.; lbukuro, E; Seki, H.; Kamo, O.; lmanari, M.; Ogura, K. J. Ant Chem. Soc. 1996, 118, 899. 39. The first synthesis of a molecular Mtibius strip: (a) Walba, D. M.; Richards, R. M.; Haltiwanger, R. C. J. Am. Chem. Soc. 1982, 104, 3219; (b) Walba, D. M.; Homan, T. C.; Richards, R. M.; Haltiwanger, R. C. New. J. Chem. 1993, 17, 661. 40. Fujita, M.; Ibukuro, E; Yamaguchi, K.; Ogura, K. J. Am. Chem. Soc. 1995, 117, 4175. 41. Fujita, M.; Aoyagi, M.; lbukuro, E; Ogura, K.; Yamaguchi, K. J. Am. Chem. Soc. 1998, 120, 611. 42. The bridging ligand used to construct catenane 31 or its constitutive ring is asymmetric since a methylene group is inserted between one and only one 4-pyridyl nucleus and the 1,4-phenylene fragment. This provides a direction to the ligand and consequently to the ring. Catenane 31 thus

Molecular-Coordinated Self-Assemblies

43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58.

59. 60.

39

consists of two oriented interlocking rings, which is a necessary requirement for a [2]catenane to be topologically chiral. Preparation of a topologically chiral catenane: Mitchell, D. K.; Sauvage, J.-P. Angew. Chem., Int. Ed. Engl. 1988, 27, 930. Desiraju, G. R. Angew. Chem., Int. Ed. Engl. 1995, 34, 2311. Fujita, M.; Fujita, N.; Ogura, S.; Yamaguchi, K. Nature 1999, 400, 52. Fujita, M.; Nagao, S.; Ogura, K. J. Am. Chem. Soc. 1995, 117, 1649. Bilyk, A.; Harding, M. M. J. Chem. Soc., Chem. Commun. 1995, 1697. Hiraoka, O., Fujita, M. J. Am. Chem. Soc. 1999, 121, 10239. "Engineering a Small World: From Atomic Manipulation to Microfabrication" (A special section of Science), Science 1991, 254, 1300. Nanotechnology; Crandall, B. C.; Lewis, J., Eds.; The MIT Press: Cambridge, 1992. Drexler, K. E. Nanosystems. Molecular Machinery, Manufacturing, and Computation; John Wiley: New York, 1992. Fujita, M.; Oguro, D.; Miyazawa, M.; Oka, H.; Yamaguchi, K.; Ogura, K. Nature 1995, 378, 469. Ibukuro, E; Kusukawa, T.; Fujita, M. J. Am. Chem. Soc. 1998, 120, 8561. Kusukawa, T.; Fujita, M. Angew. Chem., Int. Ed. Engl. 1998, 37, 3142. Kusukawa, T.; Fujita, M. J. Am. Chem. Soc. 1999, 121, 1397. (a) Wyler, R.; de Mendoza, J.; Rebek, J. Jr.Angew. Chem., Int. Ed. Engl. 1993, 32, 1699; (b) Rebek, J. Jr. Chem. Soc. Rev. 1996, 255. Lehn, J.-M. Pure Appl. Chem. 1978, 50, 871. Ito, H.; Kusukawa, T.; Fujita, M. Manuscript in preparation. (a) Iijima, S. Nature 1991, 354, 56; (b) Harada, A. In Modular Chemistry; Michl, J., Ed.; Kluwer Academic: Dordrecht, The Netherlands, 1997, p. 361; (c) Hartgerink, J. D.; Clark, T. D.; Ghadiri, M. R. Chem` Eur. J. 1998, 3, 1367. Aoyagi, M.; Biradha, K.; Fujita, M. J. Am. Chem. Soc. 1999, 121, 4757. Takeda, N.; Umemoto, K.; Yamaguchi, K.; Fujita, M. Nature 1999, 398, 794.

This Page Intentionally Left Blank

CHIRAL SELF-ASSEMBLED STRUCTURES FROM BIOMOLECULES AN D SYNTH ETIC ANALOG U ES

Martinus C. Feiters and Roeland J. M. Nolte

1.

2.

3.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1. Chirality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2. Self-Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3. Monolayers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4. History: 12-Hydroxystearic Acid . . . . . . . . . . . . . . . . . . . . . . 1.5. Scope of the Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lipids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Biological Lipids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Synthetic Phospholipids . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Phospholipid Analogues . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. Cholesterol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5. Ceramides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Amino Acids, Peptides, and Proteins . . . . . . . . . . . . . . . . . . . . . . . 3.1. Polypeptides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Amino Acid Amphiphiles . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Other Assemblies Involving Amides . . . . . . . . . . . . . . . . . . .

Advances in Supramolecular Chemistry Volume 6, pages 41-156. Copyright 9 2000 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 0-7623.0557-6 41

42 42 44 45 "50 55 58 58 65 71 79 83 85 85 92 104

42

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

4.

Carbohydrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Biological Polysaccharides . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Gluconamides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. Other Carbohydrates . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Nucleotides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1. Biological Polynucleotides . . . . . . . . . . . . . . . . . . . . . . . . 5.2. Nucleotides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3. Nucleotide Analogues . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4. Other Complimentary Hydrogen Bonding Systems . . . . . . . . . . . 6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Note Added in Proof . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

106 106 109 127 130 130 133 134 140 143 143 144 144

1. I N T R O D U C T I O N 1.1. Chirality A molecule or object is chiral when it cannot be superimposed on its mirror image. 1 In the context of this review, as in most areas of organic chemistry and biochemistry, chirality or "dissym6trie molrculaire ''2 is typically associated with the presence of one or more carbon atoms which each have four different substituents, 3 the configuration of which can be defined as R or S with the C a h n - I n g o l d - P r e l o g convention. 4 Many chiral assemblies are helical, with right-handed (P) or lefthanded (M) screw sense. For detailed discussions and definitions of chirality, including the aspects of time invariance 5 and quantitative measure, 6 the reader is referred elsewhere. 7-12 Nature is a source of inspiration for chemists to investigate the relation between the chirality of molecules and their assemblies and/or polymers. Examples are the so-called or-helix motif in proteins which is due to the chirality of the amino acid building blocks, and the double helix of deoxyribonucleic acid which is related to the chirality of the sugar components of the nucleotide building blocks. The chirality of biomolecules plays an important factor in molecular recognition phenomena, for example at the receptor level with natural compounds: 13 the (-)- or (R)-enantiomer of carvone 1 smells of spearmint, whereas the (+)- or (S)-enantiomer smells of caraway (see Chart 1).14 Chirality is also important in the case of pharmaceuticals; at a time when this was not so much realized, the drug thalonamide 2 was supplied as a racemate because one enantiomer was found to act as ananalgesic; the other enantiomer, however, turned out to have teratogenic effects. 15'16 An interesting question is how chirality in living systems may have arisen, and what has determined the occurrence of only one set of enantiomers of the possible sugars and amino acids. 17 It can be understood that, once an autocatalytic cycle using and producing one of two possible enantiomers is established, this could then maintain itself as there are various mechanisms (chiral induction due to the presence

Chiral Self-Assembled Structures

43

&el% (R)-Carvone (spearmint)

O

O

o

O O

(S)-Carvone (caraway)

(S)-Thalidonamide (teratogenic)

CI

CI

CI

N"

R

N

a) R'=-(CH2)3COOH, R = -n-CsH17 b) R'=-(CH2)3COOH R = -n-C12H25

R . Phil H

COOH

~

o-~-o

4

OCH3

7 0 H ~@

0

A H ,H -

I

H,,N .N~

(Ft)-Thalidonamide (sedative)

3

R. N

N*"'~N

Br"

o

'-Z

N. H

(s)

-6

H3CO

o~o "

I

"=

A

H

H

I

Z-

N.

N. ~N_

/'

8

/ Chart 1.

I

00iPr

H~-~OH

44

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

of chiral centers already in the metabolite, effect of chirality of the catalyst on the configuration of the new stereocenters) 1s-2~by which the presence of an excess of an enantiomer in a catalyst can lead to an excess of an enantiomer or diastereomer in the product. The formation of so-called hypercycles, collectives of two or more self-replicating species interlinked through a cyclic catalytic network, is considered as an important step in the transition from inanimate to living chemistry. 21 The fact that at some stage in the chemical evolution the balance tipped in favor of one of two possible enantiomers, which was produced in slight excess in an autocatalytic cycle and subsequently became the basis of new autocatalytic cycles evolving toward life as we know it, is probably not merely a matter of statistics. The thermodynamic stabilities of amino acid enantiomers are sufficiently different, due to the so-called weak force, 22 to become significant after a number of autocatalytic cycles. We will return to this point in the discussion of the chirality of polynucleotides. On the subject of spontaneous generation of chirality, it is of interest to know that spontaneous formation of chiral aggregates from nonchiral monomers is known to occur, e.g. the assembly of tetra-alkyl benzimidocyanins 3 as monitored by CD (circular dichroism). 23 Formation of chiral crystals from achiral monomers is also reported, e.g. by photodimerization in the solid state. 24-27In a recent example, chiral crystals of acridine 4 and diphenylacetic acid 5 give excess of the (S)-product 6 upon a photodecarboxylating condensation reaction. 2s Symmetry breaking is also known to occur for supramolecular complexes of achiral components; e.g. glutarimide 7 and the diaminopyridine 8, 29 and, as will be discussed below, in monolayers at the air-water interface. 3~ 1.2.

Self-Assembly

One can look at the process of self-assembly in various ways. In one analogy it is compared to computer programming, 31 and the molecule is considered an "algorithm" or a "set of instructions" viz. instructions to aggregate to an assembly of a particular shape. The formation of helical assemblies can then be conceived as "helical molecular programming.''32 Another philosophy is to start from the process of making compounds with covalent bonds, the process of synthesis (the Greek word for "putting together") from synthetic building blocks or synthons, and by analogy define the process of preparing assemblies held together by non-covalent interactions as synkinesis (Greek for "coming together") and the monomeric building blocks as synkinons. 33'34 In the synkinetic approach, assemblies do not "self-assemble" or "self-organize" but follow the synkinetic plans of the chemist. There is a large variety of approaches by which chemists can design and prepare self-assembling systems 35 especially when interlocked and intertwined structures are also included. 36'37 Scales of increasingly programmed molecular information can be considered, starting from relatively simple aggregates or from metal chelates (Scheme 1).36 Starting from aggregates, one arrives at interlocked and intertwined structures and superstructures by way of stacked pseudorotaxanes, stacked

45

Chiral Self-Assembled Structures

Scheme I. The conceptual progresssion from simple recognition processes--based on donor-acceptor stacks (top) and metal ion chelation (bottom)--to a family of interlocked and intertwined structures and superstructures. Reproduced from ref. 36 (Amabilino and Stoddart, Chem. Rev. 1995, 95, 2725) with permission of the American Chemical Society.

[2]catenanes, [n]pseudorotaxanes, double helices, and polycatenanes. Starting from a metal ion chelate, structures of increasing complexity can be constructed, viz. precatenates and [2]catenates, [n]pseudorotaxanes, double helicates, and polycatenates including metal ions. In this review, we will only consider assemblies from chiral molecules resulting in chiral aggregates, with an emphasis on biomolecular examples and work from the authors' own laboratory. We will discuss intertwined structures but disregard interlocked structures, in spite of the very interesting chemistry and the obvious relevance of such structures for biological polynucleotide chemistry. We will be concerned mainly with structures that are visible at the microscopic level. There are various examples of thermotropic chiral liquid crystals obtained by self-assembly, 38 but they will not be discussed here. The exploration of assembly in the gas phase has only recently started; an interesting approach is the assembly of trimers of chiral tartrates 9 with trimethylammonium ions studied by mass spectrometry. 39 By labeling the (R)-enantiomer by isotope substitution of hydrogens by deuteriums, it could be demonstrated that the formation of homochiral assemblies is favored.

1.3. Monolayers A number of parameters are typically determined in experiments with monolayers at the air-water interface as a function of the surface area A, 4~ viz. the surface pressure r~, the surface potential AV, and the viscosity 1"1.Starting at low pressure

46

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

and high molecular area, the isotherm usually passes from (1) a gas-analogous phase where the amphiphilic molecules do not interact with each other, (2) through a liquid-expanded (LE) phase where the amphiphile molecules start to interact weakly which each other, (3) to a liquid-expanded/liquid-condensed (LE/LC) coexistence phase, which contains liquid-condensed (LC) domains in a liquidexpanded (LE) environment, and then (4) to a solid phase, and (5) finally to a phase in which the monolayer has collapsed (Scheme 2). Early studies on monolayers of chiral molecules like 2-hydroxyalkanes, 41'42 amphiphilic amino acids, 43 2-methylhexacosanoic acid esters, 44 and hydroxyhexadecanoic acid 45 and its esters 46 have been reviewed. 47 The interesting question about monolayers of chiral molecules is whether the parameters which can be determined and the phase transitions 48 are different for pure enantiomers and racemates. For components of biomembranes like phosphatidylcholines 10 this appears not to be the case, 49 but for synthetic compounds like N-(ct-methylbenzylstearamide) 11 specific interactions between the molecules of the enantiomers are observed (Chart 2). 50 In recent years, advanced techniques have been developed to probe the order in monolayers at the air-water interface, 51 including surface X-ray diffraction, 52'53 and microscopic techniques, viz. fluorescence microscopy, 54'55 and Brewster angle microscopy (BAM). 56'57 The X-ray diffraction technique has been used to identify homochiral and heterochiral two-dimensional domains in monolayers of racemic amphiphilic amino acids on subphases containing glycine. 58 Fluorescence microscopy requires the introduction in the monolayer of a small

v

v

v

v

v

v

v

1N E z E

v

/ LC / LE G Area (run 2/mOlecule) Scheme 2. Phases in isotherms of monolayers at the air-water interface.

Chiral Self-Assembled Structures

R"~O~.cH2" " "

o 11 HNJ'L ~ ~"n C17H35

10 o

II

O H2C--o--O,,. .O ~

n-C16H33~ N ~ n-C16H33" j ~

47

H

H

H

H OH

H

H

H

14

-

I

N+--\

CH3

CH3 n-ClsH33~ N . ~ N.,,~ N

N,,,,~.N 12

n-C16H~''~

13

,0

15

CH3

\NO2

Chart 2.

amount of fluorescent probe, which dissolves in the LE regions but is excluded by the LC domains. Dendritic crystals are observed in the fluorescence micrographs of the LE/LCcoexistence phase of monolayers of the achiral amphiphilic histamine derivative 12 59 (Figure 1). In the monolayers of the pure enantiomers of the chiral analogue 13, the dendritic arms are bent in a direction dependent on the chirality of the molecules (Figure 2A,B). The racemate does not show lateral phase separation but gives domains in which there is no predominant direction of bending in the dendritic arms (Figure 2C). It has been pointed out that, on the basis of the definition of chirality involving superimposability of mirror images, the structure in Figure 2C must be classified as chiral. 6a The two-dimensional crystallites in monolayers are examples of diffusion-limited aggregates. It has been proposed that such structures can have inherent chirality when the growth of the structure is chirally biased (as in Figure 2A,B), or can have incidental chirality if such a bias does not exist (as in Figure 2C). The growth of the liquid-condensed domains of (R)-13 to dendritic and fractal patterns has been studied in more detail. 6~ It was found that under the influence of illumination in the fluorescence microscope, the fractal branches become thicker and faceted (Figure 3a,b), and that they gradually develop into dendrites (Figures 3c-e and 4). Spiral structures can also arise in monolayers of achiral molecules, as in the study by polarized laser excitation fluorescence microscopy of chiral defects in pentadecanoic acid. 61 Using a rigid achiral molecule with a series of chiral centers (14), separation of chiral domains in the racemate has been observed by atomic force microscopy (AFM) on monolayers transferred from the air-water interface to

48

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

Figure 1. Fluorescence micrographs of a monolayer of 12 on an aqueous subphase of pH 5.5 at various sta~es of compression at 20.0 ~ ]"l = approx. 0 (A), 8.0 (B), 8.5 (C), and 9.0 (D) mN.m- . Reproduced from ref. 59 (van Esch et al. Langmuir 1994, 10, 1955) with permission of the American Chemical Society.

Figure 2. Fluorescence micrographs of monolayers at 20.0 ~ of (3)-13 (A, ]-] = 8.5 mN.m-1), (R)-13 (B, r[ = 8.5 mN.m-1), and (R,S)-I 3 (C, rl = 8.5 mN.m -1) on an aqueous subphase of pH 5.5 in the LC/LE coexistence phase. Reproduced from ref. 59 (van Esch et al. Langmuir 1994, 10, 1955) with permission of the American Chemical Society.

Chiral Self-Assembled Structures

49

Figure 3. (a) The observed fractal pattern when the monolayer of (R)-16 is compressed to the LC-LE coexistence phase. (b)-(e) Show the evolution of the same pattern from fractal to dendrites under the illumination of the microscope light. At first the tips of the fractal branches become thicker and faceted (b) (see arrows); gradually these tips develop into dendrites with evident main stem and stable tips (c)-(e). Reproduced from ref. 60 (Wang et al., Phys. Rev. Let/. 1993, 71, 4003) with permission of the American Physical Society.

Figure 4. (a) The hexagon domain observed in the monolayer of (R)-16. The faceted domains are usually nucleated near the shade of the optical field diagram. (b)-(d) Illustrate the evolution of the morphology of the same domain from hexagon to dendritic pattern. The inhomogeneous development of the dendritic branches is due to the influence of the neighboring domains. Reproduced from ref. 60 (Wang et al., Phys. Rev. Let/. 1993, 71, 4003) with permission of the American Physical Society.

50

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

mica. 62 Compression of the monolayer on water (Figure 5) subsequently gave A (amorphous), R (rectangular), and O (oblique) phases as characterized by AFM, with chiral phase separation in the R phase (Figure 6). With the same technique, spontaneous chiral symmetry breaking has been demonstrated for calcium arachidate (Figures 7, 8). 30Enantiomorphous monolayers have been observed on graphite by scanning tunneling microscopy (STM) of 15. 63 The racemate gave coexisting enantiomorphous images of domains with chiral packing. STM has also been used to study the homochiral association of 1,2-dihydroxyoctadecane molecules in a self-assembled monolayer on graphite. 64 Dynamic scanning force microscopic techniques have recently been used to probe the long-range tilt orientation of stearic acid monolayers on an aqueous subphase containing a polymeric counterion transferred to mica (Figure 9). 65

1.4. History: 12-Hydroxystearic Acid 12-Hydroxystearic acid 66 and polybenzylglutamate 67 were the classical cases of relatively simple molecules that produced twisted fibers, till the burst of reports on other examples, viz. amino acid amphiphiles, 68'69diacetylenic phospholipids, 7~and gluconamides 71 started in the mid-1980s. The handedness of chiral assemblies can be determined from electron micrographs, provided that care is taken to manipulate grids, specimens, films, and image scanners in a consistent way. 72 It was not till the 35.4

111

-,,-R

,o II a.8131.4 29.7

Area per molecule

(A,)

Figure 5. Surface pressure isotherm for a TCA monolayer formed at 25 ~ on a conventional Langmuir through in a clean room. The subphase is Milli-Q water. The Aaq, Raq, and Oaq-phase are labeled and are delimited by narrow plateaus. Reproduced from ref. 62 (Eckhardt et al., Nature 1993, 362, 614) with permission of Macmillan Magazines.

Chiral Self-Assembled Structures

51

Figure 6. AFM images with corresponding Fourier pattern inserts of TCA monolayer films vertically transferred at 3 m.min -1 onto freshly cleaved, atomically flat, mica (muscovite 2M2 polymorph) plates. Images were obtained using a Nanoscope II AFM with selected silicon nitride scanningltips of 100 ,/k radius of curvature which were mounted on a cantilever of 0.38 N.m- force constant. The AFM was used in constant height mode with forces ranging from 10 to 50 nN. Greater tip forces caused disruption of the film but at no time was the amphiphile scraped from the surface. The instrument was calibrated with mica (muscovite 3T polymorph) which was scanned 40 times, as were the films, with subsequent determination of average lattice values. Thermal drift and tip drag were characterized and compensated for in the measurements. AFM image and Fourier pattern of (a) a 15 x 15 nm area of the monolayer Am-phase transferred to mica at 1.8 mN.m-1; (b) a 15 nm x 15 nm area of the monolayer Rm-phase transferred to mica at 7.5 mN.m-1; (c) a 10 nm x 10 nm area of the monolayer Ore-phase transferred to mica at 15.4 mN-m-1; (d) a 10 nm x 10 nm area of the same monolayer as in (c) but in a region where the packing is the mirror image of that observed in (c). The Fourier patterns in (c) and (d) differ in general shape because of relative orientation of the regions to the scan direction of the AFM tip. Patterns for the Om-phase are the best; those for the Am-phase are the worst. The larger errors in the Rm-phase are due to shorter orientational and translational coherence lengths of the film. Reproduced from ref. 62 (Eckhardt et al., Nature 1993, 362, 614) with permission of Macmillan Magazines.

52

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

Figure 7. (a) Unprocessed 10 nm x 10 nm atomic force microscope (AFM) image of trilayer LB films of calcium arachidate and (inset) its two-dimensional Fourier transform (FT). The molecular rows in the film are clearly visible, along with evidence of periodicity along the rows. The six strongest reflections in the FT (inset) form a skewed hexagonal pattern with three distinct lattice repeats of (A) 0.442 nm, (B) 0.576 nm, and (C) 0.686 nm. The sharpness of the reflections indicated long-range positional order in the films. The vertical and horizontal streaks in the FTs are the result of noise from the raster pattern of the AFM. (b,c) Mirror-image Fourier transforms of two different domains on a single LB film of calcium arachidate (CaA). The sequence of lattice spacings in (c) is the same as in (a), ABC, while the sequence in (b) is ACB, although the lattice dimensions are identical. The two FTs are not superimposable by any rotation within the plane ofthe image, bl and b2 show the reciprocal lattice vectors corresponding to the 4 x 2 unit of CaA (cf. Fig. c). Reproduced from ref. 30 (Viswanathan et al., Nature 1994, 368, 440) with permission of Macmillan Magazines.

Chiral Self-Assembled Structures

53

Figure 8. (a) Unit cell diagram of a 4 x 2 structure showing the eight-molecule unit cell dimensions al and a2 (determined from the position of the Fourier reflections bl and b2) as well as the local molecular packing dimensions within the unit cell Ul and u2 determined from the autocorrelation function (Figure B). The unit cell can be constructed by inverting every fourth local cell, corresponding to a specific type of stacking fault, as in barium arachidate films. The circles represent the position of the of the terminal methyl groups, with the shaded circles being displaced vertically by a chain repeat distance (2.54 ,~) relative to the lighter circles. (b) Structural model, based on local triclinic hydrocarbon packing (Kitaigorodskii's T[1/2 0]) with a stacking fault after four rows, that is quantitatively consistent with our data. The view is along the chain axis of the molecules. The triclinic structure is inherently chiral and we show the mirror plane (dotted line) and the right- and left-handed structures. The black lines represent the projected hydrocarbon backbone of the chain, the circles represent the radii of the hydrogen atoms. The shaded and unshaded circles represent the hydrogen atoms belonging to the even and odd carbons, respectively. The 4 x 2 unit cell can be constructed by inverting every fourth local cell, corresponding to a regular pattern of stacking faults that causes a jump in the structure both vertically and in the plane (see (a)). As in the AFM images, lighter (unshaded) molecules are higher, darker (shaded) molecules lower. Reproduced from ref. 30 (Viswanathan et al., Nature 1994, 368, 440) with permission of Macmillan Magazines.

54

MARTINUS C. FEITERSand ROELAND J. M. NOLTE

Figure 9. Different defect lines in LC domains: (a) chiral deformation in hexagonal domains; (b) point defects in quasi-pentagonal domains. Reproduced from ref. 65 (Chi et al., Langmuir 1998, 14, 875) with permission of the American Chemical Society.

Chiral Self-Assembled Structures

55

mid-1960s that it was pointed out that platinum shadowing affects the handedness of structures observed in EM (electron microscopy). 67 It is usually required to obtain more than one image of an object, by tilting of the specimen, for a three-dimensional structure, including the handedness, to be reconstructed from two-dimensional images. 73 The sense of twist in the fibers of 12-hydroxystearic acid was found to be related to the configuration of the stereocenter in the fatty acid, viz. right-handed for the Li salt of the D-acid, and left-handed for the Li salt of the L-acid (Figure 10). 74 The twisting of these helices appeared to be dependent on the metal ion, and the study of a series of alkali metals of 12-hydroxystearic acid gave unexpected results. 75 The handedness of the helix turned out to be (P) for the Li salt of the D-acid, but (M) for the Rb and Cs salts. With increasing ionic radius of the metals the handedness changed from (P) to (M) for the D-acid, while the opposite trend is observed for the L-acid. The Na and K salts represent an interesting intermediate case where left and right-handed fibers may be observed in a single sample. The free acids form fibers with handedness opposed to that of the Li salt. 76 Whereas the enantiomers of 12-hydroxystearic acid are thu~ found to form thin twisted fibers displaying enantiomorphism, the racemic acid produces platelets 77'78 and the racemic Li salt nontwisted fibers. 76 The use of higher alcohols (> C5) as solvents also led to the occurrence of twisted fibers of both helix senses for the free acid and the Li salt. 79 The use of enantiopure L-amyl alcohol as the solvent had no effect on the formation of the fibers nor on their handedness. Monolayers of the acids at the air-water interface were also studied. 78 The isotherms of the pure enantiomers were identical, and an area of 24 ]k2 per molecule

Figure 10. Left- and right-handed helices of 12-hydroxystearic acid. Adapted from ref. 76 (Tachibana and Kambara, Bull. Chem. Soc. Jpn. 1969, 242, 3422) with permission of the Chemical Society of Japan.

56

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

could be determined, corresponding to an erect conformation of the molecule in which only the carboxyl group is in contact with the water. There was a plateau extending to 95 /~2, which presumably corresponds to the molecular area of a molecule in a conformation in which both polar groups (carboxyl and hydroxyl) are in contact with the subphase. The structures in the collapsed monolayer after transfer of thin layers to a collodion-coated EM grid and Pt shadowing corresponded to those observed in solution. The fact that the plateau extends to slightly larger molecular area for the racemic mixture of the enantiomers indicates that the molecules start to interact during compression in the plateau region. The 12-hydroxystearic acid was found to show gelating behavior for a large variety of solvents. 8~

1.5. Scope of the Review In this review we will discuss the relation between the chirality of molecules and aggregates that can be prepared from them by self-assembly. The interactions between the molecules may be of solvophobic nature, or may involve hydrogen bonding. The various assemblies will be discussed in sections where they are related to molecular examples from nature, 82 in the order: lipids and biomembrane components (hydrophobic interaction), peptides and proteins (amides with hydrogen bonding), carbohydrates (multiple non-amidic hydrogen bonds), and nucleotides (hydrogen bonding complimentarity). We will consider chirality in both three- and two-dimensional systems. Some interesting areas of chiral self-assembly that are not reviewed in detail here will be briefly summarized below. An active field of research is that of self-assembly of achiral, linear multi-ligand systems into chiral assemblies by coordination to metal ions, forming the so-called coordination helicates. 83-85 The multiple helices obtained in this way are reminiscent of the double helix of DNA. Typically achiral ligands are used, and mixtures of (P) and (M) helices are obtained. Partial spontaneous resolution has been reported for the Ni helicate with ligand 16, 86 and complete resolution in the case of a helicate of ligand 17 which was formed as a Co 2§ complex and then oxidized to Co 3§ to stabilize it (Chart 3). 87 With chiral ligands, helicates of one type of handedness are obtained in excess. 88An example is the so-called chiragen ligand 18. 89Helices have also been observed in the case of metal-free ligands, 9~ emphasizing the analogy with the helicenes. 91 Because of its unique preference for tetrahedral coordination, which is like the tetrahedral geometry of sp3-hybridized carbon in organic compounds, Cu + ions are most used in these studies, 88d,92 but helical complexes are now also known for Cu 2+ 93 and for other metal ions. 94 The principle of helicate formation can also be applied to anions, e.g. the complexation of sulfate and acetate anions by the chiral oligoguanidinium ligand 19. 95 The area of chiral polymers is also a field of growing interest, with chiral structures reported for polymethacrylate esters and polyacetaldehydes, 96 polytrihaloacetaldehydes, 97 polythiophenes, 9s polyquinoxalin-2,3-diyles, 99 poly-arene-

57

Chiral Self-Assembled Structures

vinylene, 1~176 polyisocyanides, 1~ polystyrene/polyisocyanide block copolymers, 1~ poly-(m-phenylene), 1~ poly-binaphthols, 1~ polyphenylacetylenes, 1~ poly(ethyleneoxide), 1~ polyisocyanates, 1~176 and polyphthalocyaninato-polysiloxanes, ll~ but these will not be treated in detail here, and the interested reader is referred elsewhere. 96 The helical conformations of some chiral polymers bear some analogy to the ct-helix folding motif in the tertiary structure of proteins. An interesting aspect of helical polymers is the possibility to induce a predominant helical sense in a polymer of achiral monomers by copolymerization of a relatively small amount of an enantiomerically pure chiral monomer analogue, according to the so-called "sergeants-and-soldiers-principle." 109 Examples where polymerization of a racemate with one of the enantiomers in slight excess gives polymer products in which one handedness predominates ("majority rule") are also known. 1~ Another interesting related field that has recently started to develop is that of the construction of chiral hollow spheres 112 and chiral spaces in self-assembled cavitands.ll3,114

N ~N

N

NI

/

18

\

NI %

rP'C7H15",~O 0

N .~

..... S H

H

Chart 3.

~2

19

58

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

2. LIPIDS

2.1. Biological Lipids Biological amphiphiles that form membranes have been conceived as consisting of three parts: 115 two hydrophobic aliphatic chains, a hydrophilic head group, and a rigid connector (the 'hydrogen belt' region) linking the two other parts (Scheme 3). The finding that the isolated components of biological membranes spontaneously form bilayers when redispersed in water, ll6 was one of the first demonstrations that such molecules contain in their structure the information to form a well-defined aggregate. The driving force for the aggregation of amphiphilic molecules in water is considered to be the hydrophobic effect, ll7 The effect has been considered to arise from the strong attractive forces between water molecules, which have to be disrupted and distorted upon dissolving a solute; if the solute is nonpolar, the disruption of the attractive forces is not compensated for. Based on a consideration of the thermodynamics of solvation, both in water and in other solvents like hydrazine, the driving force of the hydrophobic effect has recently been redefined as a favorable overlap of the hydration shells of hydrophobic groups. 118'119 The first synthetic molecules that could be shown to form bilayers were quaternary ammonium salts with two long alkyl tails 20, which did not contain a connector part (Chart 4). 120-123 T h e bilayer membranes of these and other synthetic molecules have physicochemical properties and functions that are comparable to those of bilayer membranes that are constructed from biolipids, and they can give different aggregate shapes or suprastructures, dependent on their molecular structure and the physical circumstances, like temperature, concentration, etc. 124'125 Fluorocarbon chains can be incorporated to change the balance of hydrophobic and hydrophilic parts, or more generally, solvophobic and solvophilic parts in a molecule. Due to their extremely low surface tension, perfluoroalkyl compounds display low solubilities in both water or organic solvents, and this feature can be applied for the design of supramolecular assemblies in water 126-128 as well as in organic solvents. 129,13~ The relation between the shape of an amphiphile and the aggregate it forms has been discussed in terms of a packing parameter TM which relates the volume of an amphiphilic molecule to its length and the head group area. On the basis of packing considerations, amphiphiles with a single alkyl chain are predicted to form micelles or bilayers, those with two alkyl chains to form bilayers, and those with three alkyl chains inverted hexagonal phases. This is a rough guideline for amphiphile design, as the experimental results agree reasonably well with these predictions for amphiphiles with conventional aliphatic chains. 121 The relationships also hold for amphiphilic molecules that are much larger than the conventional ones, i.e. super amphiphiles consisting of polystyrene a tail with a polar dendrimer head group. 132 It should be noted, however, that many deviations occur. The presence of a rigid segment like the diphenylazomethine, 133the biphenyl, 134or the azobenzene moiety

Tail

Connector Spacer Head group Tail

Connector

ad group

Ln

Tail

Head

Connector Spacer Group

g P

Tail

Scheme 3. Generalized structure of amphiphiles.

Connectors

R, ,:CH3 ,,N,, R OH3

R

"~

20

Br--

O~CH 2 :

21

CH3(CH2)n ~

~(C

CH3(CH2)n ~

--

H2)m....../O-.r,u

O H",,J,,"O..a R'

I

'oI

,.-. O ~ / R 1"I2LJ" [ 1

R'~U O----~'~H O

u

II

( C H 2 ) m ' ~"OO =="'~"H O

26, DCm,n (DC8,9, m = 8, n = 9)

o

OH

o~ o

R

R'_0 0 ~-""~H 0 "ll" :." "

HO #

O

~

~

N dN

~

O-O~H O

H2C

P-. "O"~. O"

CH3(CH2) n ----

--~ (CH2)m..,._fO~r,u

CH3(CH2)n ~

~

II

.: ii OII H2C...,P..,.,~OH

O ~.u

CH3(CH2) n - -

--

.(CH2)m

CH3(CH2)n-

__

(CH2)m~N+~cH3

23

0

H2C...P...~ 0 &O

C.

H H ? . ~ : , ' . NHCH3

H"'~h "CH3

25

NO2 Chart 4.

y, ,2

(CH2)m-~"OO ~--"~ H O

27(m=8, n=9)

24a, DMPC, R = R' = n-C13H27 X=H 24b, phosphatidylserine X = CH2CHNH2COOH

R ...~O..c H2

"

O~CH 2

29 O,, N ~

:P,,

n-CeF17C2H40

L ~ O

CH 3

Br-

28

Chiral Self-Assembled Structures

61

(see examples below) may cause even single alkyl chain amphiphiles to align at the air-water interface and form bilayers. Due to n-stacking, such aromatic groups can have specific orientations with respect to one another, which manifest themselves as red or blue shifts in the UV-vis spectra, and enhancement of the CD spectra. It is of interest to know that a group of biological flower pigments, the anthocyans and flavones, also fall within this category. 135 The pigments owe their colors to specific assemblies in which n-stacking of the aromatic groups is modulated by carbohydrate substituents, and wavelength shifts as well as other changes in the UV and CD spectra are observed. Bilayer formation from single-chain ammonium amphiphiles is also possible by incorporation of moieties that can give multiple hydrogen bonding interactions, like substituted urea and acylurea groups. 136Another situation where application of the packing parameter concept can give misleading results occurs when the relative cross sections of polar head group(s) and alkyl chains are such that intercalation and/or tilting of the hydrophobic parts can occur (e.g. ref. 137). The conclusion of an extensive study on assemblies of chiral and non-chiral amphiphiles in which the number of alkyl tails was varied 121 is that the basic building block is a bilayer assembly of the amphiphilic molecules into a globule, and that the enhanced organization of globular aggregates leads to lamellae, rods, tubes, and disks (Scheme 4). The lamellar structure is a stack of bilayer membranes. A large curvature of bilayers combined with extension in one dimension leads to rod-like structures, which have a massive interior, and tubular structures, which have an aqueous cavity. A combination of flat bilayers with domains of large curvature results in disk-like structures. As will be shown in this and the following sections, some kind of intermolecular interaction is important for the expression of the molecular chirality in the shape of the aggregate. This interaction is typically between the connector regions (networks of amide groups connected by hydrogen bonding), sometimes in addition to interactions between the head groups (hydrogen bonding, shape complimentarity) and more rarely between the alkyl tails (specific orientation of diacetylene moieties). In biological membranes, the chirality of the component molecules is typically not reflected in the structure of the supramolecular assembly, e.g. the monolamellar vesicle that constitutes the biological cell membrane, the multilamellar systems of the thylakoids that are present in the photosynthetic machinery of chloroplasts, and the highly curved inner membranes of the mitochondria. It should be noted that even if the chirality of biological lipid molecules is not expressed in an assembly, the fact that the components are chiral is still important for the following reasons" 1. It allows for enantioselective interactions. 40 '47 For example, it has been shown that the interaction of L-DPPC 10 (R = R' = n-C15H31) with (-)-carvone or (R)-I at the air-water interface gives more expanded monolayers than the interaction of this compound with (+)-carvone or (S)-I. 138 This may be relevant to the mechanism of olfactory detection of these

62

MARTINUS C. FEITERSand ROELAND J. M. NOLTE

Rod

Vesicle

...~ Tube

Globule

Disk

Scheme 4. Aggregate morphologies of single chain amphiphiles. Reproduced from ref. 123 (Kunitake, ComprehensiveSupramolecular Chemistry, 1996, Vol. 9, p. 351)

with permission of Elsevier Science.

enantiomers as (-)-carvone, which smells of spearmint (as mentioned before), is a stronger odorant than (+)-carvone, which smells of caraway. 14 It stabilizes the lyotropic liquid crystalline state of biological assemblies relative to the crystalline state, due to the so-called "chiral bilayer effect, ''139 which will be discussed in more detail in Section 4.2. For example, 10-nonacosanol, extruded from the lipophilic wax layer of pine needles, forms fluid lipid tubules rather than crystals. Although it is difficult to establish the enantiopurity of the natural product, the fact that synthetic pure enantiomers produce tubules while the racemate gives platelets suggests that the biologically relevant morphology is attained because of the enantiopurity of the biomolecule. 14~ Although the chirality of biological lipids is not typically expressed in their functional assemblies, there are various examples of chiral aggregates of purified biological lipids. For example, helical intermediates are observed when a contact preparation of lecithin with water is monitored by polarizing microscopy. TM Addition of Ca 2§ ions to linear bilayers composed of 37 mol% bovine heart cardiolipin (21) and 63 mol% dimiristoylphosphatidyl choline (10, R' = R = n-C13H27) can induce both left- and right'handed helices (Figure 11).142 Two-dimensional chiral solid domains can be observed by epifluorescence optical microscopy upon rapid compression of enantiopure dipalmitoyllecithin (DPPC, 10, R' = R = n-C15H31) monolayers at the air-water interface (Figure 12), while the solid domains of racemic DPPC are found to be achiral. 143'144Addition of 2 mol% cholesterol (22)

Chiral Self-Assembled Structures

63

Figure 11. Helical liposomes derived from binary mixtures containing 37% cardiolipin 21 and 63 mol% DPMC (10, R = R' = n-C13H27) as Ca 2+ ions diffuse from a 0.01 M CaCl2 solution at the edge of a cover slip into the region where the liposomes are located. (a-c) Illustrate the formation of a double helix initiated on the left as a hairpin loop, and on the right in an extended region of membrane-membrane contact. In (a) and (b) both helices are right-handed. In subsequent photographs (not shown) the helix at the right-hand end became unwound as the helix on the left continued to form. Subsequently, as the right-handed helix to the left continued to form, a lefthanded helix formed on the right, and the two helics met in a soliton-like defect indicated by the arrow in (c). A single helix is shown in (d). Scale bars, 25 t~m. The elapsed time from (a) to (c) is approx. 5 min; double helix formation is not continuous but proceeds in a stepwise fashion. A similar length of time was observed for the continuous formation of the single helix shown completed in (d). Reproduced from ref. 142 (Lin et al., Nature 1982, 296, 164) with permission of Macmillan Magazines.

64

MARTINUS C. FEITERSand ROELAND J. M. NOLTE

Figure 12. Epifluorescence (fluorescent probe, 23) photomicrograph of a monomolecular film of the phospholipid dipalmitoyl phosphatidyl choline (10, R' = R = n-C15H31) at the air-water interface. The black regions are composed of solid-phase lipid, and the white (fluorescent) regions are fluid-phase lipid containing about 1 tool% of a fluorescent lipid probe. (Top) Micrograph showing the onset of solid phase formation; bar, 50 I~m. (Middle) Micrograph showing formation of chiral solid domains when the phospholipid is one of the enantiomeric forms (R); bar, 50 I~m. (Bottom) Micrograph showing spiral forms of enantiomeric lipid when 2 tool% of cholesterol is included in the monolayer so as to reduce the line tension; bar, 30 I~m. Reproduced from ref. 146 (McConnell and Keller, Proc. Natl. Acad. Sci. USA 1987, 84, 4706) with permission of the Academy of Sciences of the USA.

Ch iral Self-Assembled Structures

65

reduces the line tension between the LE and LC regions, leading to very fine spiral structures, again with handedness depending on the enantiomer (Figure 12). 145-147 Similar observations have been made for dimiristoylphosphatidic acid (244) in the presence of 1% cholesterol (22). 148'149 In this case, the usual dendritic growth of liquid condensed domains was observed at neutral pH, and spiral structures were formed at pH 11, presumably because the phosphatidic acid head group is larger at that pH, inducing the right degree of chain tilt in the molecules of the monolayer to allow the expression of their chirality in the domains. The aggregation behavior of dispersions of lecithin with larger amounts of cholesterol is described in the cholesterol section (cf. Section 2.4). Dispersions ofdimyristoyl-sn-glycero-3-phosphocholine (10, R' = R = n-C13H27) can give helical ribbons besides vesicles, but the requirement for considerable incubation times at elevated temperatures to obtain the chiral structures suggests that degradation products, like free myristic acid, are also involved, possibly in a stoichiometric complex (1:2) with the lecithin. 15~ In fact, myristic acid and other fatty acids have been known for a long time to give long fibers at high pH. TM Chirality, as observed by CD, can be induced in these fibers by preparing the salts with ephedrinium counter ions (25). 152

2.2. Synthetic Phospholipids Monomeric lecithins with a diacetylenic function in their fatty acid chains, like 1,2-bis(tricosa- 10,12-diynoyl)-sn-glycero-3-phosphocholine (26 with m = 8, n = 9; initially called DC23PC, now DCs,9PC ) form liposomes in aqueous dispersions above the phase transition temperature (Tc 43 ~ for DCs,9PC). 7~ Upon gradual lowering of the temperature until a few degrees under the transition (38 ~ all liposomes transform to hollow tubules. 153 The diameters of the tubules range from 0.4 to 1 ktm with aspect ratios of 10-100. The walls vary in thickness from 2 to approximately 10 bilayers (10-50 nm). A series of DCm,nPC analogues including a set in which the position of the diacetylenic group (26, m = 4-15, n = 17-6) was v a r i e d 154'155 w a s synthesized. This revealed some structural prerequisites for tubule formation, 156 viz. that the diacetylenic moiety is required for tubule formation, that the position of the diacetylene group had little or no effect on the structure of the tubules, and that symmetry in the alkyl chain (m is equal or nearly equal to n) is preferred. The effects of the head group and of additives were also investigated. A DCPC with its trimethylammonium group substituted by a hydroxy group (27) formed tubules only in the presence of certain metal ions, e.g. Cu2§ 157 Also in the case of DCa,9PC, effects of ionic strength and pH were noted. 15s Addition of NaC1 or CaC12 in 1 M concentration resulted in shorter tubules. Extreme pH values resulted in a higher aspect ratio, mainly by reduction of the diameter, but also in lipid degradation by ester hydrolysis. The tubules and accompanying helical structures of comparable diameter can also be obtained in a different way, viz. by precipitation from solutions of DCs,9PC in ethanol upon addition of water below Tc, without even a transient appearance of

66

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

liposomes (Figure 13). 158 The precipitated helices are all fight-handed when the L-amphiphile is dispersed. The fact that the tubules can also be obtained in this way indicates that they are thermodynamically stable, and not accidental products of deforming liposomes. The ratio of precipitated helices and tubules as well as the respective dimensions (length, pitch, and diameter) depend on the ratio solvent/nonsolvent. The temperature, the concentration, and the solvent were found to have an effect on the nature of the precipitate. In the case of DCa,9PC (26) the diameters of the tubules and helices range from 0.3 to 3.0 l.tm, and the lengths from 50 to 1000 ktm, with the longest tubules found at high solvent concentrations. The length of the tubules precipitated from ethanol water (75:25, v/v) also appears to depend on the cooling rate. 159 The process of tubule formation was followed with low angle XRD (X-ray diffraction) and it was found that tubule formation is driven by a reversible first-order phase transition from an intralamellar, chain-melted La phase to a chain-frozen Lly-phase. The chain-melted La phase corresponds to multilamellar vesicles with bilayer interdigitation (d = 47.6 ,A,, length of lipid chain 28/~), while the chain-frozen Lly-phase corresponds to tubules with d = 66.1 A. The

Figure 13. Scanning electron micrograph of tubules and helices formed from DC8,9PC (26, m = 8, n = 9) at 50% 2-propanol in water that were subsequently coated with copper metal. Note that all helical structures are right-handed and that the pitch of the helices is somewhat variable. Bar, 2.48 I~m. Reproduced from ref. 158 (Georger et al., J. Am. Chem. Soc. 1987, 109, 6169) with permission of the American Chemical Society.

Chiral Self-Assembled Structures

67

ordered structure of the tubules has been confirmed in FT-IR (Fourier transform infrared spectroscopy) 16~studies. The acyl chains have very high t r a n s conformational order, with no interdigitation, as established by IR, Raman, X-ray, and electron diffraction. 16~ The presence of diacetylenic groups allows fixation of the self-assembled aggregate by polymerization through UV irradiation. 166 Upon such polymerization the helices and tubules retain their structure, which provides further evidence that they have some crystalline ordering, as polymerization can only take place if the diacetylene functions are appropriately oriented. 167 Tubules obtained by cooling of liposomes show a subtle helical signature in the form of regular spiral ripples, suggesting that they consist of regularly wrapped helical bilayer strips. 153 Careful inspection of the micrographs shows that the handedness of the spiral ripples is also right for the L-enantiomer, 169 in line with the observation made for the helices. This feature is also occasionally observed in tubules obtained by precipitation from ethanol-water (55/45, v/v%) mixtures. 158 Precipitation from methanol-water, however, yields tubules with walls of only one bilayer thick, and without apparent chirality. 17~All tubular structures, i.e. both those precipitated from ethanol and methanol, have been shown to be essentially chiral in nature because they exhibit CD effects which can only arise because the molecules in the tubules are at well-defined orientations with respect to one another. 171 Interestingly, when the amount of organic solvent is raised, a larger number of nontubular, helical structures arises, and structures with a gradual transition from tubules to loosely twisted strips are also observed. The helices are generally less tightly wound toward their ends. 158 Such structures are only occasionally observed upon cooling of liposomes. 153 The explanation is 158 that the helical structures have exposed hydrophobic bilayer edges which are stabilized by the organic solvent. Whereas the presence of water is absolutely required to prepare liposomes, it is possible to prepare tubules in pure acetonitrile, albeit of slightly different morphology than those obtained by cooling or precipitation in an aqueous environment. ~72 Interestingly, a racemic mixture of DC8,9PC also gives chiral structures, viz. both left- and right-handed helices, due to a lateral phase-separation of the pure enantiomers in their respective aggregates (Figure 14). 173 If the helices would simply be incomplete tubules, then these tubules, irrespective of the mode of preparation (precipitation or cooling), must be formed by one isomer of a pair of enantiomers. The lateral phase separation on the basis of chirality can be considered as a resolution of optical isomers by crystal growth from a solution, as first reported for tartaric acid. 2 Considering the fact that a phosphatidyl choline lacking diacetylenic functions, like DPPC (10, R - R' - n-C15H31), does not show resolution or lateral phase separation, 49'174 it can be concluded that the interactions between the chiral head groups of the PC molecules are not very important for the thermodynamics of the aforementioned gel-to-liquid crystal phase transition; the molecular packing of the gel phase is not such that the weak stereospecific interactions have significant influence. 173The tubular phase of DCs,9PC is probably more ordered than the phase

68

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

Figure 14. Electron micrograph of helical structures (diameter approx. 0.5 I.tm) formed from a racemic mixture of racemic DC8,gPC (26, m = 8, n = 9). Reproduced from ref. 173 (Singh et al., Chem. Phys. Lipids 1988, 47, 135) with permission of Elsevier Science.

of the other PC's, as was confirmed by IR and X-ray diffraction. This improved ordering, or stronger stereospecific interaction, must be accounted for by the "pseudochirality" of the unsaturated fatty acid chains. The presence of the diacetylenic group makes the alkyl tails "more chiral" in comparison to the saturated chain analogue. The van der Waals surfaces of saturated alkyl chains are essentially cylindrically symmetric, but this symmetry is broken due to the kink introduced by the diacetylenic groups. The fixed arrangement of two such bent chains has a nonsuperimposable mirror image, and constitutes a chiral object in itself. 173 The kink therefore imposes steric restraints on the ways that the molecules can be packed parallel to one another. This can impart either a counterclockwise or clockwise twist to the lipid bilayer they form. Due to the chirality of the head group, one orientation is energetically preferable. 169 The enhancement of the chirality of the amphiphile by the presence of a kink in the tail which imposes steric restraints on the packing of the hydrocarbon chains bears some analogy the aforementioned mechanism of the resolution of chiral domains in the monolayers of calcium arachidate (cf. Section 1.3). 62 The explanation of the formation of left- and right-handed tubules from mixtures of enantiomers by lateral phase separation 173 is supported by the

Chiral Self-Assembled Structures

69

finding that the intensity of the CD peak is proportional with the enantiomeric excess. 169 Although the tubular morphology is often found in studies of the self-assembly of chiral molecules, sometimes accompanied by assemblies that are indeed chiral like helices, it is of course not a chiral structure, and the question arises whether it can also be formed from nonchiral molecules. The fact that an amphiphile with two diacetylenic tails but no chirality in the head group (28, di(hexacosa-12,14diynyl)dimethylammonium bromide, DHDAB) gave no helices but showed cylindrical microstructures instead was considered as a demonstration that chirality is not an essential requirement for the formation of the latter structure. ~75 In later overviews, however, it was argued that chirality is required for the formation of tubules, and that the cylinders of 28 are in fact sheet-like structures that curl. 156 There are, however, other examples of tubule formation (diameter, 0.1-0.5 ~tm; length 10-50 lxm) from achiral surfactant, e.g. the single chain fluorocarbon dimorpholinophospho-amidate 29 which forms such structures upon dispersion (6%, w/w) in water at 50 ~ Various theories and mechanisms for the formation of the helical and tubular structures from liposomes and bilayers of synthetic DC8,9PC have been proposed 176 and their merits have been discussed in detail. 156 One model considers the electrostatic interactions between the polarization charges at the edges of a ferroelectric bilayer by analogy to the properties of chiral material in the smectic C phase. 177 The electric dipoles of chiral molecules with a preference for a certain orientation with respect to the bilayer do not cancel, 178 and this gives rise to ferroelectricity which causes a polarization of charges at the edges of the bilayer 177 and leads to a curling of the bilayer. If ferroelectric polarization were the only factor, however, the diameters of the helices and tubules would be expected to depend on electrolyte concentration, but such a dependence is not observed. 179 Based on considerations of competition between edge and curvature energies, phase diagrams for nonchiral vesicles have been proposed, which reveal a number of possible morphologies, including tubules. 18~The theory predicts that the diameters of the tubules and the helices would depend on the length, which is, however, experimentally not observed for diacetylenic phospholipids. Another more generally accepted theory has been put forward for the formation of helical structures, in particular tape-like structures, from chiral bilayer membranes. 181 This theory considers the bending elasticity of bilayers and assumes that the bending energy has two components, viz. a spontaneous torsion of the bilayer edges due to the chirality of the molecules, and the bending stiffness of the bilayer. It is in fact based on a competition between the spontaneous torsion and the bending rigidity. This theory is difficult to verify due to lack of knowledge of the appropriate parameters for the chiral bilayer. It can be derived, however, that the bending energy is minimal when the gradient angle of the helix is 45 ~ which is in agreement with most experimental observations. In a more extended model 182 the possibility of an anisotropic bending force, e.g. due to tilting of the chiral bilayer molecules, is also

70

MARTINUS C. FEITERS and ROELAND I. M. NOLTE

considered. This approach also leads to a gradient helix angle of 45 ~ The intrinsic bending force can occur in layers with a C 2 symmetry like, for example, bilayers and Sr smectic layers. It was found that this force works in crystalline layers if they are anisotropic and are composed of chiral molecules. Furthermore, it could be deduced that crystalline membranes do not require a tilt to be anisotropic. Calculations showed that the elastic energy of the crystalline membranes can be minimalized by the formation of tubes or tubules with a circular cross section. For liquid membranes these tubules only represent a relative energy minimum. Here too it is found that the gradient angle must be 45 ~. In the study of aggregates of cholesterol (cf. Section 2.4), the anisotropy part of the theory was worked out in more detail to account for the occurrence of helices with other gradient angles. The theory of bending elasticity has been further developed to include the effect of thermal fluctuations 183 and to account for the helical substructure of tubules. 184The calculations agree with the experimental observations that the tubule diameter is independent of electrolyte concentrations and tubule length. The following conclusions can be drawn from the theoretical developments as far as nanostructure engineering is concerned. 156 Tubule formation is driven by the chirality of the bilayer, with the tubule diameter depending on the magnitude of the "chirality" i.e. the favored twist in the molecular packing. Furthermore, tubule formation requires tilting of the molecules with respect to the bilayer, with divergence in the tubule diameter occurring when the tilt is decreased. The predictive value of these conclusions remains to be experimentally verified. Recent studies of solutions, vesicles, and tubules of DCs,9PC by CD show that the molar ellipticity in tubules is much enhanced relative to the solutions and vesicles. 171 This supports a refined mechanism of tubule formation 184'185 in which a membrane of chiral molecules first curves into wound ribbons, due to the favored twist between the chiral molecules, and then fuses into a cylindrical tubule (Figure 15). The CD spectra show a peak at 195 nm, which is associated with the chiral packing of the diacetylene groups within a single bilayer, and a peak at 202-205 nm, associated with chiral ordering of head groups between adjacent bilayers. 186 The fact that the latter peak is weaker in tubules precipitated from methanol than in those precipitated from ethanol is in line with the single bilayer nature of the former tubules mentioned earlier. 170 Using CD it was found that a crossover from single bilayer to multiple bilayer tubules occurs at lipid concentration of 5 mg/mL in 80:20 (v/v) methanol/water.187 This information could be used to prepare tubules with double-bilayer thick walls in methanol/water 85:15, or 2 - 4 bilayer thick but with a high aspect ratio (average length 60 t.tm) in methanol/ethanol/water 64:16:20.169 Interestingly, the CD spectra of tubules prepared of a 1 mg/mL solutions in 80:20 methanol/water of a series of lipids DCm,,,PC (26) with m + n = 21 reveal an odd-even effect. The CD spectra for the series with m = even indicate an increase in multibilayer character with the value of m, whereas the spectra with m = odd are independent of m. Tubules with odd m melt at lower temperature than those with even m, both in methanol/water 169 and in water. 16~ These effects are

Chiral Self-Assembled Structures

71

la

I?,

f

Figure 15. Schematic illustration of a chiral bilayer (A), with the molecules tilted with respect to the local layer normal. (The arrows indicate the direction of the molecular tilt, projected into the layer plane.) The favored twist between the chiral molecules leads the whole membrane to curve into (B), a wound ribbon, then fuse into (C), a cylindrical tube. The observed CD spectra come from the chiral molecular packing common to all three figures, not from the micrometer-scale helical structure in (B) and (C). Reproduced from ref. 171 (Schnur et al., Science 1994, 264, 945) with permission of the American Association for the Advancement of Science.

ascribed to the fact that the orientation of the kink in the diacetylenic acyl chains with respect to the chiral head group alternates back and forth when the number of methylene groups between them alternates from odd to even. 169

2.3. PhospholipidAnalogues The aggregation behavior and spectroscopic properties of a specific single-chain chiral phosphate amphiphile have been reported, lss The amphiphile, C12-Ala-AzoCsP (30), consists of a chiral L-Ala fragment, which is linked by an ester bond to a C12-alkyl tail, and by an amide to an azobenzene group, allowing monitoring of the aggregation process by spectroscopy (Chart 5). At pH 7 and pH 10, red shifts in the UV and enhanced CD-spectra were observed, while EM showed that the dispersion contained helical fibers and helical ribbons, respectively. The changes are consistent with a change in ordering of the phosphate head groups. These are predominantly monoanionic and hydrogen bonded at neutral pH, but deprotonation occurs at high pH leading to loss of hydrogen bonding and more charge repulsion. The amphiphile now effectively has a larger head group area which can only be accommodated in a bilayer by tilting of the hydrophobic part of the molecule. The effects of exchange of the phosphate protons by protonated amines or monocations can be explained along similar lines. The effect of Ca 2§ on the molecular organization of 30 is explained in terms of exchange of the buffer (Tris) cations by this ion, which results in a more compact packing of the amphiphiles and a considerable

72

MARTINUS C. FEITERS and ROELAND J. M. NOLTE 30

H3C;.

CHa(CH2),I ,0 O ~ 31

N

32

O.oR

ONa +

Ho~~OH

~ N

N

,

~

OH a: R = --HC'n'C1~ "n-C9H19

O

33 (

OH

HO

OH b: R

OH

N'~ .....//L-"O(CH2)sO'OH P-

~

H

~

34

~

O

.OH

.O... r" - 9 -"P\/---'X OH

Na O H O " O~ ~ ~r o

H

OH =

--HC'n'C2H408F17 ,,n_CgH19 c: R = "-'HC'n-C2H406F13.n_C9H19 Chart 5.

enhancement of the stability of the aggregate. The process is reversible upon addition of EDTA (ethylene diamine tetraacetic acid). A series of pyranoside 5-phosphates 31-34, based on glucose (31,34), galactose (32), and mannose (33), combined with various hydrophobic groups on the phosphate, viz. alkyl chains or mixed alkyl/fluorocarbon chains have been studied with respect to their aggregation behavior (Chart 5). 128 All glycophospholipids gave small unilamellar vesicles above their assumed Tc (60 ~ at pH 7, but unlike the galactose and mannose derivatives, those of glucose gave elongated structures upon cooling, while the dispersion turned into a gel. The hydrocarbon-substituted glucophospholipid 31a gave hollow multibilayer tubules with a polyhedral shape (diameter, 1 l.tm; length, 100 lam) at 4 ~ The tubules of the fluorinated glucophospholipid 31b which could be prepared at room temperature had a similar length but much smaller diameter (0.1-0.2 ktm). The difference between the glucose derivatives on the one hand and the galactose and mannose derivatives on the other was tentatively ascribed to a lower degree of hydration of the glucose head group, along with a less exposed negative charge of the phosphate. A group of phosphate containing amphiphiles (35), based on C4 sugars with phosphate groups at C1 and C4, and the hydroxyl groups at C2 and C3 esterified with stearic acid, exhibits cariostatic activity 189 and is also worth mentioning here for its rich aggregation chemistry (see Chart 6). The molecules are derived from tartaric acid stereoisomers [(R,R)- and (S,S)-isomers] or from erythritol (meso compound). 137'19~ The (S,S)-amphiphile [(S,S)-35, based on D-threitol] formed

73

Chiral Self-Assembled Structures

O,, ,OH iOi H2C-o'P\o-Na * CH3(CH2)16~O~--H H--~O,,y,- (CH2)16CH3

N-'-H

Na:::P\~~ -CH2 t~)

O H"JJ'-n'C17H35

H,

H O

n-CleH37 36

(R,R)-35

~-o,

n-C3H7

CH2 H H .... I....N'..

o ,CH '

Na+-O-P-O I O'Na +

(CH2)nCH3

O

n-C3H7

O-Na+

H+O'I~,O-Na, CH3(CH2)n'x/7__N:CH2 C) O

(R)-37; a, n = 10; b, n = 16

H

(R)-38; a, n = 10; b, n = 16

~ > - - " O"CH2 H H ....[....N', ..(CH2)nCH 3 O ,,CH2" ~ Na+-O-P-O

CH2

0

I

~~--O',cH

CH3(CH2)n ,~.__ NtCH 2 e'o'Na+ O

O-Na §

(R)-39; a, n = 10; b, n = 16

2

. . + o . ONa+ H

(/:i)-40; a, n = 10; b, n = 16

••'-

O"CH2 O'Na*

H+O'F~.oM CH3(CH2)n 'X~N:CH2 (~ O

H

(R)-41; a, n = 10; b, n = 16 Chart 6.

e

74

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

stacked bilayers at high surfactant concentration (80%) in 2 mM 1,4-piperazinediethanesulfonic acid (PIPES) buffer of pH 7. The fact that the cross section of the alkyl tails is not large enough to match the area occupied by the head group was found to be compensated for by tilting of the alkyl chains. 137 At low surfactant concentration (0.1%), small platelets were observed. In these aggregates, the area of the head group was even larger than in the stacked bilayers due to hydration, which was now compensated for by intercalation of the alkyl tails. Interestingly, when the surfactant was dispersed at much higher (25%) concentration in 0.1 M ammonium formate buffer, pH 8, and left at -18 ~ for 2 months, the formation of tubules, probably consisting of rolled-up aggregates, was observed. These tubules were found to assemble into thread-like structures (diameter approx. 500 nm). Comparing the chiral (S,S)-compound with its m e s o analogue ( m e s o - 3 5 ) , an interesting stereochemical difference came to light when Ca 2§ ions were added. 19~Both compounds formed vesicles when dispersed in 0.1% (w/w) concentration (diameter 15-25 nm for the chiral compound, 50-100 nm for the m e s o compound). Addition of Ca 2§ resulted in fusion of the vesicles of the chiral compound to give larger vesicles (diameter 50-100 nm), while the vesicles of the m e s o compound broke up into smaller ones (diameter 10-25 nm). These effects are ascribed to the differences in conformations of the surfactant molecules induced by complexation of Ca 2§ to the phosphate groups. The two phosphate groups in the m e s o compound can chelate Ca 2§ ions while maintaining a more or less parallel alignment of their alkyl tails. Complexation of Ca 2§ will therefore only result in charge compensation and dehydration of the head group, making it smaller, as confirmed by monolayer studies. This complexation destabilizes the intercalation of the alkyl tails of the molecules in the bilayer which will assume a nonintercalated structure with a higher curvature, leading to the formation of smaller vesicles. For the chiral compound, however, Ca 2§ ion complexation has to involve phosphate counterions of different molecules. This will stabilize the aggregate structure allowing formation of larger vesicles. Interesting chiral structures arose when the stereoisomers of the C4-phospholipid 35 were mixed with the histidine-based surfactant 36.191 Dispersion of pure 36 at pH 2.5 led to the formation of very long thin fibers, with some tendency to curl up into right-handed, twisted structures, while addition of copper triflate in a ratio Cu:36 = 1:4 gave boomerang-like scrolls. 192 Stereoselective interactions of the histidine surfactant 36 with the stereoisomers of the C4-phospholipid 35 were observed upon dispersion at pH 6.5.193 Combination of 36 with m e s o - 3 5 (25 or 50 mol% 36) gave vesicles (diameter 150-750 nm). Higher concentrations of 36 (75 mol%) in the mixture led to the formation of extended bilayers and multilayer structures. Combinations of the histidine surfactant 36 and the (R,R)-phospholipid (R,R)-35 led to irregular rod-like structures (diameter 30 nm, length 200-2000 nm). Helical aggregates were found for the combination with (S,S)-35, with an optimum at 30% histidine surfactant 36, suggesting that 1:2 complexes of 36 and 35 are the building blocks for the helical suprastructure. Monolayer studies showed an in-

Chiral Self-Assembled Structures

75

crease in compressibility when 36 was mixed with increasing amounts of (S,S)-35 till the 1:2 stoichiometry was reached. It was proposed that the 1"1 mixtures contain highly ordered, linear arrays of alternating phospholipid and histidine surfactant molecules, which are confined to their positions in a two-dimensional lattice. On the other hand, 2:1 complexes of 35 and 36 may have'more translational and rotational freedom and are free to adopt orientations in which the molecular area is minimized, and stereochemical interactions maximized. The failure of the C4-sugar phospholipids 35 to assemble into superstructures in which the chirality of the monomers was expressed is ascribed to the absence of strategically positioned functionalities like amides that would give intermolecular hydrogen bonds. Therefore, another series of phospholipids based on C3 sugars including amino groups was designed and characterized. Starting with stereoselective conversions of enantiopure epoxides (butyryloxy-substituted glycidyl derivatives) to aziridines, the regioisomers of the amide-containing phosphatidic acid analogues (disodium (2R)-3-butyryloxy-2-octadecanoylamino-propane-1-yl phosphate 37 and disodium (2S)-3-butyryloxy-l-octadecanoylamino-propane-2-yl phosphate 38) could be obtained by acylation, nucleophilic ring opening by dibenzylphosphates, and hydrogenolytic debenzylation. 193 The supramolecular assemblies of the regioisomers proved to be highly sensitive to the apparently small difference in molecular geometry. 194'195The regioisomer with the phosphate group in the 1-position (37) gave plate-like structures upon dispersion in water (2% w/w), drying, and Pt shadowing (Figure 16). The regioisomer with the phosphate in the 2-position (38) showed left-handed helical strands (diameter 22 nm), which further assembled to give rope-like structures (Figure 16). The observed helicity is the result of chirality within the molecular units and also of the specific complementarity of intermolecular interactions, which carries this information through to the macroscopic level. FT-IR showed that both regioisomers formed networks of hydrogen bonds involving the amides in the t r a n s conformation. In the case of the 1-phosphate 37 the ester carbonyl of the butyrate was part of this network, indicating that this moiety is in the so-called "hydrogen belt" area. 115'122For the 2-phosphate 38, hydrogen bonds between the butyrate ester carbonyl of the 2-phosphate and water were found, suggesting an extended conformation of the molecule with exposure of the butyrate group into the aqueous environment. 194'195 Owing to its more linear shape, the 2-phosphate 38 can pack more tightly than the 1-phosphate 37, and hence the chiral molecular information is not lost upon self-assembly. The differences in packing and connected expression of supramolecular chirality were corroborated by monolayer experiments using BAM. 56'57 The isotherm of the 1-phosphate 37 at pH 6.5-7.0 showed a transition to a liquid-condensed state, which was absent in the isotherm of the 2-phosphate 38, consistent with a higher degree of hydrocarbon chain organization in the latter c a s e . 194'195 BAM of the monolayers revealed the presence of chiral domains with a counter-clockwise pattern for the 2-phosphate 38, and no distinct morphology for the 1-phosphate 37 (Figure 17). Interestingly, the handedness of the chiral domains

76

MARTINUS C. FEITERSand ROELAND J. M. NOLTE

Figure 16. (a-f) Electron micrographs taken of 2% (w/w) dispersions of 37 and 38. (a) Planar structures of 37 (Pt shadowing, bar 250 nm). (b-c) Left-handed helices of 38 (Pt shadowing, bar (b) 500 and (c) 100 nm), (d-e) right-handed super helix of 38 ((d) non-stained, bar 500 nm, (e) freeze fracture, bar 125 nm). (f) Schematic representations of the model proposed for the chiral packing of DNA molecules in ref. 197. (a-e) Reproduced from the Ph.D. Thesis of Dr. N. A. J. M. Sommerdijk; (f) from ref. 197 (Reich et al., Biochemistry 1994, 33, 14177) with permission of the American Chemical Society.

Chiral Self-Assembled Structures

77

Figure 17. (a-e) Brewster angle micrographs (BAM) of surface monolayers of (R)-38 at (a-d) 30 and (e) 10 ~ on a subphase of pH 6.5, spot size 600 x 600 ~m. Micrographs taken at ]-1 = 5 mN.m -1 (a), 16 mN.m -1 (b), 18 mN.m -1 (c), 30 mN.m -1 (d), and15 mN.m -1 (e). (f) Fluorescence micrograph of a monolayer of (R)-38 containing 0.5 mol% of a fluorescence probe taken at r[ = 51 mN.m -1 (T = 20 ~ pH 6.5, spot diameter 200 l~m).

78

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

in the monolayer could be inverted by lowering the pH of the subphase to 1.5. The lowering of the pH probably results in protonation of the phosphate head group and hence in a reduction of the head group repulsion. This apparently gives a different long range tilt order in the domains which would explain the observed change in handedness. 196 Small changes in salt concentration or pH could induce the lefthanded helices formed by the 2-(R)-phosphate 38 in dispersion to coil up and form large right-handed superhelices 195 in a manner analogous to the formation of supercoiled DNA. 197 In subsequent studies, 195'198involving (S)-38, it was shown that the formed structures are enantiomorphic, with right-handed helices for the (S)-enantiomer. When the dispersions of the phosphate (S)-38 were followed in time, it turned out that immediately after sonication, vesicles (diameter 25-100 nm) were formed. One hour later most vesicles had fused to ribbons (width 50 nm). Right-handed helices (diameter 20-40 nm, pitch 85 nm) were observed after 24 h. The phospholipid analogues with phenoxy groups substituted for the butyryloxy groups 193 were also studied. 195 No chiral superstructures were observed upon dispersion at pH 6.5: the 1-phosphate regioisomer (39) formed vesicles and ribbons, and the 2-phosphate (40) fibers. Upon lowering the pH to 2.5, the ribbons of the 1-phosphate 39 started to twist to give left-handed helices, and ultimately tubules. The change from ribbons to tubules led to a new DSC (differential scanning

Figure 18. Scanning electron micrograph of a 0.1 w/v dispersion of 41 in 10 mM CaCI2. (Courtesy of Dr. P. J. J. A. Buynsters).

