Advances in Physical Organic Chemistry, Volume 42

Advances in Physical Organic Chemistry ADVISORY BOARD B. Feringa University of Groningen, The Netherlands E. Fukuzumi...

Author: John Richard

138 downloads 2104 Views 3MB 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 Physical Organic Chemistry

ADVISORY BOARD B. Feringa University of Groningen, The Netherlands E. Fukuzumi Osaka University, Japan E. Juaristi CINVESTAV-IPN, Mexico J. Klinman University of California, Berkeley C. Perrin University of California, San Diego Z. Rappoport The Hebrew University of Jerusalem, Israel H. Schwarz Technical University, Berlin, Germany C. Wentrup University of Queensland, Australia

Advances in Physical Organic Chemistry Volume 42

Editor

J. P. RICHARD Department of Chemistry University at Buffalo, SUNY Buffalo, NY 14260-3000, USA

Amsterdam – Boston – Heidelberg – London – New York – Oxford Paris – San Diego – San Francisco – Singapore – Sydney – Tokyo Academic Press is an imprint of Elsevier

ACADEMIC PRESS

Academic Press is an imprint of Elsevier 84 Theobald’s Road, London WC1X 8RR, UK Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands Linacre House, Jordan Hill, Oxford OX2 8DP, UK 30 Corporate Drive, Suite 400, Burlington, MA 01803, USA 525 B Street, Suite 1900, San Diego, CA 92101-4495, USA First edition 2008 Copyrightr 2008 Elsevier Inc. All rights reserved No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333; email: [email protected]. Alternatively you can submit your request online by visiting the Elsevier web site at http://www.elsevier.com/locate/permissions, and selecting Obtaining permission to use Elsevier material Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent veriﬁcation of diagnoses and drug dosages should be made ISBN: 978-0-12-374093-9 ISSN: 0065-3160

For information on all Academic Press publications visit our website at books.elsevier.com

Printed and bound in Great Britain 08 09 10 11 12 10 9 8 7 6 5 4 3 2 1

Contents Editor’s Preface

ix

Contributors to Volume 42

xi

Cycloaromatization reactions: the testing ground for theory and experiment

1

IGOR ALABUGIN, BORIS BREINER and MARIAPPAN MANOHARAN 1 2 3 4 5 6 7

Introduction: bonds lost and bonds formed, or chemical bookkeeping 1 Cycloaromatization reactions: breaking p-bonds and breaking the rules 2 The diversity of cycloaromatization reactions 3 The relative role of s- versus p-effects at the early reaction stage 6 s-Effects on reactant stability 10 p-Effects on reactant stability 22 Transition state effects: communication of orthogonal orbitals in the transition state of radical-anionic cyclizations 23 8 Effects on product stability 27 9 Conclusion 31 References 31

N-Acyloxy-N-alkoxyamides – structure, properties, reactivity and biological activity

35

STEPHEN A. GLOVER 1 2 3 4 5 6

Introduction 35 Synthesis 39 Structure 43 Chemical reactivity 59 Biological activity of N-acyloxy-N-alkoxyamides Conclusions 115 Acknowledgements 116 References 117

97

The Interplay between experiment and theory: computational NMR spectroscopy of carbocations HANS-ULLRICH SIEHL 1 Introduction 125 2 Alkyl and cycloalkylmethyl cations 126 v

125

vi 3 4 5 6 7 8 9 10

CONTENTS Vinyl cations 133 Cycloalkyl cations 142 m-Hydrido-bridged carbocations 144 Bicyclic and polycyclic carbocations 145 p-Stabilized carbocations 150 Heteroatom stabilized carbocations 156 Final remarks 158 Conclusions 160 Acknowledgments 160 References 160

Dynamics of guest binding to supramolecular systems: techniques and selected examples

167

TAMARA C.S. PACE and CORNELIA BOHNE 1 2 3 4

Introduction 167 Techniques 169 Examples of supramolecular dynamics studies 185 Conclusions 216 Acknowledgements 217 References 217

Mechanisms of oxygenations in zeolites

225

EDWARD L. CLENNAN 1 2 3 4

Introduction 225 Zeolites 226 Experimental considerations 230 Intrazeolite photooxygenations 232 Acknowledgement 261 References 262

Metal-catalyzed alcoholysis reactions of carboxylate and organophosphorus esters

271

R. STAN BROWN and ALEXEI A. NEVEROV 1 2 3 4 5

Introduction 271 Background theory 274 Titrations in alcohol 276 Metal ion alcoholysis and titration in alcohol 278 Transition metal ion and Ln3+ catalysts of transesteriﬁcations of neutral carboxylate and organophosphate esters 284

CONTENTS 6 Transition metal ion and La3+-catalysis of the alcoholysis of phosphate diesters 7 Conclusions 324 Acknowledgements 325 References 325