Chiral Self-Assembled Structures

79

calorimetry) transition, pointing to an increase in molecular organization, and to an increase in bilayer thickness from 34 to 45/~. The monomethylated 1-phosphate 41 also gave helical structures in the scanning electron microscope when dispersed in 0.1% solution in 10 mM CaC12 (Figure 18). 199 2.4. Cholesterol The formation of cholesterol (22) monohydrate crystals after supersaturation of bile in the gall bladder is related to a cholesterol gallstone disease. 2~176 The crystallization process involves a number of metastable intermediate structures, viz. filaments ~ helical ribbons ~ tubules ~ thermodynamically stable cholesterol monohydrate crystals. 2~ Detailed studies were carried out on model biles of the composition sodium taurocholate (42) (Chart 7): lecithin (10): cholesterol (22) in the molar ratio 97.5:0.8:1.7, with variations in the lecithin component (10 with R = n-ClsH31 en R' = CH3(CH2)7CH--CH(CH2) 7- (I) and with R = R' = n-ClsH31 (II)). 2~ Sixfold dilution of a stable micellar model bile in water resulted in a cholesterol-supersaturated bile containing both micelles and vesicles (Scheme 5). Filamentous structures were observed within 2-4 hours. After a few days these coexisted with helices, tubules, and plate-like crystals, and eventually, only platelike cholesterol monohydrate crystals remained (Figure 19). Helical structures were only found in model biles containing lecithin (I); the inclusion of lecithin (II) was not sufficient to obtain these structures. Initially, only high-pitch helices (diameter 3-15 Ixm; gradient angle 54 ~ were formed, but after l0 days only low-pitch helices (smaller diameter, gradient angle 11 ~ remained. Low-pitch helical structures were also observed by video-enhanced light microscopy in aged (3-6 weeks) human gall bladder bile. 2~ The theory already mentioned in the "Synthetic Phospholipids" section (cf. Section 2.2) was further refined to explain the various helical forms of aggregates of cholesterol (Scheme 6). (see below, cf. ref. 202) It is based on the expression derived 177afor the elastic free energy associated with a small deformation ^of chiral smectic phases, as adapted 2~ to a frame of reference of the unit vectors k (along the direction normal to the ribbon plane), e (along the projection of the molecular A tilt, d, onto the ribbon plane), and the third axis, the product ~ = ~r x ~. Similar expressions for the elastic free energy per unit area of a tubule have been presented elsewhere, ls2a In the study of the helical ribbons from cholesterol crystals, the elastic moduli were taken to be anisotropic and it was found that the gradient angle ~t depends only on the ratio of the elastic moduli, K, along the ~ and ~ vectors: tan4xl/= KCc/KPp.202 Organogel formation by cholesterol derivatives has also been studied. Of a series of cholesterol derivatives with polyaromatic substituents, the anthracene derivative 43 was found to gelate solvents as diverse as dodecane and 1-octanol. 2~ Upon inspection by scanning electron microscopy (SEM), the 1-octanol gel was found to contain ribbon-like rectangular fibers (26.3 * 8.2 nm), the majority of which had a

80

MARTINUSC. FEITERSandROELANDJ.M. NOLTE

HO

OH ~

HO"

V lV H

_(~0~N~SO3Na

"OH 43

O

44

CH~N~

O

N~N~~JJ~O

I

OH3 45, CHODAMA-5,n = 5 0

Br- I

0

Chart 7.

81

Chiral Self-Assembled Structures

Growth Patterns in Bile '

, ",, ",, ~1 " "', "',

High Pitch Helices

Filaments',,,, "',,,, I ~

~

,, , , , q

"""',,i"""",i'"'"'"',,,,

0

~

low Pitch Helices

Cholesterol

Monohydrate ._ Crystals

~,~~

Time Scheme 5. Sequences and relative stability of metastable intermediates plotted as functions of time after supersaturation of bile. Less-stable structures have higher chemical potential. Solid and dotted arrows represent, respectively, observed and presumed transitions. Reproduced from ref. 202 (Chung et al., Proc. Natl. Acad. Sci. USA 1993, 90, 11341) with permission from the Academy of Sciences of the USA.

a

~'x / h

P

s w pu-.~c6s~

Scheme 6. (a) Geometry of a helical ribbon: ~, radius; ~, pitch angle; s pitch; w, width; 6, width along the z axis. (b) Local coordinate system of a symmetric tilted bilayer. (c) Local coordinate system of a symmetric tilted bilayer. Helical lines with arrows show the direction of c. Reproduced from ref. 202 (Chung et al., Proc. Natl. Acad. Sci. USA 1993, 90, 11341) with permission from the Academy of Sciences of the USA.

82

MARTINUS C. FEITERSand ROELAND J. M. NOLTE

Figure 19. Typical helical and tubular structures in bile. (a and b) High pitch helical ribbon and helically grown tubule, respectively. (c and d) Similar structures of low pitch. (e) Fracture of a low pitch tubule. (f) Subsequent growth of the low pitch tubule in (e) into a plate-like cholesterol monohydrate crystals after 12 h. Bar, 20 I~m. Reproduced from ref. 202 (Chung et al., Proc. Natl. Acad. 5ci. U5A 1993, 90, 11341) with permission from the Academy of Sciences of the USA.

Chiral Self-Assembled Structures

83

left-handed screw (pitch 119 nm). Azobenzene-substituted cholesterol derivatives like 44 have also been studied. 2~ The azobenzene group allows an investigation of the aggregation behavior by CD spectroscopy, and the sign of the CD effect in cyclohexane gels was found to depend on the rate of cooling after dispersion. Investigations by SEM showed that slow cooling of gels of 44 produced fibers with left-handed helical structures, whereas rapid cooling produced fibers of the opposite handedness. In this context, the results that were obtained with the polymerizable cholesterol derivatives CHODAMA (45, n = 2) and CHODAMA-5 (45; n = 5) are worth mentioning. 2~176 Dissolution of CHODAMA in water followed by direct polymerization resulted in polymeric bilayer vesicles. 2~ When an aqueous solution of CHODAMA-5 (0.28 mM) was mixed with 0.01 molar equivalents of CaCI 2, sonicated for 15 min at 60 ~ aged at Tc (62 ~ for 1 day, and then polymerized, tubular structures (length, 10 ktm, diameter 0.7-1.0 I,tm) and large extended ordered aggregates were observed. 2~ These structures were presumably formed by fusion of the vesicles that were formed initially. When the polymerized dispersion was left. for 4 more days, large fibers were obtained (length 1 mm, diameter 10 ~tm), that possessed right-handed helicity (pitch, 32 ILtm). 2.5. Ceramides Cerebrosides are involved in extended (high axial ratio) membrane structures like myelin and the intestinal brush border, 2~176 and deposition of these compounds is a feature of lipid storage diseases like globoid-cell leukodistrophy 211 and Gaucher's disease. 212 Early EM and 31p-NMR (nuclear magnetic resonance) studies showed that pure cerebrosides gave multilamellar tubular structures, and that they reduced the size of dipalmitoyl phosphatidyl choline (10, R = R' = n-C15H31) liposomes, an effect that was counteracted by cholesterol. 213 The polymorphism of bovine brain galactocerebroside and its two major subfractions have been investigated in some detail. 214 The galactocerebroside (GalCer) class of lipids (Chart 8) mainly consists of N-acylated C 18-sphingosine (D(+)-erythro- 1,3-dihydroxy-2-amino-4(E)-octadecene) with small contributions of the C18-dihydrosphingosine and Cl6-sphingosine. The subfractions 46-48 (Chart 8) differ in the amide-linked fatty acids, which can be either hydroxy-substituted (HFA-Gal-Cer 46, with t~-hydroxy C24 acid) or non-hydroxy-substituted (NFA-Gal-Cer, with either a saturated C22 (47) or an t~-unsaturated C24 acid (48)). Aqueous dispersions of NFA-Gal-Cer showed multilamellar vesicles with occasionally twisted elongated ribbons, while Gal-Cer and HFA-Gal-Cer contained irregular lamellar structures. Because of solubility problems, the aggregation behavior of these materials was further studied in dispersions in 1,2-ethanediol/water mixtures (95:5%, v/v) prepared by thermal cycling. Under these conditions, both Gal-Cer and the HFA-GaI-Cer subfraction gave cloudy suspensions which upon inspection by phase contrast optical microscopy were found to contain thick mats of rigid needles (length 5-30 I,tm). The needles formed

OH/''/'j

OH

//

46

OH

47

~ OH 48

Chart 8.

~ OH

49

~8~O3

jjJJ

Chiral Self-Assembled Structures

85

from HFA-Gal-Cer were found by freeze-fracture TEM to consist of cochleate cylinders with a small hollow lumen (diameter < 7 nm) surrounded by a lamellar stack of bilayers. This morphology is comparable to that of the scrolls (rolled bilayers, or bilayer cylinders) which are formed from phosphatidyl serine (24b) vesicles under the influence of Ca 2§ ions. 215 Comparison of the X-ray diffractograms of HFA-Gal-Cer in aqueous dispersion and in 95% ethanediol with those reported for HFA-Gal-Cer samples with varying degrees of hydration 216 showed that the structures consisted of hydrated (repeat distance 65/~) and dehydrated (50 /~) bilayers, respectively, both without interdigitation. 214 In 95% ethanediol, NFAGal-Cer formed a semitransparent viscoelastic gel which as its predominant microstructure contained helical ribbons. Some of these were fiat with a width of 39-84 nm, a gradient angle between 51 and 75 ~ and a helical pitch between 270 and 605 nm. Cylindrical structures were also observed and had a smaller pitch. These studies were continued in 100% ethanediol and extended to HFA-Cer and sulfatide (S-Cer, galactosylceramide-1-sulfate, 49), 217 which were found to yield multilamellar and unilamellar cylinders, respectively, with no indication of helical structures. Mixtures of HFA-Cer or S-Cer with NFA-Cer also gave tubular aggregates except at high NFA-Cer content where helical structures predominated. In general, the structures found for NFA-Gal-Cer quite resembled those observed for the diacetylenic phosphatidylcholines (ref. 158, cf. Section 2.2). A number of amphiphiles based on galactose-appended amino acid serine have been presented as models for glycosphingolipids, but these did not give chiral superstructures. 218

3. AMINO ACIDS, PEPTIDES, AND PROTEINS 3.1. Polypeptides Although not strictly self-assembled structures, proteins are of interest in the context of this review as examples of biomacromolecules that contain in their monomer sequence the information to form stable functional structures which are held together by van der Waals forces, dipole-dipole interactions, and hydrogen bonds. 219Ribonuclease was the first enzyme for which it was shown that the protein could be denatured by certain chemical agents and then refolded to an active structure. 22~Depending on the amino acid sequence, polypeptide parts of proteins can fold into a number of structural motifs, most notably the so-called a-helix and the ~-sheet (Scheme 7). 221 In the a-helix, the backbone of a polypeptide consisting of L-amino acids is wound in such a way that a right-handed helix arises, in which the amino acid side chains point outwards, and each CO group is hydrogen bonded to the NH group of the amino acid that is 4 residues further in the primary structure. There are 3.6 amino acids per turn and the pitch is 5.4/~. Due to the fact that there are approximately 7 residues per 2 turns, the amino acid sequences giving rise to the t~-helix typically contain heptads, i.e. recurring sequences of 7 amino acids long. There are various approaches towards the classification of the amino acids with

86

MARTINUS C. FEITERSand ROELAND J. M. NOLTE

', '

N-H"

H--N

.O_\

"H--N N--H"

o~-helix

H-N

N-H

\

O

N-H

N--H- - - O:==~

H-N N-H

N-H "H-N

\

parallel 13-pleated sheet

H-N

\

X~=:=O

/

antiparalle113-pleated sheet

Scheme 7. Structural motives in polypeptides.

regard to their likelihood of occurring in an ct-helix. One is the O, P, C matrix in which the occurrence of an amino acid is multiplied with its preference for a structural motif to give a correlation. 222 Another approach is to classify them according to the value of the AAG required for transfer of an amino acid from a hydrophobic phase to water. 223 Sequences rich in Ala, Met, Glu, Leu, and Lys, are very likely to give (:t-helices, but Gly (in addition to being achiral) is too flexible. The conformational space for large amino acids like Leu, Met, and Phe is limited by the (:t-helix, and the ~-branching in Val, Thr, and Ile interferes with the carbonyl oxygen atoms in previous turns of the helix. Set, Asp, and Asn interfere with the (~-helix formation due to other hydrogen-bonding interactions, and Pro interrupts the helix because of the rigid five-membered ring. Further stabilizing forces, including side chain-side chain electrostatic interactions, side chain-helical dipole electrostatic interactions, H-bonded side chain-main chain interactions, have been discussed. 224 Some proteins, in particular those conferring the fibrous structure on biological materials like hair, muscle, skin, etc. consist of (~-helices only. In these biomaterials, two such (~-helices can interwine by "knobs into holes" interlocking of strategically placed hydrophobic side chains, giving (:t-helical coiled coils. 225 This arrangement of coiled coils has also been proposed for one of the earliest (semi)synthetic systems that show a chiral superstructure, the helical fibers that are obtained from dioxane or dimethylformamide solutions of poly-),-benzyl-L-glutamate. 67 '226 '227 These structures are enantiomorphic, with left- and right-handed helical fibrils for poly-Land poly-7-benzyl-D-glutamate, respectively. 67 IR spectra indicate the presence of an o~-helix, which on the basis of known peptide structures is expected to be right-handed for a poly-L peptide; the phenomenon that a right-handed helix forms

Chiral Self-Assembled Structures

87

a left-handed supercoil would be an analogy to the coiled coil known for the aforementioned protein structures. 225 Monolayer experiments show differences in the isotherms of enantiopure poly-y-benzyl-L-glutamate and its racemate. 228 Both isotherms gave a plateau, but the pure enantiomer can be compressed to a smaller area per residue than the racemate (21 vs. 24/~,2/residue), indicating that it can be much more closely packed. On the other hand, the compressibility was smaller for the pure enantiomer up to the pressure where the plateau appears, showing that it must form a film entirely different from that of the racemate. Some recent examples of completely synthetic polymers that give structures reminiscent of the coiled coil have already been mentioned briefly, viz. the polystyrene/polyisocyanide block copolymers 1~ and polyphthalocyaninato/polysiloxanes, lllb Globular proteins, like enzymes, usually contain a variety of structural motifs, like l-sheets and o~-helices connected by ffturns in which their direction is reversed. Membrane-linked proteins typically contain sections with a number of closely linked membrane-spanning tx-helices. A common arrangement of tx-helices in such proteins is the so-called four-helix bundle. The space between the four helices, which are close to parallel, can be used e.g. as an ion channel, or a substrate-binding pocket containing a catalytic cofactor, like in enzymes. Based on the "waist constraint, ''229 six classes of four-o~-helix bundles can be distinguished: square, splinter, x, unicomate, bicomate, and splayed. The tobacco mosaic virus coat protein consists of a four-helix bundle which has a small divergence due to the presence of a small part of ~-sheet. Due to this divergence, 17 protein molecules can aggregate in a disk. Two such disks can stack and accommodate in their center genetic material of the virus, the RNA. A process of further assembly then starts which confers helicity both on the protein coat as well as on the R N A . 230 A recent example of a helix bundle has been found in HIV (human immunodeficiency virus), namely in the core structure of the gp41 peptide of the HIV envelope glycoprotein. TM The surface glycoproteins of envelope viruses play a critical role in the initiation of viral infections. The gp41 peptide mediates the fusion of viral and cellular membranes. It is a membrane peptide and the ct-helix part of the peptide consists of six t~-helices that form a six-o~-helix bundle. Three N36 peptides form a parallel trimer that forms the inner bundle. Around it three C34 peptides are packed in an antiparallel fashion in the hydrophobic grooves at the surface of the trimer. There are some examples where chirality of the ~-sheet becomes manifest in the folded protein structure. The formation of extended regions of ~sheet helices has been linked to the formation of amyloid fibrils in a group of diseases, amyloidosis, which includes transmissible spongiform encephalopathies. For a protein, transthyretin, related to one of these diseases, familial amyloidotic polyneuropathy, structures of both the globular physiological form 232 and the fibrous pathological form are known. Mutation of a single Val to a Met results in a change from the globular form, containing ct-helices, to a tetrameric 13-helical form in which 13-strands with an average twist of 15 ~ between each strand give a complete turn

88

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

every 24 strands. 233The formation of"polar pleated ~-sheets ''234 from stable helical structures of synthetic enantiopure peptides can sometimes be induced by addition of the other enantiomer, as in the case of the methyl ester of polyglutamate in organic solvents, 235 or polylysine in water at pH 11.236 The explanation for this phenomenon is that the disorganized ends of two enantiomeric cz-helical structures form insoluble 13-sheets together, which then start to grow at the expense of the helical structures. This is an example of the so-called "chiral bilayer effect," which has been proposed to explain the relative stability of enantiopure micellar fibers of N,n-alkyl-gluconamides compared to the racemates, 139and will be discussed below. The polypeptide chains of the collagen monomer, tropocollagen, consist mainly of Gly, Pro, and two amino acids which are posttranscriptionally modified, viz. 4-hydroxyproline (Hyp) and 5-hydroxylysine (Hyl), in triads starting with Gly. The chains are wound in a different type of helix, the so-called II trans helix, which has a number of residues per turn, 3.0-3.3, which is quite similar to that of the o~-helix (3.6), but the rise per residue is much larger (approx. 3.0 ,/k vs. 1.15/~). Moreover, the collagen helix is left-handed whereas the cz-helix is right-handed. The helix is stabilized by steric repulsion of the Pro residues. Three polypeptide chains combine to form a triple-stranded tropocollagen fiber of 15/~ diameter and 3000/~ length, which is right-handed. There are no hydrogen bonds within each polypeptide chain but the three strands are strongly hydrogen bonded. The array of tropocollagen fiber assemblies is stabilized as a collagen polymer by cross-linking reactions involving the Hyl residues. 237 There are various examples of biologically functional structures, involving proteins, in which the chirality is visible by electron microscopy. An example is hemoglobin S, the form of hemoglobin as expressed in humans suffering from sickle-cell anemia, which differs from the standard hemoglobin A by mutation of one Glu to a Val. The deoxy form of hemoglobin S forms 14-strand helical fibers with a diameter of 215 A (Figure 20). 238 In another example, image analysis of an electron micrograph of a transverse section of skinned, relaxed freeze-substituted scallop muscle reveals that the myosin filaments have seven evenly spaced projections protruding from the backbone which are bent in a clockwise direction around the filament axis. 239 Other examples are the tubular packing of spherical proteins in biological microtubuli, 24~ and the helical array of protein subunits with an internal cavity that constitute bacterial pili. TM Lipid tubes with a helical signature like the ones described above for synthetic phospholipids have recently been used to mimic this helical arrangement of protein molecules using the strong biotinstreptavidin interaction. 242 The supramolecular morphology of the biotin-containing amphiphile DODA-EO2-biotin (50) (Chart 9), containing amide bonds as well as polyethylene glycol fragments, 243 has been studied separately. It was found that from a series of amphiphiles, including dioctadecylamine and the DODA-EOE-biotin analogues with diacetylenic tails or missing the biotin group and the biotin as well as the polyethylene glycol part, 50 was the only one that formed tubules. These tubules had a constant diameter (27 nm) and lengths ranging from a few hundred

Chiral Self-Assembled Structures

89

Figure 20. Negatively stained image and reconstructions of fiber of HbS (hemoglobin S). (a) Micrograph of fiber of HbS prepared from a sickled cell by direct lysis with negative stain on the electron microscope grid. (b) Two-dimensional reconstruction of the fiber of HbS using computer reconstruction techniques with the maxima from the Fourier transform. The output is recorded from a Tektronix graphics terminal. (c) Two-dimensional reconstruction as in (b), but with only the maxima of the layer lines 1-6 of the Fourier transform used. Reproduced from ref. 238 (Dykes et al., Nature 1978, 272, 506) with permission of Macmillan Magazines.

H i N'~o'~N O

,•COOH

r/'--N/~COO

NH2

O~

(S)-(z-Aminoisobutyricacid, 51

9

50

O

H

cooH

CH30./L~

H

53, Don

i

O

O

H I

55

in

57

58 I

I

H

I

H Chart 9.

I

H

56

i

b,n=l c,n=2 d,n=3 e,n=5

H

NH2

H2N~~~.OOH

H jfl

O

I

H

NH2

O i

'

54

..~ O

'

~ o ~ N O ' '2

52, Acc

0 ~ O H H2N !

I

i

H

a,n=5 b,n=7

Chiral Self-Assembled Structures

91

Figure 21. Images (a,c) and Fourier transforms (b,d) of helical crystals of streptavidin formed on lipid tubules containing DODA-EO2-biotin (50). (a,c) Stain striations extend along the tubules. Protein densities are particularly visible at tube edges, corresponding to streptavidin molecules viewed edge-on. Scale bar; 40 nm. (b,d) Distribution of Fourier transform amplitudes from the tubes shown in (a,c) corresponding to about 1700 streptavidin molecules. The fine spacing between layer lines indicates a helical repeat of 47 nm. Visible diffraction peaks extend up to 1.7 nm [arrowhead in (b)]. Reproduced from ref. 242 (Ringler et al., Chem. Eur. J. 1997, 3, 620) with permission of Wiley-VCH.

92

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

nanometers up to several micrometers. 244 It was found essential to remove the detergent, 13-octyl glycoside, which was used to disperse the amphiphile, by dialysis at 4 ~ dialysis at room temperature only gave vesicles, superimposed plates, and short rod-like structures. No helical signature was visible on the outside of the tubule. Nevertheless, the biotin-functionalized tubule assembled from 50 could be used as a starting point for the helical crystallization of the biotin-binding protein, streptavidin. 242 The tubule-protein assembly had a diameter of 38 /~, which corresponds to the sum of the tubule diameter and twice the thickness of the streptavidin molecule, and showed clear striations in the electron micrographs (Figure 21). Analysis of the electron diffraction patterns showed a helical repeat of 47 nm. The gradient angle was very small. Related to the o~-helix the so-called 310-helix with a higher pitch (1.94/~/residue vs. 1.56 for cz-helix) exists (Scheme 8). 245 It involves hydrogen bonds between amino acids that are three residues apart in the primary structure (cf. (z-helix 4 residues) and is promoted by incorporation of ct-aminoisobutyric acid (Aib, 51) residues. 246A transition from cz-helix to 310-helix can be induced in a model peptide by raising the temperature. 247 The transition can be monitored by incorporating amino acids with suitable acceptor (Acc, 52) and donor (Don, 53) side chains for fluorescence studies (Chart 9). Two molecules of the ion-channel-forming peptide gramicidin, which contains alternating L- and D-amino acids, have been shown to form a membrane-spanning ~-helix. 248'249 The alternation of L- and D-amino acids is a feature of a series of cyclic peptides that self-assemble into membrane spanning nanotubes. 249 [3-Helices have now also been shown to occur as a structural element in proteins consisting only of L-tX-amino acids, viz. in the enzyme pectate lyase. 25~Related to the naturally

0(2) ),..r ...,~,~%._

~C,(1 ) 1~

~ )

C,(:))

. r , i - 0(31

5 trims, ullns 0(~)

1 * - - 4 traru (m) 3.6~j I'~ek~r

(=-helix)

3.0w helix

Scheme 8. (a) The 3.613 helix (a-helix) and its building block, one of the 1 <-- 5 trans, trans intramolecularly hydrogen bonded peptide conformations (also termed o~-bend or C13-conformation). (b) The 3.010 helix (a-helix) and its building block, one of the 1 <-- 4 trans intramolecularly hydrogen-bonded peptide conformations (also termed type III 13-bend or C10-conformation). Reproduced from ref. 245 (Toniolo and Benedetti, Trends Biochem. Sci. 1991, 16, 350) with permission from Elsevier Science.

Chiral Self-Assembled Structures

93

occurring c~-amino acids, the possibility to form [3-helical structures from synthetic [3-amino acids has recently been considered. TM Examples are trans-2-arrfinocyclohexanecarboxylic acid 54, 252 giving polymer 55, and trans-2-aminocyclopentanecarboxylic acid 56, 253 giving polymer 57, which are found to form stable 14- and 12-helices, respectively. These so-called "foldamers" are considered to be relatively stable due to the conformational rigidity of the cycloalkane rings. It is also possible, however, 254-256 to form a 14-helix from a [3-peptide hexamer (58) which is alternatingly substituted in the [3-positions by isopropyl and methyl groups.

3.2. Amino Acid Amphiphiles The first observations of chiral aggregates from double-chain amphiphiles based on glutamate were reported independently and simultaneously. 68'69 Dialkyl oligoglutamates 5969 (Chart 10) of the general formula 2CmGlUn (m = 12 and 16, n = 14), in which the alkyl tails are connected to the carboxylates of the glutamate building block by amide links, gave slightly turbid solutions when dispersed (0.1 wt%) in water at pH 8-9. These were found (TEM, molybdate staining) to contain lamellar structures with widths of 3-5 nm for 2CI2Glu14 and 6-10 nm for 2C16Glul4. Close inspection of the electron micrographs of 2C12Glu14 revealed the presence of helical structures and tubes of comparable diameter. A peptide amphiphile based on L-Glu didodecylamide with a head group of a poly(L-Asp) (13 residues, 60) ultimately (24-48 h) also gave twisted fibers (width 3-5 nm) after dispersion in water of pH > 4. 257 Electron microscopy revealed various morphologies in the growth of helical structures from two untwisted filaments. An amphiphile based on L-Glu didodecylamide with a head group of a poly(L-Pro) (3 residues, 61, n = 3) formed fine fibrous assemblies upon incubation (30 min) in water (1 mg/mL), which gradually transformed into right-handed helical superstructures (length 10-100 lam).258 In the presence of BaC12, CaCI 2, MgC12, and KC1 the length of the structures was reduced (5-10 lxm), while NaC1, NHaC1, and FeC13 had no effect. 259 It was established by CD experiments that a reduction of the length of the helix is accompanied by a better packing of the amphiphiles, and a reduction of the membrane fluidity. This suggests that a complexation of the amphiphiles with the cations is the cause for the reduction in the length. The order in which the reducing effect decreases (Ba 2+ >> Ca 2+= Mg 2+ = K + > Na § = NH~ = Fe 3§ is in agreement with selectivity of cation binding by oligopeptides. 26~ When more Ba 2+ is added, the formation of crystallization nuclei is probably stimulated. As a result of a larger number of these nuclei the crystals will be shorter. In a more detailed study, TM the helical ribbons (now left-handed) of the amphiphilic Pro trimer 61 (n = 3) were shown to be connected to tubular structures. Extension of the peptide chain to 4 Pro residues led to a molecule 61 (n = 4) which gave stable vesicles. In this study, the amphiphilic Sar (sarcosine, N-methylated Gly, 62) trimer analogue was found to give twisted ribbons (width 10-200 nm). The Gly trimer analogue 63 just gave amorphous crystals, in spite of the chirality of the bis-alkylated Glu residue in this molecule.

94

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

.N-4 ~ H CH3(CH2)m/'ICN _-~ CHa(CH2)ml/I~__ ~

O H

O

H

5 9 a , m = 12, n = 14 5 9 b , m = 16, n = 14

# O O H J n.1

CHa(CH2)11/H--~ O H H--O

c.,CH=,11.,0

CmICH~),~--~ ~ ~ O NH H

C~176I 113

CHa(CH2)11\ 61

N N

--=

H /N

)

O

CH3(CH2)11~H~-~O 1 CH3(CH2)1N1~62 ~\O l x""-~lH3

H

CHa(CH2)11N -~(

O

-N

63

R OH

H

COOH

H H

3

64a, Trp, R = (2-indolyl)methyl 6 4 b , Ala, R = M e

64c, Val, R = i-Pr 64d, Leu, R = i-Bu 64e, Asp, R = -CH2COOH

Chart 10.

Another early observation of helical structures is that on a number of 2-hydroxydodecyl amino acid derivatives 64a-d. 262 The absolute configuration of the chiral carbon atom in the side chain, which was introduced by nucleophilic opening of 1,2-epoxydodecane by the amino group of the amino acid, was not determined in this study. The Trp derivative 64a gave rod-like aggregates regardless of the

Chiral Self-Assembled Structures

95

configuration of amino acid or side chain. Helical fibers with a right-handed twist were observed for the derivatives of L-Val (64e), L-Ala (64b), and L-Leu (64d), and fibers with a left-handed twist for the D-analogues; in all cases, however, the helical pitch was very large and irregular. In another study, 263 2-hydroxydodecyl aspartate 64e was not found to form fibers upon dispersion in water. Double-chain ammonium amphiphiles based on glutamic acid of the general formula 2Cn-GlU-CmN§ (65) typically give typical bilayer vesicles by dispersion in water with sonication above the phase transition temperature (Chart 11). 264 The dispersion of 2C12-L-Glu-CIIN § (65b, n = 12)68,265 showed, a few hours after cooling to room temperature, predominantly flexible filaments (length 5-50 ktm). Vesicles (average diameter of 1-10 txm) are present to a smaller degree. The filaments show rapid Brownian movements. Further aging of the dispersion for 1 day at room temperature resulted in structures that arose from twisting of the filaments (Figure 22). In a few hours, such a twisted filament changes completely into a helix, 68'265 as demonstrated by SEM with a low acceleration voltage (Figure 23). 266 In fact, enhanced CD effects gave an indication of the chirality of the superstructures of some of the glutamate amphiphiles with azobenzene substituents or absorbed dyes before this was established by microscopy. 264'266'268The presence of helices of 2CI4-(L)-Glu-Cll N§ (65b, n = 14) is reflected in an unusually large induced circular dichroism of a dye, 9-anthracene carboxylate, which is absorbed to the helical bilayers. 268 Further structural changes take place slowly and gradually. 68'265 After approximately 1 month, rod-like structures predominate. Detailed schemes for the transformation of bilayers to helices (Scheme 9) and from helices to tubules (Scheme 10) have been proposed. It is remarkable that the helical pitch of the twisted tapes or helical ribbons remains constant during the slow and gradual change to rods. Considering this it is likely that the aging of the helices to rods occurs by way of broadening of the tape. The enantiomer 2C12-(D)-Glu-C11N § (enantiomer of 65b, n = 12) was also investigated. The supramolecular structures show enantiomorphism, i.e. the helix is always right-handed for the L-amphiphile, and left-handed for the D-amphiphile. While the separate enantiomers clearly produce defined helices (albeit with opposed handedness) upon aging of the bilayer membranes, the dispersion of the racemic 2C12-(DL)-GIu-ClIN§ (65b, n = 12) does not show helices, but elastic fibers instead. The helical superstructures are only stable at temperatures below the phase transition temperatures (Te, 34 ~ for 2C12-(L)-Glu-CllN+) 68 of the respective bilayers or membranes. The helices change to spherical vesicles instantaneously when heated at 40-50 ~ These conversions are best studied by slow heating. It appears that, at temperatures close to Te, the pitch of the helix becomes smaller (e.g. from 3.2 to 2.8 Ixm for 2C12-(L)-Glu-CllN § causing the structure to change to a tube. This tube converts immediately to vesicles by breaking up. Not only the chirality of the amphiphile plays a crucial role for the morphology, but also its chemical structure. 2C14-(L)-Glu-CI1N§ (65b, n = 12) gives helices, while 2C12(L)-Glu-C2N§ (65a, n = 12) and 2CI4-(L)-GIu-C2N§ (65a, n = 14) only yield

96

MARTINUS C. FEITERSand ROELANDJ. M. NOLTE

HN'--~O ,,

O (CH2)m.I-N---- X" I

(~ CH3(CH2)nI~N__ ~

65, 2Cn-L-Glu-CmN* aX-CI, m=2 bX=Br, m=11

O CF3(CF2)TCH2CH

O

O N.~R

66a, R =

"~O(CH2)sO(CI..12)sCH3 Z -- (CH2)7CH=CH(C1"12)7CH3

66b, R =

CF3(CF2)7CH2CH2e-~ O

X

CH3(CH2)11OCH2CH2~/~ O

~

(CH2)m'l-~--"

67, 2(C12OC2)2-L-Glu'CmN* aX=CI, m=2 bX=Br, m=11

O 68

.O(CH2)11-N--~

(/"~

O CH3(CH2)8CCCC(CH2)9O-~H-~O (k CH3(CH2)8CCCC(CH2)9~' O Chart 11.

Et

(CH2)s--N*-'-EtBr" 69

Et

Chiral Self-Assembled Structures

97

Fisure 22. Dark-field op_tical micrographs of aqueous dispersions of 2C12-L-GluCllN + (65b, n = 12) (10-3 M). Aging condition: (a) 20 ~ several hours; (b) 15-20 ~ 1 day; (c) 5-6 hours after (b); (d) after 1 month at 15-20 ~ Bars, 10 I~m. Reproduced from ref. 265 (Nakashima et al., J. Am. Chem. Soc. 1985, 107, 509) with permission of the American Chemical Society.

98

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

Figure 23. Scanning electron micrograph (acceleration voltage, 16 kV) of a 2 mM dispersion of 2C12-L-Glu-Cll N § (65b, n = 12) in water which was aged for 1 day and dried on a carbon-coated Cu grid. Reproduced from ref. 123 (Kunitake, Comprehensive Supramolecular Chemistry 1996, Vol. 9, p. 351) with permission of Elsevier Science. spherical vesicles. Apparently, long alkyl spacers (high m) are required for the formation of helical structures. 68 The dimensions, the enantiomorphism, and fact that helices occur as intermediates for the final cylindrical structures of the glutamic acid amphiphiles 65 68,265 are all features analogous to those observed for the synthetic lecithins 26.158 A large variety of other bis-alkyl glutamate amphiphiles was also found to form chiral superstructures. The perfluorinated analogues of 2C14-(L)-Glu-CllN § with hydrophobic head groups, viz. aromatic (66a, ref. 129a) and unsaturated alkyl (66b, ref. 129b) were also studied and found to give twisted tapes in cyclohexane and butanone, respectively. Ether links could be introduced into the alkyl chain to yield amphiphiles of the type 2(C12OC2)-(L)-GIu-C2 N+ (67), and beautiful tubular superstructures as well as double-helical myelin figures (diameter 3-5 Ixm) were observed when these compounds were left as aqueous dispersions on the stage of a

Chiral Self-Assembled Structures

99 Growth

.~

/

0

Bilayer Vesicle

[~ J

~tr~fonnation |

/ ,

J ,

,

Scheme 9. Folding of bilayers to helical structures. Reproduced from ref. 123 (Kunitake, cc~mprehensive Supramolecular Chemistry 1996, Vol. 9, p. 351) with permission of Elsevier Science.

dark-field optical microscope for 30 min at room temperature (above Tc) (Figure 24). 269 In the same study, glutamate amphiphiles with a single unsaturation in the alkyl tail, viz. at position 9 in the octadecyl chain, were found not to produce helical superstructures. This phenomenon was attributed to the irregular packing, as demonstrated by DSC, of the alkyl chains containing the double bond. Related to this work, the effect of incorporation of polymerizable groups, viz. photoreactive 2,4-hexadienoyl (sorboyl) groups, into the alkyl chains of a glutamate amphiphile with a pyridinium head group (68) has been investigatedY ~ The polymerizable compound produced helical supramolecular aggregates (4-6 nm thick, diameter across the helix 25-30 nm). CD indicated that the sorboyl groups were stacked below Tc and the UV-irradiation-induced photoreaction of these groups was 25 times faster than above Tc. Upon polymerization the helical structures changed to tubular aggregates below Tc, and to twisted fibrous aggregates above Tc. Glutamate amphiphiles with diacetylenic alkyl tails like 69 were also investigated, and found a

b

J_..ZI //._// Scheme 10. Schematic representation of the folding process of a helix. (a), (b), and (c) are illustrations approximately corresponding to Figure 15. Reproduced from ref. 265 (Nakashima et al., J. Am. Chem. Soc. 1985, 107, 509) with permission of the American Chemical Society.

100

MARTINUS C. FEITERSand ROELAND J. M. NOLTE

Figure 24. Dark-field optical micrograph of an aqueous dispersion of 67b at 20-25 Bar, 10 p.m. Reproduced from ref. 123 (Kunitake, Comprehensive5upramolecular Chemistry 1996, Vol. 9, p. 351) with permission from Elsevier Science.

~

to show rapid and complete polymerization, but this did not result in chiral superstructures.271,272 A glutamate amphiphile with both carboxylate groups free but its amino group acylated with dodecanoic acid (70a) was also studied (Chart 12).263 A 1% dispersion in water gave a gel but only spherical particles (diameter < 100 nm) were observed. Analogous aspartate amphiphiles, with tetradecanoyl (70b) and hexadecanoyl (70c) chains, gave helical fibers (diameter 12-20 nm, helical pitch 65 nm) upon dispersion (1%) at pH values between 5 and 6. Amphiphiles 71 based on dialkyl (L)-aspartate linked to a Tris (tris(hydroxymethyl)aminomethane) group bearing one or more (D)-galactosyl head groups have also been investigated, and found to assemble into a variety of microstructures, including helices, tubules, and vesicles. 273The stereochemically pure mono-Gal amphiphile 71a (Tc 56.5 ~ gave long (several microns) hollow, open-ended tubules (diameter 450 nm), and the di-Gal amphiphile 71b (Te 52.1 ~ hollow multilayered tubes of similar dimensions. The mono- and digalactose derivatives 71a and 71b were also synthesized using a procedure which included deliberate racemization of the chiral center in the aspartate building block. 274 Because of the chirality of the galactose head group(s),

Chiral Self-Assembled Structures

HOO%N_O (CH2)n H (CH2)mCH3 "COOR

101

70a, n =2, m = 11 70b, n= 1,m= 13 70c, n = 1,m= 15

O

HN n 012H2/5

O ~-~n'CllH23

71a, RI = R2 =

H

71b, RI = J]-D-Gal, R2 = H

CH3(CH2)110-'~

O

H

/,-.N

I

Br"

O(CH2)4-N + -

I

72

O CH3(CH2)nO--~ H

2--N

H3C

73a, n = 6

~ - ~ N ,

o

./=~

I Br

73b, n = 7 73c, n = 8 73d, n = 9

73e, n = 10 73f, n = 11 73g, n = 12

73h, n = 13 HO~

O N

~

O

74a, R = OH 74b, R =

R

HN,,,,~, N H H N~

O n'C3H7

CH2 H H .... I....N'..

o ,CH' Y

Na*-O_P_O O-Na §

H'" (CH2)nCH3

O

H2)I3CH3

HO

(CH2)I3CH3

76a, R = H

H-N

\

0 OH / (CH2)laCH 3 (CH2)13CH3

76b, R = OH 3

O~ OH

(R)-75; a, n = 16; b, n = 10

Chart 12.

77

102

MARTINUS C. FEITERSand ROELAND I. M. NOLTE

which is not affected, the resulting product is a mixture of diastereomers, and should not be confused with a racemic mixture. It is therefore not surprising that the mixture of diastereomers of the mono-Gal amphiphile 71a showed no phase transition in DSC, and was found to give rigid bilayer sheets. The mixture of diastereomers of the di-Gal amphiphile 71b gave vesicles (100-1000 nm). The results demonstrate the importance of the precise design of the stereocenters for the formation of highly ordered assemblies like tubules. A 3 mM dispersion of an ammonium amphiphile based on L-Ala in conjunction with a rigid biphenyl group 72 produced fibrils (diameter 14 nm) after 1 h, which completely transformed into helices after 24 h. 275 The observed helical pitch was between 1.0 and 1.5 mm, and the width varied from 100 to 1000 nm, which almost corresponds to closed tubes. Both the L- and the D-form of an analogous azobenzene derivative 73f' (n = 11) were synthesized and characterized. 276 Both enantiomers initially (10-30 min) formed straight fibrils (diameter 10 nm) which turned into helices upon prolonged (4-24 h) incubation (Figure 25). These structures were enantiomorphic, with a right-handed helical sense for the L- and left-handed for the D-enantiomer. The racemate initially (10 min) gave irregular fibrils (diameter 1 0 - 1 2 nm) which were not as straight as those of the pure enantiomers. After 4 h,

Figure 25. Electron micrographs of aqueous dispersions of 73f (1 mM) stained by uranyl acetate. Scale bar, 1000 A. Open arrows point to right-handed helices, solid arrows to left-handed helices. Aging conditions: (A) L-73f, 30 ~ 30 min; (B) L/D-73f, 40 ~ 10 min; (C) D-73f, 40 ~ 10 min; (D) L-73f, 30 ~ 24 h; (E) L/D-73f, 40 ~ 4 h; (F) D-73f, 40 ~ 4 h. Reproduced from ref. 276 (Yamada et al., Chem. Lett. 1989, 568) with permission from the Chemical Society of Japan.

Chiral Self-Assembled Structures

103

both types of helices became observable, indicating a lateral phase separation of the racemate into enantiomeric aggregates. This is in analogy to the synthetic diacetylene phospholipids (26), discussed in Section 2.2.173 A series of amphiphiles of the type C,-L-Ala-Azo-C10 N§ (n = 6-13, 73a-h) was studied by UV-vis, XRD, and FT-IR and found to show a remarkable odd-even effect. 277 XRD indicated a bilayer structure with chain penetration for all amphiphiles, while IR and UV-vis revealed a parallel alignment of the alkyl chains and azobenzene chromophores for n = odd, and a perpendicular arrangement for n = even. In spite of these differences, helical structures were observed for all chain lengths. 278 A number of amphiphiles based on serine (74a) and histidine (74b), with 1-yne or 1-ene groups in the alkyl tails, were investigated. 274 The N-decanoylated serine derivative produced gels in dichloromethane, chloroform, and tetrachloromethane, which upon examination by electron microscopy were found to contain tubules, helical ribbons, and spheres. The N-decanoylated histidine analogues did not show gelation or aggregation in organic solvents, but gave spherical structures and helical rods in water. The 1-yne moiety at the end of the all-.yl tails did not perturb the supramolecular structure but gave characteristic UV-vis spectra. The morphology of the bis-alkyl surfactant 36 based on histidine 191'192 has already been discussed in the section on phospholipid analogues (Section 2.3). As its behavior is related to that of the histidine-based surfactants, the imidazole analogue 75193 of the phosphate-containing amphiphile 37 (cf. Section 2.3) is worth mentioning here. Whereas the stearic acid derivative 75a in 0.1% dispersion gave planar bilayer aggregates with intercalation of the hydrocarbon chains, the lauric acid derivative 75b gave fibers with a right-handed twist. 195Addition of 0.25 molar equivalent of CuSO 4 to 75a in 0.1% dispersion initially gave vesicles which transformed to right-handed helical ribbons (pitch, 103 nm; diameter, 30 nm) upon aging (1 day) at 4 ~ The combined results of XRD and EM pointed to presence of stacks of four intercalated bilayers which were twisted. Addition of Cu 2§ triflate initially also gave small vesicles, but in this case ageing led to giant vesicles (diameter 5 I.tm), again with intercalated bilayers. The difference between the triflate and sulfate was proposed to be due to the more hydrophobic character and larger size of the triflate anion. Monolayer experiments showed that the molecular area was larger for the complex with copper triflate than for that with sulfate in both the solid-analogous and liquid-expanded phases. Although only remotely linked to amino acids, because of their synthesis from serine, the aziridine carbinols 76 and their aggregation behavior are worth mentioning here. 28~ Dispersions of 76a and 76b at pH 3 gave vesicles (diameter 72-300 nm). In the case of 76b these transformed to bilayer structures after a few days, and finally to right-handed helical ribbons (width 100 nm, length 10 Ixm, pitch 1.2 ktm), while for 76a fibers (length 2.5 ~m, width 50 nm) were found. Interestingly, codispersion of a methanolic solution of (S)-76a with aspirin gave helices, both left- and fight-handed. The products of the reaction could be identified only in part

104

MARTINUS C. FEITERSand ROELAND J. M. NOLTE

and included (30%) a compound (77) that resulted from acetylation of the aziridine nitrogen followed by opening of the aziridine ring (now activated) by the salicylate. This product was not, however, responsible for the helices in its pure form, as it gave only rods upon dispersion in water. Comparative monolayer studies showed that 76b, in spite of the additional methyl group, could be compressed to a smaller molecular area than 76a. The latter compound showed a higher collapse pressure, indicating a more rigid film. Inspection of the monolayers with BAM revealed that 76a gave dendritic two-dimensional crystals in the LF.A,C coexistence phase, whereas 76b did not. Taken together, these results suggest that the packing of 76a in monolayers is determined by H-bonding interactions between the head groups, and that of 76b by steric interactions between the alkyl tails. 3.3. Other Assemblies Involving Amides Bisacylated t r a n s - 1,2,-diaminocyclohexane derivatives provide an astonishing illustration of the assembling power of the amide functionality. 281 The (R,R)-bisdodecanoyl derivative [(R,R)-78] gelates a large variety of solvents (Chart 13). The gelation occurs because the compound assembles into fibers which are chiral as demonstrated by an enhancement of the signals in the CD spectrum, and by the observation of right-handed helical structures in dried acetonitrile gels (width 40-70 nm). Studies including the (S,S)-enantiomer [(S,S)-78] showed that the structures are enantiomorphic. The equatorial amide NH and CO can orient themselves in an antiparallel fashion perpendicular to the cyclohexane ring, and an extended molecular tape can be formed by intermolecular hydrogen bonding. The tapes are proposed to interlock through van der Waals interactions to give the helical fibers. Terephthalic acid molecules that are connected by amide links to two 2-amino5-methyl-pyridine groups (79) give achiral structures when complexed with a dicarboxylic acid of the optimum geometry to form a 1:1 complex, in this case adipic acid (80). 282 The related isophthalic acid analogue 81 and pimelic acid 82 are themselves not chiral, but a helical crystal packing of the product of the molecules in a 1:1 complex has been reported. 283 The assembly is held together by hydrogen bonds between the undissociated carboxylic acid and the aminopyridine. Solid-state helical structures of oligoanthranilamides 83 have also been reported and were characterized by X-ray crystallography. 284 Chirality directed self-assembly has been investigated for a chiral bicyclic bis-lactam 84. 285 The enantiomers were separated, and the crystal structures of one of the enantiomers and of the racemate were determined. The molecules in the racemate structure are assembled into an infinite undulating chain of alternating enantiomers, whereas in the structure of the pure enantiomer the molecules are assembled into cyclic tetrameric arrays. A new class of axially chiral lactams (85, 86) has also been studied. 286 The optically pure lactams were prepared but in this case only the crystal structures of the racemates

105

Chiral Self-Assembled Structures

O

, ~ ~ H N ' ~ n-C11H23 ~ n-C11H23 0

79

78

80 '3

O

O

oH•-N'•N "H

O

H"N- ""~O

HOOC~COOH

82

o

H

84

H HNN~~ I~NH (s)-85

(s)-8e

Chart 13.

could be obtained. One lactam (85) gave zig-zag chains of alternating enantiomers; the other (86) gave dimers of enantiomers; no helical structures were observed. Columnar stacks of 3,4,5-trialkoxy-benzoylated 3,3'-diamino-2,2'-bipyridine triamide derivatives of benzene-1,3,5-carboxylic acid (87, 88) have recently been studied in dilute alkane (dodecane) and chloroform solution (Chart 14). 287 In chloroform, neither the derivative with hexyloxy tails (88) nor the one with chiral citronellol tails (87) showed Cotton effects in the CD spectrum. The chiral derivative 87 showed such effects in dodecane, whereas the achiral derivative 88 showed them in the chiral solvent, (R)-2,6-dimethyloctane. Interestingly, addition of a small amount (2.5%) ofchira187 to stacks of achira188 in hexane induced a strong Cotton effect, similar in magnitude to that of the pure chiral derivative in hexane. This is an example of the "sergeant and soldiers" effect in a noncovalent assembly, and it is similar to the effects already established for stiff helical polymers like polyiso-

106

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

OR

RO

87, R= ~

OR

~I~N"H'N

88, R = -n-C6H13

I

~

/ NI..H,,INI.~ N

a~

"H

o

H-N

OR OR

Br" N+,/'~//--"~\ ~

89

~

~

~CH3 ,'~--N ~CH3

Chart 14.

cyanates, 1~ where small amounts of a chiral monomer in a polymer chain of achiral monomers or small enantiomeric excesses of chiral monomers can determine the helicity of the macromolecule.

4.

CARBOHYDRATES

4.1. Biological Polysaccharides The elucidation of a great number of helical structures of biological polysaccharides by fiber X-ray diffraction has been reviewed. 288 Cellulose, an example of a structural plant cell wall polysaccharide based on 13-1 --~ 4 linked D-glucopyranoside residues (Scheme 11), is known to occur in various crystalline allomorphs, I, II, III, and IV. Cellulose I and II consist of extended twofold helices with low diameter and high pitch, 289which run parallel 29~and antiparallel, 291'292respectively. Both forms have intramolecular hydrogen bonding networks (3-OH . . . O5)

107

Chiral Self-Assembled Structures B

HO ~OoZ H

HO

01~10~.,,,-~OH 0

O\

0 HO

0

\

n

CH20H

,o&~

CH:~OH OH

_o

CH:~

J

.o,-

OH

o OH

CHsOH

o

o OH

Scheme 11. Building blocks of biological polysaccharides.Top left, amylose; top right, cellulose; bottom, schizophyllan. between the helices, which confer structural stability on the material, but additional hydrogen bonds exist between the antiparallel helices in the cellulose II form, which is therefore the more stable material (Figure 26). Another important polysaccharide is chitin which is found in the cell wall of lower plants and the skeletal tissue of lower animals like arthropods and molluscs. It is a polymer of N-acetyl-13-D-glucosamine and has a twofold helical conformation much like that of cellulose, 293'294 with more hydrogen bonds between the helices due to the acetamido groups. Amylose, an important component of starch, the storage polysaccharide of plants, differs in its molecular structure from cellulose only in the stereochemistry of the glycosidic link between the D-glucopyranoside residues (Scheme 11), which is ct. This apparently small difference with cellulose, where it is [3, has a large effect on the supramolecular structure. Various types of crystalline amylose are distinguished, viz. the A-starch of cereals and the B-starch from tubules, which both give helices. 295 Amylose can also be crystallized from organic solvents yielding socalled V-amylose. High-resolution X-ray and electron diffraction studies show that the structure of A-amylose is a sixfold, left-handed, parallel double helix with a pitch of 21.38/~ (Figure 27). 296 Interchain hydrogen bonding only occurs by way of hydrogen bonding via water molecules. Apart from a difference in packing, B-amylose resembles A-amylose, giving a pitch of 20.8 fit. 297 In spite of the larger diameter of the helix compared to cellulose, there is no room for even a molecule as small as water in the helix interior. V-amylose, the form of amylose obtained by

108

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

Figure 26. Antiparallel packing arrangement of the twofold helices of cellulose II. Stereo view of two unit cells approximately normal to the ac-plane. The two corner chains (open bonds) in the back form a hydrogen-bonded sheet. The center chain (filled bonds) is linked to the corner chains by hydrogen bonds. Reproduced from ref. 288 (Chandrasekaran, Adv. Carbohydr. Chem. Biochem. 1997, 52, 311) with permission of Academic Press.

precipitation with 1-butanol, consists of a helix with six glucose residues per turn. 29s The diameter of this helix depends on the degree of hydration. 299 In anhydrous V-amylose, there is stabilization by intrahelix hydrogen bonds (3-OH... 02, 6-OH... 03). 300The diameter is 4.5/~ which allows for the formation of inclusion complexes. In the presence of KBr, a more extended form of amylose can be characterized. 3~ Interestingly, a relatively short (six glucopyranoside units) amylose segment which is blocked on one side, viz. p-nitrophenyl-o~-maltohexaose, could be crystallized as the inclusion complex (p-nitrophenyl-ot-maltohexaose2).Ba(I3)2.22 H20. 3~ Elucidation of the structure by single-crystal X-ray diffraction gave an antiparallel, left-handed double helix with two triiodide units enclosed in the central cavity. The amylose molecule can also be assembled in thin films with an alkyl-hemicyanine dye (DASPC22, 89) included in its helical cavity. 3~ This confers spontaneous polar order on the film, leading to nonlinear optical effects. Another polysaccharide that can give a double helix is carrageenan. 304A number of microbial polysaccharides consisting of 13-(1 --->3)-glucans have been found to

Chiral Self-Assembled Structures

109

Figure 27. Parallel packing arrangement of sixfold A-amylose molecules. A stereo side view of less than 2 turns of a pair of double helices 10.62 A (= a/2) apart. The two strands in each helix are distinguished by open and filled bonds, and the helix axis is also drawn for convenience. Note that atom 0-6 mediates both intra- and inter-double helix bonds. Reproduced from ref. 288 (Chandrasekaran, Adv. Carbohydr. Chem. Biochem. 1997, 52, 311) with permission of Academic Press.

give (among other forms) triple helices, like curdlan, 3~ schizophyllan, 3~ (Scheme 11) and lentinan. 3~ Another polysaccharide which give triple helices is 13-1,3-D-xylan.3~ 4.2. Gluconamides

The first chiral superstructures from N,n-alkyl aldonamides were discovered when aqueous gels of N,n-octyl-D-gluconamide (D-Glu-8 or D-90a, concentration 1-50%) were prepared by dissolving this compound at high temperature and cooling until below 80 ~ (Chart 15).71 Freeze-fracture EM as well as negative staining showed that ropes with a right-handed helical twist (diameter 125 ]k, pitch 180/~,, gradient angle 35 ~ were present. Another type of structure found in these gels was one that resembled stacks of coins (diameter 160-180 ]k, thickness of the coin 70 ]k). Shorter chain gluconamides (heptyl, hexyl) formed similar gels at lower temperatures and ropes with smaller diameters. N-Methylated gluconamides (91) did not give gels. N-Alkanoyl-N'-gluconoyl-ethylenediamines, containing 2 non-

110

MARTINUS C. FEITERS and ROELAND J. M. NOLTE OH_ OH_

H

R..N

-

O

~

OH OH

OH

D-90, D-Gluconamide a, R = n-octyl b, R = n-hexyl r R = n-heptyl d, R = n-dodecyl e, R = n-octadecyl H

CH 3 1 OH_ OH_ R'~N ~ O

OH OH

OH

D-91, D-N-alkanoyI-N-methyl-glucamide

R'

O

~NH~ N

H

a, R'= n-pentyl b, R'= n-hexyl r R'= n-heptyl d, R'=n-nonyl e, R'= n-undecyl

_OH .OH

-

OH

~ O

OH OH

D-92, D-N-alkanoyI-N'-gluconoyl-ethylenediamine

OH OH

0

R" ~ O H O OH OH

R'"~N~ N , OH3

H

OH OH

-'~~~~'~OH O OH OH

D-94, D-Galactonamide D-93, D-N-alkanoyI-N-methyI-N'-gluconoyl-ethylenediamine

OH OH O

OH OH

D-95, D-Mannonamide OH OH : : O

OH OH

OH OH

D-98, D-AIIonamide

O

OH OH

D-96, D-Talonamide

OH OH :

H O

OH OH

OH OH

D-99, D-Altronamide

O

OH OH

D-97, D-Gulonamide OH OH

H O

OH OH

D-100, D-Idonamide

Chart 15.

methylated amide groups (92) gave smooth ribbons without twist, whereas, of the investigated (C6, C7, C8, C 10) only the decanoyl derivative gave a gel, which, like that of the single non-methylated octyl gluconamide 90, was found to contain right-handed helical ropes (diameter 100 A). These findings indicate that, in addition to the van der Waals attraction between the alkyl chains, the hydrogen bonds of the intermolecular amide network must be important for the aggregation process and the formation of well-defined chiral structures, and moreover, that, in the case of multiple amide links in the monomer, a misalignment of the amide links must be avoided. Interestingly, single-crystal X-ray diffraction structures of anhydrous N,n-octyl-D-

N-alkanoyl-N-methyl-N'-gluconoyl-ethylenediarnJnes (93)

Chiral Self-Assembled 5tructures

111

gluconamide 90a showed that the crystal packing of the molecules is head-to-tail 3~ which is an example of an enantiopolar crystal structure. 25b Further investigations of gels of N,n-octyl-D-gluconamide have revealed the existence of "bulgy doublehelix" assemblies which have single strands as thin as bilayers. 139 The structure of the bulgy helices formed in aqueous gels of D-Glu-8 has been investigated in detail by electron microscopy using phosphotungstate (1%) staining and image processing.3 l0 The electron micrographs showed almost crystalline two-dimensional arrays of fibers. As a result, a structure consisting of quadrupole helices of threads was proposed. More refined image analysis involving images obtained by phosphotungstate staining followed by rapid freezing and cryoelectron microscopy later led to a proposal of a left-handed helix consisting of six ribbons rather than four threads. 3~ Interestingly, the handedness of the structures obtained by phosphotungstate staining and cryomicroscopy was the opposite to that reported for transmission electron microscopy (TEM) of Pt-shadowed dried gels. The structure of N,n-octyl-D-gluconamide has also been investigated as adsorbates on mica and graphite by atomic fluorescence microscopy (AFM). 312 Fibers with a slightly tilted striation reminescent of rolls of coins are observed, which also appear to be left-handed (Figure 28). The structures are proposed to be formed by a process of fusion of spherical gluconamide micelles into a micellar cylinder, and a subsequent fusion of micellar cylinders into a micellar block. The N,n-octyl-L-gluconamide has also been prepared, and it was found that the superstructures of the N,n-octyl-gluconamides are enantiomorphic, and that the racemate only produces nonfibrous, nontwisted platelets. 139This led the authors to develop the concept of the "chiral bilayer effect" to explain that only enantiopure amphiphiles can produce helical fibers, and that rearrangement of the bilayer fibers to enantiopolar crystal layers is slow (Scheme 12). In the micellar fiber of an enantiopure molecule, the carbohydrate head groups are in contact with the water environment, and they must be dehydrated and turned over 180 ~ for the fiber to disassemble and give enantiopolar crystals. The dehydration and rotation would both be energetically unfavorable and hence slow processes in an aqueous environment. The condition that both these processes have to occur does not apply to the formation of precipitates from the racemic micellar fiber, which can therefore readily crystallize in a bilayer packing. In connection with the chiral bilayer effect it is worth mentioning that methods have been found to stabilize the micellar fibers, even at 60 ~ by adding 0.1-0.4% SDS (sodium dodecyl sulfate). 313 SDS is proposed to act by preventing the formation of head-to-tail (enantiopolar) sheets which would act as nucleation sites for crystallization of the gluconamide. The effect of configuration of the stereocenters in the carbohydrate head group has also been investigated by a systematic study of all eight diastereomers of N,n-octyl-gluconamide with the D-configuration (90a, 94a-100a), as well as the L-enantiomers and racemates of the galactonamide (L-94a), mannonamide (L-95a), and gluconamide (L-90a). 314 For galactonamide 94a, enantiomorphic "whisker" type aggregates were found (left-handed for D-enantiomer). The mannonamide 95a

112

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

Figure 28. (a)-(b) AFM height images of the rod-like structures of N-n-octyI-D-gluconamide (90a) on graphite. The cross-section profiles determined horizontally along the middle of (b) are shown below the image, where the vertical distances between the pairs of adjacent arrows are 9.6, 7.8, and 7.7 nm going from left to right. (c) Zoomed-in part of the image in (a}. The cross-section profile determined along the B-B line in (c) is shown above this image, where the vertical distance between the arrows is 8 nm and the small corrugations along this profile has a period of 9 nm. (d) Zoomed-in part of the image in (b). The contrast covers height variations in the 0-40 nm range in (a) and in the 0-50 nm range in (b). Reproduced from ref. 312 (Tuzov et al., New. J. Chem. 1996, 20, 37) with permission of Gauthier-Villars.

Chiral Self-Assembled Structures

113

Mi~l~

Cryml

Slow

One eMntionwr

9

9~,~

Fist

Scheme 12. The chiral bilayer effect: (a) chiral micellar cylinders rearrange slowly to enantiopolar crystals; (b) the hydrophobic bilayer of achiral micellar cylinders is retained in the crystal. Crystallization is fast. Reproduced from ref. 139 (Fuhrhop et al., J. Am. Chem. 5oc. 1987, 109, 3387) with permission of the American Chemical Society.

formed cochleate cylinders, both in water and 1,2-xylene, which look like cigars, with an angle of 45 ~ between the edges of the sheet and the long axis. Both the galactonamide and mannonamide gels were much more stable than the gels from the gluconamides described earlier. Interesting phenomena were observed in experiments with N,n-octadecyl-L-mannonamide (L-95e), which is completely insoluble in water but can be dispersed in boiling 4% SDS. 313 At the first appearance of turbidity, upon cooling to 70 ~ helices with both screw senses were detected (Figure 29). This is ascribed to a high level of hydration of the head groups in the SDS micelles, inducing statistical screw dislocations in both directions, which are also due to the fact that the growth process from SDS micelles is relatively fast. The molecules in the (M) and (P) helices are proposed to have the G § and G- conformations, respectively. 33 When 0.5% dodecylmaltoside was used as the stabilizing surfactant, the gelation was retarded, and only left-handed helices were observed. 313 The talonamide 96a crystallized rapidly from water, but occasionally formed whiskers, and in xylene, upon slow cooling, helical fibers. 314 The gulonamide 97a crystallized from water and gave rolled-up sheets in 1,2-xylene. The allon- (98a),

114

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

Figure 29. Octadecylmannonamide 95e dissolves in boiling micellar solutions of SDS or dedecylmaltoside (molar ratio 1:2) P- and M- helices are formed within minutes from SDS solutions at 70 ~ At room temperature they rearrange to bilayer scrolls. Reproduced from ref. 313 (Fuhrhop et al., J. Am. Chem. 5oc. 1990, 112, 4307) with permission of the American Chemical Society. altron- (99a), and idonamides (100a) were very soluble in water. For the glucon(90a), mannon- (95a), and galactonamide (94a) racemates, the solubility was much lowered compared to that of the enantiopure compounds, and platelets precipitated for Man and Glu, in line with the chiral bilayer effect. Gal gave stable gels containing long tubules. It was concluded that the differences in aggregation behavior between the diastereomers are related to the degree of all-trans conformation that is possible for a certain diastereomer. N-Octyl-D-galacton- (94a) and mannon-amide (95a), without 1,3-syn-hydroxyl groups, prefer the nondistorted all-anti conformation of the carbon chain, leading to an extended conformation and a preference for flat bilayer aggregates. The long and uniform ribbons that they exhibit are probably whiskers with screw dislocations. These dislocations can arise even in a bilayer between two monolayer lattices, and are probably caused by the interactions of the hydroxyl groups between the layers. On the other hand, the N-octyl-D- and L-gluconamides 90a as well as N-octyl-D-talonamide 96a form micellar cylinders and no ribbons or whiskers. This is probably caused by the bend in the head group, which has been found in the crystals of the corresponding polyol, and is the result of the 2,4-syn-interaction between the hydroxyl groups, leading to a relative broadening of the head group. 315 This bend has as a consequence that an aggregate with a large curvature is formed. The amphiphiles N-octyl-o-gulon-

Chiral Self-Assembled Structures

115

(97a), altron- (99a), allon- (98a), and idon-amide (100a) are water soluble due to the syn-positioned OH groups on C3 and C5, and do not form aggregates. In their most stable conformation, these carbohydrate head groups are bent in such a way that no regular chain of amide hydrogen bonds can be formed due to excessive hydration. The crystal structures of D-Gul-8 97a 316 and D-Tal-8 96a 317 have been reported and shown to contain tail-to-tail bilayers. The assemblies formed upon gelation of water by mixtures of aldonamides with varying chain lengths [octyl (a) vs. dodecyl (d), indicated as D-Ald-8 and D-AId- 12, respectively] and carbohydrate head group stereochemistry have been compared. 318 No direct resolution of racemates like that observed for the diynic phosphatidyl choline 26173 or azobenzene alanine amphiphile 73273 was found for the gluconamides in any case. Racemic Glu-8 (90a) or Glu- 12 (90d) just gave platelets. 318 It was possible, however, to observe "chain length-induced racemate resolution." Mixing of D-i31u-8 (90a) and D-Glu-12 (90d) initially gave a clear non-turbid gel containing spherical aggregates which soon whitened to give the helical fibers and knot structures known from the studies of the pure compounds, but without a separation in Glu-8 and Glu-12 fibers. Interestingly, mixing of D-Glu-8 (90a)and L-Glu-12 (L-90d) led to the formation of P- and M-helices which could be assigned unambiguously to pure D-Glu-8 and L-Glu-12 fibers, respectively (Figure 30). The assignment was possible both on the basis of the screw senses and the differences in the diameter of the helices, that of the (M)-helix (L-GIu-12) being approx. 1.5 times that of the (P)-helix (D-Glu-8). The "enantiomeric" fibers were only transiently observable, as ultimately multilayered thin crystals appeared. The latter were proposed to contain D-GIu-8/L-Glu-12 bilayers which is in line with the estimated layer thickness from the shadowing in the electron microscope (5 + 0.5 nm). Their formation is explained by a "retarded chiral bilayer effect" (Scheme 13). This means that the amphiphiles are first demixed by chain length into homogeneous fibers whichsubsequently combine to bilayers without chiral structure. There are also examples of resolution of diastereomers: while a 1" 1 mixture of D-Glu-8 (90a) and D-Man-8. (95a) showed no tendency to separate, separate helical and tubular fibers were observed for mixtures of D-Glu-8 (90a) and L-Man-8 (L-95d). In this case, the structures were assigned by an autoradiographic method using tritiated D-Glu-8. 319 These and other results taken together, led to the following rules of thumb for the interactions between N-alkylaldonamides. 318 In the case of amphiphiles with only a difference in chain length mixed structures are formed and no separation occurs. Possible separated intermediates are short-lived. Racemic modifications crystallize together to form platelets (Glu, Man) or tubes (Gal), 314 but different diastereomers with opposed configurations at C5 and C3 are separated. The examples given illustrate the importance of the stereochemistry of the head groups for obtaining the special morphology of the aggregates. Diynoic tails instead of alkyl tails have been incorporated in a number of hexonamides, 32~ and are considered to have a number of interesting features. 322 The rigid diacetylene unit had previously been found to be an important structural

116

MARTINUS C. FEITERSand ROELAND J. M. NOLTE

Figure 30. (P)-helices (D-Glu-8, 90a) and (M)-helices (L-Glu-12, L-90d) first separate (a; bar, 100 nm) and then (b; bar 300 nm) unite to form elongated "racemic" platelets. Reproduced from ref. 318 (Fuhrhop and Boettcher,J. Am. Chem. 5oc. 1990, 112, 1768) with permission of the American Chemical Society.

Chiral Self-Assembled Structures

L L L L .

.

.

0$I ON i,I o ~ N O14O140 ON 14

N

ON 0140

.

~

LL~ "

117

I

..f'./l

~.

,t~a,.

IMll. |

N O14 O14 N OH ON ON

ON

ON O

L-GIn 12

D-C~|

Scheme 13. Schematic representations of the chain-length induced racemate reso-

lution, a,b correspond to Figure 2 I. Reproduced from ref. 318 (Fuhrhop and Boettcher,

J. Am. Chem. Soc. 1990, 112, 1768) with permission of the American Chemical

Society.

component in the tubular and helical morphology of the diacetylene phosphatidyl cholines (26) (cf. Section 2.2, ref. 70). Moreover, diacetylene polymerization proceeds in a topotactic fashion 167'168 and is accompanied by the development of red-purple color due to the conjugated enyne backbone, so that it may be followed by UV-vis spectroscopy. The polymerization reaction can therefore be used to probe the ordering of the hydrophobic tails, and also to stabilize, by the formation of covalent bonds, the self-assembled aggregate with respect to chemical or thermal degradation. Finally, the assemblies of the diacetylenic molecules can be inspected in TEM without additional staining so that no artifacts are introduced. The morphologies of a number of diacetylenic (N-dodeca-5,7-diynyl) aldonamides (f) were studied and compared with their non-acetylenic dodecyl counterparts (d), 322 which are similar to the octyl analogues described before. 314 In the discussion of the morphologies, it proved convenient to sort the amphiphiles with carbohydrate parts of varying length by the configurations of the carbon atoms closest to the amide bond (Chart 16).322 The enantiopure diynoic galactonamide 94f gave helical ribbons (Figure 31, top), which were enantiomorphic (right-handed helix for the D-enantiomer), as well as closed hollow tubules (diameter approx. 1 I.tm), whereas the racemate gave planar assemblies. 322 An indication that this result can be explained with the chiral bilayer effect 139is the observation that the crystal structure

OH O

OH

OH

OH

OH

OH

O

OH

OH

OH

OH

D,D-102, D-Glycero-D-gluconamide D,L-104, D-Glycero-L-mannonamide OH O

OH

OH

OH : R

OH

O

O

OH

OH

O

O

OH

.OH

R"N~oH O

N R"

OH

L-103, L-Threonamide

OH

OH

O

OH

O

OH

L-105, L-Lyxonamide

a, R = n-octyl

b, R = n-hexyl c, R = n-heptyl d, R = n-dodecyl e, R = n-octadecyl f, R = n-dodeca-5,7-diynyl g, R = n-cleca-2,4-diynyl h, R = n-dodeca-4,6-diynyl i, R = n-tetradeca-6,8-diynyl

Chart 16.

OH

OH OH

OH

OH

D-97, D-Gulonamide

OH OH H : ! R.N . ~ O H

D-Xylonamide

L-101, L-Arabonamide H

OH

L-95, L-Mannonamide

OH OH H -- : R..N . ~ O H

OH

OH

H

D-90, D-Gluconamide

D-94, D - G a l a c t o n a m i d e

OH

OH :

f

OH H : R-'N'~oH O OH D-106, D-Erythronamide

Chiral Self-Assembled Structures

119

Figure 31. Transmission electron micrograph of dispersions (1.0 mg/mL water) of D-Gal-diyne (94f, top) and L-Ara-diyne (101f, middle and bottom). Reproduced from ref. 322 (Frankel and O'Brien, J. Am. Chem. Soc. 1994, /16, 10057) with permission of the American Chemical Society.