vii 308

Author Index

333

Cumulative Index of Authors

351

Cumulative Index of Titles

353

Subject Index

363

This page intentionally left blank

Editor’s Preface Volume 42-Advances in Physical Organic Chemistry The six chapters in this volume each focus on a signiﬁcant topic in organic chemistry, and branch out to include themes from biochemistry, theoretical chemistry, and materials chemistry. The cyclization reaction of linear enediynes to form aromatic ring compounds has attracted the attention of organic chemists, because of the extensive electronic reorganization involved. There has been a tremendous increase in interest in the parent Bergman cyclization reaction since the discovery that this and related reactions play a central role in the in vivo activation of tumor-speciﬁc DNA cleaving agents. Igor Alabugin presents a lucid description of the interplay between experimental and theoretical studies to map the electronic reorganization that occurs as one moves from the linear enediyne reactant to the partly aromatic transition state for these complex reactions. The generation of carbocations in strongly acidic media, and the characterization of their structure by NMR in the 1950s was a breathtaking accomplishment that led to the award of the Nobel Prize in Chemistry to George Olah. Over the past 50 years NMR spectroscopy has evolved as the most important experimental method for the direct study of structure and dynamics of carbocations in solution and in the solid state. Hans-Ullrich Siehl provides an excellent review of computational studies to model experimental NMR spectra for carbocations. This chapter provides an example of how the fruitful interplay between theory and experiment has led to a better understanding of an important class of reactive intermediates. Supramolecular systems perform complex chemistry not accessible to their molecular components. These systems are extraordinarily difﬁcult to study, because of the diverse set of skills required for the design, synthesis, characterization, and mechanistic analyses of a single supramolecular complex. The chapter by Cornelia Bohne focuses on a particularly valuable skill, and presents an overview of modern techniques used for kinetic analyses of reactions that occur in supramolecular complexes, with an emphasis on the characterization of fast reactions. Zeolites are aluminosilicate solids with cavities of dimensions comparable in size to small- or medium sized organic molecules. Many fascinating questions have been raised about the effect of conﬁnement of organic reagents in these cavities on chemical reactivity. The progress towards answering these questions for oxygenation reactions in zeolites is reviewed in a chapter by Edward Clennan. ix

x

EDITOR’S PREFACE

The design of catalysts of the cleavage of RNA and DNA, and the determination of the mechanism for these cleavage reactions has considerable intellectual appeal, and the potential to produce cleavage reagents with applications in medicine. These studies are most-often carried out in water, with the ultimate aim of developing in vivo applications for particularly powerful catalysts. It has become clear in recent years that the activity of these catalysts is sometimes reduced in water compared with alcohol solvents with lower dielectric constants than water. This observation has proven useful in developing conditions for detoxiﬁcation of organophosphorus nerve agents by alcoholysis reactions. The chapter by R. S. Brown summarizes the results of recent studies on the mechanism for metal-catalyzed alcoholysis reactions of carboxylate and organophosphorus esters. In the second chapter with a biological theme, Stephen Glover summarizes what is known about the structure, physical, spectroscopic, chemical and biological properties of N-acyloxy-N-alkoxyamides. The convergence of several functional groups at a single nitrogen results in interesting chemical properties for these compounds as well as in their mutagenic activity, which is a useful probe for drug-DNA interactions. John P. Richard University of Buffalo

Contributors to Volume 42 Igor Alabugin Department of Chemistry and Biochemistry, Florida State University, Tallahassee, Fl 32306-4390, USA Cornelia Bohne Department of Chemistry, University of Victoria, PO Box 3065, Victoria, BC, V8W 3V6, Canada Boris Breiner Department of Chemistry and Biochemistry, Florida State University, Tallahassee, Fl 32306-4390, USA R. Stan Brown Department of Chemistry, Queen’s University, Kingston, Ontario K7L 3N6, Canada Edward L. Clennan Department of Chemistry, University of Wyoming, 1000 East University Avenue, Laramie, WY 82071, USA Stephen A. Glover School of Science and Technology, University of New England, Armidale, NSW 2351, Australia Mariappan Manoharan Department of Chemistry and Biochemistry, Florida State University, Tallahassee, Fl 32306-4390, USA Alexei A. Neverov Department of Chemistry, Queen’s University, Kingston, Ontario K7L 3N6, Canada Tamara C.S. Pace Department of Chemistry, University of Victoria, PO Box 3065, Victoria, BC, V8W 3V6, Canada Hans-Ullrich Siehl Institute of Organic Chemistry I, Ulm University, AlbertEinstein-Allee 11, D-89069, Ulm, Germany

xi

This page intentionally left blank

Cycloaromatization reactions: the testing ground for theory and experiment IGOR ALABUGIN, BORIS BREINER and MARIAPPAN MANOHARAN Department of Chemistry and Biochemistry, Florida State University, Tallahassee, FL, USA 1 Introduction: bonds lost and bonds formed, or chemical bookkeeping 1 2 Cycloaromatization reactions: breaking p-bonds and breaking the rules 2 3 The diversity of cycloaromatization reactions 3 4 The relative role of s- versus p-effects at the early reaction stage 6 5 s-Effects on reactant stability 10 Strain and antiaromaticity in cyclic enediynes 11 Control of strain through ligand–metal coordination 16 Steric assistance of ortho substituents 17 Control of steric repulsion with positively charged ortho substituents 20 Hybridization effects 21 6 p-Effects on reactant stability 22 7 Transition state effects: communication of orthogonal orbitals in the transition state of radical-anionic cyclizations 23 8 Effects on product stability 27 s-Effects 27 p-Effects 30 9 Conclusion 31 References 31

1 Introduction: bonds lost and bonds formed, or chemical bookkeeping Chemical reactions involve the reorganization of electron density in which bonds are broken and formed in accordance with the reaction mechanism. The energetic consequences of such reorganization are accounted for by the reaction thermodynamics. Unless a very weak bond is broken with a signiﬁcant increase in entropy, bond breaking is thermodynamically unfavorable and leads to the formation of highly reactive intermediates: radicals and diradicals through bond homolysis, anions and cations through heterolysis, and carbenes as a result of a-elimination. On the other hand, bond formation is favorable and when a strong bond is formed, serves as a thermodynamic sink that often terminates a cascade of chemical transformations, thus balancing the thermodynamic checkbook or registering a proﬁt. In a perfect world of balanced checkbooks, the bond breaking and bond forming processes are synchronized, as it happens (albeit not always perfectly) in concerted pericyclic reactions. Even when such synchronization is not perfect, unimolecular 1 ADVANCES IN PHYSICAL ORGANIC CHEMISTRY VOLUME 42 ISSN 0065-3160 DOI: 10.1016/S0065-3160(07)42001-9

r 2008 Elsevier Inc. All rights reserved

2

I. ALABUGIN ET AL.

rearrangements are still rarely accompanied with a change in the number of chemical bonds. In other words, the number of chemical bonds generally is conserved.