120

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

of the enantiopure N-dodeca-6,8-diynyl-D-gluconamide 90i 321 shows enantiopolar (head-to-tail) packing of the amphiphiles, in line with the packing in the crystal structure of the octyl-D-gluconamide 90a. 3~ It is worth noting that a comparison of the crystal structures of the diynoic gluconamides N-tetradeca-5,7-diyne-D-gluconamide 90i 321 and N-trideca-5,7-diyne-D-gluconamide, 323 which both give headto-tail packing, allows the conclusion that the diynyl gluconamides show odd-even effects as far as space group and conformation are concerned. Exposure of the dispersion of the diynoic galactonamide 94f to UV light (254 nm) gave a polymer in which the superstructure was retained. 32~The dodecyl galactonamides 94d gave similar structures. The diynoic L-arabonamide 101f gave chiral structures which consisted of fibers (diameter 30 nm) that were braided together and also could kink back on themselves (Figure 31, middle). 322 As many as four strands could be intertwined to form the braid (diameter 0.12 ktm, Figure 31, bottom). The dodecyl analogue 101d gave helices and fibers but no braids (Chart 16). The helical structures observed for the octyl gluconamide 90a 71'139 as well as for its dodecyl analogue 90d 322 were not found for the diynoic gluconamides. For the N-deca-2,4diynyl (90g), N-dodeca-4,6-diynyl (90h), and N-tetradeca-6,8-diynyl (90i) D-gluconamides as well as the N-tetradeca-6,8-diynyl L-gluconamide (L-90i) very thin tubules were found, 321 whereas the N-dodeca-5,7-diynyl derivatives of D-gluconamide 90t, the 7-carbon D-glycero-D-gluconamide 102f, and the 4-carbon D-threonamide 103f only gave planar sheets with no tendency to roll up. 322 The N-dodeca-5,7-diynyl derivative of the 7-carbon O-glycero-L-mannonamide (104f) gave helical ribbons (diameter 0.63 ktm) and tubules (diameter 0.30 btm, length 9.2 ktm), whereas the corresponding L-mannonamide (L-95f) and L-lyxonamide (105t") gave hollow tubules (diameters 0.37 and 0.27 ktm, lengths 2.9 and 6 }.tin, respectively), plus sheets for the lyxonamide. The gulonamide 97f gave tubular assemblies (diameter 0.4 lain), consisting of fibers that twisted around each other in a helical fashion. With the exception of the already mentioned gluconamide 90f/d and the arabonarnide 101f/d, there is a large similarity in the superstructures found for related diynoic and dodecyl aldonamides 322 as well as for the octyl analogues described earlier. 314 The results show that in the case of the aldonamides, the presence of the diyne is not a prerequisite for tubule formation as it is for the phosphatidyl cholines 26. 7~ Compared to the superstructures of the latter compounds, the tubules of the aldonamides are relatively robust, even if they are not polymerized. The diynoic galactonamide 94f, arabonamide 101t", glycero-mannonamide 104f and mannonamide 95f gave blue to blue-purple polymers 324 with a much higher conversion than the phosphatidyl cholines. A survey of crystal structures (of non-hydrated aldonamides) in combination with molecular modeling revealed the following. 322 Head-to-tail packing and dromic hydrogen bonding are associated with fiber-like supramolecular assemblies, whereas head-to-head packing produces planar, helical, or tubular assemblies. In the latter case, the hydrogen bonding patterns are simple, with hydrogen bonding involving the amide groups while some terminal carbohydrate oxygens are still available for interlayer hydro-

Chiral Self-Assembled Structures

121

gen bonding. This is the case for the gulonamide, where the crystal structure 316 permits interbilayer hydrogen bonding which is in line with the observed planar morphologies. 322The diyne chains have larger van der Waals distances between the alkyl chains, which can alter the head group packing, but only if the change is large enough to overcome other intermolecular associations. This happens in the case of gluconamide, where introduction of the diyne chain leads to loss of expression of the molecular chirality in the supramolecular structure. A number of D-gluconamides in which one or more hydroxyl groups are substituted or functionalized have been synthesized and characterized, 325'326 including bis(2,4;3,5)-dimethylene derivatives without (107) or with substituents in the 6 position (Chart 17). The 6-imidazolyl-6-deoxy derivative of this bis-methylene protected N-n-octyl-gluconamide (108a) showed interesting aggregation behavior in water (Figure 32). 325 At pH 4.5, below the pKa (6.28) of the compound, vesicles (diameter 160-780 nm) were found exclusively. The deprotonated compound present in dispersions in Tris buffer (pH 8.5) gave long multilayered fibers (diameter 100 nm, aspect ratio up to 500), and hollow tubules (diameter 3 I.tm). The Cu complex, formed at pH 8.5 in 1:4 Cu:ligand ratio, formed helices (diameter 300 nm), which appeared to consist of a "braid" in a coiled coil of four fibers (Figure 33). The crystal structure of the 6-imidazolyl 6-deoxy bismethylene protected N-n-octyl-gluconamide 108a showed bilayer packing of V-shaped molecules (bend at amide link) with strong intercalation. This is a deviation from the crystal structures of a variety of nonprotected N-alkyl gluconamides with various tail lengths 3~ which showed head-to-tail packing. Bilayer packing with intercalation is found in the crystal structures of other gluconamides, viz. N,n-octyl6-deoxy-D-gluconamide 329 and (1S,2S)-l,2-bis(D-gluconamido)cyclohexane. 33~ Studies of the thermotropic liquid crystalline behavior of a series of analogues of 108 with varying R TM indicated that Cu ions can be coordinated to the imidazoles between the bilayers. In the assembly of the helical braid, the role of the Cu ions would be merely to allow the fibers to coil up together. Another example of the organizing power of the Cu ion was found in the monolayer studies of the long-chain bis-methylene protected pyridine derivatives (109c) (see below, ref. 332). Organogel formation has been reported for a number of aldonamides, 314 viz. for N-octyl-D-gluconamide 90a, gulonamide 97a, and talonamide 96a. Unlike the supramolecular structures of these molecules in water (see above), the structures observed in the gels are all identical, viz. bilayer scrolls, which is an indication of the importance of head group hydration in obtaining the various morphologies in the former solvent. Functionalization of the 6-hydroxy group in D-gluconamide yields compounds that are able to gelate a surprisingly large range of solv e n t s . 325'333'334 The benzoate ester 110 recrystallizes from water, is insoluble in ether and n-hexane, and soluble in THF, but forms gels in methanol, ethanol, acetonitrile, acetone, dioxane, chloroform, ethyl acetate, dichloromethane, toluene, benzene, and 1,2-xylene (Figure 34A,B). 333 EM of the dried gels of the benzoate

1 22

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

H

O~O

R.. N

~ O

H

a

107

O

H

H

-

108

-

O

R"N

OvO

O ~ O

OH OH

110

O O

OH OH

H

OH OH

O

111

d

O

OvO

~

O

OH OH

q,..N~_.~

109

H

-OA-o

~J

H

R - , N ~ o O

O

115 O

OAO .

O

.

H

OvO

H

O~

=

OH OH

H OH ~ H ) HO~ Z,,,...r.~ N ~ ...~ O - - - ~ O H

CH3(CH2)10

O

o

I

.

H NH CF3(CF2)7CH2CH2 " - ~ O

O 114

116

N.,,ll.(CH2)nil/,

~N I

113

:

.

O

112x, R' = methyl 112y, R '= pentyl

OH OH

... H OH OH R" N ~ O O OH OH

Hi

,

OH OH

O

a, R = n-octyl b, R = n-dodecyl r R = n-hexadecyl

HO

O

R ' N ~ o ~ [ L R

O,,~, O

H

OH OH

HOOH ~

H

o.

N ~ ~": ~ 0

Chart 17.

-

~''~OH

OH OH

117

Chiral Self-Assembled Structures

123

Figure 32. Aggregates of 108a. (A) TEM of nonprotonated 108a; bar, 3.27 I~m. (B) Fiber of 108a being formed from a tape of multilayers; no staining; bar, 345 nm; inset, freeze-fracture electron micrograph of multilayer. (C) Schematic drawing of the process shown in (B). (D) SEM of hollow tubules and thin fibers of 108a; bar, 10 I~m. (E) TEM of vesicles formed of 108a at pH 4.5 (2% uranyl acetate on a hydrophilic carbon-coated copper grid); bar, 1.46 l~m. Reproduced from the Ph.D. Thesis of Dr. R. J. U. Hafkamp.

124

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

Figure 33.

Braided structures (diameter 330 nm, pitch 980 nm) formed by [(108a)aCu]. Transmission electron micrograph without staining on a formvar-coated copper grid; bar, 147 nm. Inset, overview of extended structure. Reproduced from ref. 326 with kind permission.

ester (110) in chloroform, or of the nicotinate (111) in dioxane showed fibers without twisting. Interestingly, helical structures were observed for the dried gel of the acetate (l12a) and hexanoate (l12b) esters in chloroform, while the gel of the imidazolyl compound 113 in this solvent gave twisted ribbons. 334 The cyclohexanoate 114 gel in ethyl acetate gave the most interesting structures, viz. helically wound fibers with variable pitch that were twined to form a rope with knobs at their ends (Figure 34C). It is tempting to propose that the fibers consist of antiparallel strands, and that the knobs at the end represent hairpins where the fibers make a turn. It was concluded that large aromatic substituents (apart from the imidazolyl

Chiral Self-Assembled Structures

125

Figure 34. TEM pictures (Pt shadowing) of (A,B) dried gels of 110a in chloroform, (A) network of fibers (bar, 1.5 l.tm), (B) bundles of whisker-type fibers (bar, 240 nm); (C) dried gel of 114a in ethyl acetate; (D) helical ribbon of [Pd(109c)2CI2] in THF. Reproduced from ref. 333 (Hafkamp et al., Chem. Commun. 1997, 545) with permission of the Royal Chemical Society.

group which appears to be too small) interfere with the transfer of the chirality from the molecule to the supramolecular aggregate, because formation of the H-bonding network is overruled by n-re stacking effects. Compounds which had the 6-hydroxy group functionalized and the other hydroxy groups protected as in (2,4;3,5 bismethylene derivatives) did not show gelation, but methylation of only the hydroxy

126

MARTINUS C. FEITERSand ROELAND J. M. NOLTE

group at C2 did not interfere with gelation in 1,2-xylene or benzene. Pd complexes of the 3-nicotinate ester 115 and the 4-pyridyl derivative 109 of the 2,4;3,5 bismethylene derivative were prepared and found to dissolve in THF and gelate this solvent, respectively. 333 Monolayer studies of aldonamides have also been reported. N-n-dodecyl-Lgluconamide (L-90d) did not exhibit a LC/LE coexistence phase but Brewster angle micrographs (BAM) allowed observation of dendritic LC domains with fixed angles between the stem and the branches (Figure 35A). 335 The racemate did not produce two-dimensional crystallites (Figure 35B). Chiral domains have been found for monolayers of nonprotected mannonamides 95d but not for gluconamides 90d. 336 The fact that the change in configuration of one of the hydroxyl-bearing carbon atoms appears to be essential for the expression of chirality in the superstructure suggests that a hydrogen-bonding network involving nonderivatized hydroxyl groups is important in addition to that of the amides. Bis-methylene protected hexadecyl gluconamides functionalized with 1-imidazolyl (108c) and 4-pyridyl groups (109c) have been reported. 332 BAM of the monolayers of 108e showed well-defined LC domains at low temperatures (10 and 15 ~ and pH 10, but lowering of the pH and addition of Cu 2§ ions only led to higher solubility and disruption of the structures, respectively. This is in contrast to what is observed for aqueous dispersions of the octyl analogue 108a. 325 The molecular area of 108c was determined by the cross section of the head group in all monolayers. 332 The pyridine-functionalized compound 109c was found to form stable monolayers with a large temperature stability (Figure 27). The observed value of the

Figure 35. Brewster angle micrographs (BAM) of (a) D-Glu-8 (90a) and (b) racemic Glu-8 (D/L-90a). Reproduced from ref. 33 (Fuhrhop and K6ning, The Synkinetic Approach, 1994) with permission of the Royal Chemical Society.

Chiral Self-Assembled Structures

127

molecular area (approximately 37/~k2) shows that this area again is determined by the head group and not by the alkyl tail cross section. Dendritic domains were observed in the LC/LE coexistence phase both at pH 7 and pH 10 (Figure 36, top). The domains have a particular shape as only fixed angles occur between the stem and the branches, viz. 78 and 60 ~ in remarkable agreement with the results reported for the nonprotected N-n-dodecyl-L-gluconamide L-90d. 335 Interestingly, for the pyridyl compound 109c, addition of Cu 2§ ions in the form of CuSO 4 (pH 7) or Cu(C104) 2 (pH 10) to the subphase led to the formation of chiral two-dimensional crystals (Figure 36, middle and bottom), 332 the first observation of this kind for gluconamide monolayers. As mentioned a b o v e , 336 changing the nonprotected gluconamide 90d to the nonprotected mannonamide 95d also leads to chiral LC domains, suggesting that hydrogen-bonding networks involving the free hydroxyl groups are essential for the expression of the chirality in the domain. The results for 109c, which has all the hydroxyl groups protected, indicate that other organizing principles, such as coordination to Cu 2§ ions, can give the same effect as hydrogen bonding between the free hydroxyl groups. The organizing effect of Cu 2§ ions as well as varying effects of counterions have also been noted in the studies of the assembly of the imidazole-containing surfactant 75a (see above, Section 2.4). 195 The hydrogen-bonding network of the amide together with the shape complimentarity of the cis-decalin moieties in the methylene-protected aldose and the organization induced in the assembly by coordination of the pyridines to the Cu 2+ ions are proposed to be the crucial factors in determining the expression of chirality in the monolayers of 109c. These results are also examples of how aldonamides can be covalently attached to other moieties, in this case a metal ligand like a pyridine or imidazole, and confer chirality on its self-assembled superstructure, the metal complex. This has also been shown for larger molecules, for example the construction of helical fibers of porphyrins by functionalization with gluconamides. 337

4.3. Other Carbohydrates A series of 1-glucosamide bolaamphiphiles of the general structure Glc-NC(n)NGlc (116), with n = 6, 9, 10, 11, 12, 13, and 14 has been studied (Chart 17). 338'339 An interesting odd-even effect is observed for these compounds. For n = odd, planar platelets and amorphous solids are found, whereas n = even gives rise to fibrous assemblies. The derivative 116 with n = 12 gives right-handed helical ribbons (width up to 3 I.tm, pitch l - 1 0 ~tm, aspect ratio several thousands) in water which can be observed by light microscopy (Figure 37). In some cases, two separate twisted fibers appear to grow out of one larger twisted fiber. The fibers can be isolated and dried and show extreme stability (> 1 year at temperature below 220 ~ The crystal structures of Glc-NC(11)N-Glc (116, n = 11) 34o and of the bisGal analogue of Glc-NC(12)N-Glc (116, n = 12) TM are known. In the former structure, successive layers of molecules have opposite orientations, and the directions of the amide hydrogen bonds are the same within each layer and alternate

128

MARTINUS C. FEITERSand ROELAND J. M. NOLTE

Figure 36. Monolayer isotherms of 109c on subphasesof (top) pH 10, 20 ~ (middle) pH 7, 20 ~ CuSO4; (bottom) pH 10, 20 ~ Cu(CIO4)2. Insets,corresponding Brewster

angle micrographs (BAM) of the LE/LC coexistence phase.

Chiral Self-Assembled Structures

129

Figure 37. (a) Crystalline and right-handed helical fibers made of GIc-NC(12)CN-GIc (116, n = 12) observed using polarized light microscopy (at 25 ~ in water). Periodical structures of the fibers are denoted by arrows. (b) Polarized light micrographs of representative dehydrated and right-handed fibers from GIc-NC(12)CN-GIc (116, n = 12), (top) photographed trough cross-polarized filters and (bottom) through plane-polarized filters. Reproduced from ref. 338 (Shimizu and Masuda, J. Am. Chem. Soc. 1997, 119, 2812)with permission of the American Chemical Society.

130

MARTINUS C. FEITERSand ROELAND J. M. NOLTE

between the layers. In the latter structure, the molecules are packed in parallel sheets, the carbonyl dipoles compensate each other within each molecule, and each molecule participates in antiparallel linear chains of hydrogen bonds. XRD of the dehydrated Glc-NC(n)N-Glc fibers shows that the packing of the alkyl chains is generally triclinic for n = odd, and orthorhombic or monoclinic for n = even. 338 From the low-angle XRD it can be derived that the layer thickness for n = 12 is 2.45 nm. A model for the internal molecular arrangement is proposed in which the bolaamphiphile molecules are tilted at an angle of 45-50 ~ The aggregation properties of a series of glycolipids, combining sugars, an amino acid, and both an alkyl and a perfluoroalkyl tail, have been studied and compound 117 was found to be the most interesting one. 342 As established by freeze-fracture electron microscopy, this compound displays a rich aggregation chemistry, with stacked disk-like assemblies, liposomes, helically twisted tapes, and tubules to be found in a single micrograph. No explanation for the variation in morphology was given. 5.

NUCLEOTIDES

5.1. Biological Polynucleotides There is a large variability possible in the structures of double stranded DNA due to the fact that (compared to polypeptides) many more bonds can be rotated in the backbone of each monomer (Scheme 14). The most common and physiologically most important structure is the B-DNA helix. It consists of two polynucleotide chains running in opposite direction which coil around a common axis to form a right-handed double helix. In the helix, the phosphate and deoxyribose units of each strand are on the outside, and the purine and pyrimidine bases on the inside. The purine and pyrimidine bases are paired by selective hydrogen bonds: adenine is paired with thymine, and guanine with cytosine (Scheme 15).343 The structure is very flexible and can form a supercoil with itself, 197 or around proteins. It can form a left-handed supercoil around histones to form nucleosomes which assemble in yet another helical structure to form chromatin. 344 One cause of the variability in the DNA structure is the fact that the five-membered ribose ring can be puckered in different ways, which is related to the extent of hydration of the phosphate groups. 345 The deoxy-ribose units in B-DNA are in the so-called C2.-endo conformation. In so-called A-DNA, the ribose units are in the C3,-endo conformation, which results in a tilting of the nucleotides with respect to the axis of the helix, and a lower degree of hydration of the phosphates. Due to the steric effects of the 2-OH group, the Cy-endo conformation is the only one possible for the ribose units in RNA (Scheme 14B), and double strands involving RNA therefore have a structure similar to A-DNA. For DNA itself, the A-DNA conformation is favored by low relative humidity. If the DNA sequence contains alternating pyrimidines and purines, the glycosidic bonds can alternate between anti and

131

Chiral Self-Assembled Structures

O

Me..~L.NH

HO"t~.~O~ "

.o. ~.~o

.NH2

0-%.

NH

).~ A) DNA

o

q.o o\

o. pk

B) RNA

o

o,

~..0

~

o

. Jl '<, JJ ~J. I~,~N- ""~ "NH~2 / / N ~ "NH

o, ..o

-

-

~B OHOH

~. ~L

O,,P~O"

~"~...~N N

OH

NH2CO H N - - ~

T

OH

o1-~~1

C) PNA, R~Horacridineintercalator

N~o

-

HN

II-lr~ P~'-~ ~ O

_ OH O

Scheme 14.

~

/--~

_ n N-~O

HN-R

Polynucleic acids.

syn. 346 This can lead to the formation of Z-DNA which has a dinucleotide as the repeating unit, resulting in a zigzag alignment of the backbone phosphates, and in a left-handed double helix. The Z-DNA structure is favored, among other factors, by the negative supercoiling that is possible in circular DNA. Other DNA structures, including cruciform and triple helix DNA (H-DNA) have been identified in bio-

132

MARTINUS C. FEITERSand ROELAND J. M. NOLTE H,

Base pairing:

H

Watson-Crick R

Base pairing: Hoogsteen

O.......H-N

H3C,

O

R

H Cytosine-Guanine

Thymine-Adenine H ~)....... H - N ~ N , ~ i

~ ~ N - H .... N , ~ R

0

N \

Thymine-Adenine Scheme 15. Base-pairing schemes.

logical structures. 347 The complimentary and flexibility of the DNA molecule, and its relative stability compared to RNA, make it an interesting starting point for the design of nanostructures, 348 including nanomechanical devices based on the B - Z transition. Single molecules of DNA can now be unwound and extended 349 and this process can be monitored by STM. 35~ The question why Nature would have selected polymers of furanose-based pentose nucleotides rather than pyranose-based pentose or hexose ones as genetic material has been investigated in some detail 351 and linked, among other aspects, to the chirality of oligonucleotides. 352 Considering that the first self-replicating systems probably consisted of RNA molecules (the "RNA world" hypothesis), 353 the question arises (related to that raised in the Introduction with respect to biomolecules generally) how the chirality of these systems was decided. The amplification of homochiral RNA molecules due to autocatalytic cycles in which a homochiral polynucleotide template selectively catalyzes the oligomerization of monomers of the same chirality is unlikely due to the problem of "enantiomeric cross-inhibition, ''354 i.e. the inhibition of such an oligomerization when both enantiomers of the oligomers are present. It has therefore even been proposed that chiral RNA as the first self-replicating molecule may have had an achiral precursor. 355 The peptide-nucleic acids (PNA, Scheme 14C), which will be discussed in more detail below, have also been considered as such precursors, but although they could be produced by oligo-ribonucleotide-catalyzed oligomerization, 356 this polymerization suffers from enantiomeric cross-inhibition just like that of RNA. 357

Chiral Self-Assembled Structures

133

The chiral RNA isomer pyranosyl-RNA (p-RNA, Scheme 14D), which has its ribose in the pyranosyl rather than in the furanosyl form, and phosphodiester linkages between the 2' and 4' positions, instead of 3' and 5', forms stronger and more selective hydrogen bonds than RNA. 358 Moreover, it has recently been shown 359 that p-RNA oligomerization on a p-RNA template is chiroselective for the formation of oligomers of the same handedness, which makes p-RNA a serious candidate to be considered as a possible precursor of RNA as the first self-replicating molecule.

5.2. Nucleotides Nucleotides can also have other connectivities through hydrogen bonds than the Watson-Crick base pairs considered so far (Scheme 15), although they are of limited physiological importance. Guanine-containing nucleotides like guanosine3'-phosphate can form planar tetramers, so-called G quartets, which are linked by hydrogen bonds in the so-called Hoogsteen mode (Scheme 16, top). 36~This results in aggregates which account for the long-known 361 ability of guanylic acid to form gels. The planar tetramers are stacked and the layers are rotated with respect to each other which results in a helical structure. 362 Guanosine-5-phosphate is able to form a continuous helix. Polyguanylic acid (Scheme 16, bottom) forms a right-handed quadruple-stranded helix which is held together by hydrogen bonds as in the G-quartet. 363 Guanylic acid oligomers like d(GpG), d(GpGpG), d(GpGpGpG), and d(GpGpGpGpGpG) (Scheme 16, bottom) display lyotropic liquid crystalline phases, with a cholesteric phase at low concentration and an ordered hexagonal phase at higher concentration. 364 In the former phase, the columns of molecules are relatively free to move, while in the latter the expression of the chirality of the monomer in that of the aggregate appears to be limited by space restrictions. The relation between the handedness of the helix and that of the cholesteric phase depends critically on the ratio between pitch and diameter (Scheme 17); righthanded columns are found to lead to right-handed cholesterics for all oligonucleotides mentioned except for d(GpG). Synthetic peptide-nucleic acids (PNA, Scheme 14C) consisting of poly-N-(2aminoethyl)glycine which is derivatized with nucleotides, are an interesting hybrid class of compounds, 365 as they are found to form Watson-Crick base pairs with a complimentary peptide-nucleic acid, 366 RNA, 367 and DNA. 368 The polypeptide backbone is achiral but chirality can be induced in the molecule or its assemblies in various ways. Tagging of the peptide-nucleic acid duplex with either L- or D-lysine leads to enantiomorphic structures with opposite CD spectra, 366 analogous to the so-called "sergeants-and-soldiers effect" for polyisocyanates. 1~ When paired with RNA, the peptide-nucleic acid assumes the A-structure typical of RNA (see above). 367

134

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

H I

A)

H-N

R I

/'-" Ni

.0

H :

H

H : =

.

--" :" =

H

'

- o '==~ (/NI~N.~ ....... o N -

N-H

..........

I

R

H

0 .N , ~

B)

1

,o}

C~ r o

,,R

NH

o N

a, b, r d,

NH

NH

n= n= n= n=

1, 2, 3, 5,

d(GpG) d(GpGpG) d(GpGpGpG) d(GpGpGpGpG)

n

OH Scheme 16. Guanylic acids.

5.3.

Nucleotide Analogues

Aqueous dispersions of dimyristoylphospholipids with nucleoside (cytosine 118, uracil 129, adenosine 130, and guanosine 131) head groups have been prepared enzymatically from the corresponding phosphatidyl choline and nucleoside. 369 A dispersion of the dipalmitoyl 5'-phosphatidylcytidine derivative (128, n - 14, X OH) which had been sonicated in the presence of 0.1 M KCI at 50 o for 45 min initially showed vesicles (diameter 15-30 nm). 37~Upon aging overnight at 25 ~ these were transformed into circular helical strands (diameter 50-150 nm) along with larger vesicles (diameter 50-100 nm). The strands were only formed at pH 8.0, not in acidic solution, which points to the requirement of completely ionized phosphate groups. Low (0.01-0.05 M) KC1 concentrations produced right-handed

Chiral Self-Assembled Structures d(pG)

Left-handedness

135 d(GpG)

d(GpGpG)

~GpGpGpGpGpG)

t,q..__Right-handolmtu Left-handedness

Scheme 17. A representation of the chiral columnar aggregates formed by d(pG), d(GpG), d(GpGpG), and d(GpGpGpGpGpG). The filled circles represent the sugar units (for clarity, only one sugar per tetramer is shown). The screw thread is obtained by joining the filled circles: the screw pitches are in the order d(GpG) > d(GpGpG) > d(GpGpGpGpG), indicating parallel unwinding of the helical structure. Reproduced from ref. 362 (Gottarelli et al., Comprehensive5upramolecularChemistry 1996, Vol. 9, p. 483) with permission from Elsevier Science. linear helical strands, with grooves of 10 nm and a helical pitch of 24 nm, whereas higher (0.5-1.0 M) KCI concentrations favored the formation of vesicles. In a study where the length of the hydophobic tails was varied, 371 it was found that only the didodecanoyl, ditetradecanoyl, and dihexadecanoyl 5'-phosphatidylcytidine derivatives (118, X - OH, n = 10, 12, and 14, respectively) produce superstructures (Chart 18). The ditetradecanoyl derivative (118, X = OH, n - 12) showed network structures with junctions where two single strands (diameter 6 nm, pitch 10 nm) combined to give double strands (diameter 9 nm, pitch 19 nm). The deoxy cytidine analogue dipalmitoyl-5'-phosphatidyldeoxycytidine (118, X - H, n - 14) 370 produced circular helical strands at 0.5 M KCI, linear helical strands at 0.1-0.2 M KC1 and vesicles above 1 M KC1. The grooves of the linear strands were 6.5 nm, and the helical pitch 5.2 nm. Image processing of the electron micrographs of the ditetradecanoyl derivative (118, X - H, n = 12)372 revealed that two types of helical strands were present after 1 day (Figures 38 and 39). One type is a thin helical strand, which consists of a duplex (diameter 11 nm, helical pitch 24 nm). It forms a right-handed superhelical structure with a helical pitch of 110 nm. The other is a thick helical strand, which is a double duplex with a diameter similar to that of the single duplex. This one also forms a right-handed superhelical structure, with a helical pitch of 95-110 A. All dispersions containing superhelical structures gave gels after 1 week, 372 and image processing indicated that the dimensions of the composing single strands had changed slightly (diameter 5.5 nm, helical pitch 15 nm). 369 Stacking of the nucleotide bases, hydrogen bonding, and hydrophobic interactions between the alkyl tails were all considered to be necessary for the

136

MARTINUSC. FEITERSand ROELANDJ. M. NOLTE CH3(CH2)~O~ .OH2 CH3(CH2)~O _=-_~H 0 H2C~O ',&O "O"r~c)~

,NH2 ~ " MI4 I[~. Z ' O

118

L~o~ I

I

HO X

CH3(CH2)~O" .OH2 CH3(CH2)a,~O'"~" H

O H2C~O ~R"~O

0 ,fi~NH

o ~o.~ L..~ ~ L-o-~ I

119

I

HO X

CH3(CH2)~O-cH 2

CH3(CH2)~/O __=."~ H H2C

O

CH3(CH2) O,r,u ~T" T ''2 CH3(CH2)~..O --__'~ H O H2C~_

I

120

I

HO X

Ro

- ~

O.,.~

W

N~<,,,~~ NH2

HO X

Chart 18.

O 121

Chiral Self-Assembled Structures

13 7

Figure 38. Electron micrograph of superhelical strands formed from 118 (X = H, n =

12); bar, 1000 ,/k. Strand 1, thin helical strand (duplex, diameter 110 ,/k, pitch 240 ,/k; right-handed superstructure, pitch 1100/~); strand 2-4, thick helical strands (double duplex, diameter 110/~, pitch 240/~; right-handed superhelix, pitch 950-1100/~). Reproduced from ref. 372 (Yanagawa et al., J. Am. Chem. Soc. 1989, 111, 4567) with permission of the American Chemical Society.

formation of the superstructure. 37~ A molecular model was put forward 372 where the phospholipid nucleoside conjugates are first proposed to assemble into bilayer vesicles by interaction of the hydrophobic alkyl chains, and then rearrange to bilayer cylinders by stacking of the nucleic acid base moieties. This stacking was confirmed by CD spectroscopy. The observed single strands are then proposed to consist of a helical bilayer cylinder structure with the molecules in a bent conformation. The diameter of the double strand (110/~) corresponds with double bilayers of the molecules in a bent (26/~) rather than an extended (36/~) conformation, but, although not considered by the authors, interdigitation of alkyl chains in the bilayers cannot be ruled out. The scope of the formation of helical aggregates from conjugates of phospholipids with other nucleosides was also investigated. 369 Neutral and alkaline solutions of dimiristoyl-5'-phosphatidyladenosine (120, X = OH, n = 12) produced multihelical strands consisting of several helical strands (diameter 5 nm, pitch 10 nm). In acidic solution, the adenosine conjugate was found to give cigar-like scrolls, consisting of many parallel double-helical strands. Dimiristoyl-5'-phosphatidylu-

138

MARTINUS C. FEITERSand ROELAND J. M. NOLTE

Figure 39. Formation process of superhelical strands from 118 (X = H, n = 12). The electron micrographs were taken after aging for 5 h (a), 10 h (b), 15 h (c and d), 1 day (e), and 1 week (f), respectively. All scale bars represent 2000 t~. Arrows in (d) represent positions where the strands are twisting. Inset in (f), Fourier-transformed image (bar, 100 t~) of the segment boxed by a rectangle. Reproduced from ref. 372 (Yanagawa et al., J. Am. Chem. Soc. 1989, 111, 4567) with permission ofthe American Chemical Society.

HN'~O

123

~o

o

O

0

~

~ ~

122

HN~,

~N

RO

H

O II

O

H

O' R

I'~N' H O

O.~.....NH

_..=

"'O~"~H"~~.......'LO II ....N- ~ O.....HN~ . . ~ O ~

~o

RO

O

~o~~

O

O

OR

O.~./ ~N..

I~ "'HN "~,

RO

O

H

'

o~'NH'....O/.~NI..~v..O~o~..I~N..~O.....HN.'~O --

Chart 19.

H

O

OR

%NH

....

.'o ..'H.~O~. ~o

140

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

ridine (119, X = OH, n = 12) gave crystalline platelet structures (100 nm wide) over a wide pH range. The failure of the uridine derivative to produce helical structures was ascribed to the smaller tendency, know also from studies of homopolynucleotides, 373 of the uracil moieties to stack as compared to the other bases. 369A new hybrid helical strand (diameter 30 nm, pitch 200 nm) was found, however, upon mixing dimiristoyl-5'-phosphatidyluridine (119, X = OH, n = 12) at pH 8.0 with an equimolar amount of the adenine analogue 120. Other mixtures of nucleoside phospholipid conjugates did not give helical structures. It is not clear if the complimentarity, in the Watson-Crick sense, of the adenine and uridine is essential for the induction of the helicity in the uridine aggregate by the adenine analogues, as no molecular model for this new type of helix was put forward. Nevertheless, it was proposed that the phospholipid nucleoside conjugates may serve as models for prebiotic templates 374and prebiotic assembly of nucleotides, from which some kind of polymerization may be envisaged.

5.4. Other Complimentary Hydrogen BondingSystems Systems of amphiphilic nucleotide analogues, viz. tartaric acid derivatives of uracil (U) (122) and of a synthetic 2,6-bisacetylamino-pyridine (P) (123) which has hydrogen bonding complimentarity to uracil have been studied (Chart 19). 375 The pure tartaric acid derivatives were solids, but if appropriate pairs of molecules were mixed in chloroform, thermotropic mesophases were obtained after evaporation of the solvent. Thin layers of these mixtures could be dried on electron microscope grids and characterized. The most extensively studied material, the mixture of the bis-U and bis-P derivatives based on L-tartaric acid gives right-handed helices upon drying from concentrations > 50 ktg/mL, and the mixture of D-tartaric acid derivatives gives left-handed helices. By varying the concentration of the sample, various stages in the assembly could be distinguished. Nucleation yields small particles, which expand linearly to form filaments. These associate laterally to give superhelical strings and thick fibers. The diameters of the filaments as derived from electron microscopy (45/~) are of the same magnitude as those of the columnar structures (37-38 ,/k) inferred from X-ray diffraction. 376 In mixtures of uracil and pyridine derivatives of different tartaric acid enantiomers, the uracil derivative determined the handedness of the resulting helical structures (Figure 40). The self-assembly of other systems that form arrays of complimentary hydrogen bonds, e.g. melamine and cyanuric acid derivatives, has been investigated. 377 The formation of multiple hydrogen bonds is even possible between melamine (124a) in aqueous solution with monolayers of the barbituric acid analogue 125a at the air-water interface. 378 The self-assembly of the alkylated melamine 124b and the alkylaryl barbituric acid analogue 125b in chloroform solution initially gave chiral nanofilaments (diameter, 80 nm; length, several tens of ~tm) which were transformed into chiral supercoils (diameter, 300 nm; length, 10 ktm) after prolonged staining when equilibrium had been reached. 379 Since the starting materials were

Chiral Self-Assembled Structures

141

Figure 40. Air dried chloroform solutions of racemic mixtures of LP2, LU2, DP2, and DU2(122 and 123). Notice the presence of superhelical structures showing L-L and D-D type of handedness and of extended regions of mixed handedness (arrows). Bars, 0.2 I.tm. Reproduced from ref. 375 (Gulik-Krzywicki et al., Proc. Natl. Acad. Sci. USA 1993, 90, 163) with permission of the Academy of Sciences of the USA.

both achiral, both left-handed and right-handed supercoils were observed. Complexes of either the substituted melamines 124c or 126 with either of the substituted imides, "diimide" 127 or "monoimide" 128 could be sonicated in methylcyclohexane to give transparent dispersions. 38~ TEM showed the presence of strands (diameter 100/~) for the 1:1 complex of 124e and 127, and SEM and AFM revealed multiple strands twisted into helical bundles for the 1:1 complex of 126 and 127. The assembly of bis-melamines with a calixarene spacer 129 and substituted cyanuric acid analogues 130 has been found to result in the formation of 3:6 complexes with complete asymmetric induction. TM It had been observed previously that the 3:6 complex of 129a and the substituted barbituric acid analogue 130a displayed supramolecular chirality both in solution and in the solid state because of the antiparallel orientation of the rosette motifs, but due to the absence of chirality of the constituting monomers only racemates were found. 382 According to NMR and CD, the melamines bearing enantiopure ct-phenylethyl ligands (R,R)-129b and (S,S)-129b gave the (M)- and (P) enantiomer, respectively, of the complex with 130a, whereas the combination of (R,S)-substituted analogue 129d with 130a just gave poorly defined oligomers. TM The cyanuric acid derivatives 130b-f all gave complete chiral induction in their complexes with achira1129c. Chiral selection has also been observed in the complexation of the substituted chiral melamine 131 with the achiral barbiturate 132, resulting in the formation of homochiral crinkled tapes. 383

142

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

H

O /JL.

125a

R--N'

N H ,,, \,,' I~\ //---I~ ~--N H R_N/

0

124a, R=H . 124b melamine-1 I~ ,,'p u ~ n u .....

HN

...~ . . . . HN NH / / O"~~ I O

"v12r'25v~3'~6"

124b, R= n-C12H25

H

.~

U"