2 Cycloaromatization reactions: breaking p-bonds and breaking the rules An interesting breakdown of the tendency for the conservation of chemical bonds occurs in the so-called cycloaromatization reactions1 (Scheme 1). In these reactions, a s-bond is formed at the expense of two p-bonds and, thus, the process leads to a net loss of one chemical bond that is intrinsically unfavorable thermodynamically. Formation of the new s-bond leads to ring closure, whereas the net loss of a bond leads to the formation of two radical centers, which can be either inside (the ‘‘endo’’ pattern in Scheme 1) or outside of the newly formed cycle (the ‘‘exo’’ pattern). Note that s-radicals are formed through the endo path, while exo-closures may produce either a s-radical when a triple bond is involved or a conjugated p-radical when the new bond is formed at the central carbon of an allene. The parent version of this process is the transformation of enediyne 1 into p-benzyne diradical2 (the Bergman cyclization), shown in Scheme 2. This transformation requires signiﬁcant thermal energy to overcome the activation barrier and is endergonic. The endergonicity of the Bergman cyclization is not surprising and stems logically from the net loss of a bond in the product. What is surprising is that the energy penalty (DH) for this intrinsically unfavorable

a

X

a X

c

b X b

endo, endo

endo, exo c

c′

X

X

exo, exo bis-allene

exo, exo bis-alkyne

Scheme 1 Schematic representation of cycloaromatization reactions. Double lines correspond to the out-of-plane p-systems of a bis-alkyne reagent. Only orbitals of the in-plane p-system in the reactant and of new s-bond and radical centers in the product are shown explicitly.

CYCLOAROMATIZATION REACTIONS

3 H

. 2RH

1

. 2

H

Scheme 2 Bergman cyclization of (Z)-3-ene-1,5-diyne.

transformation is only 8–9 kcal/mol.3 This low value results from the aromatic stabilization of the product provided by the completely conjugated system of out-ofplane p-orbitals at each carbon of the enediyne system. Thus, even though the outof-plane p-array does not directly participate in the bond forming/bond breaking event, it is not a simple bystander in the overall process. Developing aromatic stabilization in the conjugated out-of-plane system plays an important role in deepening the potential energy well for the diradical product to the extent that this species can be trapped by external reactions such as H-atom abstraction or addition to a p-system. The goal of this review is to dissect the relative contributions of the two orthogonal p-systems in the most general sense and provide a clear and general analysis of electronic effects in the Bergman cyclization as they apply to and originate from these two different sources. We will outline the relative role of the two systems in deﬁning the kinetics and thermodynamics of the parent Bergman cycloaromatization pathway, delineate factors that affect these systems separately and discuss ways to provide communication between the two orthogonal p-systems. This review will be selective rather than comprehensive since a number of excellent and detailed reviews are available on various aspects of the chemistry (and biochemical application) of cycloaromatization reactions in the recent literature: Rawat and Zaleski4 compared and contrasted geometric and electronic effects in thermal Bergman cyclizations to those of the recently developed metalloenediynes, while Basak and coworkers5 concentrated even more closely on the use of enediynyl ligands in chelation-controlled Bergman cyclizations. The Schreiner group6 reviewed the application of computational methods toward a variety of cycloaromatization reactions and Ko¨nig and coworkers7 provided a compendium of experimental results related to the electronic effects on the Bergman cyclization. Intricate details of biological activity of natural enediynes were most recently reviewed by Shen and coworkers,8 whereas medicinal utility of designed enediynes was expertly discussed by Jones and Fouad.9

3

The diversity of cycloaromatization reactions

Fig. 1 summarizes cyclizations of the two prototype molecules, (Z) hex-3-ene-1,5diyne and (Z) hept-3,5,6-triene-1-yne. In all of these processes, bonds are formed from in-plane p-orbitals in the presence of an orthogonal p-system. However, it is clear that the properties of the newly formed cyclic conjugated systems can be quite

4

I. ALABUGIN ET AL.

a

a

b

b

Schreiner

Bergman c

d

d c

Schmittel

Myers-Saito

Fig. 1 Cycloaromatization reactions of (Z) hex-3-ene-1,5-diyne and (Z) hept-3,5,6-triene1-yne. X

X R

R′

R

R′

X = NH, O, S, PH, + NH2 , OH+, SH+, CH-, etc.