~

~

IL,.,,.,~_..J~O.n.~

~

N,H

NH ~ ..... "T; "U 1Z~D II ' ~

N, n"ClsH33

>=,

2H25

. o

~~~12S,

N elamine-2

129

FI 2

~ HN---~\

H-

-H 0

~

O 127, diimide

HN~~,_NH2 /) Unl-'r nPrO~'

,x~---NH

H-N O

NH ~ R1 a, R1 = R2 = ~Bu b,R1=R2=-(N-a-CN(CloHT)CH3

128, monoimide

/

C, R1 = R2 = benzyl d, R1 = (R)~-CH(CsH5)CH3 R2 = (S)-o~-CH(C6H5)CH3

.NH2 N)~-N

O

H. NJ~, N.,,/~ N,, H

HN"JJ"NH

"R

"R

'

132 131 (S,S)

Chart 20.

130

O HN"~NH

o -x o a, DEB, X = C(CH2CH3)2 b, (R)-MePhCYA, X = N-(R)-(x-CH(C6Hs)CH3 r (S)-MePhCYA, X = N-(S)-o~-CH(CsHs)CH3 d, (S)-PheCYA, X = N-(S)-CHPhC(O)OCH3 e, (S)-ValCYA, X = N-( S)-CH(kPr)C(O)OCH3 f, (S)-LeuCYA, X = N-(S)-CH(kBu)C(O)OCH3

Chiral Self-Assembled Structures

143

6. CONCLUSIONS Chiral structures are ubiquitous in nature, and many examples of biomimetic self-assembly to chiral superstructures have been reported in the literature. In spite of the large variety of structures, it is still difficult to make any general statements with predictive value with regard to the possible expression of the chirality in a supramolecular system. It is virtually impossible to predict whether a racemic system will show lateral phase separation of its enantiomers, because relatively few examples of this phenomenon have been reported. Chirality of the molecules is a requirement for the chirality of the superstructure but not for achiral superstructures like tubules; chiral polymers or helicates may be obtained even if the monomers or building blocks are achiral, but the incorporation of even a small amount of chiral analogue can help to induce predominant or even exclusive formation of one of the possible enantiomeric structures. For the design of molecules that self-assemble in water, the general guidelines concerning critical aggregate concentration and shape of the molecules for the aggregation of amphiphiles apply. In order to express the chirality of the molecule in the self-assembled structure it is necessary to have in addition some organizing principle like amide bonds, diyne moieties, or metal-ligand interactions, which must be present in the molecule in such a position relative to the chiral group as to allow efficient communication of the chirality of one molecule with that of the next one in the assembly. Self-assembly in organic solvents mainly relies on hydrogen bonding. In studies of the self-assembly of chiral molecules, it is important to monitor the formation of the various aggregate structures with time from the moment that the amphiphile is dispersed, as crystallization processes are slow due to phenomena like the chiral bilayer effect. Conditions like temperature and the presence of cosolvents also need to be varied, as systems with varying degrees of solvation and packing of the molecules may be stable above or below certain transition temperatures. In view of the relative instability of most of the systems described here, the possibilities to stabilize self-assembled superstructures by incorporating appropriately oriented polymerizable groups, or to prepare a superstructure in a polymerizable solvent and then "imprint" it, continue to be of interest. As our understanding of the rationale of self-assembly of chiral superstructures and the techniques for studying them continue to be developed, this area of research holds promises of new exciting results in the new millenium.

7. ACKNOWLEDGMENTS The authors thank Mr. L. Thijs, Dr. G. W. Ariaans, and Dr. A. E. Rowan for literature searches, Ms. V. S. I. Sprakel and Messrs. D. H. W. Hubert, H. Engelkamp, and R. Rasing for their help in preparing part of this review, Drs. P. J. J. A. Buynsters, R. J. H. Hafkamp, and N. A. J. M. Sommerdijk for carrying out some of the studies included from the authors' laboratory in this review, and Mrs. A. M. Roelofsen and Mr. H. P. M. Geurts for their dedicated technical

144

MARTINUS C. FEITERS and ROELAND I. M. NOLTE

assistance with monolayer experiments and electron microscopy, respectively, in this work. The Dutch Organization for the Advancement of Research (NWO) is acknowledged for financing part of the authors' research through its chemical branch SON (now CW).

NOTE ADDED IN PROOF W h i l e this manuscript was in preparation, the first e x a m p l e of twisting of the bilayers of cationic gemini surfactants by chiral counterions like tartrate, glucarate, malate, and gluconate was reported. 384

REFERENCES 1. (a) Kelvin, J. T. J. Oxford Univ. Junior Scientific Club 1894, 18, 25; (b) Kelvin, Ld. Baltimore Lectures 1904, Clay: London, p. 618. 2. (a) Pasteur, L. C. R. Hebd. S~anc. Acad. Sci. Paris 1848, 26, 535; (b) Pasteur, L. Ann. Chem. 1848, 24, 457. 3. (a) Van't Hoff, J. H. Arch. Ne~rland. Sci. Exactes Naturelles 1874, 9, 445; (b) Le Bel, J. A. Bull. Soc. Chim. Fr. 1874, 22, 337. 4. (a) Calm, R. S.; Ingold, C. K.; Prelog, V. Angew. Chem., Int. Ed. Engl. 1966, 5, 385; (b) Prelog, V.; Helmchen, G. Angew. Chem., Int. Ed. Engl. 1982, 21,567. 5. (a) Barton, L. D. Chem. Soc. Rev. 1986, 15, 189; (b) Barron, L. D. Chem. Eur J. 1996, 2, 743. 6. (a) Katzenelson, O.; Zrabodsky Hel-Or, H.; Avnir, D. Chem. Eur. J. 1996, 2, 174; (b) Avnir, D.; Katzenelson, O.; Zrabodsky Hel-Or, H. Chem. Eur. J. 1996, 2, 744. 7. Brewster, J. H. Top. Cu~ Chem. 1974, 47, 29. 8. Prelog, V. J. Mol. Catal. 197516, I, 159. 9. Meumr, K. E; V'6gtle, E Top. Curt Chem. 1985, 127, 1. 10. Buda, A. B.; Auf der Heyde, T.; Mislow, K. Angew. Chem. 1992, 31,989. 11. Jonas, J. New. J. Chem. 1997, 21, 1251. 12. Eliel, E. L.; Wilen, S. H. Stereochemistry of Organic Compounds; Wiley: New York, 1994. 13. Enders, D.; Hoffmann, R. W. Chemie in unserer Zeit 1985, 19, 177. 14. Leitereg, T. J.; Guadagni, D. G.; Harris, J.; Mon, T. R.; Teranashi, R. Nature 1971, 230, 455. 15. Barbeau, A. Can. Med. Assoc. J. 1969, 101,791. 16. Pfeifer, S. Pharmazie 1968, 23, 225. 17. Bonner, W. A. Orig. Life Evol. Biosphere 1991, 21, 59. 18. Frank, E C. Biochim. Biophys. Acta 1953, 11,459. 19. Mason, S. E Nature 1984, 311, 19. 20. (a) Bolm, C.; Bienewald, E; Seller, A. Angew. Cheat Int. Ed. Engl. 1996, 35, 1657; (b) Shibata, T.; Mouska, H.; Hayase, T.; Choji, K.; Soai, K.J. Am. Chem. Soc. 1996, 118, 471; (c) Alberts, A. H.; Wynberg, H. J. Am. Chem. Soc. 1989, 111, 7265; (d) Sheng Jian, L.; Yaozhong, J.; Aigiaio, M.; Suishu, Y. J. Chem. Soc., Perkin Trans 1 1993, 885; (e) Wynberg, H. J. Macromol. Sci. Chem. 1989, A26, 1033. 21. Lee, D. H.; Severin, K.; Yokobayashi, Y.; Ghadiri, M. R. Nature 1997, 390, 591, and correction, Nature 1998, 394, 101. 22. Wu, C. S.; Ambler, E.; Hayward, R. W.; Hoppes, D. D.; Hudson, R. P. Phys. Rev. 1957,105, 1413. 23. De Rossi, U.; D~ihne,S.; Meskers, S. C. J.; Dekkers, H. P. J. M. Angew. Chem., Int. Ed. Engl. 1996, 35, 760.

24. Elgavi, A.; Green, B. S.; Schmidt, G. M. J. J. Am. Chem. Soc. 1973, 95, 2058. 25. (a) Addadi, L.; van Mil, J.; Lahav, M. J. Am. Chem. Soc. 1982, 104, 3422; (b) Addadi, L.; Berkovitch-Yellin, Z.; Weissbuch, I.; van Mil, J.; Shimon, L. J. W.; Lahav, M.; Leisorowitz, L.

Chiral 5elf-Assembled Structures

26. 27. 28.

29. 30. 31. 32. 33.

145

Angew. Chem., Int. Ed. Engl. 1985, 24, 466; (c) Van Mil, J.; Addadi, L.; Gati, E.; Lahav, M. J. Ant Chem. Soc. 1982, 104, 3429. Suzuki, T.; Fukushima, T.; Yamashita, Y.; Miyashi, T. J. Am. Chem. Soc. 1994, 116, 2793. Sakamoto, M. Chem. Eur. J. 1997, 3, 684. (a) Koshima, H.; Ding, K.; Chisaka, Y.; Matsuura, T. J. Ant Chem. Soc. 1996, 118, 12059; (b) Koshima, H.; Hayashi, E.; Matsuura, T.; Tanaka, K.; Toda, E; Kato, M.; Kiguchi, M. Tetrahedron Lett. 1997, 38, 5009. Smtrez, M.; Branda, N.; Lehn, J.-M.; Decian, A.; Fischer, J. Helv. Chint Acta 1998, 81, 1. Viswanathan, R.; Zasadzinski, J. A.; Schwartz, D. K. Nature 1994, 368, 440. Lehn, J.-M. Angew. Chent, Int. Ed. Engl. 1990, 29, 1304. Rowan, A. E.; Nolte, R. J. M. Angew. Chem., Int. Ed. Engl. 1998, 37, 63. Fuhrhop, J.-H.; K6ning, J. Monographs in Supramolecular Chemistry. Membranes and Molecular Assemblies: The SynkineticApproach; Stoddart, J. E, Ed.; Royal Society of Chemistry, Cambridge

(OK), 1994. 34. Fuhrhop, J.-H. In Comprehensive Supramolecular Chemistry; Atwood, J. L.; Davies, J. E. D.; Macnicol, D. D.; V6gtle, E, Eds.; Lehn, J.-M., Series Ed.; Pergamon: Elsevier Science, Vol. 9, 1996, p. 407. 35. Lawrence, D. S.; Jiang, T.; Levett, M. Chent Rev. 1995, 95, 2229. 36. Amabilino, D. B.; Stoddart, J. E Chent Rev. 1995, 95, 2725. 37. Philp, D.; Stoddart, J. E Angew. Chent, Int. Ed. Engl. 1996, 35, 1154. 38. (a) Kato, T.; Kihara, H.; Uryu, T.; Ujiie, S.; Iimura, K.; Fr6chet, J. M. J.; Kumar, U. Ferroelectrics 1993, 148, 161; (b) Tian, Y.; Xu, X.; Zhao, Y.; Tang, X.; Li, T.; Sun, J.; Li, C.; Pan, A. Thin Solid Films 1996, 284-285, 603; (c) Kihara, H.; Kato, T.; Uryu, T.; Fr6chet, J. M. J. Liquid Crystals 1998, 24, 413. 39. Winkler, E J.; Medina, R.; Winkler, J.; Krause, H. J. Mass Spectrom. 1997, 32, 1072. 40. Cadenhead, D. A. Rec. Prog. Surf. Sci. 1970, 3, 169. 41. (a) Stiillberg-Stenhagen, S.; Stenhagen, E. Arkiv Kemi Mineral. Geol. 1944, 18A, 1; (b) St~llbergStenhagen, S.; Stenhagen, E. Nature 1945, 25, 239. 42. Lundquist, M. Arkiv Kemi 1960, 17, 183. 43. (a) Zeelen, E J. Ptt D. Thesis 1956; (b) Zeelen, E J.; Havinga, E. Recl. Trav. Chint Pays-Bas 1958, 77, 267. 44. Lundquist, M. Arkiv Kemi 1962, 21,395. 45. Kellner, B. M. J.; Cadenhead, D. A. J. Colloid Interface Sci. 1978, 63, 452. 46. Kellner, B. M. J.; Cadenhead, D. A. Chent Phys. Ldpids 1979, 23, 41. 47. Stewart, M. V.; Arnett, E. M. Topics Stereochent 1981, 13, 195. 48. Andelman, D. J. Ant Chent Soc. 1989, 111, 6536. 49. Amett, E. M.; Gold, J. M. J. Ant Chent Soc. 1982, 104, 636. 50. (a) Amett, E. M.; Chao, J.; Kinzig, B.; Stewart, M.; Thompson, O. J. Ant Chent Soc. 1978, I00, 5575; (b) Amett, E. M.; Thomson, O. J. Ant Chent Soc. 1981, 103, 968; (c) Arnett, E. M.; Chao, J.; Kinzig, B. J.; Stewart, M. V.; Thompson, O.; Verbiar, R. J. J. Ant Chent Soc. 1982, 104, 389; (d) Amett, E. M.; Harvey, N. G.; Rose, P. L. Acc. Chent Res. 1989,22, 131. 51. Tredgold, R. H. Order in Thin Organic Films; Cambridge University Press: 1994. 52. Kjaer, K.; Als-Nielsen, J.; Helm, C. A.; Laxhuber, L. A.; M/Shwald, H. Phys. Rev. Lett. 1987, 58, 2224. 53. Dutta, P.; Peng, J. B.; Lin, B.; Ketterson, J. B." Prakash, M.; Georgopoulos, P.; Ehrlich, S. Phys. Rev. Lett. 1987, 58, 2228. 54. von Tscharner, V.; McConnell, H. M. Biophys. J. 1981, 36, 409. 55. (a) l./Ssche,M.; Sackman, E.; MiShwald,H. Ber. Bunsenges. Phys. Chem. 1983, 87, 848; (b) IdSsche, M.; Mtihwald, H. Rev. Sci. Instrunt 1984, 55, 1968; (c) l.~sche, M.; Rabe, J.; Fischer, A.; Reicha, B. U.; Knoll, W.; M6hwald, H. Thin Solid Films 1984, 117, 269. 56. H6non, S.; Meunier, J. Rev. Sci. Instrunt 1991, 62, 936.

146

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

57. H/Snig,D.; Mtibius, D. J. Phys. Chem. 1991, 95, 4590. 58. Weissbuch, I.; Berfeld, M.; Bouwman, W.; Kjaer, K.; Als-Nielsen, J.; Lahav, M.; Leiserowitz, L. J. Am. Chem. Soc. 1997, 119, 933. 59. Van Esch, J. H.; Nolte, R. J. M.; Ringsadorf, H.; Woldburg, G. Langmuir 1994, 10, 1955. 60. Wang, M.; Wildburg, G.; van Esch, J. H.; Bermema, P.; Nolte, R. J. M.; Ringsdorf, H. Phys. Rev. Lett. 1993, 71, 4003. 61. (a) Qiu, X.; Ruiz-Garcia, J.; Stine, K. J.; Knobler, C. M.; Selinger, J. V. Phys. Rev. Lett. 1991, 67, 703; (b) Qiu, X.; Ruiz-Garcia, J.; Knobler, C. M. Mat. Res. Soc. Syrup. Proc. 1992, 237, 263. 62. Eckhardt, C. J.; Peachey, N. M.; Swanson, D. R.; Takacs, J. M.; Khan, M. A.; Gong, X.; Kim, J.-H.; Wang, J.; Uphaus, R. A. Nature 1993, 362, 614. 63. Stevens, E; Dyer, D. J.; Walba, D. M. Angew. Chem., Int. Ed. EngL 1996, 35, 900. 64. Qian, P.; Nanjo, H.; Yokoyama, T.; Suzuki, T. M. Cheat Lett. 1998, 1133. 65. Chi, L. E; Gleiche, M.; Fuchs, H. Langmuir 1998, 14, 875. 66. Hotten, B. W.; Birdsall, D. H. J. Colloid Sci. 1952, 7, 284. 67. Tachibana, T.; Kambara, H. Kolloid-Z 1967, 219, 40. 68. Nakashima, N.; Asakuma, S.; Kunitake, T. Cheat Lett. 1984, 1709. 69. Yamada, K.; Ihara, H.; Ide, T.; Fukumoto, T.; Hirayama, C. Cheat Lett. 1984, 1713. 70. Yager, P.; Schoen, P. E. MoL Cryst. Liq. Cryst. 1984, 106, 371. 71. Pfannemtiller, B.; Welte, W. Cheat Phys. Lipids 1985, 37, 227. 72. Belnap, D. M.; Olson, N. H.; Baker, T. S. J. Struct. Biol. 1997, 120, 44. 73. De Rosier, D. J.; Klug, A. Nature 1968, 217, 130. 74. Tachibana, T.; Kambara, H. J. Ant Cheat Soc. 1965, 87, 3015. 75. Tachibana, T.; Kambara, H. J. Colloid Sci. 1968, 28, 173. 76. Tachibana, T.; Kambara, H. BulL Cheat Soc. Jpn. 1969, 242, 3422. 77. Uzu, Y.; Sugiura, T. J. CoiL Interfac. Sci. 1975, 51,346. 78. (a) Tachibana, T.; Hori, K. J. CoiL Interfac. Sci. 1977, 61,398; (b) Tachibana, T.; Yoshizumi, T.; Hori, K. Bull. Cheat Soc. Jpn. 1979, 52, 34. 79. Tachibana, T.; Kitazawa, S.; Takeno, H. BulL Cheat Soc. Jpn. 1970, 43, 2418. 80. Tachibana, T.; Moil, M.; Hori, K. Bull. Cheat Soc. Jpn. 1980, 53, 1714. 81. Tachibana, T.; Moil, M.; Hori, K. Bull. Cheat Soc. Jpn. 1981, 54, 73. 82. (a) Stryer, L. Biochemistry; Freeman: New York, 1988; (b) van Holde, K. E.; Johnson, W. C,; Ho, P. S. Principles of Physical Biochemistry; Prentice Hall: NJ, 1998. 83. (a) Lehn, J.-M.; Rigault, A.; Siegel, J.; Harrowfield, J.; Chevrier, B.; Moras, D. Proc. NatL Acad. USA 1987, 84, 2565; (b) Lehn, J.-M.; Marquis-Rigault, A. Angew. Cheat, Int. Ed. Engl. 1988, 30, 1095. 84. (a) Constable, E. C. Angew. Cheat, Int. Ed. Engl. 1991, 30, 1450; (b) Constable, E. C. Tetrahedron 1992, 48,10013; (c) Constable, E. C. Nature 1994, 367, 415; (d) Constable, E. C. In Comprehensive Supramolecular Chemistry; Atwood, J. L.; Davies, J. E. D.; Macnicol, D. D.; V6gtle, E, Eds.; Pergamon, Elsevier Science, Vol. 9, 1996, p. 213. 85. (a) Williams, A. Cheat Eur. J. 1997, 3, 15; (b)von Zelewsky, A.; Belser, P. Chimia 1998, 52, 620. 86. Kriimer,R.; Lehn, J.-M.; De Cian, A.; Hischer, J. Angew. Cheat, Int. Ed. EngL 1993, 32, 703. 87. Charbonniere, L. J.; Bemardelli, G.; Piguet, C.; Sargeson, A. M.; Williams, A. E J. Cheat Soc., Cheat Commun. 1994, 1419. 88. (a) Zarges, W.; Hall, J.; Lehn, J.-M.; Bolm, C. Helv. Chiat Acta 1991, 74, 1843; (b) Enemark, E. J.; Stack, T. D. P. Angew. Cheat, Int. Ed. EngL 1995, 34, 996; (c) Enemark, E. J.; Stack, T. D. P. Angew. Cheat, Int. Ed. EngL 1998, 37, 932; (d) Woods, C. R.; Benaglia, M.; Cozzi, E; Siegel, J. S. Angew. Cheat, Int. Ed. Eng. 1996, 30, 1830; (e) Mtimer, H.; von Zelewsky, A.; Hopfgartner, G. Inorg. Chiat Acta 1998,271,36; (f) Baret, P.; Gaude, G.; Gellon, G.; Pierre, J.-L. New J. Cheat 1997, 21, 1255. 89. Mamula, O.; von Zelewsky, A.; Bemardinelli, G. Angew. Cheat, Int. Ed. EngL 1998, 37, 289.

Chiral 5elf-Assembled Structures

147

90. (a) Hanan, G. S.; Lehn, J.-M.; Krytsakas, N." Fisher, J. J. Cheat Soc., Cheat Commun. 1995, 765; (b) Bassani, D. M." Lehn, J.-M.; Baum, G." Fenske, D. Angew. Chem., Int. Ed. Engl. 1997, 36, 1845" (c) Owens, L.; Thilgen, C.; Diederich, E" Knobler, C. B. Heir. Chiat Acta 1993, 76, 2757. 91. Laarhoven, W. H.; Prinsen, W. J. C. Top. Curr. Chem- 1984, 125, 63. 92. (a) Koert, U.; Harding, M. M.; Lehn, J.-M. Nature 1990, 346, 339; (b) Judice, J. K.; Cram, D. J. Am. Chem. Soc. 1991, 113, 2790; (c) Potts, K. T.; Keshavarz-K, M.; Tham, E S.; Abruna, H. T." Arana, C. lnorg. Chem. 1993, 32, 4422; (d) Potts, K. T.; Keshavarz-K, M.; Tham, E S.; Abruna, H. T.; Arana, C. Inorg. Chem- 1993, 32, 4450; (e) Constable, E. C.; Heirtzler, E R.; Neuburger, M." Zehnder, M. Chem- Commun. 1996, 933; (f) Constable, E. C." Heirtzler, E; Neuburger, M." Zehnder, M. J. Am. Cheat Soc. 1997, I 19, 56060; (g) Dietrich-Buchecker, C. O.; Sauvage, J.-E" Armaroli, N.; Ceroni, E" Balzani, V. Angew. Chem-, Int. Ed. Engl. 1996, 35, 1119; (h) Funeriu, D. E; Lehn, J.-M.; Baum, G.; Fenske, D. Cheat Eur. J. 1997, 3, 99; (i) Goodman, M. S.; Weiss, J." Hamilton, A. D. Tetrahedron Lett. 1994, 48, 8943" (j) (including redox-cycling) El-ghayaouy, A." Harriman, A.; de Cian, A.; Fischer, J.; Ziessel, R. J. Am. Chem. Soc. 1998, 120, 9973. 93. (a) Kawamoto, T.; Hammes, B. S." Haggerty, B.; Yap, G. E A.; Rheingold, A. L.; Borovik, A. S. J. Am. Chem. Soc. 1996, 118, 285; (b) Ohata, N.; Masuda, H." Yamauchi, O. Angew. Chem., Int. Ed. Engl. 1996, 35, 531. 94. Ni: (a) KriLrner,R." Lehn, J.-M.; Marquis-Rigault, A. Proc. Natl. Acad. Sci. USA 1993, 90, 5394; (b) Dai, Y.; Katz, T. J.; Nichols, D. A.Angew. Cheat, Int. Ed. Engl. 1996, 35, 2109; Fe: (c) Smith, V. C. M.; Lehn, J.-M. Cheat Commun. 1996, 2733; (d) Hasenknopf, B.' Lehn, J.-M.; Kneisel, B. O." Baum, G.; Fetiske, D. Angew. Chem., Int. Ed. Engl. 1996, 35, 1838; (e) Hasenknopf, B.; Lehn, J.-M.; Baum, G.; Fenske, D. Proc. Natl. Acad. Sci. USA 1996, 93, 1397; (f) Hasenknopf, B.; Lehn, J.-M.' Boumedienne, N.; Leize, E." van Dorsselaer, A. Angew. Cheat, Int. Ed. Engl. 1998, 37, 3265; (g) Saalfrank, R. W.; Trummer, S.; Krautscheid, H." Schtinemann, V.; Trautwein, A. X." Hien, S.; Stadler, C.; Daub, J. Angew. Chem., Int. Ed. Engl. 1996, 35, 2206; (h) Zurita, D.' Baret, E; Pierre, J.-L. New. J. Chem. 1994, 18, 1143; Zn: (i) Canary, J. W." Allen, C. S.; Castagnetto, J. M.; Wang, Y. J. Am. Cheat Soc. 1995, 117, 8484; (j) Barf, T.; Jansen, J. E G. A.; van Bolhuis, E; Spek, A. L.; Feringa, B. L. Recl. Tray. Chiat Pays-Bas 1993, 112, 376; Rb: (k) Saalfrank, R. W." Harbig, R.; Nachtrab, J.; Bauer, W.; Zeller, K.-E" Stalke, D.; Teichert, M. Chem. Eur J. 1996, 2, 1363; (l) Weber, G.; Saenger, W." V'ogtle, E; Sieger, H. Angew. Chem., Int. Ed. Engl. 1979, 18, 226; K: (m) Weber, G.; Saenger, W. Angew. Cheat, Int. Ed. Engl. 1979, 18, 227; Na: (n) Bell, T. W.; Jousselin, H. Nature 1994, 367, 441. 95. S~tnchez-Quesado, J.; Seel, C.; Prados, E; de Mendoza, J. J. Ant Chem. Soc. 1996, 118, 277. 96. Okamoto, Y.; Nakano, T. Cheat Rev. 1994, 94, 349. 97. Sikorski, E; Cooper, S. J.; Atkins, E. D. T." Jaycox, G. D.; Vogl, O. J. Polym. Sci. Part A Polyat Cheat 1998, 36, 1855. 98. (a) Langeveld-Voss, B. M. W.; Janssen, R. A. J." Christiaans, M. E T.; Meskers, S. C. J.; Dekkers, H. E J. M.; Meijer, E. W. J. Ant Cheat Soc. 1996, 118, 4908; (b) Langeveld-Voss, B. M. W.; Christiaans, M. E T.; Janssen, R. A. J." Meijer, E. W. Macromol. 1998, 31, 6702. 99. (a) Ito, Y.; Ihara, E.; Murakami, M.Angew. Chem., Int. Ed. Engl. 1992, 31, 1509; (b) Ito, Y.; Ihara, E.; Murakami, M.; Sisido, M. Macromolecules 1992, 25, 6810. 100. Hu, Q.-S.; Vitharana, D.; Liu, G.-Y.; Jain, V.; Wagaman, M. W.; Zhang, L.; Lee, T. R.; Pu, L. Macromol. 1996, 29, 1082. 101. (a) Nolte, R. J. M. Cheat Soc. Rev. 1994, 11; (b) Kauranen, M.; Verbiest, T.; Boutton, C.; Teerenstra, M. N.; Clays, K.; Schouten, A. J.; Nolte, R. J. M.; Persoons, A. Science 1995, 270, 966; (c) Ramos, E." Bosch, J.; Serrano, J.-L" Sierra, T.; Veciano, J. J. Am. Cheat Soc. 1996, 118, 4703; (d) Takei, E; K0ichi, Y.; Onitsuka, K." Takahashi, S.Angew. Cheat, Int. Ed. Engl. 1996, 35, 1554; (e) Amabilino, D. B.; Ramos, E.; Serrano, J.-L." Sierra, T.; Veciano, J. J. Ant Cheat Soc. 1998, 120, 9126. 102. Comelissen, J. J. L. M.; Fischer, M.; Sommerdijk, N. A. J. M.; Nolte, R. J. M. Science 1998, 280, 1427.

148 103. 104. 105. 106. 107. 108.

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

Williams, D. J.; Colquhoun, H. M.; O'Mahoney, C. A. J. Chem. Soc., Chem- Commun. 1994, 1643. Huang, W.-S.; Hu, Q.-S.; Zheng, X.-E; Anderso, J.; Pu, L. J. Am. Chem. Soc. 1997, 119, 4313. Nelson, J. C.; Saven, J. G.; Moore, J. S.; Wolynes, E G. Science 1997, 277, 1793. Janssen, H. M.; Meijer, E. W. Macromolecules 1997, 30, 8129. Bur, A. J.; Fetters, L. J. Chem. Rev. 1976, 76, 727. (a) Green, M. M.; Reidy, M. E; Johnson, R. J.; Darling, G.; O'Leary, D. T. J. Am. Chem. Soc. 1989, 111, 6452; (b) Green, M. M.; Garetz, B. A.; Munoz, B.; Chang, H. J. Am- Chem. Soc. 199S, 117, 4181. 109. Green, M. M.; Peterson, N. C.; Sato, T.; Teramoto, A.; Cook, R.; Lifson, S. Science 199S, 268, 1860. 110. (a) MUller, M.; Zentel, R. Macromolecules 1996, 29, 1609; (b) Mayer, S.; Zentel, R. Macromol. Chem. Phys. 1998, 199, 1675. 111. (a) Engelkamp, H.; van Nostrum. C. E; Picken, S. J.; Nolte, R. J. M. Chem- Commun. 1998, 979; (b) Engelkamp, H.; Middelbeek, S.; Nolte, R. J. M. Science 1999, 284, 785. 112. MacGillivray, L. R.; Atwood, J. L. Nature 1997, 389, 469. 113. (a) Rivera, J. M.; Mart/n, T.; Rebek Jr., J. Science 1998, 279, 1021; (b) Castellano, R. K.; Kim, B. H.; Rebek Jr., J. J. Am- Chem. Soc. 1997, 117, 12671; (c) Tokunaga, Y.; Rebek Jr.; J. J. Am- ChemSoc. 1998, 120, 66. 114. de Mendoza, J. Chem. Fur. J. 1998, 4, 1373. 115. Brockerhoff, H. In Bio-Organic Chemistry; van Tamelen, E. E.; Ed.; Academic Press: New York, 1977, p. 1. 116. Bangham, A. D.; Home, R. W. J. Mol. Biol. 1964, 8, 660. 117. Tanford, C. The Hydrophobic Effect, Formation of Micelles and Biological Membranes; Wiley Interscience: New York, 1973. 118. Blokzijl, W.; Engberts, J. B. E N. Angew. Chem-, Int. Ed. Engl. 1993, 32, 1545. 119. Evans, D. E; Wennerstr~Sm, H. The Colloidal Domain, Where Physics, Chemistry, Biology, and Technology Meet. VCH: New York/Weinheim, Cambridge, 1994. 120. Kunitake, T.; Okahata, Y. J. Am- Chem- Soc. 1977, 99, 3860. 121. Kunitake, T.; Okahata, Y.; Shimomura, M.; Yasunami, S.i Takarabe, K. J. Am- Chem. Sot:. 1981, 103, 5401. 122. Kunitake, T. Angew. Chem., Int. Ed. Engl. 1992, 31,709. 123. Kunitake, T. In Comprehensive Supramolecular Chemistry, Atwood, J. L.; Davies, J. E. D.; Macnicol, D. D.; V6gtle, E, Eds.; Pergamon: Elsevier Science, 1996, Vol. 9, p. 351. 124. Fendler, J. H. Membrane Mimetic Chemistry; Wiley-lnterscience: New York, 1982. 125. Ringsdorf, H.; Schlarb, B.; Venzmer, J. Angew. Chem., Int. Ed. Engl. 1988, 27, 113. 126. Kunitake, T.; Okahata, Y.; Yasunami, S. J. Am- Chem. Soc. 1982, 104, 5547. 127. Giulieri, E; Krafft, M.-E; Riess, J. G. Angew. Chem-, Int. Ed. Engl. 1994, 33, 1514. 128. Giulieri, E; Guillod, E; Greiner, J.; Krafft, M.-E; Riess, J. G. Chem- Eur J. 1996, 2, 1335. 129. (a) Ishikawa, Y.; Kuwahara, H.; Kunitake, T. J. Ant Chem- Soc. 1989, I 11, 8530; (b) Ishikawa, Y.; Kuwahara, H.; Kunitake, T. Chem. Lett. 1989, 1737. 130. Ishikawa, Y.; Kuwahara, H.; Kunitake, T. J. Am- Chem- Soc. 1994, 116, 5579. 131. (a) Israelachvili, J. N.; Mitchell, D. J.; Ninham, B. W. J. Chem. Soc., Faraday Trans. 2,1976, 1525; (b) Israelachvili, J. N.; Marcelja, S.; Horn, R. G. Quart. Rev. Biophys. 1980, 13, 121. 132. Van Hest, J. C. M.; Baal's, M. W. E L.; Delnoye, D. A. E; van Genderen, M. H. E; Meijer, E. W. Science 1995, 268, 1592. 133. Kunitake, T.; Okahata, Y. J. Am. Chem. Soc. 1980, 102, 549. 134. Okahata, Y.; Kunitake, T. Ber. Bunsenges. Phys. Chem. 1980, 84, 550. 135. Goto, T.; Kondo, T. Angew. Chem., Int. Ed. Engl. 1991, 30, 17. 136. Kunitake, T.; Yamada, N.; Fukunaga, N. Chem. Lett. 1984, 1089. 137. Sommerdijk, N. A. J. M.; Feiters, M. C.; Nolte, R. J. M.; Zwanenburg, B. Recl. Tray. (?him. Pays-Bas 1994, 113, 194.

Chiral Self-Assembled Structures

149

138. Pathirana, S.; Neely, W. C.; Myers, L. J.; Vodyanoy, V. J. Ant Chem. Soc. 1992, 114, 1404. 139. Furhhop, J.-H.; Schnieder, E; Rosenberg, J.; Boekema, E. J. Ant Chem. Soc. 1987, 109, 3387. 140. Fuhrhop, J.-H.; Bedurke, T.; Hahn, A.; Grund, S.; Gatzmann, J.; Riederer, M. Angew. Chem., Int. Ed. Engl. 1994, 33, 350. 141. Sakurai, I.; Karvabura, T.; Dakurai, A.; Kegami, A.; Setoi, T. Mol. Cryst. Liq. Cryst. 1985, 130, 203. 142. Lin, K.-C.; Weis, R. M.; McConnell, H. M. Nature 1982, 296, 164. 143. Weis, R. M.; McConneU, H. M. Nature 1984, 310, 47. 144. McConnell, H. M.; Tamon, L. K.; Weis, R. M. Proc. Natl. Acad. Sci. USA 1984, 81, 3249. 145. Weis, R. M.; McConnell, H. M. J. Phys. Chem. 1985, 89, 4453. 146. McConnell, H. M.; Keller, D. J. Proc. Natl. Acad. Sci. USA 1987, 84, 4706. 147. McConnell, H. M. Annu. Rev. Phys. Chem. 1991, 42, 171. 148. Heckl, W. M.; Mtihwald, H. Ber. Bunsen-Ges. Phys. Chem. 1986, 90, l 159. 149. Heckl, W. M.; I.~sche, M.; Cadenhead, D. A.; Mtihwald, H. Eur. Biophys. J. 1986, 14, 11. 150. Servuss, R. M. Chem. Phys. Lipids 1988, 46, 37. 151. (a) Thiessen, E A.; Ehrlich, E. Z Phys. Chem. 1933,165, 453; (b) Marton, L.; McBain, J. W.; Void, R. D. J. Am. Chem. Soc. 1941, 63, 1990; (c) Thiele, H. Kolloid-Z 1952, 125, 31; (d) Kling, W.; Mahl, H. Fette Seife Anstrichmitte11995, 57, 643. 152. Tr~iger,O.; Sowade, S.; B6ttcher, C.; Fuhrhop, J.-H. J. Am. Chem. Soc. 1997, 119, 9120. 153. Yager, E; Schoen, E E.; Davies, C.; Price, R.; Singh, A. Biophys. J. 1985, 48, 899. 154. Singh, A.; Schnur, J. M. Polym. Prepr. Am. Chem. Soc. Div. Polym. Chem. 1985, 26, 184. 155. Singh, A.; Schnur, J. M. Synth. Commun. 1986, 16, 847. 156. Schnur, J. M. Science 1993, 262, 1669. 157. (a) Markowitz, M. A.; Schnur, J. M.; Singh, A. Chem. Phys. Lipids 1992, 62, 193; (b) Markowitz, M. A.; Baral, S.; Brandow, S.; Singh, A. Thin Solid Films 1993, 224, 242. 158. Georger, J. H.; Singh, A.; Price, R. R.; Schnur, J. M.; Yager, E; Schoen, E E. J. Am. Chem. Soc. 1987, 109, 6169. 159. Thomas, B. N.; Safinya, C. R.; Piano, R. J.; Clark, N. A. Science 1995, 267, 1635. 160. Rudolph, A. S.; Singh, B. R.; Singh, A.; Burke, T. G. Biochim. Biophys. Acta 1988, 943, 454. 161. Schoen, E E.; Yager, E J. Polynt Sci. Polym. Phys. 1985, 23, 2203. 162. Schoen, E E.; Nagumo, M.; Singh, A. Chem. Phys. Lipids 1994, 69, 251. 163. Rhodes, D. G.; Blechner, S. L.; Yager, E; Schoen, E E. Chem. Phys. Lipids 1988, 49, 39. 164. Caffrey, M.; Hogan, J.; Rudolph, A. S. Biochemistry 1991, 30, 2134. 165. Lando, J.; Sudiwata, R. V. Chem. Mater. 1990, 2, 594. 166. (a) Johnston, D. S.; Sanghera, S.; Pons, M.; Chapman, D. Biochim. Biophys. Acta 1980, 602, 57; (b) Hub, H. H.; Hupfer, B.; Koch, H.; Ringsdorf, H. Angew. Chent, Int. Ed. Engl. 1980, 19, 938; (c) O'Brien, D. E; Whitesides, T. H.; Klingbiel, R. T. J. Polym. Sci. Polym. Lett. Ed. 1981, 19, 95. 167. Wegner, G. Makromol. Chem. 1972, 154, 35. 168. Baughman, R. H.; Chance, R. R. Ann. N. Y. Acad. Sci. 1978, 313, 705. 169. Spector, M. S.; Selinger, J. V.; Singh, A.; Rodriguez, J. M.; Price, R. R.; Schnur, J. M. Langmuir 1998, 14, 3493. 170. Ratna, B. R.; Baral-Tosch, S.; Kahn, B.; Rudolph, A. S 9Schnur, J. M. Chem. Phys. Lipids 1992, 63,47. 171. Schnur, J. M.; Ratna, B. R.; Selinger, J. V.; Singh, A.; Jyothi, G.; Easwaran, K. R. K. Science 1994, 264, 945. 172. Rudolph, A. S.; Calvert, J. M.; Ayers, M. E.; Schnur, J. M. J. Am. Chem. Soc. 1989, 111, 8516. 173. Singh, A.; Burke, T. G.; Calvert, J. M.; Georger, J. H.; Herendeen, B.; Price, R. R.; Schoen, E E.; Yager, P. Chem. Phys. Lipids 1988, 47, 135. 174. Boyanov, A. I.; Tenchov, B. G.; Koyanova, R. D.; Koumanov, K. S. Biochim. Biophys. Acta 1983, 732,711. 175. Singh, A.; Schoen, E E.; Schnur, J. M. J. Chem. Soc., Chem. Commun. 1988, 1222.

150

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

176. Yager, P.; Price, R. R.; Schnur, J. M.; Schoen, P. E.; Singh, A.; Rhodes, D. G. Cheat Phys. Lz'pids 1988, 46, 171. 177. (a) de Gennes, E-G. The Physics of Liquid Crystals; Oxford University Press: London, 1974; (b) de Gennes, E G. Comptes Rendues Acad. Sci. (Paris) 1987, 304, 259. 178. Meyer, R. B.; Liebert, L.; Strzelecki, L.; Keller, E J. de Phys. (Fr) 1975, 36, L69. 179. (a) ChappeU, J. S.; Yager, E Biophys. J. 1991, 60, 952; (b) Chappell, J. S.; Yager, E Cheat Phys. 1991, 150, 73. 180. Lubensky, T. C.; Prost, J. J. Phys. H 1992, 2, 371. 181. Helfrich, W. J. Chem. Phys. 1986, 85, 1085. 182. (a) Helfrich, W.; Prost, J. Phys. Rev. A. 1988, 38, 3065; (b) Fuhrhop, J.-H.; Helfrich, W. Cheat Rev. 1993, 93, 1565. 183. (a) Nelson, E; Powers, T. Phys. Rev. Lett. 1992, 69, 3409; (b) Nelson, E; Powers, T. J. Phys. II (Fr.) 1993, 3, 1535. 184. Selinger, J. V.; Schnur, J. M. Phys. Rev. Lett. 1993, 71,4091. 185. Selinger, J. V.; MacKintosh, E C.; Schnur, J. M. Phys. Rev. E. 1996, 53, 3804. 186. Spector, M. S.; Easwaran, K. R. K.; Jyothi, G.; Selinger, J. V.; Singh, A.; Schnur, J. M. Proc. Natl. Acad. Sci. USA 1996, 93, 12943. 187. Spector, M. S.; Selinger, J. V.; Schnur, J. M. J. Aat Cheat Soc. 1997, 119, 8533. 188. Kunitake, T.; Kim, J.-M.; Ishikawa, Y. J. Cheat Soc., Perkin Trans. 2 1991, 885. 189. Borggreven, J. M. E M.; Hoeks, T. H. L.; Driessens, E C. M.; Zwanenburg, B. Caries Res. 1992, 26, 84. 190. Sommerdijk, N. A. J. M.; Hoeks, T. H. L.; Synak, M.; Feiters, M. C.; Nolte, R. J. M.; Zwanenburg, B. J. Ant Cheat Soc. 1997, I 17, 4338. 191. Sommerdijk, N. A. J. M.; Lambermon, M. H. L.; Feiters, M. C.; Nolte, R. J. M.; Zwanenburg, B. Cheat Commun. 1997, 1423. 192. Sommerdijk, N. A. J. M.; Lambermon, M. H. L.; Feiters, M. C.; Nolte, R. J. M.; Zwanenburg, B. Cheat Commun. 1997, 455. 193. Sommerdijk, N. A. J. M.; Buynsters, E J. J. A.; Akdemir, H.; Geurts, D. G., Nolte, R. J. M.; Zwanenburg, B. J. Org. Cheat 1997, 62, 4955, and correction, 9388. 194. Sommerdijk, N. A. J. M.; Buynsters, E J. A. A.; Pistorius, A. M. A.; Wang, M.; Feiters, M. C.; Nolte, R. J. M.; Zwanenburg, B. J. Cheat $oc., Cheat Commun. 1994, 1941, and addition, 2736. 195. Sommerdijk, N. A. J. M.; Buynsters, E J. J. A.; Akdemir, H.; Geurts, D. G., Feiters, M. C.; Nolte, R. J. M.; Zwanenburg, B. Cheat Fur. J. 1998, 4, 127. 196. Moy, V. T.; Keller, D. J.; Gaub, H. E." McConnell, H. M. J. Phys. Cheat 1986, 90, 3198. 197. Reich, Z.; Zaidman, L.; Gutman, S. B.; Arad, T.; Minski, A. Biochemistry 1994, 33, 14177. 198. Buijnsters, E J. J. A.; van der Reijden, E Unpublished results. 199. Buijnsters, E J. J. A.; Sommerdijk, N. A. J. M. Unpublished results. 200. Donovan, J. M.; Carey, M. C. Gastroenterol. Clin. 1991, 20, 47. 201. Konikoff, E M.; Chung, D. S.; Donovan, J. M.; Small, D. M.; Carey, M. C. J. Clin. Invest. 1992, 90, 1155. 202. Chung, D. S.; Benedek, G. B.; Konikoff, E M.; Donovan, J. M. Proc. Natl. Acad. Sci. USA 1993, 90, 11341. 203. Kaplun, A.; Talmon, Y.; Konikoff, E M.; Rubin, M.; Eitan, A.; Tadmor, M.; Lichtenberg, D. FEBS Lett. 1994, 340, 78. 204. Dahl, I.; Lagerwall, S. T. Ferroelectrics 1984, 58, 215. 205. Lin, Y.-c.; Kackar, B.; Weiss, R. G. J. Aat Cheat Soc. 1989, 111, 5542. 206. Murata, K.; Aoki, M.; Suzuki, T.; Harada, T.; Kawabata, H.; Komori, T.; Ohsato, E; Uedo, K.; Shinkai, S. J. Am. Cheat Soc. 1994, 116, 6664. 207. Cho, I.; Chung, K.-C. Macromol. 1984, 17, 2937. 208. Cho, I.; Park, J. G. Cheat Lett. 1987, 977. 209. Hauser, H.; Howell, K.; Dawson, R. M. C.; Bowyer, D. E. Biochiat Biophys. Acta 1980, 602, 567.

Chiral Self-Assembled Structures

151

210. Hansson, G. C. Biochim. Biophys. Acta 1983, 733, 295. 211. Suzuki, K.; Suzuki, Y. In The Metabolic Basis oflnherited Disease II, 6th ed.; Scriver, C. M., Ed., McGraw-Hill: San Francisco, 1989, p. 1699. 212. Lee, R. E.; Worthington, C. R.; Glew, R. H. Arch. Biochem- Biophys. 1973, 159, 259. 213. Curatolo, W.; Neuringer, L. J. J. Biol. Chem. 1986, 261, 17177. 214. Archibald, D. D.; Yager, P. Biochemistry 1992, 31, 9045. 215. Papahadjopoulos, D.; Vail, W. J.; Jacobson, K.; Poste, G. Biochim- Biophys. Acta 1975, 394, 483. 216. Femandez-Bermudez, S.; Loboda-Cackovic, J.; Caskovic, H.; Hosemann, R. 7_,Naturforsch. 1977, 32, 362. 217. Archibald, D. D.; Mann, S. Chem. Phys. Lipids 1994, 69, 51. 218. Shimizu, T.; Minamikawa, H.; Murakami, T.; Hato, M. Chem. Lett. 1993, 567. 219. Branden, C.; Tooze, J. Introduction to Protein Structure, 2nd ed.; Garland: New York, 1999. 220. Anfinsen, C. B. Science 1973, 181,223. 221. Pauling, L.; Corey, R. B.; Branson, H. R. Proc. Natl. Acad. Sci. USA 1951, 37, 205. 222. Paliakasis, C. D.; Kokkinidis, M. Protein Engr. 1992, 5, 739. 223. O'Neil, K. T.; DeGrado, W. E Science, 1990, 250, 646. 224. Bryson, J. W.; Betz, S. E; Lu, H. S.; Suich, D. J.; Zhou, H. X.; O'Neil, K. T.; DeGrado, W. E Science 1995, 270, 935. 225. Crick, E H. C. Acta Cryst. 1953, 6, 689. 226. lshikawa, S.; Kurita, T. Biopolymers 1964, 2, 381. 227. Rybnikar, E; Geil, P. H. Biopolymers 1972, 11,271. 228. Isemura, T.; Hamaguchi, K. Bull, Chem- Soc. Jpn. 1954, 27, 125. 229. Harris, N. L.; Presnell, S. R.; Cohen, E R. J. Mol. Biol. 1994, 236, 1356. 230. (a) Klug, A. Fed. Proc. 1972, 31, 30; (b) Klug, A.; Caspar, D. L. D. Adv. Virus Res. 1960, 7, 274. 231. Chan, D. C.; Fass, D.; Berger, J. M.; Kim, P. S. Cell 1997, 88, 263. 232. Blake, C. C. E; Geison, M. J.; Oatley, S. J.; R6rat, N.; R6rat, C. J. Mol. Biol. 1978, 121,339. 233. Blake, C.; Serpell, L. Structure 1996, 4, 989. 234. Pauling, L.; Corey, R. B. Proc. Natl. Acad. Sci. USA 1951, 37, 729. 235. Yoshida, T.; Sakurai, S.; Okuda, T.; Takagi, Y. J. Am. Chem. Soc. 1962, 84, 3590. 236. Fuhrhop, J.-H.; Krull, M.; Btildt, G. Angew. Chem. 1987, 99, 707. 237. (a) Fraser, R. D. B.; MacRae, T. P.; Suzuki, E. J. Mol. Biol. 1979, 129, 463; (b) Bella, J.; Eaton, M.; Brodsky, B.; Berman, H. M. Science 1994, 266, 75. 238. Dykes, G.; Crepeau, R. H.; Edelstein, S. J. Nature 1978, 272, 506. 239. Craig, R.; Padr6n, R.; Alamo, L. J. Mol. Biol. 1991, 220, 125. 240. Erickson, R. O. Science, 1973, 181,705. 241. Gong, M.; Makowski, L. J. Mol. Biol. 1992, 228, 735. 242. Ringler, P.; Mtiller, W.; Ringsdorf, H.; Brisson, A. Chem. Eur. J. 1997, 3, 620. 243. Ebato, H.; Herron, J. N.; MUller, W.; Okahata, Y.; Ringsdorf, H.; Suci, P. Angew. Chem- Int. Ed. Engl. 1992, 31, 1087. 244. Huetz, P.; van Neuren, S.; Ringler, P.; Kremer, E; van Breemen, J. E L.; Wagenaar, A.; Engberts, J. B. E N.; Fraaije, J. G. E. M.; Brisson, A. Chem. Phys. D'pids 1997, 89, 15. 245. Toniolo, C.; Benedetti, E. Trends Biochem- Sci. 1991, 16, 350. 246. Otoda, K.; Kimura, S.; Imanizhi, Y. Biochim. Biophys. Acta 1993, 1150, 1. 247. Hungerford, G.; Martinez-lnsua, M.; Birch, D. J. S.; Moore, B. D. Angew. Chem., Int. Ed. Engl. 1996, 35, 326. 248. (a) Weinstein, S.; Wallace, B. A.; Blout, E. R.; Morrow, J. S.; Veatch, W. Proc. Natl. Acad. Sci. USA 1979, 76, 4230; (b) Ketchem, R. R.; Hu, W.; Cross, T. A. Science 1993, 261, 1457. 249. (a) Ghadiri, M. R.; Granja, J. R.; Buehler, L. Nature 1994, 369, 301; (b) Granja, J. R.; Ghadiri, M. R. J. Am- Chem. Soc. 1994, 116,10785; (c) Hartgerink, J. D.; Granja, J. R.; Milligan, R. A.; Ghadiri, M. R. J. Am- Chem. Soc. 1996, 118, 43; (d) Lee, D. H.; Ghadiri, M. R. In Comprehensive

152

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

Supramolecular Chemistry; Atwood, J. L.; Davies, J. E. D.; Macnicol, D. D.; Vogtle, E, Eds.; Pergamon, Elsevier Science: 1996, Vol 9, p. 451. 250. (a) Yoder, M. D.; Keen, N. T.; Jumak, E Science 1993, 260, 1503; (b) Yoder, M. D.; Lietzke, S. E.; Jumak, E Structure 1993, 1,241. 251. Dado, G. P.; Gellman, S. H. J. Am. Chem. Soc. 1994, 116, 1054. 252. Appella, D. H.; Christianson, L. A.; Karle, I. L.; Powell, D. R.; Gellman, S. H. J. Am. Chem. Soc. 1996, 118, 13071. 253. Appella, D. H.; Christianson, L. A.; Klein, D. A.; Powell, D. R.; Huang, X.; Barchi Jr., J. J.; Gellman, S. H. Nature 1997, 387, 381. 254. Seebach, D.; Ciceri, P. E.; Overhand, M.; Jaun, B.; Rigo, D.; Oberer, L.; Hommel, U.; Amstutz, R.; Widmer, H. Helv. Chim. Acta 1996, 79, 2043. 255. Seebach, D.; Overhand, M.; Kiihnle, E M. N.; Martinoni, B.; Oberer, L.; Hommel, U.; Widmer, H. Helv. Chim. Acta 1996, 79, 913. 256. Hintermann, T.; Seebach, D. Synlett. 1997, 437. 257. lhara, H.; Fukumoto, T.; Hirayama, C.; Yamada, K. Polym. Commun. 1986, 27, 282. 258. (a) Shimizu, T.; Hato, M. Thin Solid Films 1989,180, 179; (b) Shimizu, T.; Mori, M.; Minamikawa, H.; Hato, M. Chem. Lett. 1989, 1341. 259. Shimizu, T.; Moil, M.; Minamikawa, H.; Hato, M. J. Chem. Soc., Chem. Commun. 1990, 183. 260. Madison, V.; Deber, C. M.; Blout, E. R. J. Am. Chem. Soc. 1977, 99, 4788. 261. Shimizu, T.; Hato, M. Biochim. Biophys. Acta 1993, 1147, 50. 262. Hidaka, H.; Murata, M.; Onai, T. J. Chem. Soc., Chem. Commun. 1984, 562. 263. lmae, T.; Takahashi, Y.; Muramatsu, H.J. Am. Chem. Soc. 1992, 114, 3414. 264. Nakashima, N.; Fukushima, H.; Kunitake, T. Chem. Lett. 1981, 1207. 265. Nakashima, N.; Asakuma, S.; Kunitake, T. J. Am. Chem. Soc. 1985, 107, 509. 266. lshikawa, Y.; Hishimi, T.; Kunitake, T. Chem. Lett. 1990, 25. 267. (a) Kunitake, T.; Nakashima, N.; Shimomura, M.; Okahata, Y.; Kano, K.; Ogawa, T. J. Am. Chem. Soc. 1980, 102, 6642; (b) Kunitake, T.; Nakasima, N.; Morimitsu, K. Chem. Lett. 1980, 1347. 268. Nakashima, N.; Ando, R.; Muramatsu, T.; Kunitake, T. Langmuir 1994, 10, 232. 269. Kimizuka, N.; Takasaki, T.; Kunitake, T. Chem. Lett. 1988, 1911. 270. lhara, H.; Takafuji, M.; Hirayama, C.; O'Brien, D. E Langmuir 1992, 8, 1548. 271. (a) Kuo, T.; O'Brien, D. E Macromol. 1990, 23, 3225; (b) Kuo, T.; O'Brien, D. E Langmuir 1991, 7, 584. 272. Rhodes, D. G.; Frankel, D. A.; Kuo, T.; O'Brien, D. E Langmuir 1994, 10, 267. 273. Polidori, A.; Pucci, B.; Zarif, L.; Lacombe, J.-M.; Riess, J. G." Pavia, A. A. Chem. Phys. b'pids 1995, 77, 225. 274. Zarif, L.; Polidori, A.; Pucci, B.; Gulik-Krzwycki, T.; Pavia, A. A.; Riess, J. G. Chem. Phys. Lipids 1996, 79, 165. 275. Kunitake, T.; Yamada, N. J. Chem. Soc., Chem. Commun. 1986, 655. 276. Yamada, N.; Sasaki, T.; Murate, H.; Kunitake, T. Chem. Lett. 1989, 205. 277. Yamada, N.; Okuyama, K.; Serizawa, T.; Kawasaki, M.; Oshima, S. J. Chem. Soc., Perla'n Trans. 2 1996, 2707. 278. Yamada, N.; Kawasaki, M. J. Chem. Soc., Chem. Commun. 1990, 568. 279. Ragunathan, K. G.; Bhattacharya, S. Chem. Phys. Lipids 1995, 77, 13. 280. Buijnsters, P. J. J. A.; ten Holte, P.; Verhaegen, S. A. C.; Chittenden, G. J. E; Feiters, M. C.; de Gelder, R.; Pistorius, A. M. A.; Sommerdijk, N. A. J. M.; Thijs, L.; Nolte, R. J. M.; Zwanenburg, B. Langmuir. Submitted. 281. Hanabusa, K.; Yamada, M.; Kimura, M.; Shirai, H. Angew. Chem., Int. Ed. Engl. 1996, 35, 1949. 282. Garcia-Telledo, E; Goswami, S.; Chang, S. K.; Geib, S. J.; Hamilton, A. D. J. Am. Chem. Soc. 1990, 112, 7393. 283. Geib, S. J.; Vicent, C.; Fan, E.; Hamilton, A. D. Angew. Chem., Int. Ed. Engl. 1993, 32, 119.

Chiral Self-Assembled Structures

153

284. (a) Hamuro, Y.; Geib, S. J.; Hamilton, A. D. Angew. Chem-, Int. Ed. Engl. 1994, 33, 446; (b) Hamuro, Y.; Geib, S. J.; Hamilton, A. D. J. Am- Chem- Soc. 1996, 118, 7529; (c) Hamuro, Y.; Geib, S. J.; Hamilton, A. D. J. Am. Chem- Soc. 1997, 119, 10587. 285. (a) Brienne, M.-J.; Gabard, J.; Leclercq, M.; Lehn, J.-M.; Cesario, M.; Pascard, C.; Chevt, M.; Dutruc-Rosset, G. Tetrahedron Lett. 1994, 35, 8157; (b) Brienne, M.-J.; Gabard, J.; Leclercq, M.; Lehn, J.-M.; Chevt, M. Helv. Chim` Acta 1997, 80, 856. 286. Tichy, M.; Ridvan, L.; Hory, P.; Z~ivada, J.; Cisarov~i, I.; Podlaha, J. Tetrahedron: Asymmetry 1998, 9, 227. 287. Palmans, A. R. A.; Vekemans, J. A. J. M.; Havinga, E. E.; Meijer, E. W. Angew. Chem., Int. Ed. Engl. 1997, 36, 2648. 288. Chandrasekaran, R. Adv. Carbohydr. Chem` Biochem- 1997, 52, 311. 289. Meyer, K. H.; Misch, L.; Helv. Chim- Acta 1937, 20, 232. 290. Gardner, K. H.; Blackwell, J. Biopolymers 1974, 13, 1975. 291. Kolpak, E J.; Blackwell, J. MacromoL 1976, 9, 273. 292. Stipanovic, A. J.; Sarko, A. Polymer 1978, 9, 3. 293. Gardner, K. H.; Blackwell, J. Biopolymers 1975, 14, 1581. 294. Carlstrom, D. J. Biochem. Biophys. CytoL 1957, 3, 669. 295. Wu, H. C.; Sarko, A. Carbohydr. Res. 1978, 27. 296. lmberty, A.; Chanzy, H.; Perez, S.; Bul~,on, A.; Tran, V. J. MoL Biol. 1988, 201,365. 297. Imberty, A.; Ptrez, S. Biopolymers 1988, 27, 1205. 298. (a) Rundle, R. E.; Edwards, E C. J. Am. Chem. Soc. 1943, 65, 2200; (b) Rundle, R. E.; French, D. J. Am` Chem` Soc. 1943, 65, 558; (c) Rundle, R. E.; French, D. J. Am- Chem` Soc. 1943, 65, 1707. 299. (a) Rundle, R. E. J. Am- Chem` Soc. 1947, 69, 1769; (b) Zaslow, B.; Miller, R. L. J. Am- ChemSoc. 1961, 83, 4378; (c) Zaslow, B. In Starch Chemistry & Technology, Fundamental Aspects; Whistler, R. L.; Paschall, E. E, Eds.; Academic Press: New York/London, 1965, Vol. 1, p. 279. 300. Winter, W. T.; Sarko, A. Bioplymers 1974, 13, 1205. 301. Senti, E R.; Witnauer, L. P. J. Polymer Sci. 1952, 9, 115. 302. Hinrichs, W.; Biittner, G.; Steifa, M.; Betzel, C.; Zabel, V.; Pfannemtiller, B.; Saenger, W. Science 1987, 238, 205. 303. Kim, O.-S.; Choi, L.-S.; Zhang, H.-Y.; Zhang, He, X.-H.; He, Shih, Y.-H. Thin Solid Films 1998, 327-329, 172. 304. (a) Anderson, N. S.; Campbell, J. W.; Harding, M. M.; Rees, D. A.; Samuel, J. W. B. J. Mol Biol. 1969, 45, 85; (b) Rees, D. A.; Welsh, G. E.Angew. Chem., Int. Ed. Engl. 1977,16, 214; (c) Piculell, L.; Borgstrtm, J.; Chronakis, I. S.; Quist, P.-O.; Viebke, C. Int. J. Biol. Macromol. 1997, 21,141; (d) Viebke, C.; Borgstrtm, J.; Carlsson, I.; Picullell, L.; Williams, P. MacromoL 1998, 31, 1833. 305. (a) Marchessault, R. H.; Deslandes, Y.; Ogawa, K.; Sundarajan, P. R. Can. J. Chem. 1977, 55, 300; (b) Deslandes, Y.; Marchessault, R. H.; Sarko, A. Macromolecules 1980, 13, 1466; (c) Okuyama, K.; Otsubo, A.; Fukuzawa, Y.; Ozawa, M.; Harada, T.; Kasai, N. J. Carbohydr. Chem` 1991, 10, 645. 306. (a) Norisuye, T.; Yanaki, T.; Fujita, H. J. Polym- Sci. Polym- Phys. Ed. 1980, 18, 547; (b) Yanaki, T.; Norisuye, T.; Fujita, H. Macromol. 1980, 13, 1462; (c) Kashiwagi, Y.; Norisuye, T.; Fujita, H. Macromol. 1981, 14, 1220; (d) ltou, T.; Van, K.; Teramoto, A. J. AppL Polym` Sci., AppL Polym. Symp. 1985, 41, 35; (e) Van, K.; Asakawa, T.; Teramoto, A. Polym- J. 1984, 16, 61; (f) Sato, T.; Sato, Y.; Umemura, Y.; Teramoto, A.; Nagamura, Y.; Wagner, J.; Weng, D.; Okamoto, Y.; Hatada, K.; Green, M. M. Macromolecules 1993, 26, 4551; (g) Sato, T.; Nakamura, J.; Teramoto, A.; Green, M. M. Macromolecules 1998, 31, 1398. 307. Bluhm, T. L.; Sarko, A. Can. J. Chem. 1977, 55, 293. 308. (a) Atldns, E. D. T.; Parker, K. D. J. Polym- Sci., Part C 1969, 28, 69; (b) Atkins, E. D. T.; Parker, K. D.; Preston, R. D. Proc. Roy. Soc. B 1969, 173, 209. 309. Zabel, V.; Mtlller-Fahrnow, A.; Hilgenfeld, R.; Saenger, W.; Pfannemiiller, B.; Enkelmann, V.; Welte, W. Chem. Phys. la'pids 1986, 39, 313.

154

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

310. K6ning, J.; Boettcher, C.; Wirdder, H.; Zeitler, E.; Talmon, Y.; Fuhrhop, J.-H. J. Am. Chem. Soc. 1993, 115, 693. 311. Boettcher, C.; Stark, H.; van Heel, M. Ultramicroscopy 1996, 62, 133. 312. Tuzov, I.; Cr/imer, K.; Pfarmemtiller, B.; Magonov, S. N.; Whangbo, M.-H. New. J. Chem. 1996, 20,37. 313. Fuhrhop, J.-H.; Svenson, S.; Boettcher, C.; Rossler, E.; Vieth, H.-M. J. Am. Chem- Soc. 1990,112, 4307. 314. Fuhrhop, J.-H.; Schnieder, P.; Boekema, E.; Helfrich, W. J. Am- Chem- Soc. 1988, 110, 2861. 315. Horton, Z.; Walaszek, Z.; Ekiel, I. Carbohydr. Res. 1983, 119, 263. 316. Andr6, C.; Luger, P.; Svenson, S.; Fuhrhop, J.-H. Carbohydr. Res. 1992, 230, 31. 317. Andr6, C.; Luger, P." Svenson, S." Fuhrhop, J.-H. Carbohydr. Res. 1994, 240, 47. 318. Fuhrhop, J.-H.; Boettcher, C. J. Am- Chem- Soc. 1990, 112, 1768. 319. Boettcher, C.; Boekema, E. J.; Fuhrhop, J.-H. J. Microscopy 1990, 160, 173. 320. Frankel, D. A.; O'Brien, D. E J. Am- Chem. Soc. 1991, 113, 7436. 321. Fuhrhop, J.-H.; Blumtritt, P.; Lehmann, C.; Luger, P. J. Am- Chem. Soc. 1991, 113, 7437. 322. Frankel, D. A.; O'Brien, D. E J. Am- Chem. Soc. 1994, 116, 10057. 323. Andr6, C.; Luger, P.; Fuhrhop, J.-H. Chem- Phys. Lipids 1994, 230, 175. 324. Frankel, D. A.; O'Brien, D. E Macromol. Syrup. 1994, 77, 141. 325. Hafkamp, R. J. H.; Feiters, M. C.; Nolte, R. J. M. Angew. Chem-, Int. Ed. Engl. 1994, 33, 986. 326. Hafkamp, R. J. H., PhD Thesis, University of Nijmegen, 1996. 327. Jeffrey, G. A.; Maluszunska, H. Carbohydr. Res. 1990, 207, 211. 328. (a) Miiller-Fahrnow, A.; Hilgenfeld, R.; Hesse, H.; Saenger, W.; Pfannemtiller, B.; Carbohydr. Res. 1988, 176, 165; (b) Mtiller-Fahrnow, A.; Saenger, W.; Fritsch, D.; Schnieder, P.; Fuhrhop, J.-H. Carbohydr. Res. 1993, 242, 11. 329. Herbst, R.; Steiner, T.; Pfannemiiller, B.; Saenger, W. Carbohydr. Res. 1995, 269, 29. 330. Andr6, C.; Luger, P.; Nehmzow, D.; Fuhrhop, J.-H. Carbohydr. Res. 1994, 261, 1. 331. Hafkamp et al. Manuscript in preparation. 332. Feiters, M. C.; Roelofsen, A. M.; Hafkamp, R. J. H.; Nolte, R. J. M. Chem- Commun. 1999. Submitted. 333. Hafkamp, R. J. H.; Kokke, B. P. A.; Danke, I. M.; Geurts, H. P. M.; Rowan, A. E.; Feiters, M. C.; Nolte, R. J. M. Chem. Commun. 1997, 545. 334. Hafkamp, R. J. H.; Feiters, M. C.; Nolte, R. J. M. J. Org. Chem- 1999, 64, 412. 335. Vollhardt, D.; Gutberlet, T.; Emrich, G.; Fuhrhop, J.-H. l.xmgmuir 1995, 11, 2661. 336. Vollhardt, D.; Emrich, G.; Gutberlet, D.; Fuhrhop, J.-H. Langmuir 1996, 12, 5659. 337. Bindig, U.; Schulz, A.; Fuhrhop, J.-H.; Siggel, U. New. J. Chem- 1995, 19, 427. 338. Shimizu, T.; Masuda, M.J. Am- Chem- Soc. 1997, 119, 2812. 339. Shimizu, T.; Masuda, M. Mol. Cryst. b'q. Cryst. 1997, 295, 197. 340. Masuda, M.; Shimizu, T. Chem- Commun. 1996, 1057. 341. Masuda, M.; Shimizu, T. Unpublished results, cited in ref. 338. 342. Zarif, L.; Gulik-Krzwycki, T.; Riess, J. G.; Pucci, B.; Guedj, C.; Pavia, A. A. Colloids and Surfaces A. Physicochem. Eng. Aspects 1993, 84, 107. 343. Watson, J. D.; Crick, E H. C. Nature, 1953, 171,737. 344. Finch, J. T.; Klug, A. Proc. Natl. Acad. Sci. USA 1976, 73, 1897. 345. Dickerson, R. E.; Drew, H. R.; Conner, B. N.; Wing, R. M.; Fratini, A. V.; Kopka, M. L. Science 1982, 216, 475. 346. Rich, A.; Nordheim, A.; Wang, A. H. J. Annu. Rev. Biochem- 1984, 53, 791. 347. (a) Mirkin, S. N.; Lyamichev, V. I.; Drushlyak, K. N.; Dobynin, V. N.; Filippov, S. A.; Frank-Kamenetskii, M. D. Nature 1987, 330, 495; (b) Voloshin, O. N.; Mirkin, S. M.; Lyamichev, V. I.; Belotserkovskii, B. P.; Frank-Kamenetskii, M. D. Nature 1988, 333, 475; (c) Hentschel, C. C. Nature 1982, 295, 714; (d) Htun, H.; Dahlberg, J. E. Science 1989, 243, 1571; (e) Johnston, B. H. Science 1988, 241, 1800; (f) Schroth, G. P.; Ho, S. H. Nucleic Acid Res. 1995, 23, 1977.

Chiral Self-Assembled Structures 348. 349. 350. 351. 352. 353. 354. 355.

356. 357. 358. 359. 360. 361. 362.

363.

364. 365. 366. 367. 368. 369. 370. 371. 372. 373. 374. 375. 376. 377.

378. 379. 380. 381.

155

Seeman, N. C. Angew. Chem. 1998, 37, 3220. Bustamante, C.; Marko, J. E; Siggia, E. D.; Smith, S. B. 1994, Science, 265, 1599. Samorf, B. Angew. Chem., Int. Ed. Engl. 1998, 37, 2198. Eschenmoser, A.; Dobler, M. Helv. Chim. Acta 1992, 75, 218. Schwarz, A. W. Current Biol. 1997, 7, 477. (a) Woese, C. In The Genetic Code; Harper & Row: New York, 1967, p. 179; (b) Crick, E H. C. J. Mol. Biol. 1968, 38, 367; (c) Orgel, L. E. J. Mol. Biol. 1968 38, 381. Joyce, G. E; Visser, G. M.; van Boeckel, C. A. A.; van Boom, J. H.; Orgel, L. E.; van Westrenen, J. Nature 1984, 310, 602. (a) Joyce, G. E; Schwartz, A. W.; Miller, S. L.; Orgel, L. E. Proc. Natl. Acad. Sci. USA 1987, 84, 4398; (b) Visscher, J.; Schwarz, A. W. J. Mol. Evol. 1990, 31, 163; (c) van Vliet, M. J. Ph.D. Thesis 1995. B6ehler, C.; Nielsen, P. E.; Orgel, L. E. Nature 1995, 376, 578. Schmidt, J.G.; Nielsen, P. E.; Orgel, L. E. J. Am. Chem. Soc. 1997, 119, 1494. Pitsch, S.; Wendebarn, S.; Jaun, B.; Eschenmoser, A. Helv. Chim. Acta 1993, 76, 2161. Bolli, M.; Micura, R.; Eschenmoser, A. Chem. Biol. 1997, 4, 309. Gellert, M.; Lipsett, M. N.; Davies, D. R. Proc. Natl. Acad. Sci. USA 1962, 48, 1463. Bang, I. Biochem. Z 1910, 26, 293. Gottarelli, G.; Spada, G. P.; Garbesi, A. In Comprehensive Supramolecular Chemistry; Atwood, J. L.; Davies, J. E. D.; Macnicol, D. D.; Vogtle, F., Eds.; Pergamon; Elsevier Science: 1996, Vol. 9, p.483. (a) Amott, S.; Chandrasekaram, R.; Martilla, C. M. Biochem. J. 1974, 141,537; (b) Zimmermann, S. B.; Cohen, G. H.; Davies, D. R. J. Mol. Biol. 1975, 92, 181; (c) Blackburn, E. H.; Szostak, J. W. Annu. Rev. Biochem. 1984, 53, 163. Bonazzi, S.; Capobianco, M.; De Morais, Garbesi, A.; Gottarelli, G.; Mariani, P.; Ponzi Bossi, M. G.; Spada, G. P.; Tondelli, L. J. Am. Chem. Soc. 1991, 113, 5809. (a) Nielsen, P. E.; Egholm, M.; Berg, R. H.; Buchardt, O. Science 1991, 254, 1497; (b) Egholm, M.; Buchardt, O.; Nielsen, P. E.; Berg, R. H. J. Am. Chem. Soc. 1992, 114, 1895. Wittung, P.; Nielsen, P. E.; Buchardt, O.; Egholm, M.; Norden, B. Nature 1994, 368, 561. Brown, S. C.; Thomson, S. A.; Veal, J. M.; Davis, D. G. Science 1994, 265, 777. Egholm, M.; Bui:hardt, O.; Christensen, L.; Behrens, C.; Freier, S. M.; Driver, D. A.; Berg, R. H.; Kim, S. K.; Norden, B.; Nielsen, P. E. Nature 1993, 365, 566. Itojima, Y.; Ogawa, Y.; Tsuno, K.; Hand, N.; Yanagawa, H. Biochemistry 1992, 31, 4757. Yanagawa, H.; Ogawa, Y.; Furuta, H.; Tsuno, K. Chem. Lett. 1988, 269. Yanagawa, H.; Ogawa, Y.; Furuta, H.; Tsuno, K. Chem. Lett. 1989, 403. Yanagawa, H.; Ogawa, Y.; Furuta, H.; Tsuno, K. J. Am. Chem. Soc. 1989, 111, 4567. Simpkins, H.; Richards, E. G. J. Mol. Biol. 1967, 29, 349. Kanaya, E.; Yanagawa, H. Biochemistry 1986, 25, 7423. Gulik-Krzywicki, T.; Fouquey, C.; Lehn, J.-M. Proc. Natl. Acad. Sci. USA 1993, 90, 163. Fouquey, C.; Lehn, J.-M.; Levelut, A.-M. Adv. Mater 1990, 5, 254. (a) Seto, C. T.; Whitesides, G. M. J. Am. Chem. Soc. 1993, 115, 905; (b) Whitesides, G. M.; Mathias, J. P.; Seto, C. T. Science 1991, 254, 1312; (c) Mathias, J. P.; Seto, C. T.; Simanek, E. E.; Whitesides, G. M. J. Am. Chem. Soc. 1994, 116, 1725; (d) Mathias, J. P.; Simanek, E. E.; Zerkowski, J. A.; Seto, C. T.; Whitesides, G. M.J. Am. Chem. Soc. 1994, 116, 1725; (e) Mathias, J. P.; Simanek, E. E.; Whitesides, G. M. J. Am. Chem. Soc. 1994, 116, 4326. Ahuja, R.; Caruso, P.-L.; Mobius, D.; Paulus, W.; Ringsdorf, H.; Wildburg, G. Angew. Chem., Int. Ed. Engl. 1993, 32, 1033. Yang, W.; Chai, X.; Chi, X.; Chi, L.; Liu, X ; Cao, Y.; Lu, R.; Jiang, Y.; Tang, X.; Fuchs, H.; Li, T. Chem. Eur. J. 1999, 5, 1144. Kimizuka, N.; Kawasaki, T.; Hiram, K.; Kunitake, T. J. Am. Chem. Soc. 1995, 117, 6360. Prins, L.; Huskens, J.; de Jong, E; Timmermans, P.; Reinhoudt, D. N. Nature 1999, 398, 498.

156

MARTINUS C. FEITERS and ROELAND J. M. NOLTE

382. (a) Vreekamp, R. H.; van Duynhoven, J. P. M.; Hubert, M.; Verboom, W; Reinhoudt, D. N.Angew. Chem., Int. Ed. Engl. 1996, 35, 1215; (b) Timmerman, E; Vreekamp, R. H.; Hulst, R; Verboom, W.; Reinhoudt, D. N.; Rissanen, K.; Udachin, K. A.; Ripmeester, J. Chent Eur J. 1997, 3, 1823. 383. Russell, K. C.; Lehn, J.-M.; Kyritsakas, N.; DeCian, A.; Fischer, J. New. J. Chent 1998, 123. 384. Oda, R.; Huc, I.; Schmutz, M.; Candau, S. J.; MacKintosh, E C. Nature 1999, 399, 566.

SPHERICAL MOLECULAR CO NTAI N ERS" FROM DISCOVERY TO DESIGN

Leonard R. MacGillivray and Jerry L. Atwood

1.

2. 3.

4. 5.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 1.1. Supramolecular Chemistry . . . . . . . . . . . . . . . . . . . . . . . . 158 1.2. Towards Supramolecular Synthesis . . . . . . . . . . . . . . . . . . . 159 1.3. Self-Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 Elaborating the Cavities of Resorcin[4]arenes Supramolecularly . . . . . . . 160 3.1. Cavities Based Upon Rigid Extenders . . . . . . . . . . . . . . . . . . 161 3.2. A Cavity Based Upon a Flexible Extender . . . . . . . . . . . . . . . . 163 A Spherical Molecular Container Held Together by 60 Hydrogen Bonds . . . 164 General Principles for the Design of Spherical Containers . . . . . . . . . . . 165 5.1. Spheroid Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 5.2. Self-Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 5.3. Subunits for Spheroid Design and Self-Assembly . . . . . . . . . . . . 166 5.4. Platonic Solids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 5.5. Archimedean Solids . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 5.6. Models for Spheroid Design . . . . . . . . . . . . . . . . . . . . . . . 171

Advances in Supramolecular Chemistry Volume 6, pages 157-183. Copyright 9 2000 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 0-7623-0557-6

157

158

LEONARD R. MACGILLIVRAY and JERRYL. ATWOOD

6. Examplesof Spherical Containers from the Laboratory and from Nature . . . 171 6.1. PlatonicSolids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 6.2. ArchimedeanSolids . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 6.3. ArchimedeanDuals and Irregular Polygons . . . . . . . . . . . . . . . 178 6.4. IrregularPolygons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 7. Why the Platonic and Archimedean Solids? . . . . . . . . . . . . . . . . . . . 180 8. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 References and Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 1.

INTRODUCTION

During the past century, the field of X-ray crystallography has provided numerous insights into problems in chemistry. Indeed, since the discovery by Laue, in 1912, that crystals diffract X-rays, l chemists have used information gained from single-crystal studies to formulate principles for understanding properties of inorganic, organic, and biochemical compounds and to synthesize new chemical entities. The work of Pauling in 1928, for example, provided sets of rules for describing and predicting arrangements of ions in inorganic compounds such as minerals, 2 while structure determinations of deoxyribonucleic acid (DNA) and viruses by Watson and Crick and Klug, in 19533 and 1962, 4 respectively, furnished an understanding of those processes associated with the chemistry of life.

1.1. Supramolecular Chemistry With a role of crystallography in providing structural details of molecular connectivity, in the first three quarters of the century, and in "traditional" areas of chemistry thus established, it is, perhaps, not surprising that it has only been within the last two decades that chemists have, in part, used this information to turn to a multidisciplinary approach to synthesis that utilizes noncovalent bonds (e.g. hydrogen bonds, n - x interactions) for the design of highly organized chemical entities. Known as supramolecular chemistry, or chemistry beyond the molecule, this field, first conceptualized by Lehn in 1973, 5 deals with the chemist's ability to exploit the structure-directing properties of noncovalent forces for the design of multicomponent chemical species that, in a similar way to biological systems (e.g. proteins), exhibit properties (e.g. host-guest) not found in the individual components (e.g. amino acids). Ideally, such systems have been designed to gain insights into various biological phenomena. If accomplished, these insights may ultimately permit chemists to rationally design drugs, catalysts, and perhaps, processes reminiscent of life itself, as well as generate new classes of materials with unique properties.

Spherical Molecular Containers

159

Discrete

or

Infinite

Figure 1. Two modes of self-assembly of a complementary subunit.

1.2. Towards Supramolecular Synthesis Although it is now generally accepted that a supramolecular approach to chemistry can provide a route to achieving these goals, it has become clear that if chemists are to synthesize entities analogous to biological systems, they must face the realization that the dimensions of biological structures typically range from tens to thousands of angstroms 6 and that an approach to synthesis that depends exclusively upon the stepwise formation and breakage of covalent bonds is burdened with inherent limitations. 7 For example, in addition to requiring a large amount of time, such an approach often results in low product yields (e.g. natural product synthesis). As a result, an alternative method for designing large chemical systems (e.g. nanosystems) must be employed. 1.3.

Self-Assembly

Within the last decade, it has become evident that self-assembly, the single-step construction of molecular architecture using noncovalent forces, has provided an attractive means for constructing large, highly organized chemical entities. Owing to the reversibility of such interactions, noncovalent bonds, upon selection of appropriate chemical subunits, can facilitate error-free generation of either discrete or infinite supramolecular species usually in quantitative yield (Figure 1). Although the idea is not entirely new--Nature uses this approach in a number of multicomponent systems such as hemoglobin 8 and viruses 4 while crystallization can be viewed as an example of self-assembly par excellence9mchemists are just beginning to understand those principles which govern the self-assembly process. 2.

OVERVIEW

It is with these ideas in mind that we now provide an account of work performed in our laboratory during the last 2 years which focuses upon the discovery, in the

160

LEONARD R. MACGILLIVRAY and JERRYL. ATWOOD

Figure 2. (a) X-ray crystal structure of 1 and (b) the snub cube, one of the 13 Archimedean solids. The square faces of the snub cube correspond to the resorcin[4]arenes; the eight shaded triangles that adjoin three squares correspond to the water molecules of 1. solid state, of a chiral, spherical host assembly held together by 60 hydrogen bonds 110 (Figure 2a) and the development of a rational design strategy for the construction of related host systems, ll In particular, upon recognizing that the topology of 1 conforms to the structure of a snub cube, one of the 13 Archimedean solids (Figure 2b), 12 we have structurally classified a variety of organic, inorganic, and biological shells and have shown that their structures may be cataloged according to principles of solid geometry in which the five Platonic and 13 Archimedean solids may be used as models for host design. This approach, which we recognize as being driven by biology, has led us to identify containers that have yet to be synthesized or discovered and we propose such systems as targets for chemical synthesis.

0

ELABORATING

THE CAVITIES OF RESORCIN[4]ARENES S U P R A M O L E C U LARLY

We have recently shown that it is possible to extend the cavities of bowl-shaped molecules known as resorcin[4]arenes (e.g. C-methylcalix[4]resorcinarene 2), supramolecularly, using hydrogen bond acceptors as extender units. 13'14As a starting point, we chose readily available 2 as a platform for the assembly process. 15Indeed, solid-state studies had revealed the ability of 2 to adopt a bowl-like conformation

Spherical Molecular Containers

......

161

GUEST

','H H. i o - - ' " ' o - , , ~ ' ~ , o : . : o 5 . ~

H/ ,..,...

4 Scheme 1. with C2v symmetry in which four upper rim hydroxyl hydrogen atoms of 2 are pointed upward above its cavity which, in turn, effectively make 2 a quadruple hydrogen bond donor. 16 Using a resorcinol-based supramolecular synthon 17 3 for host design, we reasoned that cocrystallization of 2 with hydrogen bond acceptors such as pyridines would result in formation of four O-H...N hydrogen bonds between the upper rim of 2 and four pyridine units which would extend the cavity of 2 and yield a discrete, multicomponent host, 2.4(pyridine) (where pyriaine = pyridine and derivatives), capable of entrapping a guest, 2.4(pyridine).guest 4 (Scheme 1).

3.1. Cavities Based Upon Rigid Extenders The product of the cocrystallization of 3 with pyridine is shown in Figure 3.13 The assembly is bisected by a crystallographic mirror plane and consists of 2 and five molecules of pyridine, four of which form four O-H...N hydrogen bonds, as two face-to-face dimers, such that they adopt an orthogonal orientation, in a similar way to 3, with respect to the upper rim of 2. As a consequence of the assembly process, a cavity has formed inside which a disordered molecule of pyridine is located, interacting with 3 by way of C-H...Tt-arene interactions. Notably, the

Figure 3. X-ray crystal structure of 2.4(pyridine).pyridine.

162

LEONARD R. MACGILLIVRAY and JERRY L. ATWOOD

remaining hydroxyl hydrogen atoms of the six-component assembly form four intramolecular O-H...O hydrogen bonds along the upper rim of 3 resulting in a total of eight structure-determining O-H...X (X = N, O) forces. Indeed, the inclusion of an aromatic such as pyridine within 2-4(pyridine) is reminiscent of the ability of covalently modified calix[4]arenes, such as p-tert-butylcalix[4]arene, to form molecular complexes with aromatics such as benzene and toluene. 18 To determine whether it is possible to isolate a guest within 4 which, unlike 2.4(pyridine), is different than that of the "substituents" hydrogen bonded to the upper rim of 2, we next turned to pyridine derivatives, namely 4-picoline (monopyridine) and 1,10-phenanthroline (bipyridine). 14 In a similar way to pyridine, both molecules possess hydrogen bond acceptors along their surfaces and n-rich exteriors which we anticipated would allow these units to assemble along the upper rim of 2 as stacked dimers. As shown in Figure 4, cocrystallization of 2

(a)

.,

. ;i,'

(b)

\./

.,,j

Figure 4. ORTEP perspective of: (a) 2.4(4-picoline).MeNO2 and (b) 2.4(1,10-phenanthroline).MeCN.

Spherical Molecular Containers

163

with either 4-picoline or 1,10-phenanthroline from MeNO 2 and MeCN, respectively, yielded six-component complexes, 2.4(4-picoline)-MeNO 2 and 2.4(1,10phenanthroline).MeCN, which are topologically equivalent to the parent pyridine system. 13 Unlike the parent assembly, however, the cavities created by the five molecules were occupied by guests different than the walls of the host. Indeed, this observation illustrated that this approach to discrete, extended frameworks based upon 2 is not limited to two different components. Notably, whereas the 4-picoline moiety was observed to interact with 2 by way of conventional O-H...N hydrogen bonds, the 1,10-phenanthroline extender was also observed to interact with 3 by way of a bifurcated O-H-..N force.

3.2. A Cavity Based Upon a Flexible Extender With the realization that 4 may be exploited for the inclusion of guests different than the supramolecular extenders of 2 achieved, we shifted our focus to pyridines

Figure 5. Space-filling view of the X-ray crystal structure of (a) 2.4(4-viffylpyridine).MeNO2 and (b) resorcinol.2(4-vinylpyridine).

164

LEONARD R. MACGILLIVRAYand JERRYL. ATWOOD

that possess flexible substituents. In addition to introducing issues of stereochemistry, we anticipated that this approach would allow us to further address the robustness and structural parameters that define 3 and thereby aide the future design of analogous host-guest systems based upon 2. Our first study in this context involved 4-vinylpyridine. 19 As shown in Figure 5, four 4-vinylpyridines assemble along the upper rim of 2, as two face-to-face stacked dimers, in 2.4(4-vinyl-pyridine).MeNO 2, to form a six-component assembly. Interestingly, the olefins of this system, in contrast to resorcinol.4-vinylpyridine, adopt a parallel orientation in the crystalline state, the bonds being separated by a distance of 4.18/~. Indeed, approaches that utilize host frameworks to promote alignment of olefins in the solid state for conducting [2+2] photochemical reactions, for example, are rare 2~ and these observations suggest that similar complexes based upon 3 may provide a route to achieving this goal.

4. A SPHERICAL MOLECULAR CONTAINER HELD TOGETHER BY 60 HYDROGEN BONDS It is clear that cocrystallization of 2 with pyridine and its derivatives in the presence of a suitable guest results in the extension of the cavity of 2 in which four pyridine units assemble along the upper rim of 2, as two stacked dimers, to yield a multicomponent complex 4 in which 2 serves as a quadruple hydrogen bond donor. 13'14'19During studies aimed at cocrystallizing 2 with pyridines from aromatic solvents--potential guests for 4 (e.g. nitrobenzene)--we discovered the ability of 2 to self-assemble in the crystalline state as a spherical hexamer, along with eight water molecules, to form a spherical container assembly with idealized 432 symmetry held together by 60 O-H.--O hydrogen bonds 1. l~ The assembly, which is chiral, possesses a well-defined cavity with a maximum diameter of 1.8 nm and an internal volume of about 1.4 nm 3, a volume five times larger than the largest molecular capsule previously reported. 21 Although guest species could be located with the interior of 1 (i.e. electron density maxima), it was not possible to determine their identity from the X-ray experiment, presumably owing to the high symmetry of the host and high thermal motion within the host cavity. Notably, solution studies also revealed the ability of C-undecylcalix[4] resorcinarene to maintain the structure of I in apolar organic solvents such as benzene. 1~ Consultation of polyhedron models revealed the structure of 1 to conform to a snub cube, one of the 13 Archimedean solids, in which the vertices of the square faces correspond to the comers of 2 and the centroids of the eight triangles that adjoin three squares correspond to the eight water molecules. 12 Indeed, to us, the ability of six resorcin[4]arenes to self-assemble to form 1 was reminiscent of spherical viruses in which identical copies of proteins self-assemble, by way of noncovalent forces, to fo.rm viral capsids having icosahedral symmetry and a shell-like enclosure. In fact, owing to the fit displayed by its components, 1 exhibits a topology that agrees with the theory of virus shell structure which states that

Spherical Molecular Containers

165

octahedral systems must contain 24 asymmetric units and possess 432 symmetry.4 Moreover, these observations suggested that to design related spherical hosts, one must consider the limited possibilities available in space for constructing spherical frameworks based upon regular polygons, those being the five Platonic and 13 Archimedean solids. 0

GENERAL PRINCIPLES FOR THE DESIGN OF SPHERICAL CONTAINERS

The discovery that members of the resorcin[4]arene family self-assemble to form 1, owing to its classification as an Archimedean solid, prompted us to examine the topologies of related spherical hosts with a view to understanding their structures on the basis of symmetry. In addition to providing a grounds for classification, we anticipated that such an approach would allow us to identify similarities at the structural level, which, at the chemical level, may not seem obvious and may be used to design large, spherical host assemblies similar to 1. We thus now describe the results of this analysis which we regard as the development of a general strategy for the construction of spherical molecular hosts. 11 We will begin by presenting the idea of self-assembly in the context of spherical hosts and then, after summarizing the Platonic and Archimedean solids, we will provide examples of cubic symmetry-based hosts, from both the laboratory and Nature, with structures that conform to these polyhedra.

5.1. SpheroidDesign The strategy stems from ideas developed within host-guest 22 and supramolecular chemistry. 5 As is the case for spherical viruses 4 and fullerenes, 23 an appropriately sized, shaped, and functionalized guest is packaged within the interior of a host such that it is completely surrounded or enclosed. From a geometrical standpoint, this may be achieved by centralizing the guest within a hollow spherical shell S (Figure 6). It must be noted, however, that, from a chemical standpoint, it is impossible to mimic S since atoms and molecules are discrete entities, whereas the surface of S is uniform. Thus, to design a spherical molecular host, an alternative procedure must be employed.

5.2. Self-Assembly In a paper describing the structure of regular viruses, Caspar and Klug 4 have shown that viral capsids use self-assembly to construct spherical shells up to 100 nanometers in diameter by utilizing identical copies of proteins as chemical subunits. Indeed, such a design strategy is desirable since it employs an economy of information, 23'24 giving rise to a host whose subunits exist in identical chemical environments, exposed surface area is at a minimum, and strain energy is distributed evenly along its surface. That this process is also amenable for spherical shells at

166

LEONARD R, MACGILLIVRAY and JERRY L. ATWOOD

Figure 6.

Hollow spherical shell S.

the angstrom level was later realized with the discovery of buckminsterfullerene, or C60.25 Notably, in the extreme case, an infinite number of identical subunits placed along the surface of a sphere will lead to a shell topologically equivalent to S. Thus, to construct a spherical molecular host, one must ultimately consider the number of subunits n for spheroid design and their placement along the surface of the shell.

5.3. Subunitsfor Spheroid Design and Self-Assembly To construct a spherical shell using a single subunit, n = 1, the only structure obtainable is S. As outlined above, it is impossible to construct a spherical molecular host using a single chemical entity and therefore S does not represent a self-assembled spherical framework. For n = 2, each subunit must cover one-half of the surface of the sphere. This can only be achieved if the subunits exhibit curvature and they are placed such that their centroids lie at a maximum distance from each other. These criteria place two points along the surface of a sphere separated by a distance equal to the diameter of the shell. We assign these positions the north and south poles. As a consequence of this arrangement, there exist two structure types for n = 2. The first belongs to the point group D**h and consists of two identical subunits attached at the equator (Figure 7a). Since it is impossible to create a shell-like hemisphere that possesses oo-fold rotation symmetry using atoms and molecules, this structure is not obtainable from a chemical standpoint. The second belongs to the point group Dnd and its simplest member, n = 2, is topologically equivalent to a tennis ball (Figure 7b). Each subunit of this system is symmetrical and may be divided into four identical asymmetric units (Figure 7c) which implies that eight asymmetric units are required to design the shell. Thus, n = 2 represents the minimum number of subunits which may be used to construct a spherical molecular host via self-assembly.

Spherical Molecular Containers

)

167

)

(c)

Figure 7. n = 2 shells, (a) Doohsymmetry, (b) D2d symmetry (tennis ball), (c) subunit of D2d shell depicting the four asymmetric units. For n = 3, each subunit must cover one-third of the surface of the sphere. Following the design conditions described previously, placing three identical subunits along the surface of a sphere results in an arrangement in which their centroids (Figure 8) constitute the vertices of an equilateral triangle. As a result, there is only one structure type for n = 3 (Figure 8a). The structure belongs to the point group D3h and, as for the n = 2 system, the subunits must exhibit curvature (Figure 8b). Since each "arm" may be divided into four identical asymmetric units (Figure 8c), 12 asymmetric units are required to construct the shell. For n = 4, positioning four points along the surface of a sphere such that they lie a maximum distance from each other places the points at the vertices of a tetrahedron (Figure 9a). This is the first case in which joining the points via line segments gives rise to a closed surface container. The container, a tetrahedron, is comprised of four identical subunits, in the form of equilateral triangles where surface curvature is supplied by edge-sharing of regular polygons rather than the subunits themselves (Figure 9b). Owing to its symmetry, each triangle may be divided into six asymmetric units (Figure 9c), which implies that 24 asymmetric units (4 • 6)

Figure 8. n = 3 shell, (a) equilateral triangle from S, (b) D3h symmetry, (c) subunit of D3h shell depicting the four asymmetric units.

168

LEONARD R. MACGILLIVRAY and JERRY L. ATWOOD

Figure 9.

n = 4 shell, (a) equilateral triangles from 5, (b) Td symmetry, (c) subunit of Td shell depicting the six asymmetric units.

are required to generate the shell. The tetrahedron belongs to the point group Td and its polygons are related by combinations of twofold and threefold rotation axes.

5.4. Platonic Solids The Platonic solids comprise a family of five convex uniform polyhedra which possess cubic symmetry (i.e. 32,432, or 532 symmetry) and are made of the same regular polygons (e.g. equilateral triangle, square) arranged in space such that the vertices, edges, and three coordinate directions of each solid are equivalent (Figure 10, Table 1).12 That there is a finite number of such polyhedra is due to the fact that there exists a limited number of ways in which identical regular polygons may be adjoined to construct a convex comer. Equilateral triangles may be adjoined in three ways while squares and pentagons may be adjoined in only a single manner. Moreover, it is impossible to create a convex corner using regular polygons with six or more sides since the sum of the angles around each vertex would be greater than or equal to 360~ 12These principles give rise to five isometric polyhedra which are achiral and whose polygons are related by combinations of n-fold rotation axes. The Platonic solids include the tetrahedron, which belongs to the point group T0, possesses 32 symmetry, and requires a minimum of 12 asymmetric units: the cube and octahedron, which belong to the point group Oh, possess 432 symmetry, and require a minimum of 24 asymmetric units; and the dodecahedron and icosahedron, which belong to the point group I h, possess 532, and require a minimum of 60 asymmetric units. The number of asymmetric units required to generate each shell doubles if mirror planes are present in these structures.

5.5. Archimedean Solids In addition to the Platonic solids, there exists a family of 13 convex uniform polyhedra known as the Archimedean solids. Each member of this family is made up of at least two different regular polygons and may be derived from at least one Platonic solid through either trunction or the twisting of faces (Figure 11, Table

Spherical Molecular Containers

169

cube

tetrahedron

octahedron

icosahedron

dodecahedron

Figure 10. The five Platonic solids (see also Table 1).

2). 12 In the case of the latter, two chiral members, the snub cube, and the snub dodecahedron are realized. The remaining Archimedean solids are achiral. Like the Platonic solids, the Archimedean solids possess identical vertices, exhibit either 32, 432, or 532 symmetry, and require a minimum of either 12, 24, or 60 asymmetric units, respectively. The Archimedean solids possess a wider variety of polygons than the Platonic solids. These include the equilateral triangle, square, pentagon, hexagon, octagon, and decagon.

Table 1. Platonic Solids

Solid tetrahedron cube octahedron dodecahedron icosahedron

Vertices 4 8 6 20 12

Edges

Face Type

Faces

6 12 12 30 30

triangle square triangle pentagon triangle

4 6 8 20 20

1 70

L E O N A R D R. M A C G I L L I V R A Y and JERRY L. A T W O O D

1

2

3

4

6

7

8

I0

II

12

S

9

13

Figure 11. The 13 A r c h i m e d e a n solids, in order o f increasing n u m b e r o f vertices. Truncated tetrahedron (1), c u b o c t a h e d r o n (2), truncated c u b e (3), t r u n c a t e d octahed r o n (4), r h o m b i c u b o c t a h e d r o n (5), snub c u b e (6), i c o s i d o d e c a h e d r o n (7), r h o m b i truncated cuboctahedron

(8), truncated d o d e c a h e d r o n

(9), t r u n c a t e d i c o s a h e d r o n

(10), r h o m b i c o s i d o d e c a h e d r o n (11 ), snub d o d e c a h e d r o n (12), r h o m b i t r u n c a t e d icosid o d e c a h e d r o n (13)(see also Table 2).

Table 2. Solid

A r c h i m e d e a n Solids

Vertices Edges f(3)a

f(4)

f(5)

w

w

f(6)

f(8)

(1) truncated tetrahedron

12

18

4

(2) cuboctahedron

12

24

8

(3) truncated cube

24

36

(4) truncated octahedron

24

36

(5) rhombicuboctahedron

24

48

8

18

.

(6) snub cube

24

60

32

6

.

(7) icosidodecahedron

30

60

20

w

12

(8) rhombitruncated cuboctaheclron

48

72

~

12

~

(9) truncated dodecahedron

60

90

20

.

(10) truncated icosahedron

60

90

~

--

(11) rhombicosidodecahedron

60

120

20

30

12

w

~

(12) snub dodecahedron

60

150

80

w

12

~

~

120

180

~

30

~

20

~

(13) rhombitruncated icosidodecahedron

8

6

.

--

--

4

~

.

.

8

~

.

.

.

.

.

.

.

~

6

--

~

.

Note:. af(n)where f denotes face and n indicates the number of sides of the face.

6

~

~ 8

. 12

f(lO)

6

. 20

12 w

12

Spherical Molecular Containers

171

5.6. Models for Spheroid Design It is our contention here that the Platonic and Archimedean solids represent the limited ways in which n > 3 identical chemical subunits which correspond to regular polygons may be arranged along the surface of a sphere to approximate S. 26 As a result, these solids may be used as models for the design of spherical molecular hosts. These polyhedra provide combinations of n-fold rotation axes and subunits for self-assembly, in the form of regular polygons, which allow one to determine where chemical subunits of a host should be placed and the bonding arrangements they should adopt. In effect, the Platonic and Archimedean solids simplify the task of constructing a spherical molecular host by facilitating a priori spheroid design, and serving as targets in chemical synthesis. 4,

0

EXAMPLES OF SPHERICAL CONTAINERS FROM THE LABORATORY AND FROM NATURE

The premise that spherical molecular hosts may be constructed according to principles of solid geometry renders both organic and inorganic components viable for their design and permits the individual subunits to be held together by covalent and/or noncovalent bonds. Indeed, a common feature displayed by organic and inorganic chemists is that they must utilize space for assembling atoms into molecular frameworks. To demonstrate the utility of this approach, we will now present selected examples of spherical hosts from the laboratory and Nature. We will begin with the Platonic solids and work our way to the Archimedean polyhedra.

6.1. Platonic Solids As stated, the Platonic solids constitute a family of five convex uniform polyhedra made up of the same regular polygons and possess either 32,432, or 532 symmetry. As a result, the three coordinate directions within each solid are equivalent which makes these polyhedra models for spheroid design.

Tetrahedral Systems (Td, Th, T) The macrotricyclic spherand designed by Lehn et al. was the first tetrahedral host (Figure 12a). 27 The bridgehead nitrogen atoms, located at the comers of the tetrahedron, and ethyleneoxy units, the edges, supply the threefold and twofold rotation axes, respectively. As a result, the spherand is composed of 24 asymmetric units [(N/3)-(CH2CH2)-(O/2)]/2. Notably, this molecule and its tetraprotonated form has been shown to bind an ammonium and a chloride ion, respectively. Schmidtchen et al. have introduced similar tetrahedral cages with edges comprised entirely of methylene bridges, 2s while V6gtle et al. have demonstrated the synthesis of a hollow hydrocarbon called spheriphane (Figure 12b). 29

172

LEONARD R. MACGILLIVRAY and JERRYL. ATWOOD

Figure 12. Tetrahedral hosts with Td symmetry, (a) Lehn's spherand, (b) VOgtle's spheriphane, (c) Saalfrank's metal-based cage.

Saalfrank et al. were the first to introduce metal-based tetrahedral cages by using metal ions as comer units and bridging malonate ligands as edges (Figure 12c). 3~ Owing to a bend in each ligand, these M4L6 cages are adamantane-like. In terms of host-guest behavior, an iron based system has been shown to complex a single ammonium ion. 31 Following the introduction of these metal-based systems, similar tetrahedral cages have emerged. Raymond et al. 32 and Huttner et al., 33 for example, have revealed that linear, rather than bent, bridging ligands may be used to form such cages. In the case of the former, bidentate ligands were used to completely fill the coordination sites around each metal while, in the latter, monodentate ligands along with "stopper" units were employed. These hosts were shown to form complexes with four dimethylformamide molecules and a tetrafluoroborate ion, respectively. 32,33 Huan et al., 34 Birker et al., 35 and BiJrgi et al.36 have described tetrahedral (Figure 13a) shell-like hosts, [H12(VO2)12(C6HsPO3)s]4-', [M(I)sM(II)6{SC(Me)2CH (NH2)CO2}12CI]5- (where M(I)sM6(II) = Cu(I)sNi(II) 6, Ag(I)sNi(II) 6, Ag(I)sPd(II)6), and [ICds(HOCH2CH2S)12] 3§ that possess vanadium and sulfur atoms at the vertices of an icosahedron. These structures also possess phosphorus and metal ions at the centroids of triangular faces which correspond to the corners of a cube. As a result, these shells belong to the point group Th, the point group of a volleyball (Figure 13b).

Octahedral Systems (Oh, (9) We will now illustrate four octahedral hosts related to the Platonic solids. Three are based upon the cube while one possesses features of both a cube and octahedron. The first is a cyclophane-based system reported by Murakami et al. (Figure 14a). 37 The sides of the host consist of tetraaza-[3.3.3.3]paracyclophane units and

Spherical Molecular Containers

173

Figure 13. (a) Tetrahedral host with Th symmetry, X-ray crystal structure of the

[Ag(I)sNi(ll)6 {SC(Me)2CH(NH2)CO2}I 2CI] 5- ion showing the positions of the Ag, Ni, and S atoms, and (b) a volleyball, an object that possesses Th symmetry.

its octaprotonated cation has been shown to bind anionic guests. The molecule possesses 48 asymmetric units of [(N/3)-(CH2)- {(C6H4)/2 }]/2. The second is a polyoxovanadate, [(VO6)(RPO3)8] § (R = tBu, OSiMe3), reported by Zubieta et al. 38 and Thorn et al. 39 which consists of VO 5 pyramids and phosphonate ligands (Figure 14b). The vanadium atoms of the shell are located at the vertices of an octahedron while the phosphorus atoms are located at the corners of a cube, thus displaying the dual relationship of these polyhedra. In both cases, the host has been shown to complex a chloride ion.

Figure 14. Octahedral hosts, (a) Murakami's cyclophane-based cube, (b) X-ray crystal structure of the [(VO6)(RPO3)8]+ (R = tBu, OSiMe3)ion.

1 74

LEONARD R. MACGILLIVRAY and JERRY L. ATWOOD

The third is a gold selenide cube, [NaAu12Se8]3-, reported by Kanatzidis et al. 4~ The anion is made up of eight selenium atoms and 12 gold atoms. A sodium cation occupies the center of the complex. The fourth is a cube synthesized by Chen and SeemanMthe components of which are based upon DNA. 41 The directionality and ability of the double helix to form branched junctions are exploited for the edges and vertices, respectively. Interestingly, each face of the molecule forms a cyclic strand which is catenated with strands of adjacent faces. Molecular modeling experiments indicate the length of each edge to be approximately 6.8 nm.

Icosahedral Systems (Ih, i) Spherical viruses are icosahedral molecular hosts related to the Platonic solids (Figure 15a). 4 Consisting of identical copies of proteins which assemble by way of noncovalent forces, these hosts range from 15 to 90 nm in diameter and encapsulate strands of ribonucleic acid (RNA). Although spherical viruses require a minimum of 60 subunits, most are made up of 60n (n = 1,3, 4...) subunits owing to a reduction in symmetry of their polygons (Figure 15b). This process, known as triangulation, gives rise to quasi-equivalent positions along the surface of the shell which enable the virus particle to cover the RNA with the largest number of subunits. Since only certain triangulations are permitted by symmetry, viruses may be classified into a coherent system. 4 Kretschmer et al. have recently described the ability of 12 CpSmC12 units to form a neutral samarium-based shell, [CP12Sm12(kt3-Cl)24], in which 12 samarium atoms are located at the vertices of an icosahedron and 20 chloride ions are at the vertices

Figure 15. An icosahedral host, (a) X-ray crystal structure ofthe rhinovirus, a spherical virus linked to the common cold, (b) a schematic representation of the rhinovirus displaying triangulation.

Spherical Molecular Containers

175

of a dodecahedron. 42 The remaining chloride ions form a tetrahedron at the center of the shell.

6.2. Archimedean Solids As stated, the Archimedean solids constitute a family of 13 convex uniform polyhedra made up of two or more regular polygons and, like the Platonic solids, possess either 32,432, or 532 symmetry. As a result, the three coordinate directions within each solid are equivalent, making these polyhedra, in addition to the Platonic solids, a model for spheroid design.

Truncated Tetrahedron (1) Fujita et al., 43 Stang et al., 44 and Steel et al. 45 have recently described the synthesis of M6L4 cages which are topologically analogous to a truncated tetrahedron (Figure 16). These systems, which may be regarded as inverted MaL6 frameworks, consist of metal ions and aromatic-based bridging ligands which constitute the twofold and threefold rotation axes, respectively. Notably, owing to the presence of a chiral stopper unit, the system reported by Stang et al. possesses T symmetry and is therefore chiral. 44 In terms of host-guest behavior, X-ray crystallographic studies have revealed the assembly reported by Fujita et al. to form a complex with four adamantyl carboxylate ions, 43 while that of Steel et al. to encapsulate a molecule of dimethylsulfoxide. 45 According to mass spectrometric data, the cage reported by Stang et al. associates with four triflate ions. 44 Zubieta et al. have also demonstrated the formation of a cage topologically equivalent to a truncated tetrahedron, [Mo16(OH)12040]8-. Composed of four

P.d----..~ N...._~

n \

g

...~,~,/Pd

N~.,. N

N

]P'N

o Pd

0

/

[~ /bP'~\ N

P

Figure 16. Fujita's metal-based cage, a host based upon the truncated tetrahedron.

176

LEONARD R. MACGILLIVRAY and JERRYL. ATWOOD

Mo(VI) and 12 Mo(V) centers, this host possesses a central [M012040] 2~ core which encapsulates a proton or sodium ion. 46

Cuboctahedron (2) GonzMez-Duarte et al. have recently described the ability of eight cadium ions and 16 thiolate ligands to assemble from aqueous media to form a highly charged cage, [C1Cd8{ SCH(CH2CH~N(H)Me }16]15§ the sulfur atoms of which sit at the vertices of a cuboctahedron:' The host contains a central chloride ion and an inner tetrahedral array of cadmium ions. Interestingly, Ross et al., using MM2 molecular model simulations, have considered the existence of a molecule containing eight benzene rings and either 12 oxygen or sulfur atoms which they refer to as heterospherophane (Figure 17). 48 Although it is not mentioned in the original report, the shell exhibits a topology identical to a cuboctahedron. A large container that conforms to the structure of a cuboctahedron which is based upon the self-assembly of Cu(II) ions with a triazo ligand has also been described by Robson et al. 49 The shell possesses a cavity with a volume of approximately 816/~3 and is thought to accommodate 5-6 molecules of dimethylformamide.

Truncated Octahedron (4) Seeman et al. have constructed a DNA-based shell analogous to a truncated octahedron. The edges of this molecule, each of which contains two turns of the double helix, contain 1440 nucleotides and the molecular weight of the structure, which is an overall 14-catenane, is 790,000 Daltons. 5~ Interestingly, the design strategy relies on a solid support approach in which a net of squares is ligated to give the polyhedron. It is currently unclear what shape the molecule adopts in various media.

~X

Figure 17. A theoretical organic shell based upon the cuboctahedron (X = O, S).

Spherical Molecular Containers

177

Kretschmer et al. have described the formation of a lanthanide complex, [CPrYbrCll3]- (Cp = cyclopentadienyl), which conforms to a truncated octahedron. 42 The anion contains six ytterbium ions, located at the corners of an octahedron, and 12 bridging chloride ions. A single chloride ion occupies the center of the shell.

Rhombicuboctahedron (5) Mtiller et al. have shown that 24 oxygen atoms of the polyoxometalate [As4Mo6V7039] 2-, may be attributed to the structure of a rhombicuboctahedron. 51 Notably, a strong "tetrahedral distortion" of each ion is required to correspond each host to the polyhedron. This shell has been shown to complex a sulfate ion in the solid state.

Snub Cube (6) We have recently demonstrated the ability of six resorcin[4]arenes and eight water molecules to assemble in apolar media to form a spherical molecular assembly which conforms to a snub cube (Figure 18). 1~ The shell consists of 24 asymmetric unitsmeach resorcin[4]arene lies on a fourfold rotation axis and each H20 molecule on a threefold axismin which the vertices of the square faces of the polyhedron correspond to the comers of the resorcin[4]arenes and the centroids of the eight triangles that adjoin three squares correspond to the water molecules. The assembly, which exhibits an external diameter of 2.4 nm, possesses an internal volume of about 1.4 tl,3 and is held together by 60 O-H...O hydrogen bonds.

Figure 18. Space-filling view of the cavity of 1.

178

LEONARD R. MACGILLIVRAY and JERRY L. ATWOOD

Truncated Icosahedron (10) Buckminsterfullerene, an allotrope of carbon, is topologically equivalent to a truncated icosahedron, an Archimedean solid that possesses 12 pentagons and 20 hexagons (Figure 19).25 Each carbon atom of this fullerene corresponds to a vertex of the polyhedron. As a result, C60 is held together by 90 covalent bonds, the number of edges of the solid.

6.3. Archimedean Duals and Irregular Polygons The Platonic and Archimedean solids comprise two finite families of polyhedra in which each solid consists of identical vertices, edges, and either a single or two or more regular polygons. It is of interest to note, however, that there exists a family of spherical solids which are made up of irregular polygons which may also be used as models for spheroid design. Known as Archimedean duals, 12these polyhedra are constructed by simply connecting the midpoints of the faces of an Archimedean solid. Such a treatment gives rise to 13 polyhedra which possess two or more different vertices and identical irregular polygon faces (Figure 20). As a result, chemical subunits used to construct hosts which conform to these polyhedra cannot be based upon regular polygons. To the best of our knowledge, there is one host which conforms to the structure of an Archimedean dual. Harrison was the first to point out that the quaternary structure of ferritin, a major iron storage protein in animals, bacteria, and plants, corresponds to the structure of a rhombic dodecahedron. 52 This protein, which is approximately 12.5 nm in diameter, consists of 24 identical polypeptide subunits (Figure 21 a), and holds up to 4500 iron atoms in the form of hydrated ferric oxide

Figure 19. X-ray crystal structure of buckminsterfullerene, C60, a shell based upon the truncated icosahedron.

Spherical Molecular Containers

1

179

2

5

4

3

6

7

8

9

10

11

12

13

Figure 20. The 13 Archimedean duals derived from corresponding Archimedean solids (see Figure 9). Triakis tetrahedron (1), rhombic dodecahedron (2), triakis octahedron (3), tetrakis hexahedron (4), deltoidal icositetrahedron (5), pentagonal icositetrahedron (6), rhombic tricontahedron (7), disdyakis dodecahedron (8), triakis icosahedron (9), pentakis dodecahedron (10), deltoidal hexecontahedron (11), pentagonal hexecontahedron (12), disdyakis triacontahedron (13).

Figure 21. X-ray crystal structure of ferritin. A spherical host based upon the rhombic dodecahedron, (a) carbon trace of the polypeptide subunit, (b) the assembly displayed by the subunits. Dark ovals represent single subunits.

180

LEONARD R. MACGILLIVRAY and JERRY L. ATWOOD

with varying amounts of phosphate [Fe203(I'I20/H3PO4)n]. 53 The polypeptides, which consist of four helix bundles, assemble by way of noncovalent forces and form dimers which correspond to the faces of the solid (Figure 21 b).

6.4. Irregular Polygons It is also important to point out that if "partial" truncation is applied to the Platonic solids such that Archimedean solids are not realized, or if truncation is applied to the Archimedean solids, then the resulting polyhedra will not possess regular faces but, like the Archimedean duals, may be used as models for spheroid design as a consequence of their cubic symmetries. Indeed, it is striking to note that of the spherical hosts synthesized to date, all have been constructed using chemical subunits which either correspond to regular polygons (e.g. calix[4]arenes, VO 4 pyramids, tridentate bridging ligands) or form regular polygons (e.g. carbon-based hexagons and pentagons). Moreover, the realization that spherical shells may be constructed using polyhedra with irregular faces, as in the case of ferritin, implies that spherical shells based upon irregular polygons may be rationally designed.

7. WHY THE PLATONIC AND ARCHIMEDEAN SOLIDS? With interests in chemical synthesis moving towards the fabrication of nanometerscale molecular frameworks 6'54-56and the miniaturization of functional microstructures 57 (Scheme 2), it becomes apparent why the Platonic and Archimedean solids are appropriate models for shell design. First, for a given chemical subunit, these solids inherently give rise to larger structures. This may be illustrated by comparing three molecules which adopt the structures of two Platonic solids and an Archimedean solid: cubane, 5s dodecahedrane, 59 and C60.25 In each molecule, a carbon atom is located at the vertex of a polyhedron. Upon traversing the series, however, one observes a gradual increase in the size of these molecules. In fact, the increase is such that C60 is capable of entrapping a guest. 23 Second, we note that these solids incorporate economy of design. This may be illustrated upon considering a sphere (guest) inscribed within either a tetrahedron (Td), cube (Oh), or icosahedron (lh) (host). Upon comparing these hosts, one realizes that the icosahedron facilitates encapsulation of the guest with the largest number of subunits. Moreover, from a chemical perspective, the icosahedron allows the host to complex the guest with the smallest chemical subunits. Indeed, this feature could

covalent synthesis self-assembly

nanoscale architecture 4a miniaturization Scheme 2.

Spherical Molecular Containers

181

eliminate errors in subunit design and shell formation and, in principle, reduce certain "costs" of a given framework. Thus, the Platonic and Archimedean solids not only provide a means for host design, but a way in which to maximize chemical information, allowing the chemist to simplify the structures of complex molecular frameworks and, in effect, engineer host-guest systems.

8. CONCLUSION Using information obtained from X-ray crystallography, we have described the structure of a chiral, spherical molecular assembly held together by 60 hydrogen bonds. 1~The host, which conforms to the structure of a snub cube, self-assembles in apolar media and encapsulates guest species within a cavity that possesses an internal volume of approximately 1.4 nm 3. From this information, general principles for the design of spherical molecular hosts have been developed. 11 These principles rely on the use of convex uniform polyhedra as models for spheroid design. To demonstrate the usefulness of this approach, structural classification of organic, inorganic, and biological hostsm frameworks which can be rationally compared on the basis of symmetry---has revealed an interplay between symmetry, structure, and function. 6~ Indeed, we anticipate that the Platonic and Archimedean solids may be used for the construction of hosts which conform to those solids not yet realized and additional members of each family, where supramolecular synthesis, via selfassembly, will play a major role in their design, ushering in an era of spherical host-guest chemistry.

ACKNOWLEDGMENTS We are grateful for funding from the Natural Sciences and Engineering Research Council of Canada (NSERC) and the National Science Foundation (NSF).

REFERENCES AND NOTES 1. Friedrich,W.; Knipping,P.; von Laue, M. T. E Sitzungsber. Math. Phys. Kl. K. Bayer. Akad. Wiss. Munchen 1912, 303. 2. Pauling,L. Z Kristallogr. 1928,67, 377. 3. Watson,J. D.; Crick, E H. C. Nature 1953, 171,737. 4. Caspar,D. L. D.; Klug, A. Cold Spring Harbor Syrup. Quant Biol. 1962,Vol. 27, p. 1. 5. Lehn,J.-M. Structure and Bonding 1973, 16, 1. 6. Whitesides,G. M.; Mathias, J. P.; Seto, C. T. Science 1991, 254, 1312. 7. Lawrence,D. S.; Jiang, T.; Levett, M. Chem. Rev. 1995, 95, 2229. 8. Stryer,L. In Biochemistry, 3rd ed.; Freeman: New York, 1988. 9. Dunitz, J. D. In Perspectives in Supramolecular Chemistry. The Crystal as a Supramolecular Entity; Desiraju, G. R., Ed.; Wiley:Chichester, 1996. 10. MacGiUivray,L. R.; Atwood,J. L. Nature 1997, 389, 469.

182

LEONARD R. MACGILLIVRAY and JERRY L. ATWOOD

11. 12. 13. 14. 15. 16. 17. 18.

MacGillivray, L. R.; Atwood, J. L. Angew. Chem., Int. Ed. Engl. 1999, 38, 1018. Wenninger, M. J. Polyhedron Models; Cambridge University Press: New York, 1971. MacGillivray, L. R.; Atwood, J. L.J. Am. Chem. Soc. 1997, 119, 6931. MacGiUivray, L. R.; Atwood, J. L. Chem. Commun. 1999, 181. H6gberg, A. G. S. J. Org. Chem. 1980, 102, 6046. Murayama, L.; Aoki, K. Chem. Commun. 1997, 119. Desiraju, G. R. Angew. Chem., Int. Ed. Engl. 1995, 34, 2311. Andretti, G. D.; Ugozzoli, E; Ungaro, R.; Pochini, A. In Inclusion Compounds; Oxford University Press, Oxford, 1991, Vol. 4, pp. 64. MacGillivray, L. R.; Reid, J. L.; Atwood, J. L.; Ripmeester, J. A. Cryst. Eng. In press. Toda, E Top. Curr. Chem. 1988, 149, 211. Corm, M. M.; Rebek, Jr., J. Chem. Rev. 1997, 97, 1647. Cram, D. J. Nature 1992, 356, 29. Buckminsterfullerenes; Billups, W. E.; Ciufolini, M. A., Eds.; VCH: New York, 1993. Philip, D.; Stoddart, J. E Angew. Chem., Int. Ed. Engl. 1996, 35, 1155. Kroto, H. W.; Heath, J. R.; O'Brien, S. C.; Curl, R. E; SmaUey, R. E. Nature 1985, 318, 162. Hosts based upon n = 2 subunits possess dihedral symmetry and their structures may be considered to be based upon tennis balls, rugby balls, prisms, and antiprisms (see ref. 11). Graf, E.; Lehn, J.-M. J. Am. Chem. Soc. 1975, 97, 5022. Schmidtchen, E P.; MUller, G. J. Chem. Soc., Chem. Commun. 1984, 1115. V'6gtle, E; Seel, C.; Windscheif, P.-M. In Comprehensive Supramolecular Chemistry, Cyclophane Hosts: Endoacidic, Endobasic, and Endolipophilic Large Cavities; Lehn, J.-M.; Atwood, J. L.; Davies, J. E. D.; MacNicol, D. D.; V'ogtle, E, Eds.; Pergamon: New York, Vol. 2, 1996. Saalfrank, R. W.; Stark, A.; Peters, K.; von Schnering, H. G. Angew. Chem., Int. Ed. Engl. 1988, 27, 851. Saalfrank, R. W.; Burak, R.; Breit, A.; Stalke, D.; Herbst-lrmer, R.; Daub, J.; Porsch, M.; Bill, E.; Miither, M.; Trautwein, A. X. Angew. Chem., Int. Ed. Engl. 1994, 33, 1621. Beissel, T.; Powers, R. E.; Raymond, K. N. Angew. Chem., Int. Ed. Engl. 1996, 35, 1084. Mann, S.; Huttner, G.; Zsolnai, L.; Heinze, K. Angew. Chem., ~Int. Ed. Engl. 1996, 35, 2808. Huan, G.; Day, V. W.; Jacobson, A. J.; Goshorn, D. P. J. Am. Chem. Soc. 1991, 113, 3188. Birker, P. J. M. W. L.; Freeman, H. C. J. Am. Chem. Soc. 1977, 99, 6890. BUrgi, H. B.; Gehrer, H.; Strickler, P.; Winkler, E K. Helv. Chim. Acta 1976, 59, 2558. Murakami, Y.; Kikuchi, J.; Hirayama, T. Chem. Len. 1987, 161. Salta, J.; Chen, Q.; Chang, Y. D.; Zubieta, J. Angew. Chem., Int. Ed. Engl. 1994, 33, 757. Thorn, D. L.; Harlow, R. L.; Herron, N. Inorg. Chem. 1995, 34, 2629. Huang, S. P.; Kanatzidis, M. G. Angew. Chem., Int. Ed. Engl. 1992, 31,787. Chen, J.; Seeman, N. C. Nature 1991, 350, 631. Kretscher, W. P.; Teuben, J. H.; Troyanov, S. I. Angew. Chem., Int. Ed. 1998, 37, 88. Fujita, M.; Oguro, D.; Miyazawa, M.; Oka, H.; Yamaguchi, K.; Ogura, K. Nature 1995, 378, 469. Stang, P. J., Olenyuk, B., Muddiman, D. C.; Smith, R. D. Organometallics 1997, 16, 3094. Hartshorn, C. M.; Steel, P. J. Chem. Commun. 1997, 541. Khan, M. I.; MUller, A.; Dillinger, S.; B6gge, H.; Chen, Q.; Zubieta, J. Angew. Chem., Int. Ed. Engl. 1993, 32, 1780. Gonz/dez-Duarte, P.; Clegg, W.; Casals, I.; Sola, J.; Rius, J. J. Am. Chem. Soc. 1998, 120, 1260. Ross, R. S.; Pincus, P.; Wudl, E J. Phys. Chem. 1992, 96, 6169. Abrahams, B. E; Egan, S. J.; Robson, R. J. Am. Chem. Soc. 1999, 121, 3535. Zhang, Y.; Seeman, N. C.J. Am. Chem. Soc. 1994, 116, 1661. MUller, A.; Krickemeyer, E.;-Dillinger, S.; Brgge, H.; Stammler, A.J. Chem. Soc., Chem. Commun. 1994, 2539. Smith, J. M. A.; Stansfield, R. E D.; Ford, G. C.; White, J. L.; Harrison, P. M. J. Chem. Edu. 1988, 65, 1083.

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. 46. 47. 48. 49. 50. 51. 52.

Spherical Molecular Containers 53. 54. 55. 56. 57. 58. 59. 60.

183

Trikha, J.; Theil, E. C.; Allewell, N. M. J. Mol. Biol. B, 1995, 949. Fyfe, M. C. T.; Stoddart, J. E Acc. Chem. Res. 1997, 30, 393. Stang, E J.; Olenyuk, B. Acc. Chem. Res. 1997, 30, 502. Ozin, G. A. Acc. Chem. Res. 1997, 30, 17. Wise, K. D.; Najafi, K. Science 1991, 254, 1335. Eaton, E E.; Cole, T. W. Jr. J. Am. Chem. Soc. 1964, 86, 3157. Paquette, L. A.; Ternansky, R. J.; Balogh, D. W.; Kentgen, G. J. Am. Chem. Soc. 1983, 105, 5446. Pennisi, E. Science 1998, 279, 978.

This Page Intentionally Left Blank

SYNTHETIC PEPTIDE RECEPTORS" NONCOVALENT INTERACTIONS INVOLVING

PEPTIDES

Hans-J6rg Schneider, Frank Eblinger, and Mallena Sirish 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Covalent Models for the Observation of Noncovalent Interactions between Peptide Chains . . . . . . . . . . . . . . . ................... 3. Host Compounds for Peptide Complexation in Nonpolar Solvents . . . . . . . 4. Receptors for Peptide Complexation in Aqueous Media . . . . . . . . . . . . References and Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

185 188 194 197 214

1. INTRODUCTION The development of synthetic host compounds for the selective complexation of natural peptides is of considerable interest for possible medical applications as well as for biotechnology, where they could be used for the separation of oligopeptides from natural or industrial sources. In addition, the study of such supramolecular complexes can help to shed light on biologically important binding mechanisms

Advances in Supramolecular Chemistry Volume 6, pages 185-216. Copyright 9 2000 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 0-7623-0557-6

185

186

HANS-JORG SCHNEIDER, FRANK EBLINGER, AND MALLENA SIRISH

and particularly on factors governing protein folding. Peptides represent a particularly intriguing substrate, as reading and transmission of information in these biopolymers relies exclusively on interactions with the different amino acid side chains, 1 in contrast to nucleic acids where hydrogen bonds provide the corresponding codes. Selective recognition of amino acid sequences is further hampered by the lack of conformational order, again in contrast to the conformationally more stable nucleic acids or, for example, to terpenoids. In spite of their additional significance for the development of peptidomimetics, 2 fewer studies are reported in the literature of host compounds for peptides compared to those for amino acids or nucleotides. The studies that are available in the literature were often conducted in aprotic solvents 3'4'5 and only recently in water. 6 The present review tries to summarize what is known in this field, with an emphasis on the underlying noncovalent forces responsible for association between host and guest. To date, there are few reviews concentrating on peptide complexation. 7 In the present review, we have covered the literature through the end of 1998. The textbook picture of peptide association is in the form of 13-sheets (Figure 1). This mode of association for protein structures is well known in the solid state and suggests that hydrogen bonding should provide a straightforward approach to construct artificial receptors. In addition, the antiparalle113-sheet could be stabilized by ion-pairing between the opposite terminal charges. In reality, however, in water and even in nonpolar solvents such associations are not observed between normal oligopeptides, due to the extremely strong competition of protic media to hydrogen bonding, or to geometric mismatch of the terminal ionic groups, s Hydrogen bonds of the amide type are known to yield an association free energy AG of about 5 kJ/mol in chloroform. In a less polar solvent, such as carbon tetrachloride, one typically observes a value of AG = 10 kJ/mol, but even the presence of only 3% methanol in carbon tetrachloride lowers the association constant by a factor of 1000. 9 In line with this, free energy perturbation calculations predict the total

Figure 1. Antiparallel and parallel [3-sheet structures. (Three-dimensional structures of this as well as of structures occurring e.g. in Figures 2, 22, 26, and Scheme 4 can be seen as animated views with the help of CHIME etc. from our website http://www.uni-sb.de/maffak/fb 11~schneider).

Synthetic Peptide Receptors

187

absence of hydrogen bonds between amides in aqueous solution, l~ As a consequence, almost all biomimetic models for such noncovalent interactions between peptide chains are based on the syntheses of structures in which two peptide chains are held together by one or two scaffolds at the ends. Although such models rely on the introduction of covalent bonds in order to allow the analysis of noncovalent interactions with peptides, they provide important insight into the nature of such associations. We will therefore discuss these before we describe true host compounds for the complexation of this important class of biopolymers. The folding of linear peptide sequences to form proteins is essential for their biological activity, reflected by the formation of specific secondary structures. Especially conformational changes of the secondary structure are of much recent interest. A conformational change of a-helical peptide into a [3-sheet seems to be responsible for the deposition of so-called [3-amyloids, which play an essential role in Alzheimer disease. 11 Almost 50 years after the pioneering work of Pauling 12 the formation of 13-sheets and other structural motifs of protein architecture is still not completely predictable. It has been known for more than 30 years that the secondary structure of a protein is determined by the amino acid sequence. Even now, however, the driving forces that lead to a stable folded conformation of disordered polypeptide chains are understood only in general terms and are described largely on the basis of probabilistic or statistical evaluations from known solid-state structures, l In principle, there are two kinds of noncovalent interactions that contribute to the stability of protein structures and may also be used for the design of synthetic receptors. These interactions are of electrostatic origin (mainly of hydrogen bonds, salt bridges, and van der Waals interactions) or hydrophobic origin (traditionally believed to stem from entropic changes). In a review on weak interactions in proteins, Burley and Petsko 13 suggested that packing of amino acid side chains in the hydrophobic core of a protein is established by at least two requirements. These are: (1) the need to exclude water molecules, and (2) the need to both form and optimize a large number of enthalpically favorable electrostatic interactions, including ion pairsnwhich, however, are often entropy driven-, hydrogen-bonding, and weakly polar interactions like oxygen-aromatic, sulfur-aromatic, nitrogenaromatic (the N§ effect), and aromatic-aromatic (stacking or edge-to-face) interactions. Figure 2 represents a 13-sheet as observed, for example, in alkaline phosphatase (Chain A), 14 illustrating how interactions between the same kind of either hydrophobic or hydrophilic in amino acid residues can stabilize a 13-sheet structure. Models from force-field energy minimizations of j3-sheets such as shown in Figure 1 indeed show not too large distances for noncovalent attractive forces between the side chains. These factors are decisive also for the formation of so-called [3-turns, which have been studied thoroughly in proteins, 15 and also recently in synthetic equivalents which will be discussed in the next section.

188

HANS-JORGSCHNEIDER, FRANK EBLINGER, AND MALLENA SIRISH

Figure 2. A 13-sheet region of alkaline phosphatase (Chain A). TM 2. COVALENT MODELS FOR THE OBSERVATION OF NONCOVALENT INTERACTIONS BETWEEN PEPTIDE CHAINS During the past decade several attempts were made to develop and to investigate compounds called "artificial 13-sheets" that should serve as models for 13-sheet structures. These models promise insight into protein folding, and may be helpful for the development of peptidomimetic drugs and of artificial receptors. In almost all cases, lipophilic and/or aromatic amino acid residues were used. As most studies were done in nonaqueous solvents, the authors chose nonpolar residues to enhance the solubility. Hydrophobic interactions must play a minor role in these solvents. Feigel et al. 16 used rigid aromatic spacers as a platform to hold two peptide strands at a distance appropriate to induce the formation of 13-sheets (Figure 3). This was established first by synthesis of an unnatural amino acid (scaffold) and coupling to the tripeptide ile-val-gly (Figure 3).

ox) Z

H

0

Figure 3. A cyclopeptide with antiparalle113-sheet pattern. (After Feigel et ai.16).

Synthetic Peptide Receptors

189

NOE experiments and temperature-dependent chemical shift measurements of NH protons in DMSO-d 6 confirmed the formation of an antiparallel hydrogen bond pattern as found in natural cyclopeptides. In subsequent publications Feigel et al. presented cyclopeptides containing two scaffolds. The synthesis of an artificial, parallel 13-sheet was achieved by coupling of phenoxathiine-4,6-dicarbonic acid to phenylalanine, valine and a suitable diamino-spacer (Figure 4). 16 Low-temperature NMR ROESY analyses and force-field calculations supported the parallel hydrogen bond pattern shown in Figure 4. By use of biphenyl scaffolds Feigel et al. succeeded to synthesize an atropisomeric cyclooctapeptide with antiparallel 13-sheet structure (Figure 5). 16 Kemp et al. prepared an artificial 13-sheet which contains tetracyclic epiindolidione as molecular scaffold (Figure 6). 17 Starting from 2,8-diaminoepiindolidione the synthesis succeeded by coupling to the dipeptide pro-D-ala (to induce a 13-turn), followed by addition of an amino acid isocyanate to the free amino groups of proline residues, and coupling to the terminal amino acid. X-ray and NMR analyses (NOE, coupling constants, and temperature dependence of NH shifts in DMSO-d6) supported the antiparallel 13-sheet structure. Furthermore Kemp and Li described the formation of an antiparallel 13-sheet in various organic solvents by coupling two peptide strands to diphenylacetylene spacers (Figure 7). 18 Nowick et al. described, in a series of papers, artificial 13-sheets that were restricted neither in the length of scaffolds nor in the number of peptide chains. 19'2~ Two complementary scaffolds were developed, one having an oligo-urea scaffold, 21 designed to keep two or more peptide (or peptidomimetic) strands in appropriate proximity to each other. The other scaffold is a 13-strand mimetic, which rigidities the oligourea scaffold and largely avoids intermolecular aggregation. A synthetic 13-sheet consisting of a diurea scaffold and two peptide strands was also reported (Figure 8).22

0

I:t

]l ~

:ht

R - kFropyl

2b

R- Be~

-%

........//

\

Figure 4. A cyclopeptide with parallel 13-sheet pattern. (After Feigel et a1.16).

190

HANS-J(~RG SCHNEIDER, FRANK EBLINGER, AND MALLENA SIRISH H

~11o,,?

0

R

H

0

~--'~ .. "'-o T

i

""o o,,1

3a R= i-Propyl 3b R- Beuzyl

Figure 5. An atropisomeric cyclooctapeptide with antiparallel [3-sheet structure. (After Feigel et a1.16).

By elongation of the urea scaffold Nowick et al. were able to obtain a triplestranded [3-sheet combining parallel and antiparallel hydrogen bond pattern (Figure 9). 23,24 Downfield shifts of NH signals in CDC13 indicate that these protons are hydrogen bonded. Crosspeaks from NOESY measurements show clear evidence for the expected parallel and antiparallel hydrogen bond patterns.

o=t'o ,' A, ! ,o'

,•

N'H

I

H

I

z

0

H

N

0

I I 0

~

I I H

Rz

1

R1

I I 0

|

0

H.N

~_...,~ v

H

Figure 6. An artificial antiparallel [3-sheet. (After Kemp et al. 17).

Synthetic Pepticle Receptors

191

!

O

O

H

~ I

R1

H

H

0

Figure 7. Diacetylene [3-sheet. [After Kemp et al. (antiparallel)18].

Gellman et al. established some general rules for the formation of [3-hairpin structures, which contain two antiparallel peptide strands and a short connecting loop. The loop that induces the hairpin formation can consist of a dipeptide (e.g. D-pro-ala), 25 or of a trans-alkene 26 (Figure 10, in CD2C12). The folding properties of several types of synthetic oligomers having unnatural backbones were explored. The most thoroughly characterized type of oligomers are 13-peptides.27 All the covalently bound artificial [3-sheets described above essentially rely on hydrogen bonds and were therefore studied in non-aqueous solvents. Kelly and coworkers published a series of papers in which artificial 13-sheets are described that adopt this folded conformation in aqueous solution. 28 Similar to the approach taken by Feigel et al., Kelly and coworkers used aromatic scaffolds such as

CN

? T_.r

~ o--"

. Ph

H

O

A

i

I

H

Figure 8. First [3-sheet by Nowick et al. (parallel). 22

192

HANS-JORG SCHNEIDER, FRANK EBLINGER, AND MALLENA SIRISH

I

CN

o

S~'~

~

o

~N I

H I

(CH~)2 N (

~2

)3

Ni

Ph

I~

~

/

~ ~N'~'~f

I N

//

;

tt

H

0

I O ~-'Ph

//X ~N""~~ I H

O

,

H

0S

"~ J L~ N /

",

H

II 0

/

.~~

~

CH3

t

-

O

,,

,,H

Jt~ N

i ~~,

/ I H

c.

'

Figure 9. A three-stranded 13-sheet. [After Nowick et al. (parallel and antiparallel)23].

dibenzofuranyl-, biphenyl- and bipyridinyl-units to secure proximity of connected peptide strands (Figure 11). These water-soluble [3-sheets are acyclic compounds that are stabilized mainly by hydrophobic interactions. By use of different spacers and peptide strands, Kelly et al. synthesized a series of compounds that were characterized by NMR and CD spectroscopy. The CD

O I~I

Me2

Figure 10. Antiparallel ~-sheet by Gellman et al. 26

Synthetic Peptide Receptors

193 O

R4

H

O

N

I

N

,

I

H

O

I~$

H I

O

H

~ It"

O

.td,,,AX ~J. o

R6

'I,

'I,

O

,I

.,.,

Figure 11. Acyclic 13-sheet by Kelly et al. (antiparallel). 28

spectra showed the coexistence of random coil and ~-sheet structure. NOE measurements revealed e.g. vicinity of R3-1eu and Ra-val residues (Figure 11), which indicates the existence of a hydrophobic cavity. Substitution of these hydrophobic amino acids by the less hydrophobic alanine led exclusively to random coil structures. These studies demonstrated the essential role of hydrophobic interactions between side-chain residues for the formation of secondary structures in peptides. Ogawa et al. reported on the formation of a parallel I}-sheet in aqueous solution by coupling two val-val dipeptides to a Ru(bipy) 2§ complex (Figure 12). 29 One prerequisite for the nucleation of a 13-sheet here would be the formation of an unfavorable cis-amide bond (see Figure 12). From the energetic point of view this is not very likely, as already pointed out by Nowickl9C; the published data also do not support the claimed conformation.

o

<,

,_N-----;...~,,....... N,

,,

V

!Zc

A,,!"[",

'~''

Figure 12. Artificial 13-sheetcontaining a cis amide bond proposed by Ogawa et al. 29

194 0

HANS-JIDRGSCHNEIDER, FRANK EBLINGER, AND MALLENA SIRISH

HOST C O M P O U N D S FOR PEPTIDE C O M P L E X A T I O N N O N P O L A R SOLVENTS

IN

All of the artificial [3-sheets described above were invariably formed between covalently bound peptide strands within the formal sense of intramolecular hydrogen bonds. In line with our earlier analysis of energetic contributions of hydrogen bonds in such amide-type structures, 9 we found it desirable to measure the energetics between separate peptide strands. At the same time, our aim was to develop a biomimetic receptor that could distinguish enantiomers on the basis of hydrogen bonds between peptide strands. Computer aided molecular modeling suggested two peptide strands coupled in the para positions to a biphenyl ether spacer (or biphenylamine, or -methane) as a suitable host. Such an entity could bind a single-strand peptide in the fashion of an antiparallel (triple) 13-sheet, with for example two hydrogen bonds at each side of the guest molecule (see Figure 13).30 NMR titration experiments in CDCI 3 revealed a clear preference of the L-host compound for the incorporated L-isomeric guest. The chiral discrimination with AAG = 7 kJ/mol is in the range of the few enantioselective peptide receptors hitherto available. 31 Molecular mechanics calculations (gas phase, e = 3) with CHARMm 32

O

II

CH=

I

H

I

"---CH="C--N--CH HI|1 -!---- N-----CH=~, O H H II / CH='? --C-'-CHI -N"~'C\\ H CH,Ph p ',

o

O

/

.

II I / CH="C--"N---CH- C--- N---CH= I I II H CH= O

Figure 13. Antiparalle113-sheet structure of host I (L) and guest 2 (D- or L-isomer). 3~

Synthetic Peptide Receptors

195

shed light on the origin of the observed stereoselectivity: only with the L-isomer one does obtain, after energy minimization, a structure with the four hydrogen bonds as depicted in Scheme 1. Minimizations with the D-isomer invariably lead to structures without hydrogen bonds: unfavorable interactions of the guest (2D) and host benzyl groups lead to a deformation of the backbone unable to form hydrogen bonds. As is often the case, the observed stereoselectivity of association is the consequence of repulsions of groups that are not involved in the formation of noncovalent bonds, and are remote from the actual binding sites. Still and Liu 3 reported the enantioselective complexation of amino acids and small peptide derivatives by a bowl-shaped C 3 macrotricyclic receptor in chloroform (Figure 14). In all cases the receptor showed the largest hitherto reported chiral discrimination, with a preference for the L-configuration of the substrates (with e.g. 99% ee and a AAG ~ 13 kJ/mol for Boc-val-NHMe). The observed free energies of binding were up to AG ~ 30 kJ/mol in the case of MeO2C-ser-OtBu. Furthermore, Still et al. reported a very simple synthesis of a polycyclic receptor, which again shows high enantioselectivities for Boc-protected and N-acetylated peptides: 33The condensation of Boc-protected (R,R)- 1.2-diaminocyclohexane with pentafluorophenyl esters of trimesic acid resulted in a cyclo-oligomer in 39% yield (Figure 15). Even from the reaction of the commercially available free diamine with trimesic acid trichloride one could obtain the desired product in a one-pot reaction in 13% yield. Measured binding energies show an increasing preference for the L-configuration with an increase in the size of amino acid residues, for example, from AAG = 7.9 for gly, to 14.6 for ala and to 21.0 for val (in kJ/mol). A possible explanation could be that the side chains contribute to binding due to van der Waals interactions with the phenyl tings of the host and/or that the solvent molecules (CHC13) find insufficient space adjacent, for example, to valine residues inside the remaining cavity.

6'~Y

Y

-o

""

i'" .j

Figure 14.

o

) '."COG' "

A macrotricyclic receptor designed by Still et al. 3

196

HANS-JORGSCHNEIDER, FRANK EBLINGER, AND MALLENA SIRISH H

I

_o/

o

A~ ~N~~N"~ A

\oo..

N

',,,

'NHCO

" I f

0,,,

N

H

tl

A....,N

,,

CONH""

,,,l H

A

B

Figure 15. A polycyclic receptor designed by Still et al. 33 Kilburn et al. 4 described the synthesis and peptide binding properties of some macrocyclic compounds containing a diamidopyridine unit as a carboxylic acid binding site (see Figure 16). Complexation studies in CDC13 solution with N-protected (Z or Boc) amino acids or dipeptides were characterized by free association energies of up to 23 kJ/mol, but rather small enantioselectivity (max. AAG - 4 kJ/mol).

~

N

,W'

,H

yo o o

Figure 16. A macrocY4clic receptor incorporating a 2,6-diamidopyridine unit designed by Kilburn et al.

197

Synthetic Peptide Receptors

HN~,~~N~HN o' ' N N H O

N'H O

Figure 17. A peptidocalixarene reported by Casnati et al. s A very interesting vancomycin type of antimicrobial activity of peptidocalixarenes was found by Casnati et al. 5 These compounds (Figure 17) bind to the cell wall of mucopeptide precursors bearing the terminal sequence D-ala-D-ala. Q

RECEPTORS FOR PEPTIDE COMPLEXATION IN AQUEOUS MEDIA

Almost all of the host compounds described to date rely on the use of nonpolar solvents and actually require modification of the peptide, typically by the use of protecting groups that make them soluble in such media. Still 34 et al. have described highly sensitive and sequence-selective complexation of peptides in water; however, the association occurs on the surface of polymers, on which dyes 34a or lipophilic macrocycles are immobilized, and not with peptides in solution; in addition the peptides were acetylated. Figure 18 illustrates a possible binding mode for such an association between the dyes and tripeptide, which needs to be protected by an acetyl group. The visible color change of the polymer beads upon binding of the dye lends itself to an elegant application for screening of combinatorial peptide libraries. The screening of combinatorial libraries of peptides with synthetic receptors is usually done with biological test sets such as cell lines. More insight into the molecular basis of the underlying interactions is possible by binding studies with a defined receptor structure. One can also attach a different dye to the substrates which are to be used to screen a library of host compounds. For this, 5Met-enkenphalin was colored red by covalently attaching the dye Disperse-Red-l, and SLeuenkephalin was colored blue using another dye. Statistical combinatorial methods were then used to generate a library of hexapeptide-containing hosts, using cheno(12-deoxy)-cholic acid or an open chained 1,5-disubstituted analogue as

198

HANS-JORGSCHNEIDER, FRANK EBLINGER, AND MALLENA SIRISH O

H

,~,,

A

4

~ rro

H

"11" II O

O

I

~,..~ / PEG-PS

H,NCOCH;"

0

I

I H

H

"

'i'H 7ro I Y - ' i ' H

H N-

o

Y

=

~.,,~ ~'..,,~ / N .

j HO

0 O oI,

0

I

/N.

Y ~

HO

=:

:

H

-

-N" I H

0

I

"11"

"If

II ~ I

O

/

HzNCOCHz,/

-N H

S

Figure 18. Possible binding mode of the dye rhodamine to peptides immobilized on polyethylene glycol-polystyrene beads; the peptides bear acetoxymethyl (3, 4) or acyl glycine (5) end groups. 34a

supporting scaffold (Scheme 1). If one immobilizes the scaffolds on polymer beads, one can obtain about 10,000 different peptide sequences as hosts. The polymer beads turn either red or blue only if one of the colored substrates is bound, upon which these few beads are selected and the corresponding peptide sequence is identified (e.g. by HPLC). This approach, although only giving yes-or-no answers on the affinities, does provide qualitative insight into peptide-peptide interactions, which can also be varied by the spacing between the host peptide chains. A tripeptide library with 29 different amino acids was also prepared, with a corresponding maximum of 293 = 24,389 different peptides. As each peptide is grown on a different polymer bead, treatment of the beads with a solution of a colored nickel salenocyclodextrin (Scheme 1) gives an interesting preselection: only a few beads become colored. Gas chromatographic analysis after hydrolyzing these polymer-bound peptides showed that these beads always contained the dipeptide sequence L-phe-D-pro or D-phe-L-pro. The association of these peptides with the underlying ]3-cyclodextrin host compound in homogenous aqueous solutions was studied by NMR-shift titration, and established moderately large constants around 120 to 180 M-I. It was shown that several commercially available dyes bind with high selectivity in water to the immobilized and acetylated peptides. 34a When aqueous solutions of the dyes were equilibrated with the encoded combinatorial library of--45,000 acylated tripeptides containing 14 different acyl substituents as end groups and 15 different L- and D-amino acids, immobilized on hydrophilic, poly(ethyleneglycol)polystyrene (PEG-PS) beads for 24 h, only -0.1% of the beads turned intensely colored as a result of the staining of the dyes. These stained beads were analyzed

Synthetic Peptide Receptors

199

/=N\

C

Caa

NH AA I

NH

AA z

AA:

Ac

A~ Ac

o

AA1 /Okz

A.A=

Ac Ac

Scheme 1. Host structures used for screening affinities of combinatorial libraries. 32

after selection using electron capture gas chromatography (ECGC) to obtain the structures of the bound peptides (Table 1). As seen in the table, a peptide possessing the acetoxymethyl (AcOMe)CO-serine sequence has high selectivity towards a class of structurally similar dyes. While rhodamine methyl ester binds strongly to peptides having AcOMe- or acetylglycine as the end group, phenol red shows preferential selectivity for peptides substituted with asn followed by ser or pro. Similar studies performed on deprotected peptide library indicate less selective association compared to the side-chain-protected peptide library. Also, certain acid-bearing dyes show a strong preference for binding peptides having basic lysine residues.

HANS-JORG SCHNEIDER, FRANK EBLINGER,AND MALLENA SIRISH

200

Table 1. Side-Chain-ProtectedPeptides Binding Dye Molecules in Wated Dye

R

AA3

AA2

AA 1

Rhodamine Rhodamine Methyl Ester Azure A Crystal Violet Safranine O Phenol Red

X AcOM (100%)

asn (45%) ser (85%)

Asn (100%) X

Asn (90%) X

AcOM (100%) AcOM (85%) AcOM (80%) X

ser or asn (90%) ser or asn (95%) ser or asn (90%) asn (80%)

X X X L-X

gin (65%) ala or gin (65%) ala or gin (80%) D-pro (50%)

Phenol Red

X

asn (80%)

D-X

L-pro (30%)

Note: aHere X: no selectivity observed; R: end group; AA3: amino acid positioned 3 in a tripeptide; AA2: amino acid positioned 2 in a tripeptide; AAI: amino acid positioned 1 in a tripeptide.

In an extension, the same group has synthesized a water-soluble peptide receptor 6 containing a covalently bound rhodamine B for monitoring libraries of immobilized tripeptides (Figure 19). 34b Similar binding studies performed on 10-5 M solutions of 6 in pH 4 water produced again ~0.1% of intensely scarlet beads. After the analysis, it was found that receptor 6 binds only 75 and 30 members of the protected and deprotected peptide libraries containing 24,000 tripeptides each. Among the protected peptide library, 6 has high selectivity for (D)leu or (D)val followed by (L)gln or gly residues (Table 2). On the other hand, receptor 6 displays preferential selectivity for (D)leu in the second position flanked by more acidic asp or glu residues on either side in the deprotected library. In a recent study Breslow et al. 35 have demonstrated a pronounced chelate effect contributing to the stability of a receptor-peptide complex. They synthesized a series of receptors (7-11), based on dimeric 13-cyclodextrins connected either at primary or secondary faces via various linkers (Figure 20). The two cyclodextrin cavities can bind two hydrophobic side chains of a peptide, with a cooperative chelate effect. These receptors have been tested for binding to a variety of peptides--for example by using titration microcalorimetry (adding the peptide solutions to a solution of the cyclodextrin host in 0.2 M pH 9.0 NaHCO3/Na2CO 3 buffer in water at 25 ~ Table 3). In Table 3, K 1 is the binding constant for the formation of a 1:1 complex between host and peptide, and K2 is the constant for binding of a second molecule of peptide to the 1:1 complex. It is apparent from this table that receptors 7, 8, and 9 show cooperative binding to peptides a and d having two L-phe-D-pro sequences spaced by asp, whereas no strong binding is seen with peptides such as trp-trp, b or c. The strongest chelate effect (K 1 >> K 2) is seen in the case of receptor 7 and a linear peptide d; as the authors point out, this is due to the flexibility of d, which fits better into the two cavities of 7 than does the rigid a, corroborating results obtained from molecular modeling studies. The high association constant observed in the case of receptor 11 with short disulfide linker and a

201

Synthetic Peptide Receptors

.I.

H= N

II

"

o

~ll'NH--CO--Rhodanfine Rhodamine B-'--OC--HN

NH

H=

O

§

LH !

NH(

Figure 19. A peptide receptor 6 with attached dye for screening of peptide libraries which are immobilized on polymer beads, in analogy to Figure 18. dipeptide trp-trp indicate the selective recognition of this receptor for trp-trp (Table 3). On the other hand, linkage at the narrow primary faces of a cyclodextrin in the dimer 8 leads to weak binding towards the cyclic peptide a. Hamilton et al. 36 have shown that a tetraguanidinium-based receptor binds strongly to a peptide having four aspartate residues and stabilizes it in an tx-helical conformation. The receptor 12 binds to a peptide with four aspartate residues, 13, with an association constant of 1 to 3 x 105 M -1 in 10% H20/90% CH3OH, and stabilizes it in an tx-helical conformation, as revealed by CD spectroscopy and NMR studies (Figure 21). 36a Molecular modeling studies carried out on this complex 12.13 showed that the receptor 12 could wrap around a right-handed or-helix conformation of peptide 13 in a coiled, left-handed helix with matching being achieved between each guanidinium moiety and the corresponding carboxylate groups of the four aspartate residues. In a similar earlier study, Hamilton et al. have

202

HANS-J(3RG SCHNEIDER, FRANK EBLINGER, AND MALLENA SIRISH Table 2. Sequencesof Peptides Selectivity Bound by

Receptor 6 in Water at pH 4a'b

Entry

1 2 3 4 5

AA3

(D)leu, (D)val (L)gln X (D)gln, (o)glu, (D)asp (D,L)asp

AA2

(L)gln (o)leu, (o)val (k)gln, gly (D)leu (t),L)asp

AA1

X X (o)leu (L)asp, (L)glu (D,L)asp

Notes: aAs in Table 1. bEntries 1-3 with protected peptide library; entries 4, 5 are with deprotected peptide library.

shown a sequence-selective recognition of a pair of aspartate residues by a rigid bis(guanidinium) receptor. 36b It should be noted, that these receptors require the presence of charged amino acid residues in the peptide, which can undergo ion pairing or hydrogen bonding between the charged host and guest functions. The design of artificial receptors that are length and sequence selective and operate in water and with all kinds of natural, unprotected peptides has been achieved only recently. 37 The receptors discussed above are mostly symmetric, and therefore are unable to distinguish between the N- and the C- terminus of peptide, which limits their use for sequence selective recognition. Our strategy for sequenceand length-selective host compounds is based on the use of crown ethers as the binding element for the peptide N-terminus, and of an ammonium group for association with the C-terminus. The presence of alkyl residues at the receptor ammonium group prevents self-association of the host, and allows the preparation of either hydrophilic or lipophilic receptors, depending on the length of the chosen alkyl group (Scheme 2). With ligand 14 one observes, as expected, no selectivity with respect to a specific amino acid, but a distinct preference for the binding of tripeptides (Table 4), in accordance with computer aided molecular modeling of the complexes (Fig. 22). Although methanol is also a very polar solvent the binding constants in this medium are up to K = 10 4 M -1. Introduction of a lipophilic side group in the form of a dimethylaminonaphthalenesulfonate (dansyl, 15, Scheme 2) provides for detection by fluorescence (Figure 23), and leads for the first time to sequence selective complexation with recognition of the particular amino acid by lipophilic side-chain interactions (Table 5). The fluorescence increase is due to the occupation of the crown ether-oxygen electron pairs by complexation with the peptide N-terminus; it shows small variations as a function of the interacting amino acid. The computer-generated model of the complex (Figure 22b) illustrates such an interaction between aromatic groups of host and guest, which can be either of the edge-to-face or the (displaced) face-toface, or stacking type.

Synthetic Peptide Receptors

203

-xx

=

I

---- O z C ' ~ ' ~ c

) , (

=0;-~

Hz 10

8

X-

X-

- - - ' ~ 0 - - -

O O"

"---O'--~O"-11

D-Pro/

Hz

Y-

As ~"~'Phe

y:

Ac-Trp-Gly-Trp-OH b

-.--C ----S----S Hz

/AI% D-AI.

"NHz

C--

Hz

Phe

Ac-#~p-Phe-D-Pro-Asp.Phe-D-Pro-Gly-Gly-N Hz d

Figure 20. Structures of dimeric ~-cyclodextrin-based receptors and peptide sequences a-e. (After Breslow et ai.35).

Implementation of a crown ether unit in water-soluble pyridinium-substituted porphyrins leads to new peptide receptors of hitherto unknown sensitivities. 38 The Soret bands in these complexes provide for a conveniently measurable and highly sensitive optical signal, which can be of importance for the development of new sensors. A series of water-soluble porphyrins were synthesized in order to test their

204

HANS-JORG SCHNEIDER, FRANK EBLINGER, A N D MALLENA SIRISH

Table 3. Host

Binding Constants of 13-Cyclodextrin-Based Receptors 7-11 with Peptides a

Peptide

K1 (M-1) b

K2 (M-1) c

7 7 7

a b c

2600 130 91

1120 117 85

7

d

1100

114

7 7

e trp-trp

930 84

452 93

8 8

a

675

97

trp-trp

8

87

86

phe-phe

12 3

49

9

a

590

1O0

10

a

98

105

10 l]-cyc lodextr i n

trp-trp a

96 220

98 300

11

trp-trp

1200

Notes: aln 0.2 M, pH 9.0 NaHCO3/Na2CO3 buffer at 25 ~ ref 35. t~onstant for formation of 1:1 complex between host and peptide. CConstantfor binding of a second molecule of peptide to the 1:1 complex.

HO

"'er .S

,,,,,/S H

H

H

Cl"

H Ci"

--

2

12

DDDD(]3)

Ac-A-A-A-D-Q-L-D-A-L-D-A-O-D-A-A-Y-NH 2

Figure 21. Structures of tetraguanadinium-based receptor 12 and tetraaspartate peptide 13.

Synthetic Peptide Receptors

205

-9

~o....q,.o,-C.

-

'0o

C'o"~

o

o

-

-

c.,

..-c.~-c~;

?"'.,

14 RI=H 15

R,=S02--~ ~/~_..~--NlCH3)2

Scheme 2. Structures of the crown ether-based peptide receptors. 37

affinities towards certain protected as well as unprotected peptides (Figure 24). Interestingly, UV titrations performed on the receptors 16 and 17 (as receptors 18 and 19 aggregate severely in water) showed that they indeed bind to N-carbobenzyloxy(Z)- protected peptides with association constants shown in Table 6. As can be seen from this table, the log K values for the protected peptides vary between 2.9 and 3.7; with higher constants for aromatic amino acids. However, these receptors did not show any affinity towards unprotected peptides, possibly due to the repulsion between the positively charged pyridinium group and the peptide N-terminus. In contrast, when the crown ether-bound porphyrin receptor 20 was subjected to UV titrations as above, 20 showed with unprotected peptides binding constants (Table 7), which are higher than those observed with protected peptides and of 16 or 17. Figure 25 shows the titration curves fitted to a 1:1 calculational model. At concentrations higher than a molar ratio of [peptide]/[20] = 200, one observes a biphasic behavior of the isotherm, indicating other binding modes with likely two peptides in the complex. Computer-aided molecular modeling of the complex between 20 and gly-gly-phe using energy minimization with the CHARMm force

206

HANS-J(~RG SCHNEIDER, FRANK EBLINGER, AND MALLENA SIRISH Table 4. Binding Constants of 14 with Peptides a

Entry

Pepti de

1 2 3 4 5 6 7 8 9 10 11 12

gly-gly-gly gly-gly-phe gly-phe-gly phe-gly-gly leu-leu-leu gly-gly gly-leu leu-gly gly-phe gly-gly-gly-gly gly-gly-glyOMeH + gly-gly-gly gly-gly-phe gly-phe-gly

13

14

Solventb

W W W W W W W W W W W M M M

K/M-1 c

200 170 180 135 155 50 45 40 40 45 30 13,000 8,500 10,500

_~ G-d

13.1 12.7 12.9 12.1 12.5 9.5 9.4 9.1 9.1 9.4 8.2 23.5 22.4 23.0

Notes: aRef. 37. Measured by NMR titration of 14 with peptides at 25 ~ [peptide] 0 = 10 mM (in D20), or 1.0 mM (in CD3OD); titrations were performed with eight measurements under neutral conditions (pH = 7.0 + 0.2), error limit in K<_ 5%, in AG < + 0.5 [kJ mole-l].

bW, water; M, methanol. CValues were obtained from the average of 2 to 4 different NMR signals of peptides. UAGin [kJ]mole-:.

Figure 22. Force-field-optimized structures of the complexes of receptor 14 with triglycine (a), and of receptor 15 with gly-phe-gly. (b). Hydrogen atoms are omitted for clarity. In each case the N-terminus of the peptide is anchored to the crown ether by hydrogen bonding; the corresponding C-terminus forms an ion pair with the ammonium group. Additional x-stacking of the side chains is observed in the case of complex (b}.

Synthetic Peptide Receptors

207

1400 1200

_>, "~ 1000 @1

"E

8OO

9 GPG

g

200 0

1

2

3

o

GGG

n

GGF

"

GWG

4

5

[peptide], 10 4 x moVI

Figure 23. Fluorimetric titration curves of receptor 15 w i t h tripeptides. Excitation w a v e length = 3 2 0 nrn. (GFG = g l y - p h e - g l y ; G G G = g l y - g l y - g l y ; G G F = g l y - g l y - p h e ; G W G = gly-trp-gly). 37

Table 5. B i n d i n g Constants of Host 15 w i t h Tripeptides in W a t e r a Entry

Peptide

K,//VI-1

-~ G d

irnax/lOb 3.5

1

gly-gly-gly

210

13.2

2

gly-gly-gly

190

13.0

c

3

gly-gly-phe

215

13.3

3.2

4

phe-gly-gly

220

13.6

2.8

5

gly-phe-gly

1700

18.4

1.8

6

gly-phe~

1620 c

18.3

c

7

trp-gly-gly

260

13.8

3.0

8

gly-gly-trp

9

gly~

310

14.3

2.8

2150

19.0

2.3 3.4

10

gly-leu-gly

850

16.7

11

leu-leu-leu

700

16.3

3.3

12

gly-ala-gly

540

15.6

3.1

13

val-val-gly

720

16.3

3.3

Notes: aRef. 37. Measured by fluorescent titration of peptides with 15 at 25 ~

[15] 0 = 10-s M, Xex = 320 nm; peptide concentration from 4 x 10-3 M to 2.5 x 10-4 M. Titrations with usually 10 to 12 measurements; pH = 7.0 + 0.2, error limit in K < + 5%, in AG< + 0.5 [kJ mole-l]. blo initial fluorescence intensity, Imax is maximum (intrinsic) intensity at 100% complexation calculated from nonlinear least square fit. CMeasured by NMR titration. dAG in [kJ mole-l].

208

HANS-J(DRG SCHNEIDER, FRANK EBLINGER, A N D MALLENA SIRISH

93

R

R4

R1

16

R 1 =I~=R 3 =R 4=

~-O~-c.. Cl"

17

Rz = R 2 = R 3 =

~-O~-c., Cl"

18

Rz = R 2 =

-<~-~

CHs

R4 = "~~COOCH3

R3 = R4 =