Scheme 3 Formation of 5-membered heterocycles through cyclization reaction.

different. A new aromatic system is formed only in the Bergman and Myers-Saito cyclizations, and thus, in a sense, only these two processes in Fig. 1 are bona ﬁde cycloaromatization reactions. Note also that s,s-diradicals are produced from enediynes, whereas s,p-diradicals are formed from enyne allenes. The benzene ring is not the only aromatic moiety that can be formed in a cycloaromatization process and formation of other aromatic systems was considered as the driving force of new cycloaromatization processes. For example, Matzger (X ¼ S)10 and Schreiner11 investigated whether the 5-membered aromatic heterocyclic systems of pyrrole, thiophene, and furan can be formed in these processes (Scheme 3), while Alabugin12 (Scheme 4) and Rabinovitz13 (Scheme 5) investigated anionic species where cycloaromatization beneﬁted (albeit to a different extent) from the aromaticity of the cyclopentadienyl anion. Whitlock et al.14 discovered a reductive cyclization of enediynes promoted by lithium naphthalenide that provides substituted fulvenes and suggested a dianionic mechanism (Scheme 6). However, even now it is still unclear whether the enediyne dianion is indeed the cyclizing species or whether the initially formed acyclic radicalanion cyclizes ﬁrst to give a fulvene radical-anion which is further reduced by lithium to give the cyclic dianion. Aromaticity of the products is only one of the factors accounting for the efﬁciency of these cyclizations as evidenced by the discovery of a dianionic synthesis of nonaromatic 5-membered heterocycles by Tamao and coworkers who found that reduction of bis(phenylethynyl)dialkylsilanes with lithium naphthalenide resulted in formation of a cyclized product by endo–endo cyclization15 (Scheme 7). Under analogous conditions, cyclic 1,2-bis(silylethynyl)benzenes afford 1,4-dilithio-2,3-disylylnaphthalenes via dianionic endo–endo cyclization leading to the formation of a 6-membered ring (Scheme 8).16

CYCLOAROMATIZATION REACTIONS F

5

F

F N

N F

+.

-.

F F N

F

hv

F F

F

F F

F

F F

F

N

F

*

F

F F

F

N

N

F

F

F

H R

R R

R

R = tetrafluoropyridinyl (TFP)

H

Scheme 4 Reductive C1–C5 cyclization of enediynes.

Δ

K, 2H I2

Scheme 5 Thermal and reductive cyclizations of cross-conjugated enediynes.

Along similar lines, Wenthold applied the radical-anionic idea to the Cope rearrangement,17 which can be considered a more saturated analog of the Bergman cyclization.18 The radical-anionic Cope rearrangement was found to have an inverted potential energy surface with the cyclic intermediate being more stable than the radical-anion corresponding diene. Consequently, one-electron reduction of 2,5-dicyano-1,5-hexadiene results in spontaneous cyclization under the experimental conditions. Certainly, the role of photochemical excitation may be displayed in different ways19 and even in those cases when no aromatic system is formed, topologically

6

I. ALABUGIN ET AL.

D

1. Li / naphthalene 2. D2O D

Scheme 6 Dianionic cyclization of enediynes promoted by double lithium naphthalenide reduction.

Ph

Ph

Ph

Ph

lithium naphthalenide (4 equivalents)

Si R

Li

Li

Si

R

R

R

Scheme 7 Dianionic endo–endo cyclization of diethynylsilanes. Li

Si

Si

lithium naphthalenide Si

Si Li

Scheme 8 Endo–endo reductive cyclization of cyclic 1,2-bis(silylethynyl)benzenes.

similar cyclization photoreactions are still possible. A classic example involves the photochemical cyclization of diethynylmethanes discovered by Zimmerman and Pincock20 around the same time as the thermal Bergman cyclization was reported by Bergman. The reaction proceeded upon direct irradiation of the diyne in isopropyl alcohol or, alternatively, when triplet-sensitized by acetophenone or xanthone. As in the Bergman cycloaromatization, two hydrogen atoms are abstracted from a suitable donor (Scheme 9). Taken together these data suggest that the family of cycloaromatization reactions will continue to grow and new interesting transformations may be discovered in the future.21 Thus, understanding of general factors which can be used to control cycloaromatization process should be of signiﬁcant value for organic chemistry.

4

The relative role of r- versus p-effects at the early reaction stage

Although data presented in the previous section illustrate the diversity of cycloaromatization reactions, most of the following discussion will concentrate on the Bergman cyclization – a reaction that has been studied intensively in recent decades due to its role in the mechanism of biological activity of natural anticancer antibiotics.8,9 We will take advantage of the large body of data produced by this

CYCLOAROMATIZATION REACTIONS

7

hν 2-propanol

Scheme 9 Photochemical cyclization of diethynylmethanes. Although this reaction is topologically analogous to the cyclization reactions discussed earlier, it does not lead to an aromatic product.

research and use this archetypal reaction to illustrate the general electronic factors involved in a typical cycloaromatization process. Since both in-plane and out-of-plane p-systems are involved, a clear dissection of their respective roles in the cycloaromatization process is beneﬁcial for the analysis of electronic effects on reactivity. Galbraith et al.22 approached this problem by utilizing valence bond (VB) theory, which provided insight into the structural and electronic nature of the transition state and illustrated the role of the s- and p-frameworks. The dominant VB conﬁguration in the enediyne was found to play only a small role in the p-benzyne diradical, while the role of the dominant VB conﬁguration (accounting for the biradical) in the product was seemingly insigniﬁcant in the starting material. The avoided crossing of the energy curves of these two VB conﬁgurations leads to the transition state, but the changes in the electronic wave function and in geometry are non-synchronous. The transition state (TS) was found to be 80% product-like geometrically and 70% reactant-like electronically in an apparent violation of the Hammond postulate.23 This is a direct consequence of the high stabilization energy provided by the formation of the aromatic system, the development of which only has signiﬁcance once the reaction approaches the product. Independently, Alabugin and coworkers utilized classic molecular orbital (MO) correlation diagrams for understanding cycloaromatization processes. The advantage of this approach is that it clearly describes changes accompanying the cyclization process at the level of individual MOs and allows comprehensive analysis of electronic reorganization that accompanies the cyclization processes including the relative roles of the two orthogonal p-arrays. As an example of the latter approach, let us analyze the MO correlation diagram for the Bergman cyclization of the parent enediyne given in Fig. 2. This diagram shows the transformation of MOs of the starting material to MOs of TS and the product. The most important change at the TS stage is the dramatic increase in the energy of HOMO-1 of the enediyne, which is paralleled by the decrease in the energy of LUMO+1 due to the transformation of these in-plane orbitals (bonding and antibonding respectively) into non-bonding MOs (radical centers). Destabilization of the HOMO-1 is the most dramatic energy penalty that the molecule has to pay to reach the TS. This effect is partially offset by the s-bond forming interaction that stabilizes HOMO-2 (in phase combination of the in-plane enediyne p-orbitals). Developing aromatic stabilization of the out-of-plane p-system lags behind the

8

I. ALABUGIN ET AL.

A

S A

S

LUMO A

LUMO HOMO

S HOMO A S

A

S Reactant

TS

Product

Fig. 2 MO correlation diagrams for the Bergman cyclization (adopted from ref.24). Out-ofplane MOs are shown in black, in-plane MOs are shown in blue.

above changes but becomes progressively more and more important as the reaction coordinate approaches the p-benzyne product. The above correlation diagrams can be used to illustrate the relative timing of changes in the s- and the p-systems. It is obvious from Fig. 2 that at the TS stage the s-system is more involved in the electronic reorganization process and that changes in the out-of-plane p-system mostly occur after the cyclization TS is passed. This analysis eliminates description of the Bergman cyclization as a 6e electrocyclic reaction that involves the out-of-plane p-system (Fig. 3). The Bergman cyclization should be considered the ﬁrst step of an asynchronous [2+2] cycloaddition, interrupted by the aromatic stabilization of the intermediate 1,4-diradical, or as a Cope rearrangement interrupted by aromatic stabilization of the TS.18

CYCLOAROMATIZATION REACTIONS

9

Fig. 3 Incorrect description of the Bergman cycloaromatization as a 6p-electrocyclic ring closure (top). A better description shown for comparison (below).

It is interesting to use MO correlation diagrams for a comparison of the two alternative cyclizations of enediynes (Fig. 4): the Bergman cyclization and the C1–C5 cyclization (Fig. 1b,a), investigated ﬁrst computationally by Schreiner.25 In both these processes, a s bond is formed at the expense of the same two in-plane p-bonds in either an endo–endo (Bergman) or an endo–exo (C1–C5) fashion (Fig. 4). Not surprisingly, the changes in in-plane MOs are comparable in the two diagrams and consistent with these p-bonds being broken in order to form the new s-bond. In contrast, changes in the out-of-plane MOs are signiﬁcantly different. In contrast to the Bergman cyclization where the enediyne highest occupied molecular orbital (HOMO) is strongly stabilized, such stabilization is small in the non-aromatic didehydrofulvene product of the C1–C5 cyclization. Another interesting difference is that the LUMO+2 of the reactant correlates with the lowest unoccupied molecular orbital (LUMO) of the product in the C1–C5 cyclization instead of LUMO+1. Such changes illustrate the different properties of the out-of-plane conjugated systems formed in the two processes (aromatic for the Bergman and non-aromatic for the C1–C5 cyclization). The above analysis illustrates why it is helpful to consider the enediyne moiety as two independent p-systems. As discussed above, the conjugated out-of-plane p-system of the reactant is smoothly transformed into the conjugate p-system of the product (e.g., the aromatic system of benzene) without an overall change in the number of bonds. We will refer to this group of electrons as ‘‘out-of-plane MOs’’ or simply as ‘‘the p-system’’. In contrast, the two in-plane p-bonds are transformed in a more drastic way to the new s-bond and a pair of radical centers. We will refer to this system of orbitals as ‘‘in-plane p-bonds’’ or as ‘‘the s system’’. Since the breaking p-bonds and the developing radical centers are orthogonal to the out-of-plane p-system and the aromatic system develops mostly after the system proceeds through the transition state, the activation barrier for the Bergman cyclization is relatively insensitive to effects in the p-system. This provides a challenge for the substituent control of such reactions: it is not always clear whether hyperconjugative and inductive s-effects are sufﬁciently strong to control these reactions efﬁciently. On the other hand, the well understood p-effects are usually indirect and they display themselves mostly in reaction thermodynamics.27 Due to the different role of s- and p-effects, we will separate the general description of electronic effects into s- and p-effects in this study. s-Effects include

10

I. ALABUGIN ET AL. LUMO+2 (reactant)

LUMO+1 (reactant)

H H

Energy

Energy

H

HOMO

H

HOMO

H H

Reaction cooordinate

Reaction cooordinate

Fig. 4 The crossings of in-plane and out-of-plane frontier MOs in the radical-anionic Bergman and C1–C5 cyclizations, respectively calculated at the B3LYP/6-31G** level. Adapted from reference26.

the in-plane p- and s-orbitals. We will analyze these effects (which can be either stabilizing or destabilizing) at the three reaction stages (reactants, TS, and products) separately. All three possibilities can have a signiﬁcant effect on the reaction enthalpy and activation energy of the cycloaromatization reaction. The ability to inﬂuence the reaction at or near the transition state provides the most direct way to control the reaction rates but reactant stabilization/destabilization is also a viable approach. Effects on the products are indirect but can be readily understood using dissection of kinetic and thermodynamic contributions to the reaction barrier provided by the Marcus theory, an approach that we used in the analysis of other pericyclic reactions28 and radical cyclizations.29

5

r-Effects on reactant stability

In a seminal work, Koga and Morokuma suggested that the high activation energy of the Bergman cyclization is due to the strong electron repulsion between the

CYCLOAROMATIZATION REACTIONS

11

in-plane occupied acetylene p-orbitals.30 These authors also pointed out that the lower activation energy of the Myers-Saito cyclization can be attributed to the decrease in electron repulsion due to the transoid orientation of the interacting p-orbitals.31 Recently, we analyzed the role of electron repulsion relative to bond breaking and antiaromaticity effects on a quantitative basis using Natural Bond Orbital (NBO) analysis.24 Two other destabilizing factors were considered at the initial stage of the cyclization in addition to four-electron repulsion between the ﬁlled in-plane acetylenic p-orbitals – distortion/breaking of the acetylenic bonds as a result of their bending, and the fact that, at a distance of ca. 3 A˚, the in-plane p-orbitals become parallel and reach a geometry that resembles the antiaromatic TS of the symmetry forbidden [2s+2s] cycloaddition (vide infra).