~~~-COOCH3

CI"

19

Rz=

~~/I~I"--CH 3

R2=R3=R4=

~~'--COOCH 3

Cl"

C~ ~

__~o~o-2 O

20 R,R,=R,= ~--C2,~-C". CI"

Figure 24.

R,=

O

Structures of the water-soluble porphyrin-based peptide receptors. 38

field shows a best fit between crown and +NH 3 at one end, and between an adjacent 4-pyridinium unit and the C O 0 - groups at the other end, as well as stacking between the terminal phenyl ring and the porphyrin (Figure 26). The two carboxylate oxygens of gly-gly-phe are in van der Waals contact with the neighboring pyridyl

Synthetic Peptide Receptors

209

Table 6. Logarithm of Association Constants (M-1) of 16 and 17 with (Z)-Amino Acids and (Z)-Peptidesa'b

Amino Entry

acid/Peptide

16

17

IogK

As x 10 -s

IogK

gg. x 10-3

1

Z-ala

3.06

56.2

2.94

31.4

2

Z-phe

2.91

75.0

3.29

44.8 47.5

3

Z-tyr

2.93

80.4

3.39

4

Z-trp

3.43

94.9

3.58

49.2

5

Z-asp

3.73

63.9

3.69

48.2

6

Z-ile

3.40

64.7

2.94

35.4

7

Z-gly-phe

3.17

44.8

3.38

50.6

8

Z-ala-trp

3.20

92.4

3.60

69.6

Notes: aRef. 38. Measured by UV-vis titration of 16 and 17 with (Z)-amino acids/(Z)-peptides at 25 ~

Titrations were carried out in 5 mM phosphate buffer, pH 6.9 + 0.2 by adding concentrated stock solutions of (Z)-amino acid/(Z)-peptide [(Z)-amino acid]/[(Z)-peptide] = 10 mM) containing ca. 5 pM of 16 or 17 to an equally concentrated solution of 16 or 17 in a 10 mm Cuevette. Error limits: log K, + 5%. bAe, extinction coefficient changes at [ligand]][host] = 200, the values agree within +5% with the Ae from nonlinear fit [e(410 nm) : 16, 149,300; 17, 128,800, Error limits: + 5%].

ortho-Hs (PheCO(1) ....... HC(Pyridyl-ortho) = 2.52/~) and the ~-pyrrole carbon atom (PheCO(2) ....... C(~-pyrrole) 3.03/~) and secure sufficient ion pairing. The data in Table 7 illustrate that the binding strength increases as a function of the peptide length and with the number of aromatic units in the amino acid side chains, as expected from stacking contributions. The extremely large association constants seen in the case of tetraglycine are obviously due to additional binding effects. Preliminary investigations with some simple aliphatic amides revealed that the amide function itself contributes significantly to the association. The complexes with the porphyrin-based receptors shed light on possible interaction mechanisms in protein-porphyrin aggregation and are of particular interest as they also may lead to supramolecular catalysts for peptide modification. Metallation of the porphyrin centers, for example with zinc(II) ions, led in some cases to small affinity increases. 39 Quite recently we synthesized truly biomimetic receptors which contain, as spacers between crown ether and ammonium units, peptides instead of unnatural spacers as in 14 and 15. 4o Host compounds such as 21 not only discriminate natural peptides by length and sequence, but are capable, in principle, of chiral recognition. They allow one to study in various media the energetics of side-chain interactions in I]-sheet-like structures in a systematic way, exemplifying both the practical and the theoretically interesting aspects in the development of artificial peptide receptors.

210

HANS-JORG SCHNEIDER, FRANK EBLINGER, AND MALLENA SIRISH

Table 7. Logarithm of Association Constants (M -1) of 20 with Peptides a'b Entry

Peptide

Log K

Ar x 10-3

1 2 3 4 5 6 7 8 9 10 11

gly-gly gly-phe phe-gly ala-phe asp-phe phe-phe gly-gly-gly gly-gly-phe gly-phe-gly phe-gly-gly gly-gly-trp trp-gly-gly gly-gly-gly-gly

2.93 4.71 4.36 4.60 4.05 4.52 3.41 4.39 4.35 4.48 3.52 4.48 5.02

15.7 22.0 18.0 10.2 14.0 13.1 7.8 20.7 12.5 12.3 15.4 16.9 18.9

12

13

Notes: aSee footnotes to Table 6. b~ (410 rim); 20, 91,200; error limits: • 5%.

0,42

E

9 PG

I

0,40

PGG 9 GGGG

tO

9 TGG

I--

0,38 o c

t2

0,36

0

< 0,34

0,32

1

9

. 0

.

4

.

.

8

_=

12

16

[peptide], 10 4 x mol/I

Figure 25. UV-visible titration curves of porphyrin 20 with peptides. (PG = phe-gly; PGG = phe-gly-gly; GGGG = gly-gly-gly-gly; TGG = trp-gly-gly). 38

Synthetic Peptide Receptors

211

Figure 26. A force-field (CHARMm) optimized structure of porphyrin 20 with the peptide gly-gly-phe. Hydrogens are omitted for clarity except those of peptide NH3. The phenyl group of phe is not far from vdW contacts to the pyrrole (day = 3.95 ,~,); the-NH3 protons have distances from 1.90 to 2.05/I, to the crown ether oxygen; the carboxylate group comes close to neighboring pyridyl protons with day = 3.0 ,i,.38

The data (Table 8 and Scheme 3) illustrate again the possible length selectivity, with a maximum affinity for (gly) 3 with receptor 21, and for (gly) 4 with receptor 22, in accordance with computer-aided molecular modeling (see also our website). Host 21 shows a pronounced sequence selectivity for aromatic amino acids in position 1 of the tripeptides; which as in the case of receptor 15 systematically decreases for aliphatic amino acids with in the order val>leu>ile>ala>gly. Notably, proline in position 1 leads to particular affinity drop, likely due to geometric mismatch. A preference for aromatic amino acids in position 1 is observed also with host 22, more pronounced with trp than with phe. In methanol, large constants of nearly 104 M -l are found, but the sequence selectivity disappears almost entirely (see Table 8); this highlights the importance of hydrophobic side-chain interactions that are also visible in computer simulated structures (see website). Until recently, cyclodextrins (CD's) were used mostly for complexation of amino acids, with partially contradictive results. 41 In some cases unsubstituted cyclodextrins were used; fluorescence titrations with tyrosine containing pentapeptides showed with o~-cyclodextrin even a lower association constant (around K = 25 M -1)

212

HANS-JORG SCHNEIDER, FRANK EBLINGER, AND MALLENA SIRISH Table 8. Binding Constants of Receptor 21 with Peptides a

Entry

Peptide b

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

gly-gly gly-gly-gly gly-gly-gly-gly phe-gl y-gly gly-phe-gly gly-gly-phe val-gly-gly leu-gly-gly ile-gly-gly ala-gly-gly pro-gly-gly gly-gly-gly phe-gly-gly gly-phe-gly gly-gly-phe

Solvent c

K/'M -Id

w w w W W W W W W W W M M M M

6o 110 50 710 260 170 140 135 140 101 17 8700 8100 7100 7450

Notes: aRef. 40. bat pD = 7.0 + 0.1. cW: water; M: methanol. dError in K < 10%.

21

22

Scheme 3. Structures of ligands 21 and 22 with peptides as host units. 4~ Some selected binding constants for the complexes between ligand 22 and peptides: (I) gly-gly: Ka = 40; (2) gly-gly-gly: Ka = 65; (3) gly-gly-gly-gly: Ka = 159; (4) phe-gly-gly: Ka = 670, gly-phe-gly: Ka = 280, trp-gly-gly: Ka = 710.

Synthetic Peptide Receptors

213

23

24

25

Scheme 4. Structures of 6-deoxy-6-aminocylodextrin peptide receptors 23-25. 43

than with tyrosine alone, whereas [3-CD yielded constants around K = 50 M -1 for tyrosine, and for the peptide K = 125 or 224 M -1, depending on the sequence. 42 The application of CD's for encapsulation of peptides with chemically modified cyclodextrins has been explored only recently. Besides the previously mentioned studies by Breslow et al. 35 there are now systematic evaluations with several cyclodextrins bearing positively charged aminoalkyl groups that could ion pair or hydrogen bond to the C-terminus of peptides. 43 The results (Table 9) show that hydrophobic interactions with lipophilic side chains, particularly phenyl substituents at the amino acids are necessary for any appreciable complexation. Host 23 shows moderate sequence selectivity, with a preference for phe at the 3 position of a tripeptide. NMR analyses of complexes with Ac-gly-phe confirm the phenyl ring is inside the cyclodextrin cavity. With receptor 24, bearing seven charges at the narrow rim, one observes slightly higher constants, in particular if the positive charge at the N-terminus of the peptide guest is blocked by acetylation. Interestingly, the affinity of phe-gly-gly and gly-gly-phe

Table 9. Association Constants (Kass) of CD-Receptors 23-25 with Different Diand Tripeptides a

K~. [M-11

Peptide

23

24

25

phe-gly-gly

13

20

56

gly-phe-gly

23

36

20

gly-gly-phe

50

27

72 206

phe-gly

22

25

gly-phe

49

35

78

Ac-gly-phe

80

53

680

Note: aRef.43. In water pH = 7.0, [NaCI] = 1 x 10.2 T= 298 K, measured by fluorescence competition with

CD and 1,8-ANS as reference system.

214

HANS-JORG SCHNEIDER, FRANK EBLINGER, AND MALLENA SIRISH

to the host 24 does not differ as much as with host 23: this is in line with NOE- and shift data from NMR analyses, which show that with gly-gly-phe and 23 there is preferential inclusion from the narrow side. In contrast, NMR experiments on 23 and 25 with Ac-gly-phe show two inclusion modes, with the phenyl group entering both from the narrow and the wide side of the cyclodextrin cavity. 44 The benzylamino residue in 25 is immersed into the CD cavity prior to complexation of the phenyl ring, as evident from NMR analyses. Complexation with a peptide then requires that this intramolecular association must be given up, which explains why the affinity is not increased by possible lipophilic interactions of aromatic peptide side chains and the benzyl group outside the cavity. The conformational switching triggered by complexation with a peptide was followed by suitable NMR experiments, and is supported by computer-aided molecular modeling (see our website).

REFERENCES AND NOTES 1. Schulz, G, E.; Schirmer, R. H. Principles of Protein Structure; Springer: New York, Heidelberg, Berlin, 1979; Zvelebii, M. J. J. M.; Thornton, J. M. Q. Rev. Biophys. 1993, 26, 333. 2. Giannis, A.; Kolter, T. Angew. Chem., Int. Ed. Engl. 1993, 32, 1244, and references cited therein. 3. Liu, R.; Still, W. C. Tetrahedron Lett. 1993, 34, 2573. 4. Dowden, J.; Edwards, P. D.; Kilburn, J. D. Tetrahedron Lett. 1997, 38, 1095; Waymark, C. P.; Kilburn, J. D.; Gillies, I. ibid. 1995, 36, 3051; Hack, S. S.; Kilburn, J. D. ibid. 1995, 36, 3409. 5. Casnati, A.; Fabbi, M.; Pelizzi, N.; Pochini, A.; Sansone, E; Ungaro, R. Bioorg.Med. Chem. Lett. 1996, 6, 2699. 6. See Section 4. 7. Schneider, H.-J. Angew. Chem., Int. Ed. Engl. 1993, 32, 848. 8. Association constants of up to 107 M-1 in water were claimed for complexes between a nonapeptide and some pentapeptides (Sasaki, S.; Takagi, M.; Tanaka, Y.; Maeda, M. Tetrahedron Lett. 1996, 37, 85). The fitted UV curves described in this communication, measured at rather short wavele.ngth, show an unusual behavior. In our hands attempted UV measurements of oligopeptide associations in water were hampered by nonlinear effects even with high-quality spectrometers of small stray light. Similarly, NMR studies in our laboratory with amino acids such as phenylalanine as well as with peptides such as gly-phe-gly, gly-tyr-ala, and tetragylcine show no evidence for any association at concentrations up to 5"10 -2 M. 9. Schneider, H.-J.; Juneja, R. K.; Simova, S. Chem. Ber. 1989, 112, 1211, see also: Sartorius, J.; Schneider, H.-J. Chemistry- Eur. J. 1996, 2, 1446; Schneider, H.-J. Chem. Soc. Rev. 1994, 22, 227. 10. Jorgensen, Acc. Chem. Res. 1989, 22, 184. 11. (a) Inouye, H.; Kirschner, D.A.J. MoL Biol. 1997, 268(2), 375-389; (b) Cotman, C. W.; Tenner, A. J.; Cummings, B. J. Neurobiol. Aging 1996, 17(5), 723; (c) Soto-Jara, C.; Baumann, M. H.; Frangione, B. Patent: US 96-630645 960410. 12. Pauling, L.; Corey, R. B.; Branson, H. R. Proc. Natl. Acad. Sci. USA 1951 37, 205. 13. Burley, S. K.; Petsko, G. A. Adv. Prot. Chem. 1988 39, 125-189. 14. From the Brookhaven Database; alkl.pdb; Kim, E. E.; Wyckhoff, H. W. J. Mol. Biol. 1991, 218, 449. 15. Ramakrishnan, C.; Srinivasan, N.; Nataraj, D. V. Int. J. Pept. Protein Res. 1996, 48, 420; Mattos, C.; Petsko, G. A.; Karplus, M J. Mol. Biol. 1994, 238, 733; Richadson, J. S. Adv. Protein Chem. 1981, 34, 167; Rose, G. D.; Young, W. B.; Gierasch, L. M.; Smith, J. A. ibid. 1985, 37, 1;

Synthetic Peptide Receptors

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.