STRAIN AND ANTIAROMATICITY IN CYCLIC ENEDIYNES

The effect of cyclic constraints that bring the two ends of the enediyne system closer together is extensively studied. This is not surprising because nature utilizes the higher reactivity of cyclic enediynes in the naturally occurring enediyne antibiotics. When the enediyne moiety is incorporated in a large cycle (more than 10 carbon atoms), the Bergman cyclization is not fast enough to proceed at physiologically relevant temperatures. In contract, enediynes in 9-membered cycles are too unstable to be isolated (unless stabilized by additional electronic effects, vide infra), whereas 10-membered enediynes cyclize within hours (Fig. 5, Table 1).32 Nicolaou and coworkers33 suggested a simple empirical criterion based on ‘‘the c–d distance’’ (the proximity of the terminal carbons of the enediyne moiety) as a measure of increased ring torsion, which is relieved when the compounds undergo the Bergman cyclization. Only below the ‘‘critical upper limit’’ (the c–d distance of 3.2–3.3 A˚), the cyclization is expected to proceed at a measurable rate at ambient temperature. More elaborate analysis incorporates the relative strain energies of the ground and

c (CH2)n d

c

OH

c

OH

c

OH

c

d

OH

d

OH

d

O O

d I

OH

II

III

O

IV

Fig. 5 Cyclic model enediynes used by Nicolaou to establish the critical distance model.

12

I. ALABUGIN ET AL.

Table 1 Calculated c–d distances and stabilities of cyclic enediynes Compound

n

Ring Size

c–d Distance (A˚)

Stability

I II III IV

1 2 3 – – – –

9 10 11 10 10 10 10

2.84 3.25 3.61 3.20 3.29 3.34 3.42

Unknown t1/2 ¼ 18 h at 251C Stable at 251C t1/2 ¼ 11.8 h at 371C t1/2 ¼ 4 h at 501C t1/2 ¼ 2 h at 501C Stable at 251C

Cl

Cl

Fig. 6 Cyclization of the ﬁrst isolable a 9-membered enediyne.

transition states.34,35 In the 11-membered system, the distance is only slightly larger (ca. 3.5 A˚), yet such enediynes are rather unreactive.36 Overall, 9-membered enediynes are usually too unstable to be utilized in the design of new anticancer drugs. An interesting way to overcome this limitation was shown recently by Jones and coworkers who utilized an electronic effect in a rare example of intentional deceleration of the Bergman cyclization (Fig. 6). Substitution of a hydrogen atom with a chlorine atom at the vinylic position of a 9-membered cyclic enediyne37 decreased the cyclization rate and allowed the ﬁrst enediyne with a 9-membered ring to be isolated. Nevertheless, the half-life of this compound was found to be only 6 min at 313 K, corresponding to an activation barrier of o18 kcal/mol. An extensive computational analysis expanded the range of the c–d distances for reactive cyclic enediynes to 2.9–3.4 A˚.38 By comparing unsubstituted enediynes with dialkyl-substituted enediynes, it was found that the activation enthalpy is dependent on factors other than the c–d distance and that reactivity hinges on a subtle interplay of steric and electronic effects that accompany distortion caused by incorporation into a macrocycle. For example, since alkyl substituents stabilize acetylenic bonds to a greater extend than oleﬁnic bonds,39 such substituents stabilize the starting material, thus increasing both the activation barrier and the reaction endothermicity. Despite the contribution of the above factors, it is clear that bending of the two alkyne moieties toward each other increases the energy of the system as illustrated by the three plots in Fig. 7.24 These plots illustrate the relation between the ring size and calculated cyclization parameters in more detail and dependence of the total energy of the system from the c–d distance. Interestingly, simple bending of alkyne moiety reproduced the effect of cyclic restraints reasonably well. A similar conclusion has been reached even earlier by Kraka and Cremer.40

CYCLOAROMATIZATION REACTIONS

13

35

Energy, kcal/mol

30 25

3

20 2

15 10

1

5 0 (4.50)

(4.00)

(3.50)

(3.00)

(2.50)

(2.00)

(1.50)

C1-C6 distance, Ang

Fig. 7 Internal reaction coordinate (IRC) computations for the Bergman cyclization of model enediynes.

It is clear that although initially the effect is relatively small and can be easily masked by other steric or electronic factors, its importance increases as the bending progresses. Where does the energy increase come from? Can we offer an electronic description to the phenomenological term ‘‘strain’’? A priori, one can consider such factors as four-electron repulsion of ﬁlled in-plane orbitals and distortion/breaking of the acetylenic bonds as a result of their bending. Fig. 8 illustrates that weakening of the triple bonds is only of minor importance at the early reaction stage: changes in the C1-C2 NBO p-bond orders (both in-plane and out-of-plane) at C1–C6 distances above 3 A˚ are negligible. Interestingly, even at the TS, the in-plane p-bonds are only 30% broken. This resilience of the acetylenic p-bonds toward bending agrees well with the observation that electronic changes in the Bergman cyclization lag behind the changes in geometry.22 Electron repulsion between the ﬁlled in-plane p-orbitals and decrease in the stabilizing p–p* interactions are more important for reactant destabilization and increased reactivity of cyclic enediynes. In particular, we found that the repulsive four-electron interaction between the occupied in-plane acetylenic p-orbitals is indeed the largest and the most important contribution to the ground state destabilization, in good agreement with the suggestion of Koga and Morokuma.30 This interaction increases continuously along the IRC path and accounts for the bulk of the energy increase in the early part of the reaction path. In contrast, the stabilizing p–p* interactions that lead to formation of the C1–C6 s-bond and to cleavage of the C1–C2 and C5–C6 p-bonds – the changes most directly associated with the Bergman cyclization – display a more complicated pattern. Their magnitude decreases at ﬁrst, vanishes at the Nicolaou’s threshold and rapidly increases at distances shorter than 3 A˚.

14

I. ALABUGIN ET AL. R 1

πi(C1-C2)

0.95

Bond order

0.9

πo(C1-C2)

0.85 0.8 TS

0.75 0.7 0.65 (4.50)

(4.00)

(3.50)

(3.00)

(2.50)

(2.00)

(1.50)

C1-C6 distance, Ang

Fig. 8 The changes in the NBO p-bond order of in-plane (pi, circles) and out-of-plane (po, diamonds) acetylenic bonds along the IRC pathway for the Bergman cyclization of (Z)-hex-3-ene-1,5-diyne.

Due to the lack of increase in stabilizing p–p* interactions at the reaction stage where the destabilizing four-electron repulsive interactions increase steadily, the inward bending of alkyne moieites in unstrained enediynes leads to continuously developing reactant destabilization without any compensation from the increased C1–C6 bonding. Only in the 9-membered enediyne the decrease in the C1–C6 distance results in an immediate increase in the extent of C1–C6 s-bond formation. Fig. 9 illustrates that the two acetylenic systems become nearly parallel at C1–C6 distances close to 3 A˚ where the constructive overlap of the p-orbital with one of the p*-nodes is compensated by a destructive overlap with the other p*-node (Fig. 9, bottom). From a conceptual point of view, the properties of the in-plane p-system at the 3 A˚ threshold bear a striking resemblance to the interaction of the two p-bonds in D2h cyclobutadiene where the p–p* interaction is zero and the p–p repulsion is considerable, thus accounting for the extreme instability of this antiaromatic molecule.41 Even more relevant is a comparison with the TS of the symmetry forbidden thermal [2s+2s] cycloaddition (Fig. 10) which prompted us to call this region ‘‘antiaromatic’’.42 The ‘‘antiaromatic region’’ is not important for the reactivity of the parent enediyne because the activation energy is determined only by the energy difference between the reactant and the TS. However, for the cyclic enediynes in Fig. 7 in which the C1–C6 distances are 3.39 and 2.92 A˚, respectively, ‘‘antiaromaticity’’ of the reactant should be relevant to the reaction kinetics. In addition, the role of repulsion between the in-plane ﬁlled orbitals is accentuated by a parallel decrease in the attractive two-electron interaction between the p and p* orbitals which vanishes at the 3.2 A˚ distance between the terminal carbon atoms.

CYCLOAROMATIZATION REACTIONS

15

Fig. 9 NBO contours of orbitals involved into the in-plane p–p and p–p* interactions for the reactant (a), IRC point at 3 A˚ C1–C6 separation (b) and TS (c). Repulsive p–p interaction is shown on top. Bond-forming p–p* interaction is shown at the bottom.

H H

H H

H H

H H

Fig. 10 Comparison of the ‘‘antiaromatic region’’ (i.e., the in-plane p–p* interaction pattern at 3 A˚ C1–C6 distance) with the antiaromatic TS of the [2s+2s] cycloaddition.

This analysis conﬁrms that the effect of cyclic constraints is not purely steric but also has an electronic component. Another aspect of this dichotomy is shown in Fig. 11 which illustrates the decrease in the energy gap between the frontier in-plane p-MOs. The decrease in the C1–C6 distance destabilizes the occupied MO where the interaction between the end orbitals is antibonding and, at the same time, stabilizes the empty MO where the p*-orbitals overlap constructively. As a result, the efﬁciency of the photochemical Bergman cyclization should increase and, indeed, the most efﬁcient photo-Bergman cyclizations reported in the literature involve cyclic enediynes.43 Again, the analogy with interrupted [2+2] photocycloaddition is instructive.

16

I. ALABUGIN ET AL.

Fig. 11 Comparison of the energy gap between highest occupied in-plane MO and lowest unoccupied in-plane MO in acyclic (left) and cyclic (right) benzannelated enediynes. Incorporation of the enediyne moiety into a cyclic structure simultaneously increases the energy of the occupied MO and lowers the energy of the unoccupied MO. CONTROL OF STRAIN THROUGH LIGAND– METAL COORDINATION