215

Gunasekaran, K.; Gomathi, L.; Ramakrishnan, C.; Chandrasekhar, J.; Balaram, P. J.Mol. Biol. 1998, 284, 1505, and references cited therein. (a) Gailer, C.; Feigel, M. J. Comput.-Aided Mol. Des. 1997, 11(3), 273-277 and therein cited references; (b) Feigel, M. J. Am. Chem. Soc. 1986, 108, 181; (c) Brandmeier, V.; Sauer, W. H. B.; Feigel, M. Helv. Chim. Acta 1994, 77, 70. (a) Kemp, D. S.; Bowen, B. R. Tetrahedron Lett. 1988, 29, 5077 and 5081; (b) Kemp, D. S.; Bowen, B. R.; Muendel, C. C. J. Org. Chem. 1990, 55, 4650. Kemp, D. S.; Li, Z. Q. Tetrahedron Lett. 1995, 36, 4175 and 4179. (a) Nowick, J. S.; lnsaf, S. J. Am. Chem. Soc. 1997, 119(45), 10903-10908; (b) Smith, E. M.; Holmes, D. L.; Shaka, A. J.; Nowick, J. S. J. Org. Chem. 1997, 62, 7906-7907; (c) Nowick, J. S.; Lee, I. Q.; Mackin, G.; Pairish, M.; Shaka, A. J.; Smith, E. M.; Ziller, J. W. NATO ASI Ser., Set C 478; Mol. Des. Bioorg. Cataly. 1996, 111-136; (d) Holmes, D. L.; Smith, E. M.; Nowick, J. S. J. Am. Chem. Soc. 1997, 119, 7665-7669; (e) Nowick, J. S.; Pairish, M.; Lee, I. Q.; Holmes, D. L.; Ziller, J. W. J. Am. Chem. Soc., 1997, 119, 5413-5424; (f) Nowick, J. S.; Mahrus, S.; Smith, E. M.; Ziller, J. W. J. Am. Chem. Soc. 1996, 118, 1066. Soth, M. J.; Nowick, J. S. J. Org. Chem. 1999, 64, 276-281. Wilson, M. E.; Nowick, J. S. Tetrahedron Lett. 1998, 39, 6613-6616. Nowick, J. S.; Smith, E. M.; Noronha, G. J. Org. Chem. 1995, 60, 7386. Nowick, J. S.; Smith, E. M.; Pairish, M. Acc. Chem. Res. 1996 401. See also Soth, Michael J.; Nowick, James S. J. Org. Chem. 1999, 64, 276. Stanger, E. H.; Gellman, S. H. J. Am. Chem. Soc. 1998, 120, 4236. Gardner, R. R.; Liang, G.-B.; Gellman, S. H. J. Am. Chem. Soc. 1995, 117, 3280. (a) Gellman, S. H. Acc. Chem. Res. 1998, 31, 173 and references cited therein; (b) Guichard, G.; Seebach, D. Chimia 1997, 51,315 and references cited therein. (a) Nesloney, C. L.; Kelly, J. W. J. Org. Chem. 1996, 61, 3127-3137; (b) Nesloney, C. L.; Kelly, J. W. J. Am. Chem. Soc. 1996, 118, 5836-5845; (c) Schneider, J. P.; Kelly, J. W. Chem. Rev. 1995, 95, 2169; (d) Diaz, H.; Espina, J. R.; Kelly, J. W. J. Am. Chem` Soc. 1992,114, 8316 and references therein cited. Gretchikhine, A. B.; Ogawa, M. Y. J. Am. Chem. Soc. 1996, 118, 1543. Eblinger, E; Schneider, H.-J. Chem. Comm. 1998, 2297. (a) Yoon, S. S.; Still, C. W. Tetrahedron Lett. 1994, 35, 8557; (b) Yoon, S. S.; Still, C. W. Tetr. Lett. 1994, 35, 2117; (c) Yoon, S. S.; Still, C. W. J. Am. Chem. Soc. 1993, 115, 823; (d) For a short review see: Schneider, H.-J.Angew. Chem., Int. Ed. 1993, 32, 848; Angew. Chem. 1993, 105, 890. (a) Brooks, B. R.; Bruccoleri, R. E.; Olafson, B. D.; States, D. J.; Swaminathan, S.; Karplus, M. J. Comp. Chem. 1983, 4, 187. (b) Brooks, C. L.; Karplus, M. Methods Enzymol. 1986, 127, 369; Brtinger, A. T.; Karplus, M. Acc. Chem. Res. 1991, 24, 54. Yoon, S. S.; Still, W. C. J. Am. Chem. Soc. 1993, 115, 823. (a) Wennermers, H.; Still, W. C. Tetrahedron Lett. 1994, 35, 6413; (b) Torneiro, M.; Still, W. C. J. Am. Chem. Soc. 1995, 117, 5887. Breslow, R.; Yang, Z.; Ching, R.; Tzrojandt, G.; Odobel, E J. Am. Chem. Soc. 1998, 120, 3536. (a) Peczuh, M. W.; Hamilton, A. D.; Sanchez-Qesada, J.; de Mendoza, J.; Haack, T.; Giralt, E. ibid. 1997,119, 9327; (b) Albert, J. S.; Goodman, M. S.; Hamilton, A. D.J. Am. Chem. Soc. 1995, 117, 1143. Hossain, M. A.; Schneier, H.-J. J. Am. Chem. Soc. 1998, 120, 11208. Sirish, M.; Schneider, H.-J. Chem. Comm. 1999, 907. Sirish, M.; Schneider, H.-J. Unpublished results. Yu, L.; Schneider, H.-J. Manuscript in preparation. Review: Connors, K. A. Chem. Rev. 1997, 97, 1325; Comprehensive Supramolecular Chemistry, Atwood, J. L.; Davies, J. E. D.; Macnicol, D. D.; Vtigtle, E; Szejtli, J.; Osa, T., Eds.; Pergamon/Elsevier: Oxford, Tokyo 1996, Vol. 3, and references cited therein; for a large data

216

HANS-JORG SCHNEIDER, FRANK EBLINGER,AND MALLENA SIRISH

compilation see K. A. Connors J. Pharr~ Sci 1995, 84, 843; available also from our website at http ://www. uni- sb. de/ma~ak/Jb 11/sctmeider/Links/download.html. 42. Bekkos, E. J.; Gardella, J. A.; Bright, E V. J. Incl. Pheno~ Mol. Recogn. 1996, 26, 185. 43. Hacket, E; Schneider, H.-J. Manuscript in preparation; Hacker, E Ph.D. Dissertation, Universit~t des Saarlandes. In preparation. 44. Simova, S.; Hacker, E; Schneider, H.-J. Manuscript in preparation.

RATIONAL DESIGN OF SYNTHETIC ENZYMES AN D TH El R POTENTIAL UTILITY AS H U M A N PHARMACEUTICALS"

DEVELOPMENT OF MANGANESE(II)-BASED SUPEROXIDE DISMUTASE MIMICS

Dennis P. Riley

1. 2. 3.

4. 5.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Initial Structure-Activity Studies: Catalyst/Drug Design . . . . . . . . . . . . Development of Improved Sod Mimics Derived from I . . . . . . . . . . . . 3.1. Stability Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. In Vivo Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mechanistic Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Computer-Aided Design (CAD) . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References and Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Advances in Supramolecular Chemistry Volume 6, pages 217-244. Copyright 9 2000 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 0-7623-0557-6 217

218 219 222 227 231 232 233 242 242

218

DENNIS P. RILEY 1.

INTRODUCTION

Low-molecular-weight catalysts which mimic a natural enzymic function (synzymes) have potential utility for the treatment of diseases characterized by the overproduction of a potentially deleterious metabolic by-product or foreign gene product. The discovery and development of pentaaza macrocyclic ligand complexes of manganese(II) as functional mimics of superoxide dismutase (SOD) enzymes and their potential utility as human pharmaceutical agents is described. Many disease states afflicting mankind can be broadly characterized as ones in which the body fails to adequately contain the overproduction of an undesired metabolic by-product. An example of this inability to control and limit the concentration of a potentially harmful agent occurs as a result of our need to metabolize energy stores. While all mammalian life consumes oxygen as the ultimate oxidant supporting cellular respiration, a considerable portion of the oxygen is metabolized through its one-electron reduction product, superoxide anion. Under normal circumstances in healthy individuals this radical burden is contained by the complement of superoxide dismutase (SOD) enzymes present in our cells (mitochondria, Mn-based) and in the plasma (Cu/Zn-based) and the extracellular spaces (Cu/Znbased). This control of the free-radical flux derived from oxygen is jeopardized in many circumstances in which superoxide (SO) anion production is excessive. This overproduction of SO can overwhelm the body's ability to catalytically dismute superoxide and eliminate the radical burden and lead to a variety of superoxideinitiated disease states. For example, reperfusion diseases such as those following acute myocardial infarct and stroke, or inflammatory processes involved in diseases such as arthritis, are diseases associated with the overproduction of superoxide anion. We have been investigating for several years the development of synthetic enzymes (synzymes) as an approach to managing the types of diseases outlined in the previous discussion. This review describes the discovery and subsequent development of highly stable Mn(II) complexes which possess high SOD activity. Key to success in designing and synthesizing highly stable and highly active SOD mimics was the development of a detailed understanding of the mechanism of action of this class of SOD mimics, and the subsequent development of a computeraided design (CAD) paradigm based on molecular mechanics (MM). A detailed overview of this work is presented here. Superoxide dismutases are a class of oxido-reductase enzymes which contain either Cu, Fe, or Mn at the active site and catalyze the dismutation of superoxide (1), the one-electron reduction product of molecular oxygen (Eqs. 1 and 2, where M n is the metalloenzyme in the reduced state and M n+l is the enzyme in the oxidized state) to oxygen and hydrogen peroxide. 1 The SOD enzymes have been shown to have efficacy in animal models of disease states proposed to be, in part, mediated by superoxide, such as myocardial ischemia-reperfusion injury,2-4 inflammation, 4-7 and cerebral ischemia-reperfusion injury. 8-1~Evidence for superoxide as a mediator

Developmentof Manganese(l#Complexes

219

of disease states continues to accrue, such as in Parkinson's disease and neuronal apoptosis,l 1-15 cancer,16-20 and AIDS. 21'22 O~

0 2 + M "+1 ~ 0 2 + M"

n+

(1) (2)

HO~ + M " - - > H202 + M"+l Due to the limited therapeutic applications of the SOD enzymes arising from a number of negatives associated with their use, such as their (1) lack of oral activity, (2) inability to gain access to the intracellular space of cells where superoxide is produced, (3) immunogenicity when derived from nonhuman sources, (4) bellshaped dose response curves, (5) short half-lives, and (6) cost, we have pursued the concept of designing synthetic, low-molecular-weight mimetics of the SOD enzymes which could overcome such limitations, and hence could serve as pharmaceutical candidates affording a new and promising approach to the treatment of disease. @

INITIAL STRUCTURE-ACTIVITY STUDIES: CATALYST/DRUG DESIGN

Our initial efforts focused on the synthesis of manganese-based complexes as low-molecular-weight SOD mimics. This decision to pursue Mn complexes was based largely on considerations of toxicity. Of the three metals known (Fe, Mn, Cu) to catalyze SO to hydrogen peroxide and oxygen, manganese is the least toxic to mammalian systems as the free aquated metal ion and is also the least likely of the three M 2§ ions to react with hydrogen peroxide to generate hydroxyl radicals (Fenton chemistry). In our early design efforts we focused on the synthesis and screening of complexes of Mn(II) which could possess high chemical and thermodynamic stability. The reason for this is obvious, but needed in pharmaceuticals is a stable complex which will not deposit a free redox-active metal ion in a biological compartment where it may interact with the local biochemistry in unnatural ways. Such a situation is clearly capable of leading to toxicity problems. Consequently, we focused much of our synthetic efforts on complexes of Mn(II) with macrocyclic ligands so that we could gain needed stability by incorporating the enhanced kinetic stability observed with such cyclic ligands. From our initial structure-activity relationship (SAR) studies, we have discovered that a class of manganese(II)-based complexes, I, incorporating the macrocyclic ligand 1,4,7,10,13-pentaazacyclopentadecane (where R can be any number of defined substituents with known stereochemistry), effectively catalyze the dismutation of superoxide. 23

220

DENNIS P. RILEY R R R R RH ~ H a R ' ~ ' - - ' Z " N / ~, ~N " ~ R

H'N

"~" I R R

N.-H a R

R

The discovery of this class of Mn(II) complexes as functional mimics of the SOD enzymes resulted from synthesis and screening studies utilizing a large number of different macrocyclic ligand complexes of Mn(II). 23 Even though over 40 complexes with different macrocylic ligands containing Mn(II) were synthesized, only the Mn(II) complex of the unsubstituted ligand derived from I (the 1,4,7,10,13-pentaazacyclopentadecane ligand), [Mn([15]aneNs)Cl2], afforded a measurable catalytic activity as monitored by stopped-flow kinetic analysis of superoxide decay. 35 This specificity of superoxide dismutase activity was somewhat surprising; e.g. the Mn(II) complex with the analogous 16-membered ([16]aneNs) ligand was 2 orders of magnitude less active and simple "changes to the ligand such as replacing one NH with an ether oxygen eliminated all activity. This complex, [Mn([15]aneNs)Cl2], is an excellent catalyst for the dismutation of superoxide to oxygen and hydrogen peroxide, possessing a kea t of 4.1 x 10+7 M -1 s -1 at pH = 7.4. 23 The complex crystallizes in either a seven-coordinate geometry, as exemplified by the trans-dichloro derivative with a planar macrocyclic ligand conformation or as a six-coordinate complex, as in the case of a nitrato nitrate complex which possesses a folded ligand geometry and a six-coordinate pseudo octahedral geometry. 23'36 In each case the complexes are high-spin d 5 white materials possessing an average Mn(II) to N bond distance of 2.28/~. All of the Mn(II) complexes derived from I show a remarkable reaction specificity that is likely due to the high oxidation potential [> 0.7 v (SHE)] that is exhibited by this class of Mn(II) complexes. While the complexes react efficiently with superoxide, they are unreactive to other relevant biological oxidants at 37 ~ such as hydrogen peroxide, nitric oxide and peroxynitrite. This lack of reactivity with hydrogen peroxide is especially important since this indicates that the complexes will not exacerbate hydrogen peroxide toxicity via the promotion of the Fenton reaction-producing hydroxyl radical. In addition, the lack of oxidation with other biological small molecule oxidants means that these complexes can be valuable as tools to probe biological mechanisms with high specificity.

Developmentof Manganesefll)Complexes

221

H/ ci\H HN~cI"~N\

-) NH

The Mn([ 15]aneNs)C12 complex possesses reasonable thermodynamic stability at pH = 7.4 (log K = 10.7) and excellent kinetic stability with the complex completely intact at pH = 7.4 with no metal dissociation for up to days even in the presence of EDTA, 23 which possesses a higher binding affinity for Mn(II) than does the [ 15]aneN 5 macrocycle. Additionally, this complex possesses excellent oxidative stability with Mn(II) to Mn(III) oxidation occurring at +0.77 v (SHE). 23 The observed stability of this complex was thus adequate to assess this synthetic SOD mimic in a variety of in vitro and in vivo models of superoxide-mediated injury. Most notably this complex exhibits efficacy in in vitro 24 and in vivo models of inflammation, 25 myocardial ischemia-reperfusion injury, 26-29 and vascular relaxation and restenosis. 3~ In addition, this and other complexes of this class of SOD mimetics derived from I have superior properties to the SOD enzymes in regard to their normal dose-response curve (no deleterious effects observed at high doses in animal models), cellular permeability (dependent on the nature of the R groups), extended in vivo stability, nonimmunogenicity, and projected lower cost. With positive biological concept data in hand, it was concluded that such synzymes may prove useful as therapeutic agents for the treatment of inflammation and myocardial ischemia-reperfusion injury, and other diseases mediated, in part, by overproduction of superoxide. 32'33 As an added potential benefit these SOD mimics also have the unique property of potentiating levels of nitric oxide, a vasorelaxant and antithrombotic, 31'32 by reducing or eliminating the diffusioncontrolled reaction between superoxide anion and nitric oxide to give peroxynitrite, another toxic metabolite derived from oxygen. 34 This nitric oxide reaction with superoxide has been measured to proceed with a nearly diffusion controlled rate, but we have observed that the SOD mimics, I, limit the formation of peroxynitrite in vivo and hence potentiate the lifetime of nitric oxide in vivo. This provided the basis for testing the compound in an animal model of platelet-mediated (NO blocks platelet aggregation) thrombosis in injured and stenotic arteries. The complex was shown to be protective in that restenosis of the damaged vessel was inhibited by the agent when administered over a 2-week period. 31 Subsequent studies reveal that the complex is effective in blocking reperfusion injury in an isolated rabbit heart model, 27 and in vivo in a canine model of cardiac reperfusion injury. 28

222

DENNIS P. RILEY 0

DEVELOPMENT OF I M P R O V E D SOD M I M I C S DERIVED FROM I

Having confirmed that the synthetic low molecular weight functional SOD mimic, [Mn([ 15]aneNs)C12], possesses biological activity ofpharmacological relevance to human disease states, provided the basis for pursuing this class of molecules for the development of a pharmaceutical agent. There are two key features necessary for improving the [Mn([15]aneNs)Cl2] complex for use as a human pharmaceutical agent: (1) increase the chemical stability via increasing the preorganizational rigidity of the macrocyclic ligand and thereby increasing the inherent kinetic stability of the complex, 37 and (2) increase the SOD activity, so that with a mechanism-based mode of drug action increased catalytic activity would translate into a lower dose of agent and hence a diminished exposure to the metal-based drug. In our attempts to develop more active and more stable complexes we probed the role that substituents (both on the N and C atoms of the parent macrocyclic ring) would exert on both the catalytic SOD activity and the overall chemical stability of the resultant complexes. Those structural factors that would affect these two key parameters are not immediately obvious since it was not known at the outset how derivatized ligand systems would affect catalytic activity. Thus, the number of substitutents, their placement, and their stereochemistry could all be critical design elements for maximizing catalytic activity and chemical stability. At the outset we found that the more readily synthesized nitrogen-substituted complexes possess no catalytic activity. Thus, we focused on the synthesis of C-substituted derivatives, I. Unfortunately, standard procedures for the synthesis of aza crown macrocyclic ligands employing Atkins, Richman 38 chemistry (Scheme 1) is inadequate with these C-substituted pentaaza crown ligands as both the cyclization yields are poor, owing to higher oligomer formation, and poor yields from the acid hydrolysis/detosylation, owing to stabilized carbonium ion induced ring cleavage/elimination. While this chemistry is sufficient for the unsubstituted and even monosubstituted macrocycles, the yields for polysubstituted ligands were unworkably low. Consequently, we set out to devise better synthetic methods to such C-substituted ligands. New routes were successfully developed which utilized, in addition to toluenesulfonamides, amide moieties as protecting groups for the amines during the cyclization. 4~ Such methods included acid chloride cyclization methodology (Scheme 2), bis-chloroacetamide methodology (Scheme 3), and cyclic peptide cyclization techniques (Scheme 4) which made it possible to run high concentration reactions and achieve high yields of macrocycle (> 50%, in most cases). 38-4~ Additionally, combinations of these methods could be successfully incorporated into a single ligand preparation. 4~ Employing such new synthetic methodologies made it possible to synthesize highly substituted and stereochemically defined ligands, and hence made it possible to probe the effect that substituents exert on the stability and catalytic activity. For these studies we employed both methyl and fused-cycloalkyl substituents on the

223

Developmentof Manganese(ll)Complexes -...

//O

1. LiAIH4,THF, reflux 2. TsCI, NaOH, H20

H2N NH2 oHCl Na+

R = H

R=Ts

O

Na+ tTs

Ts.

RO " ~ N

o,,U.o 1. k...J 2. TsCI, Et3N

Ts TsO~

-,,.. /--k R--NH HN-R

..OTs

/

Ts~ N

N.Ts

N/(OR Ts R=H R = OTs

CH2012

k

..

;

+

Ts-'NH NN-"Ts

Ts TS,, ~ cN

/Ts N-"~

Ts-~./NU-Ts Ts

9HCI

H,, ~ HBr, Phenol=

NaH DMF 100 ~

w

/H

cN NTH H"~N.v.J " H Overall Yield=4%

Richman-Atkinstype synthesisof the monomethyl-substituted ligand 16 accomplished in an overall yield of-4%. Scheme 1.

macrocyclic ring carbons as our tools to elicit the effect of stereochemical variations on stability and rate. Examples of ligands employed are shown in Figure 1. All of the resultant Mn(II) complexes were characterized and shown to be similar to the parent unsubstituted complex; e.g., high-spin d 5 Mn(II) trans-dichloro complexes with Elt2(SHE ) in the range of +0.74-0.78 v.4~Both the thermodynamic and kinetic stabilities of the resultant complexes were assessed, 4~ and one very significant feature emerged. Namely, that increasing the number of C-substituents always increased the thermodynamic and kinetic stabilities of the complexes and in a nonlinear manner; i.e. stability increased geometrically with the number of substituents so that, for example, the Mn(II) complex of the monomethyl ligand 16 was about twice as stable as the unsubstituted complex; the Mn(II) complex of the pentamethyl ligand 10 was over 160 times more stable than the unsubstituted

224

DENNIS P. RILEY Na+

OT

R

Na+

RyO

1. CICH2CO2Me Ts,~N,~ . ~ .JN~'Ts ... = ~,;~/ 2. NaOH/H20/THF Ts 3. SOCI2 R = OMe R=OH

Ts

R =CI

R R_2 _R3 r~ 1/s'.fl'~~ TM H2N NH2

+

OTCI

CI~ O

Tsar.

Et3N

N ~ N.. I Ts

CH2CI2

Ts R_2 R3 R1/,.~~..-- R4

Ts-L.-T' Ts

Scheme 2.

R2 R3 R 1 / ~ R4 DME

H~N

N-. H H

Generic acid chloride method of synthesis of ligands 11-15.

complex. 4~ Additionally, it was observed that the trans-cyclohexano complex, [Mn(trans-cyclohexano[15]aneNs)C12], derived from ligand 11 (see Table 1 for

representative examples of complexes with kcat and k~iss values), was more than twice as active an SOD catalyst (kcat = 9.1 • 10 +7 M -1 s -1 at pH = 7.4) 42 and possessed both an improved thermodynamic stability (log K of 11.6) and kinetic stability (2• over the unsubstituted complex. An additional important result was that the stereochemistry of the methyl substituents has a big effect on the catalytic rate, but little effect on the stability of

225

Development of Manganese(ll) Complexes

R2 _93

R1/~R4

H2N NH2 .2HCI

_ R2 R3 _ CICH2COCI, K2CO3= H1/']" ~.~H4 H20/CHCI3 C,__~N~HN~cI

R1/,.r R2 R3_ .---~H4

O ~ _ N~H N~,O CI'--' ~CI

Na+ + Ts~N

Ts

Rl/,]R2 -93 .-~.~R4

O~N/" HN Scheme 3.

and 7.

Na+ N~Ts

Rly~R4 Ov-~'T s

LiAIH4= DME

H. /---~ ,H NH~ ,. N~-~ ~H

Ts H Genericbis(chloroacetamide)methodof synthesisof ligandssuchas I

the complex as long as the number of substituents remains constant. A most remarkable aspect of the effect that substituents exert on the catalytic rate is revealed with the complexes of the macrocycles containing two trans-fused cyclohexano groups differing in the stereochemistry of their substitution. 42 Such a change can

NOR5 ~ R6 NH..~O HN-R~o'TOH DPPA,Et3N DMF -20 to 0 ~ HCI 9 O R1 R2 R8 R7

H2N'~'~ H~

R1/,~2/9 Rll

..

R:,,,.~ ,H/Rs R1, , ~ N LiAIH4 H--N THF

Scheme 4.

10.

N---~,,R6 N~ H - H R9

Re

Genericcyclic peptidemethodof synthesisof suchligandsas2-6, 8, and

226

DENNIS P. RILEY

C-", "--) C-", "-3

H...-N

N-..H

H-.-N

k~,=vJ

Lv,=,cJ

H

"

N-..H

H

~

~,

N--.H

Lv,=.J ....,, H

C=-D

=-= H

10 ~

H-.-N

" H

H~N

H

H

9 N--~ N~ H

N---H

Lv,=..J

H

' "Z =-L,=.~_.

cN H~N

C-", "-3 C , "-3

H~N

_-~ H

11

N--u

-H

==

~

H.-.-N

H

12

N'---H

H

_V

?

.-"

N

N- H

H

H.--N

N"H

H

H' ' N

el,?

. ""H

H" ~ v ~ N ~ / I

-H

H

Figure 1. Ligands utilized as their Mn(ll) complexes to probe the effects ofthe number and stereochemistry and placement of substituents on the stability and catalytic rate of SO dismutation. dramatically increase stability and have profound effects on activity. For example, the kcat for the Mn(II) complex of the all R-ligand 17 (or its all S-enantiomer), [Mn(2R,3R,8R,9R-bis-cyclohexano[15]aneNs)Cl2], at pH 7.4 is 1.2 x 10s M -1 s -l and possesses an increased thermodynamic stability (log K = 13.3) and an enhanced kinetic stability (at any pH over a 100-fold slower rate of dissociation results when compared to the unsubstituted complex). In contrast, the isomeric complex containing the (2R,3R,8S,9S)-bis-cyclohexano[15]aneN 5 ligand, 18, possesses a similar stability profile, but has virtually no catalytic SOD activity. 42 Clearly, the stereochemical orientation of the hydrocarbon (nonchelating) substituents has a major role in determining the ability of a Mn(II) complex of this class to function as an SOD catalyst. Thus, our synthesis and characterization

Developmentof Manganese(ll)Complexes

227

efforts, which were focused on the role that C-substituents exert on both the catalytic SOD activity and the overall chemical stability of the resultant complexes, had revealed that the interplay between position, number, and stereochemistry of substituents in dictating catalytic activity was subtle and unobvious.

-,,,,, 9

...o,,,,

17

~,,,.-

18

It should be noted that the complex derived from ligand 18 was characterized by X-ray crystallography and shown to have the Mn(II) arranged in a pentagonal bipyramidal geometry with trans-dichloro ligands and a planar macrocyclic orientation. 36 This complex and that of the parent unsubstituted complex, as well as the Mn(II) complex of ligand 7 (both also characterized by X-ray crystallography), show the same geometry and orientation of NH's. In each case the NH pattern is such that they alternate about the ring in an up:down:up:down:up orientation (Figure 2), so that the two sides of the macrocyclic ring in the complex are chemically distinct. We also observed this same NH pattern with the Cd(II) complex of the unsubstituted complex. 41 In general, we have pursued the goal of designing highly chemically stable complexes, as their end use would be as human pharmaceuticals. Our initial design goal was to maximize the number of substituents on the macrocyclic ring, thereby increasing the preorganizational rigidity of the macrocyclic ligand and thus increasing the stability of the Mn(L) complex to dissociation. 37But it emerged that we had no way of knowing in advance whether increasing the number of substituents on the parent macrocyclic ring would affect the catalytic activity in a beneficial or deleterious fashion. This necessitated that we develop a detailed mechanistic understanding of how these Mn(II) macrocyclic ligand complexes function as catalysts. This was predicated on the strong belief that once such information is in hand the compounds could be subjected to computer-aided design (CAD) techniques such as molecular mechanics (MM) calculations in order to gain a quantitative and hence predictive understanding of how position, number, and stereochemistry of the substituents could effect the catalytic rate.

3.1. Stability Studies The effect of C-substitution on the stability of these macrocyclic Mn(II) complexes is such an important aspect of the design of optimized structures suitable for human pharmaceutical development that it is worthwhile to describe the stability of these complexes and their mechanism of demetallation/dissociation in greater

228

DENNIS P. RILEY

y

n

N N

c, 7-coordin=te Side view

N

Top View

Omit~ngA1ialIAIpmdn

Figure 2. View of the Mn(ll) complexes with the [15]aneN5 ligands and the orientation of the NH's relative to the plane of the macrocyclic ring. detail. Using a combination of techniques including Cu(II) ion competition studies and direct hplc analysis of free ligand and intact complex, 4~the kinetic stability of all of the complexes, I, have been routinely assessed. The thermodynamic stability constant (K) of these complexes were also routinely measured by potentiometric titration methods. Although it should be noted that highly substituted complexes, exhibiting enhanced preorganizational rigidity of the macrocyclic ligand, are so kinetically stable that the classic potentiometric methods are unable to adequately assess the stability due to extremely slow response to pH changes. Nevertheless, the kinetic stability is actually a more useful index as that reflects a real issue for drug design efforts; namely, is the rate of excretion of a metal-based drug faster than the rate of dissociation of the complex? If so, the complex would be excreted intact and possess the desired chemical stability. Thus, it is important to understand the mechanism of metal dissociation for a metal-based drug and assess this stability ultimately in an in vivo model. In all cases the Mn(II) complexes, I, exhibit a pure first-order rate of loss of Mn(II) ion from the ligand at a fixed pH. 4~ The dissociation of the ligand does exhibit a first-order dependence in [H§ This observation allows one to describe the kinetics of ligand dissociation by its second-order dissociative rate constant, kdiss, where multiplication of the kcussvalue determined for a complex by the [H§ gives the first-order rate of Mn(II) release from the complex at any pH, and hence allows one to conveniently calculate the tlr2 for the complex at any pH. To convert the reported kdiss value to a half-life at any pH, Eq. 3 applies. tit2 = 0.694/(kdiss x [H§

(3)

Since in all cases at pH > 7, the complexes are completely intact in water based on the potentiometric titration data, the goal is to drive the stability profile to lower pH's so that even at pH-5 or less the complex will be stable.

Developmentof Manganese(ll)Complexes

229

Finally, we also observe that the kinetics of dissociation at any pH are independent of added metal ions and other ligands, even though they may form more stable complexes with the [ 15]aneN 5 macrocycle or with Mn(II), respectively. This has the consequence that in vivo kinetic stability will not be at risk in the presence of any potential biological chelating agent or in the presence of other endogenous free metal ions. Since the sole pathway for dissociation is a dissociative one for these macrocyclic ligand complexes, they possess a high inherent stability. The actual pathway for dissociation in aqueous systems though, as stated above, involves participation from protons. Thus, the basicity of the ligand itself is critical to determining the kinetics of proton-driven dissociation. An excellent example of this effect is observed with the Mn(II) complex of the pyridino ligand 19. This complex is nearly 175 times more kinetically stable at any pH than the parent unsubstituted ligand complex with Mn(II). This is undoubtedly due to some contribution from the rigidity conferred on the ligand by the pyridine substitution, but its complex with Mn(II) is even more stable than those of the bis-cyclohexyl ligand (17 and 18) complexes with Mn(II). This enhanced stability is most likely attributable to the much lower basicity of the ligand itself; namely, a measure of the affinity of the ligand for proton. For the aliphatic substituted ligands, three pKa values are observed, with the values generally in the range: 10.5-11.0, 9.2-9.5, and 5.0-5.9. With the pyridino ligand, 19, these values are lower, reflecting the weaker basicity of the ligand: 9.4, 8.8, and 5.3, 40 and its lower affinity for a proton.

NH

19 \

HN

/

Since the Mn(II) pyridino complex of ligand 19, [Mn(pyridin[15]aneNs)C12], possessed such an excellent stability profile and since the complex exhibited a SOD catalytic rate constant equivalent to the unsubstituted complex, [Mn([ 15]aneNs)Cl 2] at pH = 7.4, the complex was subjected to a complete characterization including the X-ray crystallographic determination of its structure. The structure as shown in Figure 3 is similar to other complexes of this class in that it crystallizes as a trans-dichloro pentagonal bipyramidal Mn(II) seven-coordinate complex with a planar array of the five nitrogens of the macrocyclic ligand. 46 In Table I are shown some selected bond lengths and bond angles. A few features stand out with this complex. First, the Mn to pyridine nitrogen bond distance is shorter than the four other Mn-N bonds by about 0.08-0.09 /~ and the NH pattern alternates as

230

DENNIS P. RILEY

H5n

C6

~~iL~ Mn

~)

N2

CI0

N1 . . ~

C2

C11 ~,~r~ C12

C13

C4 H2n

Figure 3. ORTEP drawing for the [Mn(pyridin[15[aneNs)CI2] showing the labeling scheme and the 50% probability ellipsoids for non-hydrogen atoms. up:down:up:down, just as observed in all the complexes we have investigated by X-ray crystallography in this class. The difference with this ligand is that there is an even number of NH's, thus, the two sides of the macrocyclic plane are chemically equivalent and hence the two axial coordination sites are thus equivalent.

Table 1. List of Representative Complexes and Their Catalytic Rate Constants for the Dismutation of Superoxide at pH = 7.4 and Their Kinetic Stabilities as a Function of [H §

Complex Mn([15]aneN5)CI2 Mn(10)CI2 Mn(11)CI2 Mn(17)CI2 Mn(18)CI2 Mn(19)CI2 Mn(20)CI2 Mn(28)CI2 Mn(29)CI2 Mn(30)CI2 Mn(31)CI2

kcat

(pH =

7.4) x 10 -7 M -1 $-1

4.1 3.9 9.1 12.1 <0.1 3.7

2.9 6.1 15.0 <0.1 4.56

kdis s (M-1 $--1)

2814 18 1375 28 26 16 23 4.8 4.0 5.9 1.42

Development of Manganese(ll)Complexes 3.2.

231

In Vivo Stability

A convenient means of assessing the "intactness" of these high-spin d 5 Mn(II) complexes both in vitro or in vivo is by electron spin resonance (ESR). 47 Both the Mn(II) complexes and free Mn(II) ion have distinctive ESR spectra. Therefore, the presence of free Mn(II) and intact complex can be determined and quantified in biological fluids and tissues with little or no sample manipulation. We have established that Mn(II) complexes with high kinetic stabilities (i.e. low kdiss) do indeed have high in vivo stabilities with excellent correlation. For example, the Mn(II) complex of ligand 20 has a high a kinetic stability (kaiss = 23 M -1 s-1) and possesses a log K > 13.48 By ESR analysis, we have shown that complex 20 has high in vivo stability with complete rat plasma stability at 37 ~ for 10 h and is greater than 90% intact in the liver of rats 30 min after intravenous injection. 46

\

cN H\ /

,,, .....

\ /H/-~

2@ It should be noted that the X-ray structure determination of the Mn(II) complex with ligand 20 has been carried out, 48 and it shows that this complex, as all the others that have been structurally characterized in this family of C-substituted [ 15]aneN 5 Mn(II) complexes, possesses the alternating NH pattern described above and depicted schematically in Figure 2. It is of interest to note that this complex crystallizes with a water in one coordination site trans to a chloro ligand with an ionic chloride in the crystal lattice. Further the water is bound on the side of the macrocycle which possesses two cis nonadjacent NH's and the chloro ligand on the side with three cis NH's. At the time this complex was synthesized, its design was based on the empirical premise that more C-substitution is desirable for stability, but with no understanding of how to arrange the substituents to achieve a fast rate for the SOD reaction. In fact, the catalytic activity of this complex is only about 70% of the parent unsubstituted complex at pH = 7.4 and almost 6 times less active than the Mn(II) complex of ligand 17; thus, this complex represents an excellent example of the dilemma that was faced for the design of high-activity catalysts; namely, while it is very stable, it has lower catalytic activity. Clearly, a sound mechanistic footing is necessary to solve the problem of how to design a maximally

232

DENNIS P. RILEY

effective SOD mimic with its catalytic activity being at least as good as that of the Mn(II) complex with ligand 17, but with an enhanced stability.

4.

M E C H A N I S T I C STUDIES

As noted above the need to understand the mechanistic details for the catalytic superoxide dismutation reaction was a critical element for the design of highly substituted complexes possessing both high stability and high activity. From kinetic and mechanistic studies we have determined that the Mn(II) complexes of these pentaaza crown ligands function via a catalytic cycle in which the rate-determining step is oxidation of Mn(II) to Mn(III). 23'36 Two independent pathways operate for most of these complexes: (1) an outer-sphere proton-coupled electron transfer from a bound water to an incoming hydroperoxy radical (Eq. 4), Mnn(L)(OH2) + HO 2 --~ Mnm(L)(OH) + HOOH

(4)

and (2) an inner-sphere substitution involving coordination of superoxide anion to Mn(II) in a vacant axial site (Eq. 5), followed by fast protonation of the bound superoxo creating a pseudo-octahedral Mn(III)-hydroperoxo complex (Eq. 6). Mnn(L) + 0 2 --4 Mnm(L)(OO 2-)

(inner-sphere substitution)

(5)

k2

MnlI(L)(O0 2-) +

H§

-~ Mnm(L) (O2H) (oxidation)

(6)

fast

Isotope studies using D20 as the reaction solvent were particularly helpful in providing further insight into these pathways. Utilizing the fast catalyst derived from ligand 17 in both H20 and D20, rate constants for both the proton-independent and proton-dependent pathways for oxidation of Mn(II) were measured. In H20 the pH independent exchange rate-limited path (Eq. 2) possesses a rate constant of 1.58 x 10§ s-1, consistent with known ligand exchange rates on Mn(II). 43 In D20 this rate is increased by ~ 10%, consistent with water exchange rates on the Mn(II) ion having a role in dictating this rate. Other complexes showed similar small effects on the magnitude of this rate constant when measured in D20. The measured rates for the outer-sphere pH dependent pathway of Mn(II) oxidation with this complex and that of the Mn(II) complex with ligand 11 were both dramatically lowered in D20 showing an isotope effect of nearly 6, consistent with H-atom transfer for the proton-dependent pathway of oxidation of Mn(II). 36 For either pathway to be maximally efficient requires that the barrier to electron transfer must be minimal; i.e. the ligand reorganizational barrier to electron transfer must be small.44'45Thus, the precursor Mn(II) complex should adopt a six-coordinate pseudo-octahedral geometry similar to that required by the corresponding Mn(III) complex. Thus, loss of an axial ligand followed by folding of one of the secondary

Development of Manganese(ll)Complexes

233

amine NH's of the macrocyclic ring so that it occupies an axial site would generate a pseudo-octahedral complex, [MnU(L)X] n§ accommodating this requirement for a six-coordinate Mn(II) complex. If the macrocyclic [ 15]aneN 5 ligand possesses C-substituents which, due to intramolecular steric repulsions and angle strains, could force the ligand to adopt this folded pseudo-octahedral geometry about the spherically symmetrical Mn(II) ion, then the Mn(II) complex would be poised to undergo facile electron transfer as the ligand reorganizational barrier would be minimized. It is in this manner that the stereochemistry of the substituents would be expected to exert a major effect on the rate of electron transfer via either pathway. This need to rearrange the ligand from a planar geometry into a folded conformation, stabilizing a pseudo-octahedral geometry on Mn(II), may be the reason why some complexes of this family, for example, show no measurable rate for one or the other of the two pathways. It is intriguing to speculate that the steric constraints imposed on the folding by the presence of a substituent positioned on a carbon of the ring may favor or inhibit a particular fold and thereby promote or block a particular pathway for electron transfer. Indeed, this implies that the two mechanisms may well operate by different folding motifs. Clearly, there are some very intriguing results that any theory must be able to explain. For example, the two ligands 17 and 18 afford complexes which are either extremely active or inactive. Further the Mn(II) complex of the pentamethyl ligand 10 has no pH dependence to its catalytic rate, even though it has the highest pH-dependent rate measured in this class of complexes, with kca t = 3.90 x 10§ M -1 s-1.4~ Thus, any theory which attempts to rationalize the observed reaction rate constants for catalytic dismutation of superoxide must predict why the Mn(II) complex of ligand 10 has no pH-dependent outer-sphere rate.

5. COMPUTER-AIDED DESIGN (CAD) In an attempt to develop a theory of the details of this SOD catalysis which can correctly predict the effects that substitutents exert on the catalytic rate, molecular modeling was utilized. The premise for the modeling paradigm is that the macrocyclic [15]aneN 5 ligand possesses C-substituents which, due to intramolecular steric repulsions and angle strains, force the ligand to adopt various degrees of folded pseudo-octahedral geometry about the spherically symmetrical Mn(II) ion. If the Mn(l/) complex is constrained in a geometry which resembles a six-coordinate pseudo-octahedral geometry, then the Mn(II) complex would be poised to undergo facile electron transfer as the ligand reorganizational barrier would be minimized. It is in this manner that the stereochemistry of the substituents could be expected to exert a major effect on the rate of electron transfer via either pathway. In order to better understand the effects that substituents exert, we utilized a combination of molecular mechanics calculations and synthesis with the goal of rational design of highly substituted chemically stable synthetic enzymes. It needs to be stressed that while the prediction of properties and reactivity of small

234

DENNIS P. RILEY

carbocyclic ring systems based on conformational control is widely practiced and relatively straightforward, s~ the conformational analysis of large macrocyclic ring metal complexes undergoing redox chemistry offers a more difficult challenge. To effectively develop a rational modeling paradigm not only requires a detailed understanding of the mechanism of the rate-determining step in the catalytic process, but also a comprehensive database of chemical structures with their rate data so that the theoretical model could be subjected to stringent testing. In addition to the ligands 1, 7, 8, 9, 10, 11, 17, 18, and 20, the ligands shown in Figure 4 were also synthesized. The Mn(II) complexes of these ligands were then utilized as our source of rate and stability data for this CAD study. The Mn(II) complex of each ligand was completely characterized in terms of physical properties (e.g. the kinetic stability to dissociation in aqueous media) and the kinetics of their superoxide dismutase catalytic activity. 49 In all cases, as we noted previously, a~ increasing the number of the substituents on the macrocyclic ligand increases the kinetic stability of the complexes to dissociation in water. The catalytic activities of these database complexes follow no apparent trend regarding the number of substituents or their stereochemistry, other than that they both clearly have effects on the overall rate and on the two competing pathways for oxidation of Mn(II) during the catalytic cycle. Most striking is that for some of these complexes there is either no measurable inner-sphere or no outer-sphere rate. Complexes with the ligands 10, 18, and 21 have no measurable/Cos. While these complexes were utilized as the original database to construct the modeling paradigm, complexes 28-31 (Figure 5) were synthesized based on predictions of the model and provided tests of the predictive power of the MM calculations. For the complexes, 28-31, their kinetic stabilities and catalytic rate constants, kcat (pH = 7.4), were also measured and are listed in Table 1. While there are a number of computational methods available that could be used to predict structures and calculate the thermodynamic properties of a given structure for a coordination complex, the simplest level of calculation would be the use of molecular mechanics (MM) assuming appropriate parameters are available. 53 We have completed the X-ray structure determinations and refinements of many complexes in this class and reported five crystal structures of this class of Mn(II) pentaaza crown complexes, 1'2'9and discuss that of the pyridino complex of Mn(II) with ligand 19 herein. This extensive X-ray structural database has allowed us to refine the commercial parameters and utilize an improved Ravg for the Mn-N distance, 2.283 A, and also an Ravgfor the Mn-CI distance, 2.616 ~. Additionally, as noted above, in all cases the NH stereochemical pattern is alternating; i.e. each side of the plane of the macrocyclic ring is chemically unique, since one side will have two nonadjacent NH's, while the other side possesses three (two in the case of pyridino ligands) NH's (Figure 2) in a cis orientation. This motif provides a basis for probing the structural effects utilizing MM calculations. The design premise, which we set out to test, was based on the concept that the ligand dictated catalytic activity by promoting or preventing a particular folded

Development of Manganesefll) Complexes

25

HN

r

NN --

-

235

%.. NN

26

11-I

o

&

Figure 4. Ligands utilized to form the database for substituted Mn([15]aneNscomplexes for probing the effects of substituents for molecular modeling (MM) studies. structure. Thus, a good catalyst will be one in which the Mn(II) center is constrained in a geometry which promotes rapid electron transfer; i.e. a pseudo-octahedral geometry preferred by Mn(III). Consequently, the folding of an NH out of the plane defined by the metal and the five nitrogens of the ligand into an axial site governs the ability of the corresponding complex to function as a catalyst. The MM

236

DENNISP.RILEY

--~HNIEI,..~/~ ~...,//NNH

28 HN4ft'~J"

"~'.o,./NH

HN~ ~ J "

29

g

~~NH HNll,,,.~'~ /---N

',,t/NH

30

HN4fr

,,,

;'NH

HN

,,,,,,: Figure S. Ligands utilized for testing the predictive ability of the folding paradigm for MM calculations.

calculations can then be utilized to determine the relative energies of all the possible folds for each complex (five possible, since each nitrogen can fold into an axial site, but only on the side of the macrocycle which the NH is located) and their relative energies compared. Although it should be emphasized that with some classes of ligand due to the ligand symmetry, there may be fewer unique folding modes. For example, the pyridin[ 15]aneN 5 ligand 19 possesses a higher symmetry than other ligands of this class displayed here. It possesses four NH's and the two axial trans coordination sites are equivalent. As a consequence there exists only two unique folding motifs with this Mn(II) complex. Our initial foray into modeling using MM calculations involved trying to probe the relative energetics of two types of six-coordinate structures generated via loss of a ligand from a trans seven-coordinate Mn(II) as depicted in Figure 6. In solution and solid state both the six- and seven-coordinate geometry exist for these complexes; thus, in solution a dynamic equilibrium of such structures is expected. 36 Two initial six-coordinate structures could be generated via the loss of an axial ligand from the seven-coordinate structure; i.e. the complex labeled A in Figure 6 which has the vacant site on the side in which three NH's are cis, and the structure labeled B which has the vacant site on the side with two nonadjacent NH's cis. Clearly, the two intermediates are chemically quite distinct and for each an energy can be calculated using an MM approach. Invariably, the structure A with the three NH side containing the vacant site was found to be at a much lower energy than the

Development of Manganese(ll)Complexes 7-coordinate

side view

6-coordina/e s/de view H

H

vN--N--NMn--N---.N ," - Y ~ I ' H n A

X

237 6-coordinaXe side view

H Inl N--N--NI~III---'INI--'INI x

-x

-

R~

R

X whmex-ly poutJ,d 6a,,,. lJpad

~

-x .~

9-

x

H v ,"

]

-

H Y

N-'-N--N1~r -'INF--N H B

Figure 6. Depiction of the equilibrium between two possible six-coordinate intermediate structures generated via ligand dissocciation from a seven-coordinate intermediate Mn(ll) complex.

structure B (in the range of 4-8 kcal for chloro as axial ligand and 8-12 kcal for aquo). The magnitude of the difference suggests that structure A should be a logical starting point as a common intermediate leading to catalysis for all complexes; i.e. the starting point for productive chemistry during a catalytic cycle. Using this approach and assuming that the six-coordinate aquo complex is the Mn(II) complex in water which leads to productive chemistry, l'2b the simplest entry to the outersphere pathway presents itself as the aquo complex of structure A (where X = H20). This suggests that for this path to be viable (fast rate) for a given complex, the substituents would force one of the three cis NH's on the side of the vacant coordination site out of the plane defined by the Mn(II) and the five nitrogen atoms (in effect, folded) into the axial pseudo O h site. When this is investigated for these complexes generated from the ligands specified above, we find, in general, that there is a most preferred NH (lowest energy conformer) folding for one of the three cis NH nitrogens into an axial pseudo O h site. This is the modeling paradigm which evolved for the outer-sphere correlation. In contrast, the inner-sphere, superoxide binding pathway must utilize a folding motif generating a six-coordinate intermediate in which one of the NH's on the side with two cis-NH's folds into the pseudo-axial site. For all the complexes, we observed that there exists a unique lower energy folded conformer. This is consistent with the possibility that the inner-sphere pathway employs a folded structure which is unique and different from the outer-sphere pathway, and it is consistent with the inner-sphere path being accessed by an intermediate common to the outer-sphere path. For example, one such scenario is that the outer-sphere path could involve folding one of the three cis-NH's folds into an axial site, and the inner-sphere path could utilize a vacant site on the opposite side of the plane of the Mn(L) complex and hence fold one of the two nonadjacent cis NH's into the axial site. Chemically this could arise if the aquo six-coordinate complex (A in Figure 5), with a vacant coordination site on the side of the macrocyclic ring with three NH's cis, binds superoxide anion in this vacant site. Loss of the bound trans-water and subsequent

238

DENNIS P. RILEY

folding of either of the NH's on the side of the two cis-NH's would generate a folded six-coordinate pseudo-octahedral complex. This is the paradigm which evolved and was validated by the modeling efforts. 49 At the outset we could make no claim as to the validity of this paradigm, other than it is consistent with the general mechanistic., understanding and characterization of the chemistry of these complexes. For all the complexes utilized in the database (as described above), an evaluation of the energy difference (zkE) of the folded six-coordinate structure for the Mn(II) and the corresponding Mn(III) complex were performed for each NH folded into an axial site. From this evaluation of different folded structures each complex will possess a lowest energy folded structure for the six-coordinate complex derived from an NH on the side of the three NH's occupying a pseudo-axial site, and also a lower energy structure for the six-coordinate complex derived from an NH on the side of two cis nonadjacent NH's occupying a pseudo-axial site. Thus, a correlation of the energy difference, &E (EMn(In)-EMn(n)), for a series of complexes for each of the two types of folds (potentially correlating with unique inner-sphere and outer-sphere folded structures) can be performed, where the ground-state energy of both Mn(III) and Mn(II) are approximated by the summation of the various contributing energy terms45: ~Mn(III) = Uelectrostatics + UH-bonding + Ubond stretch + Uangle strain + Utorsional strain + Udihedrai strain + Uvan der Waals + Uangle deformation;

(7)

and p

p

p

p

~Mn(II) = Uelectrostatics + UH-bonding + Ubond stretch + Uangle strain + t

P

P

Utorsional strain + Wdihedral strain + Ovan der Waals +

U t

angledeformation

(8)

For a series of complexes, a correlation of &E for the folded geometries (both the inner-sphere and the outer-sphere) is simplified by comparing a modified energy &E # in which the contribution of (Uelectrostatics-Uelect~osutic~) + (UH_bonding-p UH.bonaing) are removed since they are maintained invariant among the series of complexes. Plots of &E# (EMn(ili)-Y'.Mn(ii))inner_sphereVS. kis values (inner-sphere rate constant), and &E# (EMn(lii)-~Vln(ii))outer.spher%VS. kos values (outer-sphere rate constant) can be generated. Linear plots of AE~ vs. kis or kos would indicate that the correlation is valid. In the cases described here with the ligands utilized for the database, excellent correlations were obtained relating the folding of one of the three cis-NH's as the outer-sphere H-atom transfer path and the folding of one of the two cis-NH's as the inner-sphere SO binding pathway. In general, we find that there is a clear energy preference for a single, unique NH to fold on each side of the macrocycle for each pathway; i.e. the energy differences between possible NH folding modes were large (sometimes as large as 70-100 kcal); thus, &E (EMn(m)-EMn(n)) for the three outer-sphere folds and for the two inner-sphere folds for each complex always affords a clear lowest energy choice

Developmentof Manganese(ll)Complexes

239

for each type of fold with each complex. It is also clear from these modeling analyses that the orientation of the NH's is an important aspect of the strain energies required to fold the ligands. Implicit in this exercise is the assumption that the orientation of the NH's relative to the substituents is known. For those complexes in which crystallographic data exists, the orientations are known. Using this data, certain structural relationships (e.g. a trans-cyclohexano substituent dictates that the NH's t~ to the substituted carbons of the macrocyclic ring must also be trans), and MM calculations of the free ligands and for the Mn(II) planar seven-coordinate complexes, the NH pattern can be established with a high level of confidence for any complex of this family. The goal of such correlations is to determine if (1) the folding paradigm is indeed consistent, (2) there are unique folding patterns for the inner-sphere and separate unique folding patterns for the outer-sphere pathways, and (3) if the MM calculation/modeling paradigm can be successfully used to predict structures to be synthesized so as to optimize the number of substituents for maximum stability while retaining high catalytic activity. This modeling paradigm based on MM calculations successfully addressed all three points. This meant that progress in developing highly substituted complexes with high catalytic activity could be made without the need to rely on trial and error synthesis, which can be exceedingly time consuming for the synthesis of such highly substituted and stereodefined ligands. Indeed, the excellent correlations allowed us to test the model in a predictive manner utilizing the four ligands of Figure 5. From the standpoint of ease of synthesis and the desire to maximize the stability with the minimum number of substituents, we chose to further elaborate the bis-cyclohexano structure of the high-activity catalyst derived from ligand 17. Note that this complex of this ligand with four of the macrocyclic ring carbons bearing a substituent achieves equivalent stability to those complexes bearing five methyl substituents. Thus, synthetically it was attractive to add an additional methyl to this ligand. Intuitively, to increase stability most efficiently, we chose to put the additional substituent on one of the chelate rings that did not possess a substituent. Thus, structures 28 and 29, each bearing an additional methyl substituent, were modeled. The two complexes only differ in the stereochemistry of the added methyl substituent: 28 (S-Me) and 29 (R-Me). The predictions based on the use of these correlations are that the all-R complex 29 should be a very good catalyst, especially via the outer-sphere pH-dependent path, while complex 28 with the S-Me would be much less active via this mode, and they should possess comparable activities via the inner-sphere path. As predicted the complex 29 is in fact very active with a large pH dependence; about 3x that of complex 28. In fact, complex 29 possesses a catalytic rate at pH = 7.4 in excess of 1.5 x 10§ M -1 s-1 and was the most active catalyst which we had synthesized up to that point. Based on this level of success, it was intriguing to consider more highly substituted structures and whether they, in fact, would have catalytic activity. One type of substituent which we have observed to enhance stability is the gem-dimethyl;

240

DENNIS P. RILEY

thus the gem-dimethyl complex 30 was subjected to the modeling paradigm. The results were quite clear; namely, there did not exist any folds that are of a low energy; i.e. neither the inner- or outer-sphere pathway would be predicted to operate with this structure. The ligand and complex were synthesized and, indeed, only a trace of catalytic activity was observed. The desire to increase stability by increasing the number of substituents on the macrocycle while retaining good catalytic activity led us to consider whether there were carbons of ligand 29 on which one could add a substituent and retain good activity. This exercise utilizing the modeling paradigm led to the prediction that, of the five carbon atoms of the macrocyclic ring of 29 which were devoid of substituents, the 14-position could bear a methyl (or other substituent) and retain reasonable activity, but only if the stereochemistry of that additional substituent is R--generating the all-R hexasubstituted ligand structure 31. The modeling, in fact, predicts that the catalytic activity for such a complex 31 should be midrange within this family via both pathways. The actual activity was not quite as high as predicted, but was nevertheless very good considering that the complex bore six substituents. In fact, this complex has a better activity and has a kinetic stability of over 2000 times that of the original unsubstituted parent complex, [Mn([ 15]aneNs)Cl2]. An important structural insight into the driving force for the folding can be made with this complex. The 14R-Me substituent of the ligand 31 is in an axial position in the planar seven-coordinate Mn(II) complex structure. This high-energy strained conformation is dramatically relieved when the outer-sphere fold occurs generating the six-coordinate structure. This folding places the 14R-Me into an equatorial orientation; thus relieving strain and providing a driving force for the fold. Two complexes require comment: the Mn(II) complexes of ligand 10 and ligand 18, since they exhibited unique inactivity. For the complex with ligand 10 no outer-sphere H-atom pathway was observed in its SOD catalysis. From the modeling it was quite clear that such a path could not occur because the fold corresponding to the outer-sphere fold with this ligand was extremely unfavorable; i.e. the fold of one of the NH's on the side of the three cis-NH's. The complex with the bis-cyclohexyl ligand 18 was also quite distinctive in that little or no catalytic activity was observed. Again, the modeling was quite predictive of this; i.e. all NH's were of very high energy indicating that this ligand preferred a planar geometry and hence would be a poor catalyst as observed. In general, the linear correlations developed for this system of Mn(II) complexes lends support to the theory that folding is a critical aspect of catalytic activity for these synthetic enzymes (synzymes) and that substituents (relative position and stereochemistry) influence the favorability of the folding in a predictable manner. Thus, we are able to model any given complex for both the inner-sphere or the outer-sphere folding energetics and calculate the expected rate constant for each path for any ligand pattern! This paradigm has worked extremely well at correlating and predicting activity and has made it possible to design highly complicated molecules and test their potential utility as a catalyst without the extraordinary effort

Development of Manganese(ll) Complexes

241

required to first synthesize such a structure. We have been able to literally test hundreds of combinations of structures in this fashion and been able to gain a fair degree of certainty regarding a structure prior to embarking on the complicated synthesis. In no case have we yet observed this MM modeling paradigm to fail to predict the approximate activity of a complex. Indeed, we have modeled and synthesized many more structures and have been able to synthesize synzymes with increased stability (more than a factor of 100x that of complex 21) and with catalytic rates exceeding 10+9 M -l s-l. 51 Recent biological results generated with complexes which were designed using the MM technique and methods described here have provided additional confidence that these methods hold great promise for the design of improved human pharmaceutical agents. For example, the Mn(II) complex derived from ligand 31, has been shown to be very efficient in protecting from reperfusion injury in animal models, 52 and in lowering the free radical production following ischemia as monitored via in vivo ESR. 53 Additionally, using a combination of the in vivo ESR measurement and hplc analysis of blood samples for free ligand and complex, we have established that complexes in this stability range exist completely intact in vivo in rat organs and blood for up to 10 h (period of measurement) and are excreted intact via dual excretion pathways (monitored in both urine and feces).

Table 2. RepresentativeBond Lengths and Bond Angles for the Mn(ll) Complex of Ligand 19 as Determined by X-ray Crystallographic Determination Type Length, ,~ Type Angle, deg. Mn-Cll Mn-C 12 Mn-N1 Mn-N2 Mn-N3 Mn-N4 Mn-N5 N2-H2 N3-H3 N4-H4 Ns-H5

2.596 2.593 2.260 2.349 2.330 2.323 2.329 0.73 0.87 0.87 0.86

CI1MnCI2 CI 1MnN1 CI1MnN 2 CI1MnN 3 CI1MnN4 CI1MnNs CI2MnN 1 CI2MnN2 CI2MnN3 Cl2MnN4 Cl2MnN s N1MnN2

178.7 90.7 93.4 81.0 96.9 87.3 90.5 86.7 97.8 82.3 93.4 71.4

N2MnN 3 N3MnN 4 N4MnNs NsMnN 1

73.4 73.1 71.5 73.4

242

DENNIS P. RILEY

Continuing efforts aimed at developing catalysts with activities approaching that of the Cu/Zn SOD enzymes (kcat > 1 • 10+9 M -I s -1) and with enhanced stabilities (as monitored by the kdiss < 0.1 M -I s -1) are in progress. It should be noted though that the complexes discussed here and the methodologies described here are aimed solely at the design of a core structure for an optimized catalyst. The compounds presented here only present a hydrocarbon surface; where, in fact, the optimum drug for a synthetic enzyme for any particular disease may well require a different log P or different type of functional substituent to optimize the localization of the compound to optimize drug efficacy, minimize toxicity, and facilitate clearance of the drug. The chemistry presented here for synthesizing such core catalyst structures which are optimized for stability and catalytic activity are amenable to the construction of complexes which have pendant functionality; e.g. pendant alcohols, amines, amides, esters, acids, etc. Such chemistry is readily available from standard peptide methods using the available chiral pool and unnatural chiral synthetic amino acids. Thus, by adapting a promising core catalyst structure, such as that achieved with the Mn(II) complex with ligand 31, it will be possible to develop optimized structures with varying functional substituents tailored to meet the need of the disease state. It is in such a manner that human pharmaceutical agents with minimal toxicity and maximal efficacy could be realized.

ACKNOWLEDGMENTS The author thanks his collaborators from Monsanto Co., K. Aston, W. Rivers, H. Rahman, S. Henke, P. Lennon, and W. Neumann for their many contributions, and thanks Drs. C. Day and V. Day of Crystallytics for their assistance with X-ray structure determination.

REFERENCES A N D NOTES 1. (a) McCord, J. M.; Fridovich, I. J. Biol Chem. 1969, 244, 6049; (b) Fridovich, I. J. Biol. Che~, 1989, 264, 7761. 2. Werns,S. W.; Simpson, P. J.; Mickelson,J. K.; Shea, M. J.; Pitt, B.; Lucchesi, B. R. J Cardiovasc. Pharmacol. 1988, 11, 36. 3. Omar,B. A.; McCord, J. M. Mol. Cell. Cardiol. 1991, 23, 149. 4. McCord,J. K. J .Free Radicals Biol. Med. 1986, 2, 307. 5. Oyanagui,Y. Biochem Pharmacol. 1976, 25, 1465. 6. Droy-Lefaix,M. T.; Drouet, Y.; Geraud, G.; Hosford, D.; Braquet, P. Free Radical Res. Commun. 1991, 12-13, 725. 7. Shingu, M.; Takahashi, S.; Ito, M.; Hamamatu,N.; Suenaga, Y.; Ichibangasr Y.; Nobunaga, M. Rheumatol Int., 1994, 14, 77. 8. Ando, Y.; Inoue, M.; Hirota, M.; Morino, Y.; Araki, S. Brain Res. 1989, 477, 286. 9. Chan, P. H.; Yang, G. Y.; Chen, S. E; Carlson, E.; Epstein, C. J. Ann Neurol, 1991, 29, 482. 10. Yang,G.; Chan, P. H.; Chen, J.; Carlson, E.; Chen, S. E; Weinstein, P.; Epstein, C. J.; Kamii, H. Stroke 1994, 25, 165.

Developmentof Manganese(ll)Complexes

243

11. Deng, H. X.; Hentati, A.; Tainer, J. A.; lqbal, Z.; Cayabyab, A.; Hung, W. Y; Getzoff, E. D.; Hu, B.; Herzfeldt, R. P.; Roos, C.; Wamer, G.; Deng, E. Soriano, C. Smyth, H. E.; Parge, A.; Ahmed, A. D.; Roses, R. A.; Hallewell, M. A.; Pericak-Vance, M. A.; Siddique, T. Science 1993, 261. 12. Rosen, D. R.; Siddique, T.; Patterson, D.; Figlewicz, D. A.; Sapp, P.; Hentati, A.; Donaldson, D.; Goto, J.; O'Regan, J. P.; Deng, H. X.; Rahmani, Z.; Krizus, A.; McKenna-Yasek, D.; Cayabyab, A.; Gaston, S. M.; Berger, R.; Tanzi, R. E.; Halperin, J. J.; Herzfeldt, B.; Van den Bergh, R.; Hung, W.-Y.; Bird, T.; Deng, G.; Mulder, D. W.; Smyth, C.; Laing, N. G.; Soriano, E.; Pericak-Vance, M. A.; Haines, J.; Rouleau, G. A.; Gusella, J. S.; Horvitz, H. R.; Brown Jr., R. H. Nature 1993, 362, 59. 13. Brown, Jr., R. H. Cell 1995, 80, 687. 14. Troy, C. M.; Shelanski, M. L. Proc. NatL Acad. Sci. USA 1994, 91, 6384. 15. Greenlund, L. J. S.; Deckwerth, T. L.; Johnson, Jr., E. M. Neuron 1995, 14, 303. 16. Bravard, A.; Sabatier, L.; Hoffschir, E; Ricoul, M.; Luccioni, C.; Dutrillaux, B. Int. J. Cancer 1992, 51,476. 17. Church, S. L.; Grant, J. W.; Ridnour, L. A.; Obedey, L. W.; Swanson, P. E.; Meltzer, P. S.; Trent, J. M. Proc. Natl. Acad. Sci. USA 1993, 90, 3113. 18. St. Clair, D. K.; Oberley, T. D.; Muse, K. E.; St. Clair, W. H. Free Radical Biol. Med. 1994, 16, 275. 19. Safford, S. E.; Oberley, T. D.; Urano, M.; St. Clair, D. K. Cancer Res. 1994, 54, 4261. 20. Yoshizaki, N.; Mogi, Y.; Muramatsu, H.; Koike, K.; Kogawa, K.; Niitsu, Y. Int. J. Cancer 1994, 57, 287. 21. Flores, S. C.; Marecki, J. C.; Harper, K. P.; Bose, S. K.; Nelson, S. K.; McCord, J. M. Proc. Natl. Acad. Sci. USA, 1993, 90, 7632. 22. Miesel, R.; Mahmood, N.; Weser, U. Redox Report 1995, 1, 99. 23. Riley, D. P.; Weiss, R. H. J. Am. Chem. Soc. 1994, 116, 387. 24. Hardy, M. M.; Flickinger, A. G.; Riley, D. P.; Weiss, R. H.; Ryan, U. S. J. Biol. Chem. 1994, 269, 18535. 25. Weiss, R. H.; Fretland, D. J.; Baron, D. A.; Ryan, U. S.; Riley, D. P. J. Biol. Chem. 1996, 271(42), 26149-26156. 26. Zweier, J. L. J. Biol. Chem. 1988, 263, 1353. 27. Kilgore, K. S.; Friedrichs, G. S.; Johnson, C. R.; Schasteen, C. S.; Riley, D. P.; Weiss, R. H.; Ryan, U.; Lucchesi, B. R. J. Mol. CelL Cardiol. 1994, 26, 995. 28. Black, S. C.; Schasteen, C. S.; Weiss, R. H.; Riley, D. P.; Driscoll, E. M.; Lucchesi, B. R. J. Pharmacol. Exp. Therapeut. 1994, 270, 1208. 29. Venturini, C. M.; Sawyer, W. B.; Smith, M. E.; Palomo, M. A.; McMahon, E. G.; Weiss, R. H.; Riley, D. P.; Schasteen, C. S. In The Biology of Nitric Oxide 3: Physiological and Clinical Aspects; Moncada, S.; Feelisch, M.; Busse, R.; Higgs, E. A.; Eds.; Portland Press: London, 1994, p. 65. 30. Kasten, T. P.; Settle, S. L.; Misko, T, P.; Riley, D. P.; Weiss, R. H.; Currie, M. G.; Nickols, G. A. Proc. Soc. Exp. Biol. Med. 1995, 208, 170. 31. Meng, Y. Y.; Trachtenburg, J.; Ryan, U. S.; Abendschein, D. R. J. Am. Coll. Cardiol. 1995, 25, 269. 32. Weiss, R. H.; Riley, D. P. In Inorganic Chemistry in Medicine; Farrell, N., Ed.; Royal Society of Chemistry: In press. 33. Weiss, R. H.; Riley, D. P.; Rivers, W. J.; Aston, K. W.; Flickinger, A. G.; Hardy, M. M.; Ryart, U. S. In The Oxygen Paradox; Davies, K. J. A.; Ursini, E, Eds.; CLEUP University Press: Padova, Italy, 1995, p. 641. 34. Beckman, J. S.; Carson, M.; Smith, C. D.; Koppenol, W. H. Nature, 1993,346, 584. 35. (a) Weiss R. H.; Flickinger, A. G.; Rivers W. J.; Hardy, M. M.; Aston, K. W.; Ryan, U. S.; Riley, D. P. J. Biol. Chem. 1993, 268(31), 23049-23054; (b) Riley, D. P.; Rivers, W. J.; Weiss, R. H. Anal. Biochem. 1991, 196(2), 344-349. 36. Riley, D. P.; Lennon, P. J.; Neumann,W. L.; Weiss, R. H. J. Am. Chem. Soc. 1997, 119(28), 6522-6528.

244

DENNIS P. RILEY

37. Lindoy, L. E The Chemistryof Macrocyclic la'gand Complexes; Cambridge University Press: Cambridge, U.K., 1989. 38. Lennon, P. J.; Rahman, H.; Aston, K. W.; Henke, S. L.; Riley, D. P. Tetrahed. Lett. 1994, 35, 853. 39. Aston, K. W.; Henke, S. L.; Modak, A. S.; Riley, D. P.; Sample, K. R.; Weiss, R. H.; Neumann, W. L. Tetrahed. Lett. 1994, 35, 3687. 40. Riley, D. P.; Henke, S. L.; Lennon, P. L.; Weiss, R. H.; Neumann, W. L.; Rivers, Jr., W. J.; Aston, K. W.; Sample, K. R.; Rahman, H.; Ling, C.-S.; Shieh, J. J. -J.; Busch, D. H.; Szulbinski, W. Inorg. Chem. 1996, 35(18), 5213. 41. Franklin, G. W.; Riley, D. P.; Neumann, W. L. Coord. Chem. Rev. 1998, 174, 133. 42. Weiss, R.; Neumarm, W.; Meeh, L.; Brown, T.; Lennon, P.; Sample, K.; Zweier, J. A.; Samouilov, A.; Wang, P.; Riley, D. Proceedings of the Society of Magnetic Resonance and the European Society for Magnetic Resonance in Medicine and Biology, Vol. 1, Abstract 43. 43. Eigen, M. Pure Appl. Chem. 1963, 6, 105. 44. Marcus, R. A. Annu. Rev. Phys. Chem. 1964, 15, 155. 45. (a) Macartney, D. H.; Thompson, D. W. Inorg. Chem. 1989, 28, 2199; (b) Endicott, J. E; Kumar, K.; Ramasami, T.; Rotszinger, E P. In Prog. Inorg. Chem.; Lippard, S. J., Ed.; Wiley: New York, 1983, Vol. 30, p. 141; (c) Creaser, I. I.; Harrowfield, J. M.; Herlt, J.; Sargeson, A. M.; Springborg, J.; Geue, R. J.; Snow, M. R. J. Am. Chem. Soc. 1977, 99, 3181. 6. Crystal data for [Mn(pyridin[15]aneNs)Cla] 90.5 HaO: colorless, monoclinic at 21 ~ space group P21/n; a = 9.008(3) A, b = 21.955(7) A, c = 9.929 A; (x = 90.00", ~ = 117.57 ~ u = 90.00 ~ V = 1741(1) A,3, Z = 4, ~ = 0.71073 A, and p = 1.466 g-cm-3; ~ta(MoK(x)= 1.07 mm -l. 47. Zweier, J. L.; Brodericks, R.; Kuppusamy, P.; Thompson-Gorman, S.; Lutty, G. A. J. Biol. Chem. 1994, 269, 24156. 48. Neumann, W. L.; Franklin, G. W.; Sample, K. R.; Aston, K. W.; Weiss, R. H.; Riley, D. P. Tetrahed. Lett. 1997, 38, 779. 49. Riley, D. P.; Henke, S. L.; Lennon, P. J.; Aston, K. W. L. Inorg. Chem. 1999, 38, 1908. 50. Burkert, U.; Allinger, N. L. Molecular Mechanics, ACS Monograph 177; American Chemical Society: Washington, DC, 1982. 51. (a) Hancock, R. D. Acc. Chem. Res. 1990, 23, 253; (b) Wade, P. W.; Hancock, R. D.; Boeyens, J. C. A.; Dobson, S. M. J. Chem. Soc., Dalton Trans. 1990, 483; (c) Luckay, R.; Chantson, T. E.; Riebenspies, J. H.; Hancock, R. D. J. Chem. Soc., Dalton Trans. 1995, 1363. 52. Salvemini, D.; Riley, D. P. Science 1999, 286, 304. 53. Zweier, J. Unpublished results. 54. Salvemini, D. Unpublished results, and ibid., Brit. J. Pharmacol. 1999, 127, 685.

DESIGNING ACTIVE SITES OF SYNTHETIC ARTIFICIAL ENZYMES

Junghun Suh 1. 2. 3. 4. 5.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Synthesis of Small Molecules Equipped with Multiple Catalytic Elements . . Improvement of Catalytic Properties by Microdomains of Polymers . . . . . Design of Active Sites on Polymers by Molecular Imprinting . . . . . . . . . Design of Active Sites Containing Both Binding and Catalytic Sites by Random Functionalization of Polymer Skeletons . . . . . . . . . . . . . . . . 6. Design of Active Sites by Site-Directed Functionalization of Polymer Skeletons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7. Design of Active Sites by Cross-Linkage of Preassembled Catalytic Elements with Macromolecular Spacer . . . . . . . . . . . . . . . . . . . . . . . . . . 8. Design of Active Sites by Self-Assembly from Catalytic Elements . . . . . . 9. Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Advances in Supramolecular Chemistry Volume 6, pages 245-286.

Copyright 9 2000 by JAI Press Inc. All fights of reproduction in any form reserved. ISBN: 0.7623-0557-6

245

246 248 250 253 255 263 267 275 282 283 283

246

JUNGHUN SUH 1.

INTRODUCTION

Design of artificial enzymes (enzyme-mimicking catalysts) has been actively studied in various areas of chemistry and biology. In the area of molecular recognition, ~ for example, catalysis by selective recognition of transition states is one of the major targets and several host molecules have been designed as biomimetic catalysts. By utilizing versatile catalytic repertories of metal ions, 2'3 catalysts for various reactions have been devised by using mononuclear or multinuclear metal centers, a : Catalytic antibodies are examples of artificial enzymes based on biopolymers. 8-11 Artificial enzymes have also been designed by using synthetic polymers including synthetic polypeptides. 12-17 Artificial enzymes may be divided into two categories: semisynthetic artificial enzymes and synthetic artificial enzymes. Semisynthetic artificial enzymes are partly prepared by biological systems. Catalytic antibodies are typical examples of semisynthetic artificial enzymes. Semisynthetic artificial enzymes are also prepared by modification of a known protein or enzyme at a defined site with a cofactor or new functional group. 18-22 Synthetic artificial enzymes are prepared totally by synthetic methods. Synthetic artificial enzymes may be either relatively small molecules with well-characterized structures or macromolecules. The term synzymes has been coined to designate synthetic polymers with enzyme-like activities. 12 In addition, synthetic artificial enzymes are also obtained with molecular clusters such as micelles 23'24 and bilayer membranes 25-28 formed by amphiphiles. The goal in the area of artificial enzymes may be summarized as design of enzyme-like catalysts that equal the effectiveness of enzymatic catalysis and overcome limitations of enzymatic system such as instability toward heat, incompatibility with organic solvents, inapplicability to abiotic reactions, and too-narrow selectivity with respect to the substrate structure. At the present stage, however, studies on artificial enzymes are primarily aimed to reproduce characteristics of enzymatic action such as complex formation with substrates, high degrees of acceleration, and high selectivity. Active sites of enzymes are responsible for the effectiveness of enzymatic catalysis. High selectivity in a reaction type catalyzed by each enzyme as well as high stereoselectivity such as regioselectivity and enantioselectivity originate from complexation of the substrate by the active site. Enzymes recognize their own substrates with typical formation constants of 104105. For molecular recognition of the substrates, enzymes utilize various types of intermolecular forces 29'3~such as ionic interaction, coordination, hydrogen bonding, and hydrophobic interaction. Upon complex formation, the substrate loses its freedom and the reaction between the enzyme and its substrate becomes an intramolecular process instead of an intermolecular one. Sometimes, strain can be induced in the structure of substrate upon complexation with the enzyme. 29 The main function of enzymes is not the recognition of substrates but the reduction of free energy of activation. If complexation of the substrate is not coupled with stabilization of the transition state, effective molecular recognition and stabi-

Designing Artificial Enzymes

247

lization of the substrate result in rate retardation instead of rate enhancement. The main object for the design of artificial enzymes, therefore, can be summarized as the molecular recognition of substrates and the more effective molecular recognition and stabilization of transition states. To design effective artificial enzymes, it is highly desirable to construct artificial active sites comprising both binding sites for the substrates and multiple catalytic groups to be used for transformation of the complexed substrates. The catalytic groups must be located in productive positions to maximize the catalytic efficiency as illustrated by the cartoon of 1, which represents the structure of the transition state. Construction of such active sites with small synthetic molecules would be very difficult. Several catalytic elements are to be placed on the molecular framework. Furthermore, those catalytic elements should take productive positions and the conformational freedom of the molecular framework should be controlled to maintain the productive conformation. Thus, a large amount of laborious computational and skillful synthetic work is needed to synthesize such active sites. Instead, synthetic as well as natural macromolecules have been frequently chosen as the backbone of artificial enzymes. Nature has adopted polypeptide as the backbone of the catalysts for fine tuning of the positions and the reactivity of the convergent catalytic elements. Detailed mechanisms of enzymatic actions were difficult to understand in the early stage of enzymology; similarly information on the structure and mechanism of artificial enzymes built on skeletons of natural or synthetic macromolecules is not easy to obtain. At present, in the area of artificial enzymes such as catalytic antibodies or synzymes, major efforts are made to develop new strategies for designing the active site. Whether the strategy is successful is judged by the activity of the resulting artificial enzymes, although the structure of the active site may not be fully characterized and the mechanism of catalysis may not be understood on the molecular level. Strategies developed to date for designing active sites of catalytic antibodies are well documented. 8-11 In this chapter, methodologies to design active sites of synthetic artificial enzymes are reviewed.

248

JUNGHUNSUH

Efforts to design small molecules equipped with multiple catalytic elements, improvement of properties of various catalytic elements by microdomains mimicking those of enzymes, and design of active sites by molecular imprinting will be discussed first. Then, various new strategies developed in our laboratory for construction of active sites of synthetic artificial enzymes will be presented. They include (1) design of active sites containing both binding and catalytic sites by random functionalization of polymer skeletons, (2) design of active sites by site-directed functionalization of polymer skeletons, (3) design of active sites by cross-linkage of preassembled catalytic elements, and (4) design of active sites by self-assembly from catalytic elements. For construction of active sites of synthetic artificial enzymes, it is necessary to combine various concepts and methodologies developed in supramolecular chemistry with the principles of enzymology. Synthetic artificial enzymes as effective as natural enzymes may not be obtained in the near future. Even synthetic artificial enzymes with a modest degree of efficiency, however, may be useful for practical applications, and some primitive forms of synthetic artificial enzymes have been commercialized. 0

SYNTHESIS OF SMALL MOLECULES EQUIPPED WITH MULTIPLE CATALYTIC ELEMENTS

In the area of molecular recognition, many attempts have been made to design host molecules capable of complexation of guest molecules. 31 Host molecules that can recognize and stabilize transition states would behave as effective artificial enzymes. Attempts to design artificial enzymes based on host molecules with two or more catalytic elements will be briefly mentioned here before discussing construction of active sites of artificial enzymes on macromolecular skeletons. To mimic many phosphoesterases activated by two or more metal ions, several artificial phosphoesterases have been devised by linking two metal centers with a spacer group. 32-41 For example, two triaza ligands were linked by hydroxo (2) 36 or a biphenyl (3) 37 spacer. The dimetallic complexes of the bistriaza ligands

R = CH 3 or CH2CH(CH3)(OH )

O2N

/

N

NH ( ~ ~ 1 7 6 Ar

Ar

3

Designing Artificial Enzymes

249

accelerated hydrolysis of phosphoester linkages of RNA or DNA models remarkably. The degree of catalysis was much greater for the dinuclear metal complexes comparedwiththecorrespondingmononuclearspecies,suggestingparticipationof two metal centers in catalysis. In the mechanism of 2, double Lewis acid activation of the phosphate linkage by the metal centers is proposed in addition to the nucleophilic attack by a metal-bound oxide ion. In the mechanism of 3, one metal center anchors the anionic phosphate diester, whereas the other metal center delivers a nucleophilic hydroxide ion. Various repertories 2'3 of metal ions acting as Lewis acid catalysts such as recognition of anions, masking of anions to facilitate approach of anionic nucleophile, activation of electrophiles, and ionization of weak acids are involved in the mechanisms of 2 and 3. In 3, one metal ion may be considered as a binding site and the other as a catalytic group. So far, cyclodextrin (CD: 4) derivatives have been the most successful artificial enzymes based on small synthetic host molecules. Since CD derivatives form inclusion complexes with various hydrophobic molecules, they have been utilized as binding pockets. 42-5~ Examples are illustrated by 5 and 6. The CD derivative indicated in 5 43 is a mimic of ribonuclease in which the two histidyl imidazoles of the active site are believed to act as a general acid and a general base catalyst. The CD dimer indicated in 6 50 acts as an artificial metalloesterase manifesting selectivity toward esters with two hydrophobic groups.

HO 0

0

_--

4 (CD)

H

(n: 6-8)

o

5

6

250

JUNGHUN SUH IMPROVEMENT OF CATALYTIC PROPERTIES BY M I C R O D O M A I N S OF POLYMERS

Q

Medium properties of active sites of enzymes are distinctly different from those of the bulk water. This leads to improvement of abilities of organic or inorganic functional groups of the active site to complex the substrate and to transform the bound substrate to the product. For artificial enzymes, therefore, improvement of catalytic properties in the microenvironments of the active sites makes a significant contribution to overall catalysis. Micelles 23'24 and vesicles 25 create apolar microdomains in water by aggregation of amphiphiles. Complexation of apolar substrates and improvement of some catalytic functional groups have been reported with micelles and vesicles. Unless active sites comprising multiple catalytic groups are covalently attached to the molecular skeleton, design of artificial enzymes with micelles and vesicles is not easy. In this regard, micelles or vesicles have limited scope for the design of artificial active sites. A wide range of linear and branched polymers have been examined as backbones of synthetic artificial enzymes. Derivatives of linear polymers such as poly(4vinylpyridine) (7), poly(N-vinylimidazole) (8), poly[4(5)-vinylimidazole] (9), poly[5(6)-vinylbenzimidazole] (10), and copolymers such as poly[N-vinylpyrroline-co-4(5)-vinylimidazole] have been tested as synthetic artificial enzymes. 51-56 Branched poly(ethylenimine) (PEI; 11) contains ethylamines as the repeating unit. About 25% of the amino groups of PEI are primary amines, about 50% are secondary amines, and about 25% are tertiary amines. 57 The tertiary amino nitrogens represent the branching points on PEI and, therefore, PEI is highly branched and globular. PEI has been intensively exploited as the backbone of synthetic artificial enzymes due to its branched structure, high solubility in water, and easy modification by alkylation, acylation, or imine-formation of the nitrogen atoms. 2'3'12-15'58'59Active sites comprising several catalytic elements would be much more easily constructed on branched backbones than on linear backbones. Hydrophobic domains are created on synthetic polymers by attaching apolar groups to the polymer backbones. Polymers such as the N-alkylated forms of poly(4-vinylpyridine) and poly(N-vinylimidazole), 52'6~ or the N-alkylated or Nacylated derivatives of PEI contain both hydrophobic groups and charged centers

C--C ~,~

C--C,

13

7

n

8

n

9

10

Designing Artificial Enzymes

251 H

H

H H HN... . . . . . . . NS~, N~/'N% N,/" N'~, N,~ N% N,/~ N'%, Ni"I2 H H

11 (],EI)

H ~M

I~N'~,N,/~N,/~N~NH . . . . . . . . . . . . . . . . . . H

and, thus, may be considered as analogues of micelles without the basic instability element of micelles. The apolar microdomains created on the polymer attracts many organic substrates from the bulk water. Complexation of substrates often leads to the increases in the effective concentrations of the substrates toward the catalytic groups attached to the polymer. Ionic sites on a polymer attract species with opposite charges from the bulk water. For example, amide hydrolysis of 12 is catalyzed by PEI derivatives mainly due to the increased local concentration of hydroxide ion in the vicinity of cationic sites of the polymers. 61 Some reactions such as decarboxylation of 6-nitrobenzisoxazole-3-carboxylate (13) and nucleophilic aromatic substitution of 4-chloro-3,5-dinitrobenzoate (14) with azide ion are greatly accelerated simply by reducing the polarity of the medium. These reactions are accelerated by 100-1300 times by PEI derivatives containing lauryl (C12H25) chains. 62'63 Anionic charges are dispersed in the transition states for the reactions of 13 and 14, whereas they are localized in the respective reactants. Reduction of polarity of the medium can greatly accelerate such reactions. 64 The Kemp elimination, 15, is a well-characterized model for proton transfer from carbon and the rate is sensitive to solvent polarity. A subset of several hundred water-soluble polymers were prepared by alkylating PEI with different combinations of lauryl, benzyl, and methyl groups. The Kemp elimination was accelerated by up to 106 times. 15 Depending on the type of organic reactions, therefore, pronounced catalytic effects are manifested by artificial enzymes simply due to medium effects.

O COOH CF3"--~~NO 2

c, NO2

NO 12

"

+O CO 2-

13

14

JUNGHUN SUH

252

~OAr

~')H

15

16

On the other hand, reactions for which the transition states have more localized charge distribution compared with the reactants (e.g. 16) are retarded when the microenvironment becomes less polar. 65 Active sites of some enzymes possess both hydrophobic walls and ionic functional groups. 66 This unique medium has been mimicked with PEI by attaching both multipositive cationic sites and hydrophobic lauryl groups. 65 In this unique microenvironment built on PEI, the intrinsic nucleophilicity of N,N-dialkylaminopyridyl groups toward a nitrophenyl carboxylate (16) was considerably enhanced. Various functional groups, related to the polar groups present in the side chains of amino acids, have been attached to artificial enzymes as catalytic groups. Pyridyl, N,N-dialkylaminopyridyl, imidazolyl, or benzimidazolyl groups of linear polymers 7-10 as well as those attached to PEI are related to histidine. 6~ Nucleophilicity of nitrogens of pyridyl, imidazolyl, or amino groups are masked when the nitrogens are protonated. Hydrophobic microenvironments on artificial enzymes suppress protonation of nitrogen atoms, 65'68raising the fraction of the reactive basic form at neutral or acidic pHs. This sometimes counterbalances the medium effects which reduces the intrinsic nucleophilicity of nitrogen nucleophiles toward active esters aforementioned with 16. In relation to glutamate or aspartate, a carboxyl group has been attached to artificial enzymes. For example, hydrolysis of a Schiff base was catalyzed by a PEI derivative containing both lauryl group and carboxylate anion 17. 69 The reactivity of the carboxylate group acting as a general base was considerably enhanced in the apolar microdomain of the PEI derivative, presumably due to the medium effects. As oxygen nucleophiles related to serine or tyrosine, oximes (e.g. 18), or hydroxamic acids (e.g. 19) have been attached to artificial enzymes. In cationic domains

'H 17

~

H 18

C'H~ 19

253

Designing Artificial Enzymes R I

0

(X or Y: polymer backbone) 20

of polymers, ionization of oximes are enhanced, raising the fraction of the nucleophilic forms. 70-72

When positioned in hydrophobic microenvironments, anionic nucleophiles are partially desolvated. 6~ Whether this results in acceleration, however, depends on the type of the reaction involved. 29 If the charge is more delocalized in the transition state than in the reactant, as in 13 and 14, desolvation of the nucleophile would selectively destabilize the reactant, leading to rate enhancement. On the other hand, if the charge is more localized in the transition state, the reaction would be retarded in hydrophobic microdomains. 73'74 Sulfhydryl groups attached to polymers are easily ionized to form the nucleophilic anionic form in cationic domains, manifesting enhanced reactivity toward labile esters. 75'76When labile esters are used as substrates, acylation of the amines, oximes, or sulfhydryls forms intermediates which are more stable than the ester substrates, seldom leading to overall catalysis in the ester hydrolysis. Some enzymes require participation of coenzymes or metal ions for catalytic activity. Flavin derivatives (20) can be considered as an example of coenzyme analogues attached to artificial enzymes. The reactivity of ravin derivatives attached to a PEI derivative or a cationic polystyrene derivative was considerably greater than that of monomeric ravin analogues. 77'78 Efforts have been made to design artificial enzymes with enantioselectivity by introducing chiral microenvironments on polymers. For example, PEI derivatives containing L-histidine moieties were prepared, and up to 3.6-fold rate difference was observed for the hydrolysis of D- and L-N-carbobenzoxyamino acid p-nitrophenyl esters. 79

Q

DESIGN OF ACTIVE SITES ON POLYMERS BY MOLECULAR IMPRINTING 16 80-82

Binding sites have been produced in polymers by molecular imprinting. ' As illustrated by the cartoon of 21, monomers with functional groups are bound by a template and then copolymerized under conditions leading to the formation of

254

IUNGHUN SUH

highly cross-linked polymers with chains in a fixed arrangement. After removal of the template, a binding site that recognizes the template can be generated. Hosts for organic compounds have been designed by molecular imprinting as exemplified by 22. 80 Here, a host for N-benzoyl-D-valine was prepared by using the Co(III) complex of the guest molecule as the imprinted unit. Artificial enzymes have been prepared by using transition-state analogues as the template through molecular imprinting as exemplified by 23. 81 Here, the original geometry of the cavities is apparently conserved even after release of the template molecule, readily accommodating the substrate with similar geometry. The amino group introduced to the cavity wall is proposed to act as a general base catalyst for dehydrofluorination of the bound substrate. Molecular imprinting is not applicable to production of soluble synthetic polymers. Most of the active sites prepared by molecular imprinting may be embedded

o...jd

r

"

-

ttCI

/Bz-l)-Val

recognizes Bz-D-Val

22

Designing Artificial Enzymes

255

~

6

co,-

0

c.o~-

*H3N

[~NH

,,,,,

o

AIBN

DMF

\\

o'j'-

1

.
23

,,\

NO2

within the insoluble polymeric body and, therefore, substrates would have limited accessibility to the active sites. Moreover, the functional groups attached to the active sites would have very limited flexibility. Rigidity of the polymeric backbone holding the functional groups may be helpful for recognition of the guest molecules which are related structurally to the template used for preparation of the polymer. For catalysis, however, flexibility of the catalytic groups could be very important.

5. DESIGN OF ACTIVE SITES CONTAINING BOTH BINDING AND CATALYTIC SITES BY RANDOM FUNCTIONALIZATION OF POLYMER SKELETONS Major obstacles faced in the early stage of design of synthetic artificial enzymes were the limited solubility of polymer derivatives in water and the lack of specific binding sites on polymer skeletons. Linear polymers containing quaternary ammonium ions and PEI derivatives are reasonably soluble in water. Without specific binding sites for substrates, however, these polymer derivatives may be simply regarded as polymicelles. Macrocyclic metal centers have been constructed on PEI through metal-template condensation of PEI with dicarbonyl compounds (24a-d). 83-86 For example, 24d

IUNGHUN

256

SUH

02N I

24a

24b

24c

24d

25

was prepared by condensation of PEI with butanedione in the presence of Ni(II) ion. The macrocyclic metal center provides the polymer with binding sites capable of recognition of anionic esters (e.g. 25). The anionic ester included in 25 was effectively bound by the macrocyclic metal centers of 2 4 a - d with formation constants of 300-400, whereas no kinetic evidence was obtained for complexation of the corresponding neutral ester (o-nitrophenyl acetate). Due to the anchoring effect described in 25, deacylation through the nucleophilic attack of the amino groups of PEI was up to 100-times faster for the anionic ester compared with the neutral ester. 83 The macrocyclic metal centers were also effective catalysts for hydrolysis of bis(p-nitrophenyl) phosphate (BNPP) and p-nitrophenyl phosphate (NPP). DNPP has been extensively studied as a DNA model. Among various macrocyclic metal centers built on PEI, the one (26) obtained by condensation of PEI with glyoxal in the presence of Co(II) ion was particularly effective in the hydrolysis of the phosphoesters. 86 The half-lives for spontaneous hydrolysis of BNPP and NPP are reported as 2000 years and 4 years, respectively, at neutral pHs and 25 ~ In the presence of 1 mM metal centers of 26 at pH 7 and 50 ~ the half-lives for hydrolysis of BNPP and NPP were 80 min and 10 min, respectively. PEI derivative 26 was a far better catalyst than any other synthetic catalyst containing dipositive metal ions reported for hydrolysis of phosphoesters prior to this work. The mechanism of 27a-c may be operative for 26 by analogy with other 4'87'88 transition metal catalysts. The effectiveness of 26, a PEI derivative containing many macrocyclic metal centers, may be due to collaboration between two adjacent metal centers as indicated by 27d. Collaboration between two or more metal ions is believed to

/ ! - _1.-o ~ ,N ~co" \

26

/

I

H, v , _0\ 0 ~ A r Ar 27a

/ ",~_

0~0 - - H Ar

\

Ar

27b

27c

/ "O~Ar \Ar

o

27d

257

Designing Artificial Enzymes

operate in several metallophosphoesterases. 89 In 27a-d, the metal ion acts as both the binding site and the catalytic site. As binding sites for certain hydrophobic moieties, [3-CD has been attached to PEI. 9~The PEI derivative (CD-PEI: 28) containing ~ C D possesses several amino groups (represented by gray circles in 28) located above the CD cavity. Deacylation of 29 and 30 in the presence of CD-PEI manifested saturation kinetic behavior, which revealed kcat = 3.5 x 10-3 s-1 and Km = 6.0 x 10-5 M (catalyst concentration was expressed as concentration of the CD moiety) for 29 and kcat - 6.8 x 10-2 s-1 and Km = 6.0 x 10-4 M for 30 at pH 7.65 and 25 ~ Under identical conditions, deacylation of 29 or 30 by PEI or CD was insignificant. The kinetic data indicate effective complexation of the esters by CD-PEI, and, therefore, the CD cavity of CD-PEI acts as a specific binding site for the tert-butylphenyl derivatives (31). Complexation of the esters by the CD cavity increases the effective molarity of the amino groups of PEI towards the ester linkage of the substrate, leading to enhanced deacylation rates. Random functionalization of CD-PEI or PEI containing macrocyclic metal centers led to PEI-based artificial enzymes containing both catalytic functional groups and binding sites. For example, a PEI derivative (32) containing lauryl groups (content: 11 residue mol%), N,N-dialkylaminopyridyl groups (content: 5 residue mol%), and macrocyclic Ni(II) centers (content: 1 residue mol%) has been prepared. 65 In addition, the primary and the secondary amino groups of the PEI

COOH

1-i-.o,

COOH

28

29

[~NO=

30

31

Ac\ ('" PEI~/Ac N

N~'-N

C12H25 H

H

32

258

JUNGHUN SUH

backbone were blocked by acetylation. Kinetic data of the hydrolysis of the anionic nitrophenyl ester included in 25 were collected for 32 and its analogues lacking lauryl groups, NJV-dialkylaminopyridyl groups, and/or macrocyclic Ni(II) centers. Saturation kinetic behavior was observed for 32 with kcat = 7.4 x 10-2 s-1 and K m = 1.5 x 10-4 M (catalyst concentration was expressed as the concentration of N,N-dialkylaminopyridyl moiety). For the analogues of 32 lacking the metal centers, saturation kinetic behavior was not observed. For the analogues of 32 lacking the pyridyl groups, the ester hydrolysis was not appreciable. For the analogues of 32 that contain the pyridyl groups but lack lauryl groups and/or the metal centers, kcat/Km was 50-250 times smaller than 32. The kinetic data indicate that the macrocyclic metal center acts as a binding site, anchoring the anionic ester. The aminopyridyl group makes nucleophilic attack at the bound ester. The lauryl and the cationic groups exert favorable medium effects on both complexation and nucleophilic attack. Since the metal centers and the aminopyridyl groups are randomly attached to the polymer skeleton, nucleophilic attack by the aminopyridyl group at the anionic ester anchored by the metal center would involve conformational change of the polymer backbone as illustrated by 33. A CD-PEI derivative containing macrocyclic metal centers have been prepared by random functionalization (34). 91 Here, the secondary and the tertiary amino groups were blocked by acetylation. Hydrolysis of a nitrophenyl carboxylate ester

O,N L ,~

34

02N

(M: NiI,Co n,or Zn") (R: H or CH3)

33

Designing Artificial Enzymes

259

R HO. ,,~-~N_

O2N

ra ' / N- "M~n- .-N

O2N 0

DR

35

containing tert-butylphenyl group was considerably catalyzed by 34. No catalysis was observed by the analogue of 34 lacking the metal centers. Effective complexation of the ester substrates by 34 was manifested by Km of (0.2-1.3) x 10-3 M (catalyst concentration was expressed as concentration of the CD moiety). Thus, the CD cavity acts as the binding site and the metal center as the nucleophile. Since the metal centers and the CD cavities are randomly attached to the polymer skeleton, collaboration of the metal center and the CD cavity would involve conformational change of the polymer backbone as illustrated by 35. As a new macromolecular backbone of immobilized artificial enzymes, we have prepared poly(chloromethylstyrene-co-divinylbenzene) (PCD) (36). 92 Here, chloromethylstyrene monomer contains ca. 70% and 30%, respectively, of meta- and para-isomers and divinylbenzene is a mixture of isomers. Divinylbenzene serves as a cross-linking group and, therefore, PCD is highly branched. The shape of PCD synthesized with 2 mol% of divinylbenzene taken by scanning electron microscopy is illustrated in Figure 1. PCD is a copolymer of styrene and divinylbenzene where all the styrene groups are chloromethylated. When the content of divinylbenzene was 2 mol%, the chloromethyl groups were found to remain almost intact during suspension polym-

cl Ar Ar Ar Ar Ar A r ~ 1 %

~

%,,,,,._/

At

Cl

36

-,.,

Ar Ar Ar~ A r ~ , - - J

I

Ar ._ ~ / ~ , ~ C l

(m or p)

PCD

JUNGHUN SUH

260

Figure 1. Scanning electron micrograph of typical beads of PCD (a) and PCD derivative 39 (b).

erization in water. This is based on the amount of triethylamine that reacted with the resulting PCD. Merrifield's peptide resin is another example of copolymer of styrene and divinylbenzene. In Merrifield's peptide resin, the benzene tings are only partially chloromethylated. The chloromethyl group of PCD is readily modified, allowing incorporation of various catalytic elements. In addition, the properties of the bead can be controlled by substitution of the chloromethyl group. By random substitution of the chloro groups of PCD with various amines, cyclen (37) and guanidinium were attached to PCD. 92 After treatment of the remaining chloromethyl group with methoxide and complexation of Cu(II) ion to the cyclen, PCD derivatives 38 and 39 were prepared. The contents of the Cu(II)-cyclen moiety were 1.7 and 2.6 residue mol%, respectively, for 38 and 39 and that of the guanidinium group in 39 was 2.6 residue mol%. The shape of 39 taken by scanning electron microscopy is illustrated in Figure 1.

H3CO

("")

co A, IN

N H

37

38

39

Designing Artificial Enzymes

261

The guanidinium ion recognizes anions such as carboxylate or phosphate ester anions. For example, the guanidinium ion of Arg-145 of carboxypeptidase A plays an essential catalytic role by recognizing the carboxylate anion of the substrate. 93 Complexation of the phosphate ester anion or carboxylate ion by the guanidinium ion in water becomes considerably stronger as the microenvironment is changed from bulk water to micelles, bilayer membranes, or air-water interface. 94'95 The ability of the guanidinium ion to bind carboxylate anion or phosphate ester anions would be improved at the interface between water and the beads of PCD derivatives. The guanidinium ion therefore can be utilized as binding sites for substrates containing carboxylate or phosphate ester anions such as proteins or nucleic acids. Metal ions play versatile roles as Lewis acid catalysts in organic reactions. 2'3 For example, metal complexes of cyclen derivatives have been used as catalytic centers for hydrolysis of phosphate esters including DNA, RNA, and cAMP. 96-99 Several metal complexes that catalyze hydrolysis of peptide bonds are reported and the mechanisms of the catalytic peptide hydrolysis are well documented. 2'3 In this regard, the metal centers created by attachment of the Cu(II) complex of cyclen to PCD can be utilized as catalytic groups for peptide hydrolysis. The value of log Kf for the Cu(II) site of 39 was measured by a competition experiment using dipicolinic acid as the chelating reagent of Cu(II) ion. Among the Cu(II) ions present in 39, 22.7% were bound to the resin so strongly that they were not extracted by dipicolinic acid. For the remaining Cu(II) ions complexed to the resin, log Kf was estimated as 14.66 + 0.04 at pH 7.00 and 25 ~ As illustrated in Figure 2, electrophoresis (SDS-PAGE) revealed facile cleavage of both the heavy (MW 50 kDa) and the light (MW 25 kDa) chains of bovine serum y-globulin (Gbn) during incubation with 38 or 39. The rate of protein cleavage was measured by following the decrease in the density of the bands corresponding to the heavy and the light chains. Rate data were collected by varying the amount of the catalyst (Co). Here, COwas expressed as the concentration of the Cu(II) complex of cyclen obtainable when the resin is assumed to be dissolved. As illustrated in Figure 3 for the dependence of ko on Co, saturation kinetic behavior was observed at pH 4.5-7 for the cleavage of both the heavy and light chains of Gbn by 39. On the other hand, ko was proportional

Figure 2. Results of electrophoresis performed with Gbn incubated with 39 at 4 ~ and pH 7.00 (So = 8.54 • 10-7 M, Co = 8.24 • 10-4 M). The upper and the lower bands correspond to the heavy and light chains, respectively, of Gbn.

262

IUNGHUN SUH

_

a

I=

im

b

E

o

0

/,.I ..*- ............. * d...,,.-:::::::::::::::::::::::: ..... : ....... . ......... : - - - -

-

0

.~ o~II ....

1

2

"~176176176

3

Co, 10 "3 M Figure 3. The plot of ko against Co. for the hydrolysis of the light (a; A) and the heavy (b; a) chains of Gbn catalyzed by 39 and of the light (c; 9 and the heavy (d; .) chains of Gbn catalyzed by 38 at 4 ~ and pH 7.00

to Co at pH 3 and 8-10. Saturation kinetic behavior was not manifested by 38. From the dependence of ko on C o, kca~Km was estimated, which revealed that 39 was 12-15 times more reactive than 38. The pH dependence of kca~Km indicated that the optimum pH range is 5-7 for both 38 and 39. Half-lives for peptide bonds under the conditions of spontaneous hydrolysis at 25 ~ and pH 7 is approximately 1000 years. 1~176176 The fastest data point indicated in Figure 3 is associated with a half-life of 40 min at 4 ~ This half-life is about 10s-times shorter than that of the spontaneous hydrolysis of peptide bonds. Analysis of the reaction products by MALDI-TOF mass spectrometry revealed that proteins smaller than 5kDa were obtained from the hydrolysis of the heavy (50 kDa) and the light (25 kDa) chains. Initial cleavage of the heavy and the light chains of Gbn by 39 should produce intermediate proteins greater than 25 kDa and 12 kDa, respectively. Accumulation of such intermediate proteins was not observed by SDS-PAGE electrophoresis. Cleavage of the intermediate proteins into peptides smaller than 5 kDa is, therefore, much faster than the cleavage of the parent proteins. Saturation kinetic behavior was manifested by 39 at pH 4.5-7, indicating much stronger complexation of Gbn to 39 compared with 38. Thus, the guanidinium moieties of 39 acted as the binding site for Gbn recognizing the carboxylate ions of Gbn. The reactivity of the Cu(II) complex of cyclen toward Gbn was enhanced by > 104 times upon attachment to PCD containing guanidinium ions. The microenvironment of Cu(II)-cyclen moieties in the gel phase of the polymer surface contains both hydrophobic and ionic characters. The enhanced reactivity of the Cu(II) center may be related to the unique medium properties.

Designing ArtificialEnzymes

263

Several homogeneous synthetic artificial enzymes 1~176 and catalytic antibodies 1~176 with proteinase activity have been reported. The monoclonal catalytic antibody prepared with a phosphinate hapten exhibited optimum activity at pH 9.5. The kcat measured with an amide substrate at pH 9 and 37 ~ was 1.65 x 10-7 s-1.1~ Thus, the half-life is 49 days when the substrate is fully complexed to the active site of the catalytic antibody. A much more improved antibody catalyst for amide hydrolysis has been elicited very recently by a joint hybridoma and combinatorial antibody library approach. 1~ The kcat measured with a primary arnide substrate at pH 9 and 25 ~ was 5 x 10-5 s-1 for this new antibody. This corresponds to a half-life of 4 h when the substrate is fully complexed to the active site. The half-lives for the amide hydrolysis catalyzed by the antibodies are much longer than that (10-30 min at pH 4.5-7 and 4 ~ when the substrate is fully complexed to the active site) of the light chain of Gbn hydrolyzed by 39. The fastest protein cleavage recorded so far with artificial proteinases is the cleavage of chymotrypsin by a coordinatively polymerized bilayer membrane which will be discussed later in this review. The coordinatively polymerized bilayer membrane achieved half-life as short as 3 min at 4 ~ and pH 5.5-9.5.1~ This is several times faster than the hydrolysis of Gbn by 39. In terms of utility in practical applications, however, 39 is more useful than the artificial enzyme based on the bilayer membrane due to the immobile nature of the former as well as the intrinsic instability of bilayer membranes.

6. DESIGN OF ACTIVE SITES BY SITE-DIRECTED FUNCTIONALIZATION OF POLYMER SKELETONS For designing effective artificial enzymes or receptors on synthetic macromolecules, it is necessary to develop a methodology to introduce an additional functional element in the vicinity to the functional group already present on the molecular backbone. As suggested by 33 and 35, introduction of an additional catalytic element in the vicinity to the initially attached element would raise the population of catalytically productive conformations. In this regard, we have developed the methodology of site-directed introduction of the second group by using the first group as an anchor. The idea of site-directed functionalization of the polymer skeleton is illustrated by the cartoon of 40. Site-directed functionalization of CD-PEI was achieved with tert-butylphenyl ester 41 which contains a precursor of 1,5,9-triazacyclododecane (TC). 1~ By taking advantage of the recognition of tert-butylphenyl moiety by CD, PEI was acylated with 41 in 9% (v/v) DMSO-water. After a few additional steps of modification, PEI derivative 42 which contains both the metal (Cu(II), Ni(II), or Zn(II)) complex of TC and CD cavity was prepared. The content of CD and TC in 42 was 1.2 residue mol%. The primary and the secondary amino groups of 42 were blocked by acetylation. In a separate preparation, TC was attached to PEI randomly by

264

IUNGHUN SUH

acylation of PEI in DMSO with an analogue of 41 containing the phenyl leaving group instead of the t e r t - b u t y l p h e n y l group. That the TC group is positioned near the CD moiety in 42 was confirmed by measuring kinetics of deacylation of 43 promoted by 42 [M: Cu(II), Ni(II), Cu(II)] and the analogue of 42 prepared by random functionalization. Kinetics of reactions catalyzed by the PEI derivatives followed the Michaelis-Menten scheme. Parameter 1/K m is close to the formation constant for the most stable complex formed between the polymer and 43. For the PEI derivative prepared by site-directed functionalization, 1/K m was (3.7-6.4) x 103 M -1 at 25 ~ being 5.8-8.7 times greater than those for the analogue prepared by random functionalization. This indicates that an extra binding force is present in the complex formed between 43

41

__._>

LL_

42

265

Designing Artificial Enzymes

43

44

and 42. On the other hand, the ester fully complexed to the catalysts was hydrolyzed 3.7--4.8 times more slowly by the PEI derivatives prepared by the site-directed functionalization compared with those by random functionalization as revealed by kcat ((2.4--4.4) x 10-3 s-1 for the randomly prepared polymer). The stronger but less productive binding of 43 by 42 compared with that prepared by random functionalization suggests the binding mode of 44. Interaction of the metal-bound water molecule of the TC complex with the carbonyl group of the bound ester facilitates the complexation. On the other hand, the assembly of 44 may sterically protect the ester linkage from attack by nucleophiles. The difference in 1/Km between 42 and its analogue prepared by random functionalization corresponds to a decrease in the free energy for complexation of 1.0-1.3 kcal/mol (-AAGf). Hydrogen bond energies in an enzyme-substrate complex have been estimated by mutagenesis. 1~ The energy for the hydrogen bond between a good hydrogen-bond donor on the enzyme and an uncharged group on the substrate was estimated as 0.5-1.5 kcal/mol. The value of 1.0-1.3 kcal/mol for -AAGf estimated in the study involving 42 agrees with the existence of one extra hydrogen bonding in 44 which is absent in its analogue prepared by random functionalization. In an effort to establish the methodology for construction of active sites by site-directed functionalization, 2,6-diacetylpyridineketoxime was attached to CDPEI. 1~ The site-directed functionalization of CD-PEI was carried out by acylation in 13-19%(v/v) DMSO with tert-butylphenyl ester 45 of a carboxylic acid containing 2,6-diacetylpyridine followed by conversion of the carbonyl groups to oximes by treatment with hydroxylamine, leading to the formation of 46. By acylation in DMSO of CD-PEI with the phenyl ester, instead of the tert-butylphenyl ester, an analogue of 46 was prepared by random functionalization.

266

JUNGHUN SUH

45

0

0

P

NH20H >

HO/N

N~OH

"N---i I

46 Ester hydrolysis of 4'-acetoxyphenylazobenzenesulfonate (47) was considerably enhanced by the Ni(II) or Zn(II) complex of 46 or its analogue prepared by random functionalization. Analysis of the kinetic data measured at various pHs revealed that kcat for 46 (kr = 1--4 x 10-3 s-1) was 3-6 times greater than that for its analogue obtained by random modification. This was taken to indicate that the 2,6-diacetylpyridineketoxime moiety was introduced to CD-PEI in vicinity to the CD moiety, but that the orientation of the 2,6-diacetylpyridineketoxime moiety and the CD cavity in 48 was not very productive for deacylation of 47 complexed by the CD cavity.

N ~N

0''0

i--

o*S o47

48

Designing Artificial Enzymes

267

The kinetic parameter 1 / K m or kcat for ester hydrolysis indicated by 44 or 48 is not very much larger than that manifested by the corresponding PEI derivatives prepared by random functionalization. The difference is, however, significant enough to support the introduction of the second catalytic groups in vicinity to the CD cavity by site-directed functionalization. The major obstacle to overcome at the present stage is how to suppress the conformational freedom of the resulting artificial active site to achieve effective cooperation among the catalytic elements positioned in proximity on the macromolecular skeletons. 0

DESIGN OF ACTIVE SITES BY CROSS-LINKAGE OF PREASSEMBLED CATALYTIC ELEMENTS WITH MACROMOLECULAR SPACER

In enzymes, several functional groups participate in catalysis and the conformation of the resulting transition state comprising the enzyme and the substrate is optimized to achieve effective catalysis. Polypeptide skeletons of enzymes are capable of fine alignment of convergent catalytic groups in the vicinity to the complexed substrates. In this regard, the polypeptide backbone of an enzyme may be regarded as a macromolecular spacer that connects various catalytic elements in productive positions in the transition state. If a template is used to assemble various functional groups and the preassembled functional groups are cross-linked with a synthetic macromolecular spacer, removal of the template would produce a site comprising those functional groups positioned in proximity as illustrated by the cartoon of 49. If those functional groups occupy proper positions to stabilize transition states of certain chemical reactions, an effective artificial active site would be obtained. This idea is somewhat different from that of molecular imprinting. In molecular imprinting, a macromolecular backbone is built by copolymerization using monomers preassembled around a template as indicated by the cartoon of 21. A polymer is prepared first and then used as a spacer in 49, whereas the final step is formation of cross-linked polymers in 21. Only insoluble powders are obtained by molecular imprinting. On the other hand, both soluble and insoluble materials are obtained by the method of 49, depending on the nature of the macromolecular spacer employed.

268

JUNGHUN SUH OH

H•O OH

N

o~

SO

o. o 1~~ ~176

o o~ OH

The idea of cross-linkage with a macromolecular spacer to obtain an effective host molecule according to the scheme of 49 was initially tested in the design of an enterobactin analogue. Enterobactin (50) is the strongest microbial siderophore, containing three catechol units connected by a spacer. As outlined by the scheme of 51, three molecules of an activated ester of 2,3-dihydroxyterephthalate were preassembled around Fe(III) ion and then cross-linked with PEI (11). ll~ The resulting enterobactin analogue formed a very strong Fe(III) complex, being rated among the best synthetic siderophores containing three catechol moieties. Therefore, the original geometry of the coordination sphere was effectively conserved during the cross-linkage step. The amide linkages contained in the enterobactin analogue built on PEI were subjected to hydrolysis upon treatment with acid to release the bound Fe(III) ion. In subsequent studies, cross-linkage of preassembled

O

oX

X

O

1

51

Designing Artificial Enzymes

269

Br N 52

Br

units with PEI was carried out by alkylation, instead of acylation, of the amino groups of PEI. The design of highly effective host molecules for metal ions by the cross-linkage method was demonstrated with 2,9-bis(bromomethyl)-l,10-phenanthroline (52). 111Cross-linkage with small spacers of two molecules of 2,9-diaryl-1,10-phenanthrolines coordinated to a metal template such as Cu(I) has been used in preparation of many catenands. 112 By the reaction of PEI with Cu(I)(52) 2 in a DMSO-methanol mixture, two molecules of 52 preassembled by Cu(I) ion were cross-linked with PEI to obtain [Culphen2]PApEI. The Cu(I) content of [CuIPhen2]PApEI was estimated as 0.56 residue mol% by ICP analysis. For [Culphen2]PApEI, the content of the phenanthroline moieties was estimated separately as 1.2 residue mol% by NMR. This corresponds to the presence of 2.1 phenanthroline moieties for each Cu(I) ion. By treatment with NaCN, the Cu(I) ion was removed to produce [Phen2]PApEI. The primary and secondary amino groups of [PhenE]PApEI were acetylated by treatment with excess acetic anhydride to obtain [PhenE]vgAcPEI. Preparation of PEI derivatives in which two preassembled phenanthroline moieties are cross-linked by PEI is schematically presented in the scheme of 53. By the reaction of 52 with PEI in the absence of Cu(I) ion, phenanthroline moieties were attached randomly to PEI to obtain [Phen]RanpEI in which the content of phenanthroline was 0.81 residue mol%. [Phen]RanPEI was converted to [Phen]aanAcPEI by acetylation of the primary and the secondary amino groups. The two preassembled phenanthrolines cross-linked by PEI would remain in close proximity when the conformational freedom of the resulting two macrocycles is sufficiently suppressed. If the two phenanthroline moieties are very close to each other, they will affect protonation of each other considerably. For [Phen]RanPEI and [Phen]RanAcPEI, the phenanthroline moieties attached randomly to the polymer behaved as monobasic species (54) with pKa of 1.2. For [PhenE]vgPEI or [PhenE]PAAcPEI, the pair of phenanthrolines preassembled by Cu(I) and then cross-linked with PEI behaved as one unit (55) in ionization with pKa of ca. 1.3 and ca. 10. The PKa2 of [Phen2]VgPEI or [PhenE]vAAcPEI is greater by ca. 9.5 pKa units than the corresponding PKal. This pKa difference is attributable to destabilization of the diprotonated form of 55 by electrostatic interaction between two cations and

270

JUNGHUNSUH

~..N~

N

Cu11115212

[Cu~

PAPEI

~f

Cu(,) ~ ~ ' ~ 1 N.

[Phen=]P,*PEI

[Phenz]PAAcPEI

9

m'"

'

53

[Cu"Phen2]pAAcPEI

pK.

N I

54

N

55

stabilization of the corresponding monoprotonated form by hydrogen bonding between two phenanthroline moieties. Both of these two factors originate from the close proximity between the two preassembled phenanthrolines. Close proximity between the phenanthrolines in [Phen2]PAAcPEI would lead to strong binding of metal ions. For the Cu(II) complex of [Phen2]PAAcPEI, the log Kf was 17.0 when the phenanthrolines were fully deprotonated. Effectiveness of

Designing Artificial Enzymes

2 71

cooperation between the two phenanthroline moieties located within a binding site for Cu(II) ion of [Phen2]PAAcPEI may be expressed in terms of effective molarity (EM). The idea of EM has been originally introduced as a measure of efficiency of intramolecular catalysis in comparison with intermolecular catalysis. 113The EM of a phenanthroline toward Cu(II) ion bound to the other phenanthroline in the phenanthroline pair (56) was estimated as 106 M. The EM measured for [Phen2]PaAcPEI may be compared with that for enterobactin. The EM of a catechol unit toward a Fe(III) ion bound to another catechol unit contained in enterobactin is estimated at 3 • 104 M. ~~ Enterobactin contains three catechol units, whereas the Cu(II) binding site of [Phen2]PAAcPEI consists of only two phenanthrolines. Nevertheless, the EM value observed for [Phen2]PAAcPEI is extraordinary for an artificial system. The stabilization of a phenanthrolinium cation by the adjacent phenanthroline is reflected by the large increase in pK a. Close proximity between the two phenanthrolines contained in a Cu(II) binding site is revealed by the EM value of 106 M. This indicates that the geometry of the coordination sphere of Cu(I)(52) 2 is effectively conserved during cross-linkage with PEI followed by removal of Cu(I) ion and acetylation of the primary and secondary amines of the polymer. The effective conservation of the preassembled geometry in [PhenE]PApEI or [Phen2]PAAcPEI is due to the highly branched structure of PEI which suppresses the conformational freedom of the two macrocycles. In the preparation of PEI derivatives 51 or [PhenE]PApEI, each building block (2,3-dihydroxyterephthalate or 52) is linked to PEI by double attachment. This would result in better conservation of the original geometry of the preassemblage during the cross-linkage step. When the resulting structure is utilized as the active site of an artificial enzyme, however, the double attachment may not lead to effective complexation of the substrate and fast structural conversion of the bound substrate. In this regard, cross-linkage of preassembled building blocks through single attachment may be more useful for designing active sites of artificial enzymes. Three molecules of 5-(bromoacetyl)salicylic acid (57) complexed to Fe(III) ion were cross-linked with PEI in DMSO by alkylation of amino groups of PEI with 57, leading to the formation of (FeSaI3)PEI, a water-soluble polymer, as indicated

56

272

JUNGHUN SUH

Be

0

0

57

OH "OH

o

0

0

0 "'~

:

O"

H

t:~

"0"

(FeSal3)PEI

0 .

.

58

apo(Sala)PEI

by the scheme of 58. Upon demetallation of (FeSala)PEI with HCI, apo(Sal3)PEI was obtained. 114 By random alkylation of PEI with 57, (Sal)ranPEI was prepared. Analysis of the Fe(III) binding data indicated that each Fe(III) binding site in (FeSal3)PEI contained three salicylate moieties. In addition, the log Kf revealed that the EM of the salicylate group toward Fe(III) ion bound to another salicylate moiety in each trisalicylate unit of apo(Sal3)PEI was ca. 1000 M. This is smaller than the EM demonstrated by phenanthrolines of [Phen2]PAAcPEI, reflecting the difference between single attachment and double attachment during the cross-linkage step. The EM of 1000 M estimated for apo(Sala)PEI shows that the geometry of the coordination sphere was well conserved during the cross-linkage with PEI of three molecules of 57 preassembled around Fe(III) ion. The trisalicylate units in apo(Sal3)PEI contains three carboxyl groups and three phenol groups as well as the amino groups of the PEI backbone. The carboxyl group is the functional group of aspartate or glutamate residue of enzymes, whereas phenol is that of the tyrosine residue, and the amino group is that of the lysine residue. Since these functional groups are positioned in proximity in apo(Sal3)PEI, we have tested whether they can catalyze organic reactions as those present in enzyme active sites. Many proteinases, such as serine or aspartic proteinases, utilize only the functional groups of amino acid residues to catalyze hydrolysis of proteins. For aspartic proteinases (also called as acid proteinases or carboxyl proteinases), such as pepsin, penicillopepsin, renin, or HIV protease, it is generally proposed that two carboxyl groups play crucial catalytic roles. 115 Electrophoresis (SDS-PAGE) of 7-globulin (Gbn) incubated with apo(Sala)PEI revealed facile cleavage of Gbn. 116Both the heavy (50 kDa) and the light (25 kDa)

Designing Artificial Enzymes

273

trans-1,2-cyclohexylened-

chains of Gbn were cleaved. Addition of Fe(III) ion or initrilotetraacetic acid, a powerful sequestering agent of Fe(III) ion, to apo(Sala)PEI did not affect the proteinase activity, excluding the possibility of participation of contaminating metal ions in protein cleavage. The rate of protein cleavage was measured by following the density of the electrophoretic band corresponding to the parent protein. Pseudo-first-order kinetic behavior was observed up to 50-80% of the reactions. The initial pseudo-first-order rate constant (kin = was estimated from the initial linear portion in the plot of log [S] against time. Dependence on CO (concentration of the trisalicylate moieties) of kin for cleavage of both chains of Gbn was examined under the conditions of CO> > SOand the results obtained at pH 6.00 are illustrated in Figure 4. The much slower rates observed with (Sal)ranPEI indicates that the activity of apo(Sal)3PEI originated from the trisalicylate moieties which had been originally preassembled by Fe(III) ion. Proportionality between kin and Co was observed for hydrolysis of both the heavy and the light chains catalyzed by apo(Sal)3PEI, indicating weak complexation of Gbn to apo(Sal)3PEI. The proportionality constant (k2) was estimated from the plot of kin against COat various pHs. The pH dependence of k2 is illustrated in Figure 5 for the two chains. The pH profiles indicate that the optimum pH range of apo(Sal)3PEI is 5-7. The fastest data point included in the plots of Figure 4 corresponds to half-life of 1 h at 50 ~ This half-life is more than 106 times shorter

vo/So)

~

.."~

2-

"

iii ~.

O'

0

2

b

4

Co, 10

6

-4

0

8

M

Figure 4. Plot of kin against Co for the hydrolysis of the heavy (o; line b) and the light (m; line a) chains of Gbn (5.35 x 10-6 M) catalyzed by apo(Sal)3PEI at pH 6.00 and 50 ~ Also included are the rate data for cleavage of the heavy (e) and the light (E3)chains of Gbn in the presence of (Sal)ranPEI at pH 6.00 and 50 ~ For the data points of (Sal)ranPEI, the total concentration ofalicylate residues on the polymer was divided by 3 in view of the definition of Co for apo(Sal)3PEI.

274

JUNGHUN SUH

.5~ .4

"

93

-

........

.

v•lr• '7,

0

&

0.0 3

4

5

6

7

8

9

pH Figure 5. pH dependence of/<2 for the hydrolysis of the heavy (o) and the light (m) chains of Gbn catalyzed by apo(Sal)3PEI at 50 ~

than that for the spontaneous hydrolysis of peptide bonds, since half-lives ofpeptide bonds in spontaneous hydrolysis at neutral pH and room temperature are about 1000 years as mentioned earlier. Electrophoresis of the reaction mixtures indicated that no intermediate proteins accumulated during the hydrolysis of both the heavy and the light chains of Gbn. MALDI-TOF mass spectrometric analysis of the product solution indicated that the products obtained were smaller than 5 kDa. Thus, cleavage of intermediate proteins into peptides smaller than 5 kDa is much faster than the cleavage of the parent proteins, with the degree of acceleration being considerably greater than that achieved with the two chains of Gbn. Apo(Sal)3PEI is the first organic artificial proteinase based on synthetic materials. Organic artificial proteinases based on biotic materials have been obtained with catalytic antibodies. 1~176 As discussed above, the best antibody catalyst for amide hydrolysis reported to date has been elicited recently by a joint hybridoma and combinatorial antibody library approach. 1~ The keat measured with a primary amide substrate at pH 9 and 25 ~ was 5 x 10-5 s-l. This corresponds to a half-life of 4 h when the substrate is fully complexed to the active site. The half-lives for the amide hydrolysis catalyzed by the antibodies may be compared with that (ca. 1 h at pH 7 and 50 ~ when the substrate is only partially complexed to the active site) by apo(Sal)3PEI. Effectiveness of apo(Sal)3PEI in proteolysis can be compared with that of amide hydrolysis by intramolecular catalysis. The most efficient intra-

Designing Artificial Enzymes o

275

OH

R'

NHR

R"

OH

O

~

~

R'

0 ~

O

~

R'

OH

R"

OH

O

+ RNH2

59

59a

molecular catalysis by organic groups in the hydrolysis of unactivated amide has been achieved with maleamic acid derivatives. Maleamic acid derivatives are hydrolyzed through nucleophilic catalysis by the intramolecular carboxyl group (59). 113'117 Half-life for hydrolysis of N-methyl maleamic acid (59a with R' = R " = H and R = CH3) at 39 ~ and optimum pH (< 3) is 3 h. 117 This is similar to that (ca. 1 h at 50 ~ of hydrolysis of Gbn with apo(Sal)3PEI (C O= 0.4-1 mM) at neutral pHs. The catalytic group is tethered to the substrate to raise the effective molarity in the maleamic acid whereas catalytic action proceeds as an intermolecular process in the case of hydrolysis of Gbn by apo(Sal)3PEI. The proteolytic activity of apo(Sal)3PEI originates from the three molecules of salicylates originally preassembled by the Fe(III) template. The three molecules of salicylates contained in the trisalicylate unit were linked to PEI by single attachment. The three proximal salicylates effectively cooperated with each other, leading to the fast hydrolysis of Gbn. The result indicates that the novel methodology employed here can become a valuable tool in the design of active sites of artificial enzymes.

0

DESIGN OF ACTIVE SITES BY SELF-ASSEMBLY FROM CATALYTIC ELEMENTS

The easiest way to construct active sites of artificial enzymes is self-assembly of the active sites as illustrated by the cartoon of 60. If the catalytic groups take highly productive positions in the self-assembled active site, cooperation among them can result in effective catalysis. It is difficult to induce the self-assembly in aqueous solutions by using hydrogen bonding or electrostatic interactions as intermolecular

276

]UNGHUN SUH NO 2

61

62

forces. On the other hand, hydrophobic interactions and coordination bonds can induce the self-assembly in water readily. Transesterification of 2-hydroxypropyl-p-nitrophenylphosphate (HPNPP), an RNA model, undergoes effective catalysis by participation of several metal ions (61). 33'118 Since the leaving phenol of HPNPP does not require activation, catalysis by metal ions of transesterification of HPNPP has been explained in terms of the activation of the two phosphoryl-oxygen bonds and the general base assistance for the intramolecular attack of the hydroxyl group. 1is Metal complexes of terpyridine (62: R = H) derivatives exhibit catalytic activity in transesterification of HPNPP lls as well as hydrolysis of RNAs. ll9 If several moieties of a metal-terpyridine complex assemble and take positions suitable for stabilization of the transitions state for transesterification of HPNPP as illustrated by 61, cooperation among them can lead to effective catalysis. Long alkyl chains such as the lauryl (Lau) group or bulky aromatic residues such as azobenzene derivatives attached to PEI form clusters to minimize exposure to water.12~ If several moieties of a metal-terpyridine complex assemble to form clusters on the PEI backbone to take positions suitable for stabilization of the transitions state for transesterification of HPNPP as illustrated by 61, cooperation among them can lead to effective catalysis. The Ni(II) complex (Ni(II)TP) of a terpyridyl derivative (TP: 62 with R = C6H4CH 2) and Lau group were attached to PEI in random combinations. 122 The catalytic activity per the Ni(II) center for transesterification of HPNPP varied by up to several thousand times or more depending on pH as the content of Ni(II)TP or Lau was changed. The best catalyst obtained was [Ni(II)TP]sLaul2PEI in which the contents of Ni(II)TP and Lau groups were 5 and 12 residue mol%, respectively. Although [Ni(II)TP]sLaul2PEI was obtained through a simple combinatorial approach, its catalytic activity expressed in terms of kr t (the first-order rate constant for conversion of the substrate fully complexed by the catalyst) was much better than those of previously reported dinuclear metal complexes (e.g. 63, 64) 33'41 whose structures were devised through deliberate planning. The fast transesterification of HPNPP in dinuclear complexes 63 was attributed to double Lewis acid activation of the phosphoryl group upon coordination to the

Designing Artificial Enzymes NO z ~,~"

277

CH=

L../ .

|

f5 " ' C u="

2*

~_Zn

. . .

Zn

\...j

I

(~u=" '~N

R =-CH2CH2OEt 63

64

two metal centers. 33']23For catalysis by other metal complexes in transesterification of HPNPP, the general base assistance for the attack of the hydroxyl group by the hydroxide ion bound to a metal center was proposed in addition to double Lewis acid activation of the phosphoryl group, as indicated by 61.118 A similar mechanism (65) may be proposed for the transesterification of HPNPP by [Ni(II)TP]sLaUl2PEI. In [Ni(II)TP]sLaUl2PEI, Ni(II)TP moieties and lauryl groups would form hydrophobic clusters. Although the terpyridyl groups are attached to PEI randomly, several terpyridyl groups can be positioned in proximity in the hydrophobic clusters. The two phosphoryl oxygen atoms of HPNPP may interact with two Ni(II) centers in the cluster, and the nucleophilic hydroxyl group may be subjected to general base assistance by hydroxide ion attached to another metal center. An amino group on the PEI backbone may play the role of general base catalyst instead of the Ni(II)-bound hydroxide ion. The effectiveness of [Ni(II)TP]sLau12PEI suggests that these catalytic groups occupy productive positions in the clusters of the polymer leading to efficient cooperation among the multiple catalytic groups. Thus, self-assembly of active sites on macromolecular

278

JUNGHUN SUH

backbones from catalytic elements and optimization of cooperative action among the catalytic elements through combinatorial attachment of pendants is achieved with [Ni(II)TP]sLaul2PEI. As a novel methodology for designing the active sites of artificial enzymes, we have developed self-assembly of active sites from metal ions and a ligand equipped with two chelating sites that can lead to coordination'polymers or molecular clusters containing a multiple number of metal centers positioned in proximity. Metal ions have been utilized as catalytic groups in many biomimetic catalysts since they are far more versatile than organic functional groups in catalyzing organic reactions. 2'3 The metal-bound water or hydroxide ions are also effective catalytic groups. Depending on the positions of adjacent metal ions, metal ions as well as metalbound _water or hydroxide can achieve effective cooperative catalysis. If organic functional groups are present in vicinity to the metal centers, collaboration among them are also possible. For many metalloenzymes, collaboration among one or more metal centers, metal-bound water or hydroxide, and organic catalytic groups is responsible for catalysis. 93'124-127 In previous studies, 128'129o,o'-dihydroxyazobenzene (66) mixed with Fe(II) and Fe(III) ions was found to form solid materials with intrinsic conductivity. To explain the intrinsic conductivity, formation (e.g. 67) of coordination polymer was proposed. Sonication of 68 or 69 in the presence of transition metal ions produced coordinatively polymerized bilayer membranes (CPBMs) as schematically illustrated in 70. 26'27'130 Upon complexation with the transition metal ions, stability of bilayer membranes was remarkably improved. 26 Amphiphiles 68 and 69 form complexes with transition metal ions with 1"1 stoichiometry. The pronounced enhancement of the stability of bilayer membranes suggests that each metal ion of the resulting CPBM is bound to two nitrogen atoms of the two adjacent azobenzene moieties as indicated by 70. Some metal ions of the CPBMs may be attached to one azobenzene unit. Even ifmetal ions are singly attached to the azo ligand, a polymeric cluster of metal ions bound to the dihydroxyazobenzene ligand is obtained upon sonication of the amphiphile with transition metal ions.

N; N

66

67

Designing Artificial Enzymes

279

C12HisNH N~

"N I

C12H2sNH N. "N

HOaS

68

HOOC~ /

~..--/

69

Sonication of 68 in the presence of an equivalent of FeCI 3 or [Co(NH3)sC1]C12 produced a solution of Fe(III)-CPBM or Co(III)-CPBM. The transmission electron micrograph of Fe(III)-CPBM of 68 is indicated by 71. 26 The CPBM is much greater than ordinary enzymes having diameters of 20-30 !k. The proteolytic activity of the CPBMs was observed with chymotrypsin (ChT) and carboxypeptidase A (CPA). 28 When the Co(III)-CPBM of 68 ([68]o = 5.12 raM) was incubated with ChT (2.12 x 10-5 M) at pH 7.5 and 4 ~ ChT was inactivated within a few minutes. Although ChT was not cleaved by Fe(III)-CPBM of 68, CPA was inactivated by both Fe(III)-CPBM and Co(III)-CPBM of 68. The inactivation of CPA was slower compared with ChT with half-live of inactivation process being about 3 h at 4 ~ Electrophoresis (SDS-PAGE) of the inactivated ChT revealed that the protein was cleaved into fragments smaller than 3 kDa within 10-20 rain at 4 ~ Electrophoresis performed for CPA indicated accumulation of several protein fragments. The

280

JUNGHUN SUH

fragments were subjected to N-terminal sequencing by Edman degradation. The sequencing identified Val(141)-Asp(142) and Met(22)-Asp(23) as the cleavage sites. The half-life observed for cleavage of ChT with Co(III)-CPBM was about 3 min at 4 ~ This is 109 times shorter than that for spontaneous hydrolysis of peptides. The intermediate peptides were cleaved much faster than ChT, and ChT was multiply cleaved into small fragments. The degree of catalysis in peptide hydrolysis for the intermediate peptides is, therefore, considerably greater than that achieved with ChT. Although the CPBMs manifested remarkable catalysis in peptide hydrolysis, the bilayer membranes have inherent limitations due to instability in organic solvents. Moreover, phase transitions at various temperatures complicate examination of the catalytic activity of CPBMs at various temperatures. In attempts to design a new type of PAM derivatives that can overcome these problems, a PAM was conjugated with poly(allylamine) (PAA: 72). In DMSO, 73 formed a 1:1-type complex with a Fe(III) ion. A DMSO solution of PAA (av. degree of polymerization = ca. 600) was added to a DMSO solution of PAM obtained by mixing 73 with an equivalent of Fe(III) or Co(Ill) ion. Acylation of the amino groups of PAA with the activated ester linkages of PAM produced a PAM-PAA conjugate. Amino groups of the resulting polymer were partially lauroylated with the N-hydroxysuccinimide ester of lauric acid leading to the formation of Fe(III)-PAM-Lau-PAA or Co(III)-PAM-LauPAA as summarized in the scheme of 74. TM The shaded rectangle in 74 stands for PAM regions. The amount of 73 attached to PAA and that of lauroyl group were 9 residue mol% relative to PAA. By random acylation of 72 with 73 followed by lauroylation, DHAB-Lau-PAA was obtained. The contents of 71 and lauryl moieties were almost identical for PAM-Lau-PAA and DHAB-Lau-PAA. In 75, the progress of cleavage of BSA by Com-PAM-Lau-PAA is illustrated with the results of electrophoresis. The time indicated in 75 represents the period of incubation. Accumulation of several intermediate proteins was observed with Co(III)-PAM-Lau-PAA. For the cleavage of BSA by Fe(III)-PAM-Lau-PAA, on the other hand, no intermediate protein was observed. The two largest intermediate proteins shown in 75 were isolated and subjected to N-terminal sequencing

j3.o H2NH2NH2NH2NH2NH2NH2NH2N 72

HO~~' N.. N

0 73

Designing Artificial Enzymes

281

by Edman degradation. The cleavage sites were identified as Arg(24)-Asp(25) and Trp(237)-Ser(238). Analysis of density of the parent band revealed that the decay followed pseudofirst-order kinetics. The pseudo-first-order rate constant was proportional to C O(the initially added concentration of the catalyst; expressed in terms of the DHAB monomer) and the proportionality constant (k2) measured at the optimum pH (= 8.5) was 63 + 3 or 410 + 30 M -l m -l, respectively, for com-PAM-Lau-PAA or Feln-PAM-Lau-PAA. When C Owas 4 x 10-5 M or 9 x 10-5 M, respectively, for Fem-PAM-Lau-PAA or com-PAM-Lau-PAA, the half-life of 40 or 140 min at 50 ~ respectively, was observed for the cleavage of BSA. These half-lives are about 106 times shorter than that of the spontaneous hydrolysis of peptides. When the PAM-PAA conjugates were not lauroylated, cleavage of BSA was much slower. Lauroyl groups apparently provide auxiliary binding sites that facilitate complexation of BSA to the catalyst, leading to faster reactions. Cleavage of BSA by the Fe(III) or Co(Ill) complexes of DHAB-Lau-PAA was much slower than that by com-PAM-Lau-PAA or Fem-PAM-Lau-PAA. This indicates that

282

JUNGHUN SUH

76

preassemblage of 73 by the metal ions is crucial to designing the active site comprising multiple metal centers. The cleavage sites identified so far for the PAM-catalyzed hydrolysis of proteins are Arg(24)-Asp(25) and Trp(237)-Ser(238) for the hydrolysis of BSA by Co(III)PAM-Lau-PAA as well as Val(141)-Asp(142) and Met(22)-Asp(23) for the hydrolysis of CPA by Co(III)-CPBM of 68. Among the four cleavage sites identified so far for the PAM-catalyzed protein hydrolysis, three sites involve Asp as the amine portion and one contains Ser as the amine moiety. This suggests selectivity of the PAM derivatives toward Asp and, possibly, Ser. A molecular mechanics calculation (MM+, HyperChem TM) showed that the three-point contact between Asp or Ser peptide with PAM is possible as illustrated in the scheme of 76. Various catalytic roles have been identified for transition metal ions in hydrolysis of carboxyl derivatives. 2'3'132-137 Addition of water to the carbonyl group of an amide is facilitated by metal ions either through activation of the carbonyl group coordinated to the metal ion or by nucleophilic attack by metalbound hydroxide ion at the carbonyl group. In addition, protonation of leaving amines is required for facile expulsion of the amines. Metal-bound water molecules can act as general acids to protonate the leaving amines. 137In 76, the first metal ion stabilizes the tetrahedral intermediate formed by addition of water to the amide and the second metal ion provides metal-bound water that protonates the leaving amine. The extra interaction between Asp carboxylate or Ser hydroxyl with the third metal center of PAM accounts for the selectivity of the PAM derivatives toward Asp or Ser peptides.

9. PERSPECTIVES Although artificial enzymes are prepared with various kinds of molecules, only a few strategies have been developed so far for designing the active site. Thanks to recent advances in supramolecular chemistry, molecular biology, and immunology, as well as molecular design, various novel methodologies will become available in the near future for the design of active sites comprising several functional moieties on synthetic or semisynthetic molecules. The degree of rate acceleration achieved to date with artificial enzymes has been only modest compared with that of enzymes. Other characteristic features of enzymatic action such as regio- and enantioselectivity have been seldom achieved

Designing Artificial Enzymes

283

with artificial enzymes based on fully synthetic materials. Moreover, artificial enzymes have been designed for only a limited number of chemical reactions. Future activities in the area of artificial enzymes will be directed toward raising the degree of rate acceleration and achieving high selectivity with respect to the structures of substrates and products. Moreover, efforts will be m a d e to design new artificial e n z y m e s for various types of chemical reactions. Artificial e n z y m e s can be useful for c o m m e r c i a l applications even if they have low catalytic activity. W h e n artificial e n z y m e s are designed for chemical reactions such as photo-reduction of carbon dioxide or conversion of cellulose to glucose, they w o u l d contribute significantly to sustaining vitality of the Earth.

ACKNOWLEDGMENTS Support of this work by a grant from Lotte Fellowship (1997) is gratefully acknowledged.

REFERENCES 1. Feiters, M. C. In Comprehensive Supramolecular Chemistry; Atwood, J. L.; Davies, J. E. D.; MacNicol D. D.; Vogtle, E; Lehn, J.-M. Eds.; Pergamon: Oxford, 1996; Vol. 10, p. 267. 2. Suh, J. Acc. Chem. Res. 1992, 25, 273. 3. Suh, J. In Perspective in Bioinorganic Chemistry; Hay, R. W.; Dilworth, J. R.; Nolan, K. B. Ed.; JAI Press: Greenwich, CT, 1996; Vol. 3, p. 115. 4. Chin, J. Acc. Chem. Res. 1991, 24, 145. 5. Finn, G.; Sharpless, K. B. J. Am. Chem. Soc. 1991, 113, 113. 6. Shibasaki, M.; Sasai, H.; Arai, T. Angew. Chem., Int. Ed. Engl. 1997, 36, 1236. 7. Corey, E. J.; Bakshi, R. K.; Shibata, S. J. Am- Chem- Soc. 1987, 109, 5551. 8. Lerner, R. A.; Benkovic, S. J.; Schultz, P. G. Science 1991, 252, 659. 9. Schultz, P. G.; Lemer, R. A. Acc. Chem- Res. 1993, 26, 391. 10. Schultz, P. G.; Lemer, R. A. Science 1995, 269, 1835. 11. MacBeath, G.; Hilvert, D. Chem. Biol. 1996, 3(6), 433. 12. Klotz, I. M.; Royer, G. P.; Scarpa, I. S. Proc. Natl. Acad. Sci. USA 1971, 68, 263. 13. Klotz, I. M. In Enzyme Mechanisms; Page, M. I.; Williams, A., Ed.; Royal Society of Chemistry: London, 1987, Chap. 2. 14. Suh, J. In Polymeric Materials Encyclopedia; Salamone, J. C. Ed.; CRC Press: Boca Raton, 1996, p. 8230. 15. Hollfelder, E H.; Kirby, A. J.; Tawfik, D. S.J. Am. Chem. Soc. 1997, 119, 9578. 16. Wulff, G. Angew. Chem., Intl. Ed. Engl. 1995, 34, 1812. 17. Broo, K. S.; Nilsson, J.; Flodberg, A.; Baltzer, L. J. Am- Chem. Soc. 1998, 120, 4063. 18. Kaiser, E. T.; Lawrence, D. S. Science 1984, 226, 505. 19. Corey, D. R.; Schultz, P. G. Science 1987, 238, 1401. 20. Wu, Z. P.; Hilvert, D.J. Am. Chem. Soc. 1989, 111, 4513. 21. Germanas, J. P.; Kaiser, E. T. Biopolymers 1990, 29, 39. 22. Davies, R. R.; Distefano, M. D. J. Am. Chem. Soc. 1997, 119, 11643. 23. Fendler, J. H.; Fendler, E. J. Catalysis in Micellar and Macromolecular Systems; Academic Press: New York, 1975. 24. Menger, E M. Acc. Chem. Res. 1979, 12, 111. 25. Murakami, Y.; Kikuchi, J.; Hisaeda, Y.; Hayashida, O. Chem. Rev. 1996, 96, 721. 26. Suh, J.; Lee, K. J.; Bae, G.; Kwon, O.-B.; Oh, S. Langmuir 1995, 11, 2626.

284

JUNGHUN SUH

27. Suh, J.; Shim, H.; Shin, S. Langmuir 1996, 12, 2323. 28. Suh, J.; Oh, S. Bioorg. Med. Chem- Lett. 1996, 6, 1067. 29. Jencks, W. E Catalysis in Chemistry and Enzymology; McGraw-Hill: New York, 1969, Chaps. 5-9. 30. Hamilton, A. D. Adv. Supramol. Chem. 1990, 1, 1. 31. Vogfle, E Supramolecular Chemistry; Wiley: New York, 1993. 32. Chung, Y.; Akkaya, E. U.; Venkatachalam, T. K.; Czarnik, A. W, Tetrahedron Lett. 1990, 31, 5413. 33. Wall, M.; Hynes, R. C.; Chin. J. Angew. Chem., Int. Ed. Engl. 1993, 32, 1633. 34. Wahnon, D.; Lebuis, A.-M.; Chin, J. Angew. Chem-, Int. Ed. Engl. 1995, 34, 2412. 35. Young, M. J.; Chin, J. J. Am. Chem. Soc. 1995, 117, 10577. 36. Williams, N. H.; Cheung, W.; Chin, J. J. Am. Chem. Soc. 1998, 120, 8079. 37. Chapman, W. H., Jr.; Breslow, R. J. Am. Chem- Soc. 1995, 117, 5462. 38. Yashiro, M.; Ishikubo, A.; Komiyama, M. J. Chem. Soc., Chem- Commun. 1995, 1793. 39. Ragunathan, K. G.; Schneider, H.-J. Angew. Chem., Int. Ed. Engl. 1996, 35, 1219. 40. Koike, T.; Inoue, M.; Kimura, E.; Shiro, M. J. Am. Chem. Soc. 1996, 118, 3091. 41. Molenveld, E; Kapsaelis, S.; Engberson, J. E J.; Reinhoudt, D. N. J. Am- Chem- Soc. 1997, 119, 2948. 42. Tabushi, I.; Shimizu, N.; Sugimoto, T.; Shiozuka, M.; Yamamura, K. J. Am. Chem- Soc. 1977, 99, 7100. 43. Breslow, R.; Doherty, J. B.; Gullilot, G.; Lipsey, C. J. Am. Chem. Soc. 1978, 100, 3277. 44. Breslow, R.; Czarniecke, M. E J. Am. Chem- Soc. 1978, I00, 7771. 45. Breslow, R.; Bovy, E; Hersh, C. L. J. Am- Chem. Soc. 1980, 102, 2115. 46. Akkaya, E. U.; Czarnik, A. W. J. Am. Chem. Soc. 1988, 110, 8553. 47. Nakamura, M.; Ikeda, T.; Nakamura, A.; Ikeda, H.; Ueno, A.; Toda, E Chem. Lett. 1995, 343. 48. Breslow, R.; Schmuck, C. J. Am. Chem- Soc. 1996, 118, 6601. 49. Breslow, R.; Zhang, X. J. Am. Chem. Soc. 1996, 118, 11678. 50. Breslow, R.; Zhang, B. J. Am. Chem. Soc. 1997, I 19, 1676. 51. Letsinger, R. L.; Savereide, T. J. J. Am- Chem- Soc. 1962, 84, 3122. 52. Overberger, C. G.; St. Pierre, T.; Yaroslavksy, C.; Yaroslavsky, S. J. Am. Chem. Sot:. 1966, 88, 1184.

53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70.

Kopple, K. D. Biopolymers 1968, 6, 1417. Kunitake, T.; Shimada, E; Aso, C. J. Am- Chem. Soc. 1969, 91, 2716. Overberger, C. G.; Salamone, J. C. Acc. Chem. Res. 1969, 2, 217. Kirsh, Y. E.; Pluzhnov, S. K.; Shomina, T. S.; Kabanov, V. A.; Kargin, V. A. Vysokonol. Soedin. Ser. A. 1970, 12, 186; Chem. Abstr. 1970, 72, 86675. Dick, C. R.; Ham, G. E. J. Macromolecular Sci. 1970, A4, 1301. Klotz, I. M.; Royer, G. E; Sloniesky, A. R. Biochemistry, 1969, 8, 4752. Suh, J. In Polymeric Materials Encyclopedia; Salamone, J. C. Ed.; CRC Press: Boca Raton, 1996, p. 4210. Kunitake, T.; Shinkai, S. Adv. Phys. Org. Chem. 1980, 17, 435. Suh, J.; Klotz, I. M. J. Am. Chem. Soc. 1984, 106, 2373. Suh, J.; Scarpa, I. S.; Klotz, I. M. J. Am. Chem. Soc. 1976, 98, 7060. Suh, J.; Klotz, I. M. Bioorg. Chem. 1977, 6, 165. Jones, R. A. Y.Physical and Mechanistic Organic Chemistry; 2nd ed.; Cambridge University Press: Cambridge, 1983, Chap. 5. Lee, K. J.; Suh, J. Bioorg. Chem. 1994, 22, 95. Suh, J.; Kim, Y.; Lee, E.; Chang, S. H. Bioorg. Chem. 1986, 14, 33. Delaney, E. J.; Wood, L. E.; Klotz, I. M. J. Am. Chem. Soc. 1982, 104, 799. Suh, J.; Paik, H.-j.; Hwang, B. K. Bioorg. Chem. 1994, 22, 318. Suh, J.; Klotz, I. M. J. Pol. Sci. Pol. Chem. Ed. 1978, 16, 1943. Okahata, T.; Kunitake, T. J. Pol. Sci. Pol. Chem. Ed. 1977, 15, 2571.

Designing Artificial Enzymes 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. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105.

285

Kirsch, Y. E.; Lebedeva, T. A.; Kabanov, V. A. J. Pol. Sci. Pol. Lett. Ed. 1975, 14, 207. Klotz, I. M.; Drake, E. N.; Sisido, M. Bioorg. Chem. 1981, 10, 63. Suh, J.; Moon, B. S. J. Org. Chem. 1989, 54, 2009. Suh, J.; Heo, J. S. J. Org. Chem. 1990, 55, 5531. Birk, Y.; Klotz, I. M. Bioorg. Chem. 1971, 1, 175. Murachi, T.; Okamura, K. J. Pol. Sci. Pol. Lett. Ed. 1976, 14, 361. Spetnagel, W. J.; Klotz, I. M. Biopolymers 1978, 17, 1657. Shinkai, S.; Ando, R.; Kunitake, T. Biopolymers 1978, 17, 2757. Kimura, Y.; Nango, M.; lhara, Y.; Kuroki, N. Chem. Lett. 1984, 429. Fujii, Y.; Matsutami, K.; Kikuchi, K. Chem. Commun. 1985, 415. Beach, J. A.; Shea, K. J. J. Am. Chem. Soc. 1994, 116, 379. Spivak, D.; Gilmore, M. A.; Shea, K. J. J. Am. Chem. Soc. 1997, 119, 4388. Suh, J.; Cho, Y.; Lee, K. J. J. Am. Chem. Soc. 1991, 113, 4198. Suh, J.; Kim, N. J. Org. Chem. 1993, 58, 1284. Suh, J.; Kim, N. Bull. Korean Chem. Soc. 1993, 14, 292. Kim, N.; Suh, J. J. Org. Chem. 1994, 59, 1561. Chin, J.; Banaszczyk, M.; Jubian, V.; Zou, X. J. Am. Chem. Soc. 1989, 111, 186. Kim, J. H.; Chin, J. J. Am. Chem. Soc. 1992, 114, 9792. Dietrich, M.; Munstermann, D.; Suerbaum, H.; Witzel, J. Eur. J. Biochem. 1991, 199, 105. Suh, J.; Lee, S. H.; Zoh, K. D. J. Am. Chem. Soc. 1992, 114, 7917. Zoh, K. D.; Lee, S. H.; Suh, J. Bioorg. Chem. 1994, 22, 242. Jang, B.-B.; Lee, K. P.; Min, D. H.; Suh, J. J. Am. Chem. Soc. 120, 12008 (1998). Lipscomb, W. N. Acc. Chem. Res. 1970, 3, 81. Onda, M.; Yoshihara, K; Koyano, H; Ariga, K.; Kunitake, T. J. Am. Chem. Soc. 1996, 118, 8524. Kamino, A.; Koyano, H.; Ariga, K.; Kunitake, T. Bull. Chem. Soc. Jpn. 1996, 69, 3619. Linkletter, B.; Chin, J. Angew. Chem., Int. Ed. Engl. 1995, 34, 472. Kimura, E.; Kodama, Y. J. Am. Chem. Soc. 1995, 117, 8304. Komiyama, M.; Sumaoko, J. J. Chem. Soc., Perkin Trans. 2 1997, 75. Hettich, R.; Schneider, H. J. Chem. Soc., Perkin Trans. 2 1997, 2, 2067. Bryant, R. A. R.; Hansen, D. A. J. Ant Chem. Soc. 1996, 112, 5498. Radzicka, A.; Wolfenden, R.J. Am. Chem. Soc. 1996, 118, 6105. Zhu, L.; Qin, L.; Parac, T. N.; Kostic, N. M. J. Am. Chem. Soc. 1994, 116, 5218. Hegg, E. L.; Burstyn, J. N. J. Am. Chem. Soc. 1995, 117, 7015. Suh, J.; Oh, S. Bioorg. Med. Chem. Lett. 1996, 6, 1067. Martin, M. T.; Angeles, T. S.; Sugasawara, R.; Aman, N. I.; Napper, A. D.; Darsley, M. J.; Sanchez, R. I.; Booth, P.; Titmas, R. C. J. Am. Chem. Soc. 1994, 116, 6508. 106. Gao, C.; Lavey, B. J.; Lo, C.-H. L.; Datta, A.; Wentworth, P. Jr.; Janda, K. D. J. Am. Chem. Soc. 1998, 120, 2211. 107. Kim, S. M.; Hong, I. S.; Suh, J. Bioorg. Chem. 1998, 26, 51. 108. Fersht, A. R.; Shi, J-P.; Knill-Jones, J.; Lowe, D. M.; Wilkinson, A. J.; Blow, D. M.; Brick, P.; Carter, P.; Waye, M. M. Y.; Winter, G. Nature 1985, 314(21), 235. 109. Suh, J.; Kwon, W. J. Bioorg. Chem. 1998, 26, 103. 110. Suh, J.; Lee, S. H.; Paik, H.-j. Inorg. Chem. 1994, 33, 3. 111. Suh, J.; Lee, S. H. J. Org. Chem. 1998, 63, 1519. 112. Dietrich-Buchecker, C. O.; Sauvage, J.-P. Chem. Rev. 1987, 87, 795. 113. Kirby, A. J. Adv. Phys. Org. Chem. 1980, 17, 183. 114. Suh, J.; Park, H. S. J. Polym. Sci. Polym. Chem. 1997, 35, 1197. 115. Fersht, A. Enzyme Structure and Mechanism; Freeman: New York, 1985, pp. 422-426. 116. Suh, J.; Hah, S. S. J. Am. Chem. Soc. 1998, 120, 10088. 117. Kirby, A. J.; Lancaster, P. W. J. C. S. Perla'n H 1972, 1206. 118. Liu, S.; Hamilton, A. D. Bioorg. Med. Chem. Lett. 1997, 7, 1779.

286 119. 120. 121. 122. 123. 124. 125. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137.

JUNGHUN SUH Stem, M. K.; Bashkin, J. K.; Sall, E. D. J. Am. Chem. Soc. 1990, 112, 5357. Johnson, T. W.; Klotz, I. M. Macromolecules 1974, 7, 618. Suh, J.; Kim, M.- J. Bioorg. Chem. 1992, 20, 366. Suh, J.; Hong, S. H. J. Am. Chem- Soc. 120, 12545 (1998). Williams, N. H.; Chin, J. J. Chem. Soc., Chem- Commun. 1996, 131. Pelletier, H.; Sawaya, M. R.; Kumar, A.; Wdson, S. H.; Kraut, J. Science 1994, 264, 1891. Kim, Y.; Eom, S. H.; Wang, J.; Lee, D.-S.; Suh, S. W.; Steizt, T. A. Nature 1995, 376, 612. Str~iter,N.; Klaubunde, T.; Tucker, P.; Witzel, H.; Krebs, B. Science 1995, 268, 1489. Barford, D. Trends Biochem. Sci. 1996, 21,407. Suh, J.; Oh, E. Synth. Met. 1990, 39, 177. Suh, J.; Oh, E.; Kim, H. C. Synth. Met. 1992, 48, 325. Suh, J.; Shin, S.; Shim, H. Bull. Korean Chem. Soc. 1997, 18, 190. Suh, J.; Moon, S. J. Bioorg. Med. Chem. Lett. 1998, 8, 2751. Suli; J.; Cheong, M.; Suh, M. P. J. Am- Chem- Soc. 1982, 104, 1654. Suh, J.; Cho, W.; Chung, S. J. Am. Chem. Soc. 1985, 107, 4530. Suh, J.; Han, O.; Chang, B. J. Am- Chem. Soc. 1986, 108, 1839. Suh, J.; Chun, K. H. J. Am. Chem. Soc. 1986, 108, 3057. Suh, J.; Kim, J.; Lee, C. S. J. Org. Chem. 1991, 56, 4364. Suh, J.; Park, T. H.; Hwang, B. K.J. Am. Chem. Soc. 1992, 114, 5141.

THE RELEVANCE OF SU PRAMOLECU LAR CH EMISTRY FOR THE ORIGIN OF LIFE

Pier Luigi Luisi

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

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Self-Assembly as a Spontaneous Ordering Process . . . . . . . . . . . . . . . Growth and Self-Reproduction . . . . . . . . . . . . . . . . . . . . . . . . . Supramolecular Surfactant Aggregates and Chemical Autopoiesis . . . . . . . Selection and Polymerization Assisted by Vesicles . . . . . . . . . . . . . . . Catalysis Aided by Supramolecular Aggregates . . . . . . . . . . . . . . . . . As a Way of Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

287 288 290 295 297 300 304 306

1. I N T R O D U C T I O N Research on the origin of life for many still has an exotic f l a v o r m t h e search for an improbable event in time long gone. As a matter of fact, the basic question on the origin of life is a very sound one, and one which is quite actual" how can molecular

Advances in Supramolecular Chemistry Volume 6, pages 287-307. Copyright 9 2000 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 0-7623-0557-6

287

288

PIERLUIGI LUISI

complexity and specificity be built from simple (prebiotic) compounds and on the basis of robust, unsophisticated chemical reactions? In fact, the main assumption in the field is still the one proposed by Alexander Oparin in the late 1920s that life originated on this planet from a prebiotic molecular evolution, namely a series of successive spontaneous chemical steps that brought about a gradual increase of size and specificity (complexity) up to the staggering level of the living cell. The detailed pathway is not known yet. It is known that a great number of biological molecules (amino acids, sugars, nucleic bases, fatty acids) can be obtained under prebiotic conditions; we know much less on how proteins and nucleic acids or polysaccharides have originated from the monomer mixture, and even less on the events which led to the genetic code and to the organization of the early cell (for a discussion about these points see refs. 1-5). In this chapter, I wish to examine the early steps of prebiotic molecular evolution, and emphasize the fact that supramolecular chemistry of surfactant aggregates is likely to be important for the early stages of the origin of life. Since this view is the main gist of our research at the ETH-Zuerich, this article will also review the main philosophy and the experimental approach we have pursued over the last 10 years. As it will be seen, this is tantamount to emphasize the importance of hydrophobic interactions in the early stages of the origin of life. Partly, this view is not new, as it has been considered in the literature, in particular by those authors who emphasize the importance of protocellular structures on the origin of life. 1'2'6 0

SELF-ASSEMBLY AS A SPONTANEOUS ORDERING PROCESS

Let us begin with the question as to whether chemical process(es) exist which can bring about a spontaneous formation of order and molecular complexity. One answer that immediately comes to mind is self-assembly (polymerization may be the second answer). Many amphophilic molecules, which also include relatively simple molecules such as long-chain carboxylic acids and alkyl phosphates, have the capability to aggregate into ordered structures. The possible relevance of self-assembly for the early prebiotic processes may become apparent from this simple"Gedankenexperiment": let us assume a prebiotic soup containing a chaotic mixture of several compounds: sugars, amino acids, metal ions, aromatic bases, fatty acids, and a variety of simple organic compounds. Suppose that their concentration steadily increases over time so that at a certain point the CAC (critical aggregate concentration) for the fatty acids self-assembly is reached. Then, a quite remarkable spontaneous ordering process takes place: bilayer vesicles form spontaneously. The formation of ordered vesicles produces well-defined compartments which entrap a certain amount of hydrosoluble molecules that are present in the prebiotic soup, such as amino acids, metal ions, or sugars. Also, owing to the hydrophobic character of the aggregate bilayer, a series

5upramolecular Chemistry and the Origin of Life

289

of hydrophobic, poor water soluble compounds will spontaneously bind to the membrane. Furthermore, amphophilic compounds can be incorporated into the bilayer. This is illustrated in Figure 1. It is actually rather easy to perform such an experiment in the laboratory and visualize by light scattering and spectroscopy the formation of the aggregates and the entrapment. The so-formed structure has an aqueous core separated from the aqueous environment by a lipid bilayer, and the analogy with a biological cell is apparent. As already mentioned, all these processes are basically driven by hydrophobic interactions. Formation of the vesicles is therefore a process which is accompanied by a decrease of free energy, as the coming together of lipophilic molecules to form a lipid cluster "liberates" water molecules from an energetically unfavorable situation and this brings about an increase of entropy. A similar situation is encountered in the binding processes of hydrophobic compounds onto the bilayer. It is a remarkable ordering process, a very basic one, and one which is under thermodynamic control. The question then automatically arises" why should not Nature have used these processes as a start for the prebiotic evolution? In fact, this

~

9

9I

9

5 II

9

surfactant

hydrophobic amphiphilic ~9 9

hydrophilic

Figure 1. Ordering effect brought about by supramolecular surfactant aggregation in a hypothetical mixture of many different reagents (a model for a "prebiotic soup"). At a given surfactant concentration (CAC) vesicles form, incorporating a certain amount of water-soluble components in the inner cavity; hydrophobic compounds will bind at the surface ofthe vesicles, and amphophilic reagentswill be incorporated in the bilayer. Despite the apparent increase of rigidity and order, the process of aggregation is generally attended by an overall decrease of free energy (due to hydrophobic interactions).

290

PIERLUIGI LUISI

very question forms the inspiration of many researchers in the field of the origin of life--those at least who belong to the "compartimentalistic" view of the origin of life.l,2, 5 How large are these compartments which form spontaneously? In the case for example of oleic acid, vesicles are formed which range between 50 and 1000 nm in diameter; 7'8 thickness is around 2-3 nm, aggregation number around 100,000, and molecular weight reaching the value of several million dalton. Fatty acids can also form micelles, usually in a higher range of pH values, where all carboxylic groups are deprotonated. 7 In addition to inducing a process of ordering and compartimentation, can supramolecular surfactant aggregation induce other phenomena which are relevant for the prebiotic molecular evolution? The answer is affirmative, as will be seen in the following sections.

3.

G R O W T H A N D SELF-REPRODUCTION

The bottleneck in the origin of life is the formation of the functional biopolymers-enzymes and nucleic acids. The answer cannot be the random polycondensation from a chaotic mixture of the monomers, as this process would afford an astronomic number of different chainsmca. 1070 for chains with a polymerization degree of 60. Given that, the probability that the same chain is produced more than once by a random polymerization process is in first approximation equal to zero; the single active individual macromolecule, even if formed, would decompose before it could be made again by another chance event. How then can active macromolecules be formed? The general consensus is to assume that the first active macromolecular species would be capable of self-reproduction, namely capable to induce the synthesis of copies of themselves via an autocatalytic step. Self-reproduction is in fact always to be connected with autocatalysis. Shnerior Lifson provides a nice numerical example 9 to understand the importance of self-reproduction and autocatalysis: suppose having a process which produces one molecule of product each microsecond. In a heterocatalytic reactions scheme, to produce one mole of product, 6 x 1023 microseconds would be necessary, i.e. more than the age of the universe. Conversely, in an autocatalytic reaction scheme, where the number of reacting species would double every microsecond, the total time to attain a mole of product would be 79 microseconds. There are still several uncertainties, even from the conceptual point of view, about the notion of self-replication. However the advantages of one such autocatalytic scheme for producing a significant concentration of active macromolecular species are commonly accepted. It is easy, at least on paper, to see how a self-replicative mechanism can give rise to the next important steps for the origin of life: suppose in fact that during the self-replication cycles the compound undergoes some

5upramolecular Chemistry and the Origin of Life

291

chemical modification: then we will have evolution associated with self-replication, and a series of novel macromolecular species can be originated. The question then is how to produce the first self-replicating species. Several text books and researchers in the field accept the notion that such pristine self-replicating species was a family of RNA which by self-replication and mutation gave rise to ribozymes which then originated proteins and DNA. This is the so-called RNA world scenario of the origin of life. The problem here is to accept the idea that a self-replicating RNA family formed spontaneously by itself--an assumption which, at the state of our knowledge on the prebiotic biosynthesis of mononucleotides and their polymerization, comes close to blind faith. It is possible that the research of the next future will show that the spontaneous formation of a RNA self-replicating family has some scientific basis. At the present time, instead of waiting, it is perhaps fight to look for other alternatives. The capability of micelles and vesicles to bind hydrophobic compounds has been exploited in our group to induce self-reproduction of surfactant aggregates (we use here the term self-reproduction instead of self-replicationl~ This has been described first with reverse micelles, ll'12 later with caprylate micelles in water 13 in one experiment as summarized in Figure 1. The process is the following: caprylate ethyl ester(CE) is practically water-insoluble and when added to a basic aqueous solution it forms an oily supernatant. This is slowly hydrolyzed at the water interphase affording caprylic acid which, once the cmc is reached, forms caprylate micelles. The micelles, due to their hydrophobic nature, are able to efficiently bind the supernatant CE. The bound ester is hydrolyzed on the surface of the micelles and the hydrolysis produces caprylate which assembles spontaneously and rapidly into new micelles. The more micelles are formed, the more CE is bound and the more new micelles are formed in a typical autocatalytic process. In fact, the growth of the micelles population follows a sharp exponential behavior (see Figure 2). 13 It can be defined as self-reproduction since the formation of new micelles is induced by the micelles themselves. In this way, starting from a very simple chemical system--an oily ester on top of a water solution--supramolecular aggregates are formed, which then are able to multiply their population number. These processes all take place spontaneously from the initial conditions, including the important step of self-reproduction. This has brought us to ask again the question: 13 why Nature should not have used such simple, available systems as a way to start the processes of molecular evolution. The same kind of process can occur with vesicles. We have described this with oleate vesicles and other carboxylates, 14 including giant vesicles. 15 The case of vesicles is analytically more difficult, owing to the fact that upon hydrolysis of the water-insoluble ester or anhydride on the bilayer, two processes occur simultaneously: the growth in size and the increase in population number. 8'16 Both effects contribute to the increase of turbidity of the vesicle suspension, which is the most convenient way to measure the change of concentration of vesicles. As part of the

292

PIER LUIGI LUISI

quantitation difficulty, qualitatively vesicles do behave as micelles in terms of autocatalytic self-reproduction processes. How does a vesicular system behave if more monomer surfactant is made available in the aqueous solution? This situation has been studied in detail in the case of oleate vesicles, and very interesting observations have been made. 8'16 Two possibilities can be envisaged: either the added monomers rapidly and spontaneously assemble into new vesicles; or the added monomer binds to the already existing vesicles. This second case has been observed for oleate vesicles, and the kinetics and the size distribution has been studied in detail as a function of the amount and type of pre-added vesicles. Interestingly, when the pre-added vesicles are monodisperse ones (obtained by extrusion through 50 or 100 nm filters), the size distribution after addition of fresh surfactant monomer (even though in excess) closely reflects the monodisperse size distribution of the pre-laid vehicles. Conversely, when no pre-laid vesicles are present, addition of oleate to a water solution affords a very broad size distribution, ranging from 20 to 100 nm in diameter. In other words, it seems as if the pre-laid monodisperse vesicles exert a "matrix effect" on the formation of the new ones. 8'16 An example is given in Figure 3.

EC ~

OH"

cmc

Caprylate

> micelles

.5

-

1.2

Ethylr, aprylate . . ~

(~c)

j

CeWy~e

0.9

I~

0.6

,,....0 1

0.3-

t

...o, water

p~

9EtOH

A

~v

p-e-e

I

2 raM-

~,~.,.

biphasic System

clear micellar

0

'

,,,

I

10

[ more

mlcelles (self.repllcatlon)

m~ullon

,

,

._..d 0.0

accellereted hydrolysis o f EC

[MIC]

I

~

3 raM-

I

20 time ( h )

I

30

'

1 mM.

I

4O

Figure 2. Micellar autocatalysis. The biphasic system with the autocatalytic self-reproduction of aqueous caprylate micelles. 13 Hydrolysis of supernatant ethylcaprylate (EC) takes place first at the microscopic interphase of the biphasic system and is very slow, until the cmc is reached. Then the process becomes autocatalytic (see text).

Supramolecular Chemistry and the Origin of Life

293

24

21 r.-.

18

o

N 9 ~5

~

~2

~ o

9

~

E

24,~ A

02

04

06

0.8

1

02

04

06

0.8

1

1

21

C:

~-

15,

.O

~-.

~

12,

~

9

E

6,

C

3, O-

24'

21-.

C

c 18i

.o

Is~

.~

12:

n

E

'

6'

C

'

3,

0-1m

0.0

0.2

0.4 diameter

0.6

(run)

0.8

1 2

Figure 3. Freeze-fracture electron microscopy analysis of the vesicle size distribution

in the case of the spontaneous vesiculation of oleic acid/oleate. (A) Vesicles formed from the hydrolysis of 25 mM oleic anhydride (overall concentration) at 30 ~ yielding 50 mM oleic acicl/oleate. (B) Vesicles extruded throughout 50 nm diameter filters. (C) Vesicles formed upon hydrolyzing 20 mM oleic anhydride (same conditions as in A) in the presence of pre-added extruded vesicles B--all in 0.2 M bicine buffer pH 8.5. For details see ref. 8.

294

PIER LUIGI LUISI

This is again an ordering effect taking place with supramolecular surfactant aggregates. It has not been fully understood, and it may have some biological importance: a given size distribution of protocell appears to bear the information (purely based on physical principles) to make protocells of about the same size. In conclusion, it is possible to have self-reproduction processes associated with surfactant supramolecular aggregates which can be implemented with rather simple reaction conditions. One important limitation of these systems is their lack of evolution capability as well as the capability of Darwinian competition. In fact, these aspects have not been studied to date with supramolecular structures. To simulate Darwinian processes with surfactant aggregates is not impossible, however. For that, we have to consider the self-reproduction of vesicles (or micelles). Let us go back to the experiment of Figure 1 and suppose now that the prebiotic soup also contains several surfactants, A,B, C . . . . each with its own precursor (e.g. the corresponding anhydride in case of carboxylates). In this way, as the concentration of the surfactant monomers increases, the three types of aggregate will start to assemble, and then self-reproduce, but not simultaneously: the one with the lowest critical aggregate concentration (CAC)mlet us say A--will be formed first. This first type of surfactant aggregate (A n) will sequester the other surfactant molecules B and C, since those will be generally more soluble in the A n vesicle than in water. This simulates a situation in which three "organisms" compete for growth and reproduction, whereby the structure which reproduces more rapidly will take over and will inhibit the formation of the slower species. One can conceive interesting developments of this scheme. For example, assume that one of these three surfactant aggregates, say C,,, forms a stable complex with a prebiotic molecule in the soup, say Z. Then, there would be a tendency to form this complex which would shift the chemical equilibrium towards the formation of C,, so as toform (Cn)-Z, which is thermodynamically more stable. In this case, C n would be formed preferentially to the other aggregates. This would simulate again a kind of biological competition and selection mechanism. As already mentioned, this kind of experiment has not be carried out yet, and it would probably be worthwhile to pursue. We have mentioned before the possibility of combining chemical evolution with self-replication. In principle, chemical evolution can be associated to self-reproducing micelles or vesicles. There are in principle two ways to conceive this in this case: on the one hand, the surfactants of the self-reproducing vesicles could be chemically transformed during their reproduction cycles into compounds which may give rise to more efficient cell-like compartments. This possibility has been discussed theoretically some time ago. 17 On the other hand, the supramolecular structure can help and determine the evolution of internalized compounds--i.e. permitting certain reactions and avoiding others thanks to the semipermeable character of the membrane. As already mentioned, studies of this type with vesicles still remain to be initiated.

5upramolecular Chemistry and the Origin of Life 0

295

SUPRAMOLECULAR SURFACTANT AGGREGATES AND CHEMICAL AUTOPOIESIS

Self-reproducing micelles or vesicles have been considered within the framework of autopoiesis, since the chemical reactions responsible for autocatalysis and growth take place within the boundary of the system and are caused by the bounded system itself. Autopoiesis, as introduced in the early 1970s by Maturana and Varela, 1s'19 defines the logic of life based on the observation of the living cell. An autopoietic unity is the simplest organizational structure which works according to a living cell and represents therefore minimal living. Accordingly, autopoiesis is the property of a bounded system that is self-maintaining thanks to a process of component regeneration from within the boundary, the boundary being also selfmade. It goes without saying that each autopoietic system is provided with energy/nutrients from the outside medium. All living systems must be autopoietic, and by inference any autopoietic system can be said to be living. The interested reader is referred to the primary literature (refs. 18, 19) and to subsequent work (refs. 17, 20). It is important here to emphasize that the theory of autopoiesis, in addition to offering a clear distinction between the living and the nonliving, permits one to conceive experimental systems which respond to the definition of minimal life. Since those systems must be provided with a physical, semipermeable boundary, micelles and vesicles immediately come to mind. This is the gist of some of the work carried out in our group in Zurich under the headline of "chemical autopoiesis,,,11-13,17,20,22 which started from a collaboration with E Varela. 29 In order to understand this point, and in particular the relation between autopoiesis and supramolecular aggregates, consider Figure 4 which represents the minimal autopopietic system. 17 It is defined by a spherical semipermeable boundary constituted by only one component S. There are only two chemical reactions in this system: the first one transforms the nutrient A, which permeates inside (or can simply bind to the membrane) into S, which then migrates into the boundary. The second reaction is a decomposition reaction, and transforms S into some product P which then is eliminated, e.g. dispersed in the surrounding medium. One reaction produces S, the other destroys S. If the two velocities are numerically equal, the system is in homeostasis, i.e. it maintains its concentration and chemical identity; if the velocity of production of S is much lower than the decomposition rate, the system will eventually disappear. If the rate of production of S is larger than the decomposition rate, S will accumulate on the surface and will give rise to growth, or to self-reproduction. The latter case will occur when the bounded system is under the constrain of size monodispersity, i.e. it cannot grow. Then the accumulated S will produce new bounded structure. The very simple system therefore simulates the various kinetic pathways of a living cell: homeostasis, death, and growth/self-reproduction. It is the simplest possible chemical rendering of the chemical behavior of a living cell.

296

PIER LUIGI LUISI

A

Figure 4. The minimal autopoietic system. A closed boundary formed by only one molecular component S, with a reagent A which enters through the semipermeable boundary and is transformed into S with rate Vp. A competitive destruction reaction with rate Vdtransforms S into product(s) P which are eliminated. Depending upon the relative value of vp and Vd, three limit cases ofthe time development of the autopoietic system will occur, which simulate the three possible state of occurrence of a living cell.

The self-reproduction micelles or vesicles described above are thus particular cases of this minimal autopoietic system. Recently, however, a vesicular system which also works homeostatically has been realized in our laboratory. 2~ It is based on the self-reproducing oleate vesicles described above, but, while the vesicles increase in number, a competitive oxidative reaction breaks the oleate double bond destroying the vesicles. This oxidative reaction, based on the couple OsO4/[Fe(CN)6 ]+++, takes place on the bilayer thanks to an aromatic water-insoluble base which binds the osmiate. In this way, both the reaction of vesicle formation and the reaction of vesicle destruction takes place within the boundary of the vesicle system, which meets the geometrical conditions for autopoiesis. The system is operated continuously, i.e. there is a continuous supply of anhydride and of oxidizing medium by external microsyringes. In this way, by controlling the rate of the two competitive reactions, a stationary state can be maintained for long time. 2~ If one of the two

Supramolecular Chemistry and the Origin of Life

297

competitive reactions takes over, then the system will either disappear or grow/selfreproduce.

0

SELECTION A N D POLYMERIZATION ASSISTED BY VESICLES

As mentioned in the introduction, the bottleneck for the research on the origin of life is not so much the prebiotic synthesis of small molecules, but the formation of polypeptides and nucleic acids. And the problem here is not so much the question of an unspecific polymerization, but rather to explain how specific sequences can be formed, or selected out from a random mixture of polymers. One line of our research on the origin of life is linked to the question, whether and to what extent a polymerization which takes place on the membrane of liposomes can regulate the structure of the polymers.

conde

)

POLYPEPTIDE

.~_

condensation /

(PEP2) x Figure 5. A schematic representation of the hydrophobic selection operated by lipid vesicles towards a library of peptides. The hydrophobic peptide(s) is selected out (bound) and in the presence of a membrane-bound condensing agent, can be polymerized on the membrane surface.

298

PIER L U I G I LUISI

For that, let us consider the well-known liposome forming agent 1-palmytoyl, 2 oleyl-sn-phosphatidylcholine (POPC), a zwitterionic compound which forms highly hydrophobic liposomes. Suppose that a library of dipeptides is present in the aqueous solution containing the POPC liposomes, of the type H-Trp-X-OH, where X is another amino acid residue. With X = Trp we will have a rather hydrophobic dipeptide, whereas with X = Glu or Lys we will have charged, water-soluble compounds. It is clear that POPC liposomes will select out and bind the hydrophobic dipeptide(s), ignoring the hydrophilic ones. If now a hydrophobic condensing agent is present on the liposome membrane, an oligomerization of the bound hydrophobic peptide, e.g. H-Trp-Trp-OH, can take place. This mechanism of selection is illustrated in Figure 5. This is indeed what happens, 21 as shown in the experimental data illustrated in Table 1. A selective, membrane-aided polycondensation of amino acids can also be carried out starting from activated amino acids. We have chosen the NCA method,

Table 1. POPC Liposomes-Aided Product Selectivity for the Cooligomerization of Dipeptides a

Products (X, Y: Asp, Glu, or Gly)

% of Dipeptide(s) Reacted LiplMix. J~

ReflMix: c

Lip/1-rp2 .~

Ref/l-rp2 .~

0.8

0.9

Peptides: Cyclo(-Trp2-) H-Trp4-OH

0.9 43

H-Trp6-OH

2.1

H-Trps-OH

0.03

Cyclo(-Trp-X-)/H-Trp-X-Trp-Y-O H

8.0

H-Trp3-X-OH/H-Trp-X-Trp2-OH

7.0

1.6 0.6

62

1.0

0.03

5.9

0.00

0.00

0.3

0.00

13 6.4

~ ~

Derivatives EtO-CO-Trp2-OH

11

33

15

EtO-CO-Trp-X-OH

22

43

~

EtO-CO-Trp4-OH

1.6

0.1

4.3

EtO-CO-Trp6-OH

0. I

0.00

0.7

Rest

4.2

2.5

11

93 0.3 0.00 4.8

Notes: a All solutions were incubated for 26 h on a vortex at room temperature. Products were analyzed as

described in text (from ref. 21). b 5 mM H-Trp2-OH, 5 mM H-TrpGIy-OH, S mM H-TrpAsp-OH, 5 mM H-TrpGlu-OH, "100 nm liposomes" (25 mM POPC), 50 mM phosphate, pH 5.90, 2 mM EEDQ, 1.2% vJv ACN. c As b but no liposomes. d 5 mM H-Trp2-OH, "100 nm liposomes" (25 mM POPC), 50 mM phosphate, pH 5.90, 2 mM EEDQ, 1.2~ vlv ACN. e As d but no liposomes.

Supramolecular Chemistry and the Origin of Life

299

and again the NCA-Trp is selected out and is transformed into relatively long chains which grow on the membrane matrix. 21 This is shown in Figure 6. Notice that rather long oligomers are produced on the membrane, up to n = 29, although in trace amounts. Long Trp-oligomers cannot be produced in a classic polymerization in aqueous solution, since with n greater than 4 they are no longer water soluble. In other words, the membrane-aided polymerization is operating a certain degree of selectivity, as only certain monomers are selected out; and it does not afford random chains, but chains which are unique to the system. The main driving forces in all these examples are again hydrophobic forces. Surfactant aggregates can, however, also display electrostatic interactions if charged surfactants are used. One can prepare for example positively charge vesicles from DDAB (dimethyl, dioctyl ammonium bromide), and these can attract negatively charged NCA-amino acids, such as Glu, which can be trans100

.

.

.

.

.

B

A 8O

60

i~ ~ 40' s O

I

is 2~ 7met

20

0

10

20

10

30

20

30

40

Time/(min) Figure 6. Membrane assisted condensation of amino acids. 21 HPLC analysis (at 289 nm) of the products of the oligomerization of NCA-Trp assisted by POPC liposomes. There are several products (for reaction conditions and other details see ref. 21 ). It is important to notice the difference between the liposome-assisted oligomerization (A) and the control experiment, in the absence of liposomes (B). In this second case, the highest product has a oligomerization degree n = 7; in the case of liposomes we reach n = 29, although in minimal amounts.

300

PIER LUIGI LUISI

formed into glutamate chains. 34 By adding DDAB during the preparation of POPC liposomes, mixed liposomes are obtained having a variable charge density depending on the ratio POPC/DDAB. Such mixed liposomes can display both hydrophobic and electrostatic binding properties. For example, glutrp is preferentially bound and oligomerized in the presence of hydrophobic condensing agents. 34 In this way, polypeptide chains containing both hydrophobic and charged residues can be synthesized. Particularly interesting is the case of his-trp, which is also condensed into oligomers by the same mixed-liposome system, affording a peptide which, possessing histidine residues, may be provided with catalytic activity. 21 This links to the next argument I would like to discuss in this review, namely the membrane-assisted catalysis.

6. CATALYSISAIDED BY SUPRAMOLECULAR AGGREGATES The micellar self-reproduction is an autocatalytic process, and as such a case of micellar catalysis. This is however a point which requires some attention. As explained in the original paper 13 and in more detail in subsequent oneswparticularly in a theoretical treatment, 22 the main factor in this catalysis is the increase of active surface coming from the increase of micelle concentration. Let us explain this with a numerical consideration: In the experiment illustrated in Figure 2, hydrolysis starts at the macroscopic interphase of the biphasic system. Suppose this to be 1 cm 2 and the total water volume to be 10 mL. If now 1 mg of caprylate ester is hydrolyzed and transformed into micelles (i.e. ca. 10-4 M caprylate in water, or approximately 10-6 M micelles, considering that the surface of a micelle is approximately 2 • 10-13 cm 2) the total active micellar surface in the 10 mL aqueous solution will be around 104 cm 2. There is therefore an increase of active surface of several orders of magnitude, with a corresponding increase in the availability of the water-insoluble reactant to hydrolysis" even in the absence of any other chemical factor, there should be a corresponding increase of the hydrolysis rate. In fact, it has been argued in the mentioned theoretical paper 22 that one should expect a significant increase of the hydrolysis rate because of this effect, even if there is a negative chemical catalysis. This phenomenon, often forgotten, is the basis of most of the chemistry and physics of surfactant aggregates. The chemical catalysis (i.e. the decrease of the activation energy of the particular hydrolysis of the caprylic ester) may in fact be positive or negative, but it is probably negligible with respect to the physical catalysis (increase of the active surface where the hydrolysis takes place). Lack of understanding of this concept has led to confusion 23 and to the search of complicated mechanism to explain micellar autocatalysis. The same type of argument holds of course for vesicles. Having clarified this point, one may ask whether the same principle of"micellar catalysis" can be applied to other types of chemical functionality. Consider for

5upramolecular Chemistry and the Origin of Life

301

example a peptide which is endowed with some catalytic power, but which is practically water-insoluble, and as such unable to show such catalysis in aqueous solution. If we are now able to solubilize this peptide by the help of micelles or vesicles, then we should be able to see a good catalytic effect in aqueous solution. Here again, the driving forces for such an effect would be the hydrophobic interactions, which permit the water-insoluble peptide to bind to and be solubilized by the lipophylic aggregates. This effect has been studied in our group by Kenichi Morigaki in his dissertation. 24 He utilized the hydrophobic tripeptide Z-Phe-His-Leu-NH 2 (Z = carbobenzyloxy). This compound has been shown before in the literature to display catalytic properties towards the hydrolysis of certain esters. 25'26 The authors were particularly interested in the stereoselectivity of the process of L- towards D-amino acids and less in the enhancement of catalysis operated by the micelles. The substrate chosen was a very lipophylic ester, nitrophenylpalrnitate. Morigaki used oleate vesicles, and later POPC liposomes, obtaining qualitatively similar results. 24 Figure 7A shows some typical data. It is apparent that the rather modest hydrolytic effect of the tripeptide in aqueous solution is considerably increased when liposomes are used. It can be argued that this catalytic effect is "only" due to the increased solubility of both substrate and peptide catalyst due to the presence of the lipid bilayer. This may well be so, but it does not decrease at all the importance of the observation: the presence of supramolecular aggregates brings about a significant catalytic effect that otherwise would have been not present. This hydrophobic catalysis can be obtained by simple physical means, without any enzymatic magic. Therefore we may have exerted an important role in the early chemistry. Particularly interesting is the fact-evidenced in Figure 7A that the kinetics operated by the liposomes follows a Michaelis-Menten behavior. The corresponding turnover is illustrated in Figure 7B; to what an extent a turnover is really present in these experiments is however not ascertained yet. The hydrophobic effect of the vesicles can be combined with their self-reproduction property. This is possible if during the catalytic hydrolysis reaction the oleate vesicular system is provided with oleic anhydride which is hydrolyzed simultaneously to the hydrolysis of p-nitro palmitate. 24 Results are shown in Figure 8A and B. Note the autocatalytic pattern of the formation of oleate vesicles, and note that the same behavior is obtained by following the rate of the hydrolysis of nitropalmitate. We are now dealing with a catalytic system which is capable of reproducing during its chemical activity~a primitive model for a self-reproducing enzyme. It is as if one would add in Figure 7B a doubling of the catalyst concentration by each cycle. These observations are preliminary~it is apparent however that this type of physical catalysis, possibly combined with a self-reproducing mechanism, is an interesting field of inquiry for the origin of life. In the design of a cell model, the enzyme-like catalyst should eventually operate inside the vesicle. The question then

302

PIER LUIGI LUISI

"7 C-

O >

.c

,A

o

s

lo is

20 2s ao as 40 4s

so ss

[C,6-ONp] (pM)

vesicle

i

|

a~,mciafionof tile l~l)fidc

s,~;ociationof the suhsln~tc Figure 7. Liposome-assisted catalysis. (A) Dependency 24 of the initial hydrolysis rate of C16-O Np (nitrophenyl-pamitate) catalyzed by I mM carbobenzoxy-Phe-His-LeuOH on the substrate concentration, in 0.05 M borate buffer pH 8.5. The filled circles are relative to the self-hydrolysis (no peptide, no liposomes). Open triangles are without liposomes, open squares with liposomes. (B) The pseudo-enzymatic turnover of the catalytically active liposomes. The catalytic activity results primarily from the binding (and solubilization) of a very hydrophobic histidin-containing peptide and the very hydrophobic substrate.

Supramolecular Chemistry and the Origin of Life . . . .

_2o ~

15

__~

10

m ID

~

w*-'--v

....

i ....

w. . . .

v

I ....

9....

. . . .

w . ' - - - i

. . . .

A

II

E

s 0

....

1.2

,

.

.

.

.

, ....

/

,...w

B

1.0

~E

v ....

303

0.8

~p Q -9

0.6

>

0.4

....

"

v ....

w:;::

~

0.2

0.0 0

.................... 5 10

15

20

- ........ 25

30

9............. '35

40

45

time / [h]

Figure 8. (A) Autocatalytic hydrolysis of oleic anhydride in the presence of 1.9 mM Carbobenzoxy-Phe-His-Leu-OH and C16-ONp in 0.02 mM borate buffer pH 8.5. Here the progress of the concentration of formed oleate/oleic acid is reported as a function of time. (B) The simultaneous initial velocity of the hydrolysis of C16-O Np is measured.

is how to bring these catalytically active peptides inside the liposomes, and possibly by a natural process. It is known that several hydrophobic peptides can be uptaken by the membranes to form channels. 27'28 As far as hydrophilic peptides are concerned, the uptake and internalization by lipophilic liposomes appears more difficult. In fact it is well known that the liposome bilayer has per st a very low permeability towards all kinds of compounds. One way to circumvent this problem has been reported by Schubert at al. 30'31 These authors have shown that sodium cholate has the capability of increasing the permeability of phosphatidyl liposomes towards low- and high-molecular weight compounds. The mechanism of this process is still under debate, and it is not important to dwell upon this point here. Here it suffices to say that at relatively high concentration cholate induces the

304

PIERLUIGI LUISl

disruption of liposomes and the formation of mixed micelles (phosphatidyl choline/cholate micelles). It appears that the high permeability is induced somewhat before the point at which liposomes begin to transform into micelles. We have repeated the data by Schubert, confirming that nucleotides as well as enzymes and polynucleotides can be promoted inside the liposomes at high enough concentration of cholate in the medium. 32'33Other molecules have a similar cholate effect, and on this basis it seems possible to conceive a scenario according to which a certain degree of internalization may have been induced by external molecules. Until now, the classic way of having enzymatic effects with liposomes is to incorporate enzymes in them. The use of modern enzymes is of course not very appropriate for the research on the origin of life since in prebiotic times the present-day enzymes were not present. In fact, the real origin of life research is a "bottom-up approach": 35 one wishes to explain how enzymes and nucleic acids came about starting from very simple prebiotic compounds. The use of enzymes and nucleic acids is however very important in one very closely related area, the one dealing with the origin of cells. Origin of life and origin of cellular life do not need to be the same thing. In fact, following the scenario of the RNA-world, one can assume that self-replicating RNA families originated first, giving rise to a form of macromolecular life (self-replicating and mutating RNA quasi species giving rise to ribozymes which then produced proteins and DNA)mand all this possibly before the constitution of the cellular spatial organization (for a discussion about this point see refs. 5, 36). Under this perspective, the study of the origin of the cell is in principle independent from the origin of macromolecules, and can be carried out according to a "top-down approach". 35 Accordingly, one can try to build the "minimal living cell" by utilizing liposomes containing enzymes and/or nucleic acids. This research has been carried out in our laboratory and is still actively pursued-this is part of our enterprise in the direction of the "minimal living cell." It is not the aim of this review to dwell upon this part. The interested reader is referred to our work concerned with enzymes in liposomes, in particular to the work dealing with lecithin-producing enzymes in lecithin liposomes, 37 or the following work on enzymatic and molecular biological reaction taking place in vesicles and liposomes.32'33'38-41

7. AS A WAY OF C O N C L U S I O N The main point of this chapter is to show that supramolecular aggregates can play an important role in the early history of the origin of life. The main driving force in all processes we have illustrated is the hydrophobic interaction---this is responsible for the self-assembly of vesicles, for the binding of hydrophobic substances to the vesicle membranes and the corresponding autocatalytic self-reproduction of micelles and vesicles, as well as for the corresponding chemical events of polymerization. Since the hydrophobic forces take place generally spontaneously and with

5upramolecular Chemistry and the Origin of Life

305

a negative free energy change, we are dealing with processes that most likely have occurred by themselves in many different prebiotic scenario. Nobody knows whether this has been important to start the early events of life, but the valid argument can however be done that it would be unlikely that Nature has not utilized these available and simple self-ordering processes. In this way, by letting hydrophobic forces work, one can proceed a good deal in the latter of prebiotic chemical evolution, up to the first forms of enzyme-like peptides. This is illustrated in Figure 9. According to this "hydrophobic start" scenario, there first would have been the ordering and compartmentalization, followed by membrane-aided oligomerization of amino acids to form catalytically active peptides, which may have been internalized giving rise to very first primitive cell-like metabolizing entities. These structure were also provided with the capability of self-reproduction, which may have afforded a mechanism for the choice of the best-fitting compartment. According to this view, the early events in life-ordering, compartimentation, self-reproduction and condensation may have been governed by hydrophobic forces. In other words,

;arraCk~ hydrophobic binding

self-reproduction

~

membrane-aicled condensation

internalisation <

.

Figure 9. The hypothetical "hydrophobic start" in the origin of life. The hydrophobic, spontaneously formed vesicles can undergo self-reproduction if they bind the corresponding precursor; they can scavenge hydrophobic peptides and condense them into longer chain once a hydrophobic condensing agent is also present; and they can also bind water-soluble peptide catalyst (or any other potential hydrophobic catalyst) and induce an enzyme-like turnover. The catalyst can eventually be internalized, thus giving rise to a protocellular structure capable of a primitive metabolism.

306

PIER LUIGI LUISI

the early steps of life would be due to substances which are not well compatible with water. Water-soluble chemistry for life may have occurred only later. The self-reproduction mechanism illustrated here for vesicles is of course extremely primitive with respect to the self-replication mechanisms of RNA. However, one should recognize that the first prebiotic self-replicators cannot have been chemically well-developed and sophisticated macromolecules. Most likely, the first replicators were primitive structures. In this sense, the self-reproducing supramolecular structures have a distinct relevance, since the reproduction mechanism happens almost spontaneously. What is missing in this scheme are two elements which are essential for a molecular Darwinist approach to the origin of life, namely evolution of the replicators and their competition to produce the survival of the best fit. We have shown here that in principle it is possible to implement these mechanisms with vesicles. It is now needed to show that this is indeed the case. This is one fruitful direction for the work in the field of supramolecular surfactant aggregates, so as to give a dynamic aspect to the chemistry of micelles and vesicles.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

Morowitz, H. J. Beginnings of Cellular Life; Yale University Press: London, 1992. Dyson, E J. Origins of Life; Cambridge University Press, 1985. de Duve, C. Blueprint for a Cell: The Nature and Origin of l.afe; Neil Patterson: Birlington, NC. Oro; J.; Lazcano, A. In Prebiological Self Organization of Matter, Ponnamperuma, C.; Erich, E R., Eds.; Deepack Publishing: Hampton, VA, 1990. Luisi, P. L. In Astronomical and Biochemical Origins and the Search for Life in the Universe; Cosmovici, C. B.; Werthimer, D., Eds.; Editrice Compositori; Bologna, 1997, pp. 461-468. Deamer, D. W. Microbiol. Mol. Biol. Rev. 1997, 61,239-261. Hargreaves, W. R.; Deamer, D. W. Biochemistry 1978, 17, 3759-3768. Bloechiger, E.; Blocher, M.; Walde, P.; Luisi, P. L. J. Phys. Chem. 1998, 102, 10383-10390. Lifson, S. J. Mol. Evol. 1997, 44, 1-8. Luisi, P. L In Self-Production of Supramolecular Structures, Eds; Fleischacker, G. R.; Colonna, S.; Luisi, P. L.; Kluwer Academic: 1994. Bachmann, P. A.; Walde, P.; Luisi, P. L.; Lang, J. J. Amer. Chent Soc. 1990, 112, 8200-8201. Bachmann, P. A.; Walde, P.; Luisi, P. L.; Lang, J. J. Amer. Client Soc. 1991, 113, 8204-8209. Bachmann, P. A.; Luisi, P. L.; Lang, J. Nature 1992, 357, 57-59. Walde, P.; Wick, R.; Fresta, M.; Mangone, A.; Luisi, P. L. J. Amer. Chent Soc. 1994, 116, 11649-11654. Wick, R.; Walde, P.; Luisi, P. L. J. Amer. Chent Soc. 1995, 117, 1435-1436. Lonchin, S.; Luisi, P. L.; Walde, P.; Robinson, B. H. J. Phys. Chem. In press. Luisi, P. L. In Thinking about Biology; Stein, W.; Varela, E, Eds.; Addison-Wesley: 1993. Varela, E J.; Maturana, H.; Uribe, R. Biosystems 1974, 5, 287-296. Fleischacker, G. R. Biosystems 1988, 22, 37-49. Zepik, H.; Bloechiger, E.; Luisi, P. L. Submitted. Blocher, M.; Daojun, Liu; Luisi, P. L. Macromolecules. 1999, 32, 7332-7334. Mavelli, E; Luisi, P. L. J. Phys. Chem. 1996, 100, 16600-16607. Bushe, T.; Nagarajan, R.; Lavabre, D.; Micheau, J. C. J. Phys. Chem. A 1997, 101, 3910-3917. Morigaki, K. Diss. ETH-Z No. 1274, ETH-Ztirich, 1998.

5upramolecular Chemistry and the Origin of Life 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39.

307

Ueoka, R.; Mori, S.; Moss, R. A. Langmuir 1994, I0, 2892. Ueoka, R. et al. J. Amer. Chem. Soc. 1988, 110, 1588. Watts, A. Bioch. Biophys. Acta 1998, 1376, 339-352. Seelig, J. Biochim. Biophys. Acta 1997, 1331, 103-116. Luisi, E L.; Varela, E Origin of Life Evol. Biosphere 1989, 19, 633-643. Schubert, R.; Wolburg, H.; Schmidt, K.-H.; Roth, H. J. Chem. Phys. Lipids 1991, 58, 121. Schubert, R.; Beyer, K.; Wolburg, H.; Schmidt, K-H. Biochemistry 1986, 5263. Monnard, E A.; Oberholzer, T.; Luisi, E L. Biochim. Biophys. Acta 1997, 39, 1329. Oberhozer, T.; Abrizio, M.; Luisi, E L. Chem Biol. 1995, 2,677-682. Blocher, M.; Daojun, L.; Luisi, E L. In preparation. Luisi, E L.; Walde, E; Oberholzer, T. Current Opinion in Coll. Interface 1999, 4, 33-39. Luisi, E L. Origin ofLife Evol. Biosphere 1998, 28, 613-622. Schmidli, E K.; Schurtenberger, E; Luisi, E L. J. Amer. Chem. Soc. 1991, 113, 8127-8130. Luisi, E L.; Walde, E; Oberholzer, T. Ber. Bunsenges. Phys. Chem. 1994, 9, 1160-1165. Walde, E; Goto, A.; Monnard, E-A.; Wessicker, M.; Luisi, E L. J. Amer Chem. Soc. 1994, 116, 7541-7544. 40. Oberholzer, T.; Wick, R.; Luisi, E L.; Biecricher, C. K. Biochem. Biophys. Res. Comm. 1995, 207, 250-257. 41. Wick, R.; Luisi, E L. Chem. Biology 1996, 3, 277-285. 42. Walde, E Current Opinion Coll. Inte~ Science 1996, 1,638-644.

This Page Intentionally Left Blank

INDEX Active sites of synthetic artificial enzymes, designing, 245-286 catalytic properties, improvement of by microdomains of polymers, 250-253 flavin derivatives, 253 Kemp elimination, 251 micelles, 250 PEI, 250-253 vesicles, 250 by cross-linkage of preassembled catalytic elements with macromolecular spacer, 267-275 amino groups, 272 carboxyl groups, 272 in enterobactin analogue, 268, 271 maleamic acid derivatives, 275 molecular imprinting, difference with, 267 PEI, 267-275 phenol groups, 272 polypeptide backbone of enzyme, 267 introduction, 246-248 enzyme-like catalysts, 246-247 enzymology, combining with principles of, 248 goal, 246 semisynthetic and synthetic, 246 synzymes, 246

perspectives, 282-283 on polymers by molecular imprinting, 253-255 polymer skeletons, random functionalization of, 255-263 chymotrypsin (ChT), cleavage of, 263,279 PC, 259-261 PEI, 255-263,264-275 polymicelles, 255 by self-assembly from catalytic elements, 275-282 chymotrypsin (ChT), 279 CPBMs, 278-280 Edman degradation, 280, 281 HPNPP, 276-277 from metal ions, 278 PAM, 280-282 by site-directed functionalization of polymer skeletons, 263-267 PEI, 264-267 synthesis of small molecules equipped with multiple catalytic elements, 248-249 cyclodextrin (CD) derivatives most successful, 249 Lewis acid catalysts, 249, 261 Alzheimer disease, ffamyloids and 187 Amyloidosis, 87-88 309

310

INDEX

Artificial enzymes, active sites for, 245-286 (see also Active sites) Atomic force microscopy (AFM), 47-50, 51 Autopoiesis, chemical, theory of, 295-297 (see also Origin of life) Biomolecules and synthetic analogues, self-assembled structures from, 41-156 amino acids, peptides, and proteins, 85-106 adipic acid, 104 amino acid amphiphiles, 91-104 amyloidosis, 87 chiral bilayer effect, 88, 111, 113 Cotton effect, 105 DODA-EO2-biotin, 88-91 enantiomorphism, 95-98 folding process of helix, schematic, 99 "foldamers," 91 gramicidin, 91 a-helix, 85-87, 91 310-helix, 91 four-helix bundle, 87 hemoglobin S, 88-89 HIV (human immunodeficiency virus), 87 other assemblies involving amides, 104-106 polypeptides, 85-91 "sergeants-and-soldiers" effect, 105-106 13-sheet, 85, 87-88 sickle-cell anemia, 88 II trans helix, 88 tropocollagen, 88 carbohydrates, 106-130 aldonamides, 119-121,126 amylose, 107-108, 109 biological polysaccharides, 106-109

carrageenan, 108-109 cellulose, 106-107 chain length-induced racemate resolution, 115, 117 chiral bilayer effect, 88, 111, 113 chitin, 107 Cu2+ ions, organizing effect of, 127 gluconamides, 109-127 other, 127-130 retarded chiral bilayer effect, 115, 117 SDS (sodium dodecyl sulfate), 111-114 conclusions, 143 hydrogen bonding, 143 introduction, 42-57 atomic force microscopy (AFM), 47-50, 51 Brewster angle microscopy (BAM), 46, 77, 126, 128 chiragen ligand, 56 chiral polymers, 56-57 chirality, 42-45 coordination helicates, 56 "dissymrtrie molrculaire" 42 fluorescence microscopy, 46-50 Fourier patterns, 51, 52, 53 helical polymers, 56-57 12-hydroxystearic acid: history, 50-55 hypercycles, 44 "majority rule" 57 monolayers, 45-50 review, scope of, 55-57 self-assembly, 44-45 "sergeants-and-soldiers principle," 57, 105-106 synkinetic approach, 44 X-ray diffraction technique, 46 lipids, 58-85 amphiphiles, generalized structure of, 69

Index

biological, 58-65 ceramides, 83-85 chiral bilayer effect, 62 cholesterol, 79-83 Gaucher's disease, 83 globoid-cell leukodistrophy, 83 hydrophobic effect as driving force, 58 lecithin, 65 packing parameter concept, 58, 61 phospholipid analogues, 71-79 synthetic phospholipids, 65-71 nucleotides, 130-142 B-DNA helix, 130 biological polynucleotides, 130-133 C2,- and Cy-endo conformation, 130 enantiomeric cross-inhibition, 132 G quartets, 133 guanylic acid, 133, 134 Hoogsteen mode, 133 hydrogen bonding systems, other complementary, 140-142, 143 nucleotide analogues, 134-140 nucleotides, 133 peptide-nucleic acids (PNA), 132, 133 "RNA world" hypothesis, 132 "sergeants-and-soldiers effect" 133 Z-DNA, 131 Brewster angle microscopy (BAM), 46, 77, 126, 128 Buckminsterfullerene, 166, 178 Cahn-Ingold-Prelog convention, 42 CD spectroscopy, 70, 81, 137 Chirality, 42-45 Cholesterol, 79-83 Chymotrypsin (ChT), 263, 279 Cotton effect, 105 Cryoelectron microscopy, 111

311

Crystallization as self-assembly par excellence, 159 "Dissym6trte mol6culaire, "42 Edman degradation, 280, 281 Electron capture gas chromatography (ECGC), 199 Electron microscopy (EM), 55, 93, 102, 111,137 Electron spin resonance (ESR), 231 Electrophoresis, 261,262, 272-274, 279-280 Enantiomeric cross-inhibition, 132 Enantiomorphism, 95-98 Enterobactin, 268, 271 Fast atom bombardment mass spectrometry (FABMS), 8 Fenton chemistry, 219, 220 Ferritin, 178-180 Fluorescence microscopy, 46-50 "Foldamers," 91 Fourier transform infrared spectroscopy (FT-IR), 67, 75, 89, 90 Freeze-fracture electron microscopy, 130, 293 G quartets, 133 Gaucher's disease, 83 Globoid-cell leukodistrophy, 83 Hemoglobin S, 88-89 HIV (human immunodeficiency virus), 87 Hoogsteen mode, 133 Kemp elimination, 251 Lewis acid catalysts, 249, 261 MALDI-TOF mass spectrometry, 262

312

Maleamic acid derivatives, 275 Manganese (II)-based superoxide dismutase mimics, development of, 217-244 (see also Synthetic enzymes) Micellar catalysis, 300-304 (see also Origin of life) Micelles, 250, 290, 291,295-297 Michaelis-Menten scheme, 264, 301 Molecular imprinting on polymers, design of active sites by, 253-255 (see also Active sites) Nanometer-scale structures, precise control of, 24 Oparin's assumption about origin of life, 288 Origin of life, relevance of supramolecular chemistry for, 287-307 catalysis, membrane-assisted, 300-304 hydrophobic interactions, 300-306 liposome-assisted, 301,302 conclusion, 304-306 growth and self-reproduction, 290-294 autocatalysis, 290, 292 caprylate micelles, 291,292 chemical evolution, 294 Darwinian processes, 294 functional biopolymers, formation of, 290 oleate vesicles, 292-293 RNA scenario, 290-291 vesicles, 291-293, 295-306 introduction, 287-288. Oparin's assumption, 288 membrane-assisted catalysis, 300-304

INDEX

selection and polymerization assisted by vesicles, 297-300 DDAB, addition of, 300 his-trp, 300 POPC, 298 self-assembly as spontaneous ordering process, 288-290 "Gedankenexperiment" 288-289 hydrophobic interactions, 289 micelles, 290, 291 "prebiotic soup," 288-289, 294 vesicles, formation of, 288-289, 291-293 supramolecular surfactant aggregates and chemical autopoiesis, 295-297 PEI, 250-253, 255-263, 264-275 (see also Active sites) Peptide receptors, synthetic, 185-216 (see also Synthetic) Phosphotungstate staining, 111 Polymicelles, 255 "PrebiotiC soup," 288-289, 294 (see also Origin of life) Reperfusion diseases/injuries, 218, 221, 241 (see also Synthetic enzymes) Rhinovirus, 174 "RNA world" hypothesis, 132 of origin of life, 290-291 Scanning electron microscopy (SEM), 78, 79, 98 SDS (sodium dodecyl sulfate), 111-114 SDS-PAGE, 261,262, 272-274, 279-280 Self-assembly, definition, 159 as spontaneous ordering process, 288-290 (see also Origin of life)

Index

Self-assemblies through coordination: macrocycles, catenanes, cages and tubes, 1-39 [2] catenane, self-assembling, 12-21 irreversible formation of from two preformed tings, 14-15 "Molecular lock" concept, 14 molecular-scale devices, potential as, 12 preformed rings:molecular magic rings, formation from, 12-14 self-assembly from rectangular molecular boxes, 15-19 from three-component system, 17-18 three-dimensionally interlocked, 19-21 coordination capsule assembled from 24 components, 33-36 nanometer-sized molecular hexahedron, 33 coordination nanotube, 30-33 " reversible process, formation as, 32 introduction, 2-3 coordination bonds, potentialities of, 2 Pd(II) and/or Pt(II) complexes, use of, 3-7 M3L2 complex, self-assembly of, 21-24 guest-induced assembly of, 21-22 guest-selected cage formation from dynamic receptor library, 22-24 M6L4 complexes, self-assembly of hollow, nano-sized, 24-30 catalysis and acceleration of chemical reactions in, 29-30 guest-templated synthesis of kinetically stable cage, 26-27 hydrophobic dimers in cage, formation of, 27-28

313

molecular recognition of large guests by cage, 27 octahedral cage complexes, structure of, 24-26 Pt(II)-Py bond, 27 macrocycles, self-assembling, 8-12 fast atom bombardment mass spectrometry (FABMS), 8 M6L4 macrotricyclic complexes, 8-12 structure and molecular recognition, 8 tetranuclear Pd(II) macrocycle, self-assembling, 3-7 association constants, 5 Pd(II)-(4,4'-bipy) square complexes, 3-5 Pd 4 square complexes, accumulation of, 6-7 stability of square complexes, 6 Semisynthetic artificial enzymes, 246-286 (see also Active sites) "Sergeants-and-soldiers" principle, 57, 105-106, 133 13-sheets, 186-188 artificial, 188-193 Sickle-cell anemia, 88 Snub cube, 160, 164, 168-170, 177 Snub dodecahedron, 168-170 Spherical molecular containers, 157-183 conclusion, 181 design of, general principles for, 165-171 Archimedean solids, 168-170 buckminsterfullerene, 166, 178 Platonic solids, 168, 169 self-assembly, 165-166 snub cube, 168-170 snub dodecahedron, 168-170 spheroid, 165, 166 spheroid, models for, 171

314

subunits for spheroid design and self-assembly, 166-168 examples of from laboratory and nature, 171-180 Archimedean duals, 178-180 Archimedean solids, 175-178 cuboctahedron of Archimedean solid, 176 ferritin, 178-180 heterospherophane, 176 icosahedral systems of Platonic solids, 174-175 irregular polygons, 178-180 octahedral systems of Platonic solids, 172-174 Platonic solids, 171-175 rhombicuboctahedron of Archimedean solids, 177 snub cube of Archimedean solids, 177 spheriphane, 171-172 tetrahedral system of Platonic solid, 171-172, 173 triangulation, 174 truncated icosahedron of Archimedean solid, 178 truncated octahedron of Archimedean solids, 176-177 truncated tetrahedron of Archimedean solid, 175-176 held together by 60 hydrogen bonds, 164-165 pyridine, 164-165 snub cube, 164 introduction, 158-159 crystallization as self-assembly par excellence, 159 self-assembly, 159 supramolecular chemistry, definition, 158 towards supramolecular synthesis, 159

INDEX

overview, 159-160 snub cube, 160 resorcin[4] arenes, elaborating cavities of supramolecularly, 160-164 cavities based on flexible extender, 163-164 cavities based on rigid extenders, 161-163 hydrogen bond acceptors as extender units, 160-161 monopyridine and bipyridine, 162-163 pyridine, 161-165 4-vinylpyridine, 164 why Platonic and Archimedean solids, 180-181 Superoxide dismutase (SOD) enzymes, 218 Supramolecular chemistry, definition, 158 Synthetic artificial enzymes, 246-286 (see also Active sites) Synthetic enzymes, rational design of, and potential utility as human pharmaceuticals, 217-244 computer-aided design (CAD), 233-242 for improved pharmaceutical agents, 241 improved SOD mimics, development of, 222-232 acid chloride method of synthesis of ligands, 224 bis(chloroacetamide) method of synthesis of ligands, 225 electron spin resonance (ESR), 231 molecular mechanics (MM) calculations, 227 pyridino ligand, 229-230 stability studies, 227-230 in vivo stability, 231-232

Index

introduction, 218-219 computer-aided design (CAD) paradigm, 218, 233-242 diseases in which superoxide is mediator, 21 klimitations of superoxide, 219 manganese(II)-based superoxide dismutase (SOD) enzymes, 218 reperfusion diseases/injuries, 218, 221,241 synzymes, 218, 240 mechanistic studies, 232-233 structure-activity studies, initial, 219-221 Fenton chemistry, 219, 220 hydrogen peroxide, lack of reactivity with, 220 manganese least toxic to mammalian systems, 219 nitric oxide reaction, 221 thermodynamic and kinetic stability, 221,228 Synthetic peptide receptors, 185-216 covalent models to observe noncovalent interactions between peptide chains, 188-193 artificial ~sheets, 188-193 host compounds for peptide complexation in nonpolar solvents, 194-197 macrocyclic compounds, 196 introduction, 185-188 Alzheimer disease, 187

315

[3-amyloids, 187 hydrogen bonds, 186-187 for medical and biotechnological applications, 185 noncovalent interactions, 187 protein folding, 186, 187 13-sheets, 186-188 13-turns, 187 for peptide complexation in aqueous media, 197-214 cyclodextrins (Cds), 211-213 titration microcalorimetry, 200 Synzymes, 218, 240, 246 (see also Synthetic enzymes) Titration microcalorimetry, 200 Transmission electron microscopy (TEM), 111,117, 118, 123, 124, 125 Triangulation, 174 Truncated driven nuclear Overhauser effect (TOE), 14 UV-vis spectroscopy, 117 Vesicles, 250, 288-289, 291-293, 295-306 Watson-Crick base pairs, 133, 158 X-ray crystallography, 13-14, 15, 25, 32, 36, 181,227 X-ray diffraction (XRD), 46, 66

This Page Intentionally Left Blank