An elegant way to control both strain and electronics is to take advantage of metal coordination to an enediynyl ligand – a topic that has been reviewed intensively44 and, thus, will be discussed rather brieﬂy here. The inﬂuence of metal complexation on enediynes can be divided into several aspects: It can work through either s-donor coordination or p-complexation, resulting in any combination of geometric changes (analogous to the change in c–d distance proposed by Nicolaou) or by their inﬂuence on the electronic environment of the enediyne moiety. Strain-based systems work the same way as the cyclic enediynes: they reduce the c–d distance in the molecule. The ﬁrst example was provided by Buchwald and coworkers,45 who used metal complexation of a PPh2 substituted enediyne to produce a species with considerably lower activation barrier (compared to the noncoordinated enediyne) (Scheme 10). While in the unbound enediyne the c–d distance is 4.1 A˚, this distance is diminished upon metal complexation: 3.3 A˚ for Pt(II) and Pd(II), and 3.4 A˚ for Hg(II). The Pt and Pd species cyclize in the solid state at only slightly elevated temperatures, and give Bergman products below ambient temperature in solution. While the change in reactivity was attributed to the change in distance between the alkyne termini, an accelerating inﬂuence of the metal cannot be ruled out. Since both s acceptors and p donors at the alkyne termini are known to facilitate the Bergman cyclization, Zaleski and coworkers established a model46 in which the coordination of a Lewis acid (metal ion) would change the electronic environment in favor of diradical formation (Scheme 11). Metal coordination was assumed to accelerate the reaction by inductive effects. A comparison of the reactivity of the benzylated compound in presence of Mo was used to rule out any p inﬂuence on the acetylenes. Zaleski and coworkers47 later expanded upon this line of work by using tetradentate enediyne ligands, in which the reactivity could be modulated by metal complexation (Scheme 12). The advantage of tetradentate ligand systems is the fact that they help to avoid dimerization, and that with the right choice of metal (and therefore the right

CYCLOAROMATIZATION REACTIONS PPh2

17 Ph2 P MCl2 P Ph2

MCl2 PPh2

Ph2 P MCl2 P Ph2

Δ Pt: 61°C Pd: 81°C

Scheme 10 Metal coordination activates enediynes by drastically reducing the distance between the acetylenic carbons. S

SK

S

SK

S

S

S

S

S

S

S

S

S S

[MoCp2Cl2], CHD, 60°C 30 min, 40 %

S S

S Cp Mo S Cp

Ph [MoCp2Cl2], CHD

no reaction

60°C, THF/MeOH Ph Ph

DMSO, CHD, 180 °C

S

S

24 h, 60 %

S

S

Ph

S

S

S

S

Ph

Ph

Ph DMSO, CHD Ph

180 °C, 24 h

S DMSO, CHD, 120°C Mo Cp Cp 5 h, 15% S

no reaction

S S

S Cp Mo S Cp

Scheme 11 Model systems illustrating enediyne activation through metal complexation.

coordination geometry), tuning of both the c–d distance and electronic effects is easily accomplished.

STERIC ASSISTANCE OF ORTHO SUBSTITUENTS

An additional way to use strain for control of reactivity is steric assistance. This effect can be easily introduced through a choice of appropriate ortho substituents in benzannelated enediynes (Scheme 13). Whereas activation energies for the neutral para substituents lie within a range of only 0.6 kcal/mol,48 the presence of ortho-NO2, NH2, CHO, CF3, and syn-OMe groups results in large changes in the activation energy (Fig. 12).49 Depending on the substituent, three different factors account for these changes: steric assistance (decrease in steric destabilization in TS), extra stabilization of the TS, and decrease in TS stabilization. In particular, the NO2, CF3, syn-CHO, and syn-OMe groups were predicted to decrease the activation energy for Bergman cyclization by destabilizing reactants through steric repulsion between the ortho-substituent and the in-plane acetylenic

18

I. ALABUGIN ET AL.

S

S N

Δ

N M

Mn+

S

N

−Mv+

N

S

N

S

N

S

Scheme 12 Tetradentate enediyne ligands used to achieve enediyne activation by metal complexation while avoiding dimerization at the metal.

Repulsion or Attraction X

X

H

X

.

Heat

H-donor

. H +

X = H, F, Cl, CH3, CN, CF3, OH, NO2, CHO, OCH3, NH2, NH3

Scheme 13 Bergman cyclization of enediynes bearing ortho-substituents. H

R

H

H

H

O

H

H

R

33

o-OMe(a,e)

o-OMe(a,s)

o-Me

p-OH

o-OH(a) o-Cl

32 o-CN

p-NO2

p-CN

R2 = 0.9631

30

p-CHO

p-NH2(p) p-OMe p-Me

p-CF3 p-Cl

H O

o-NH3

0.015

0.02

0.025

H

H

o-CHO(s)

0.03

H

o-OMe(s,s) O H H

o-CHO(a)

o-NO2

27 0.01

p-NH2(np) o-NH2(np)

+

28

p-

o-CF3

29

o-NH2(p)

o-OH(s)

o-F

31 p-NH3

O

o-OMe(s,e)

0.035

0.04

0.045

0.05

Fig. 12 Correlation between the calculated activation energy of the Bergman cyclization and the product of natural charges at the terminal acetylenic atoms of benzannelated enediynes. Only para substituents obey the correlation. Adapted from reference49.

orbitals. This interaction becomes less signiﬁcant in the TS when the acetylene moiety is bent away from the ortho-functional group. As a result, the TS and the product are destabilized to a lesser extent than the starting material and the activation barrier is decreased through the classic steric assistance mechanism.50 The steric nature of this effect is clear from the comparison with the respective para

CYCLOAROMATIZATION REACTIONS

19

isomers. For example, the ortho-nitroenediyne is less stable than the para isomer by 5.8 kcal/mol at B3LYP/6-31G** level because of the steric repulsion discussed above. The difference in stability decreases to 2.8 kcal/mol in the TS and to 2.0 kcal/ mol in the product accounting for the 3 kcal/mol decrease in the cyclization barrier and 3.8 kcal/mol decrease in reaction endothermicity (Fig. 13). Sterically compact ortho substituents such as Me, OH, anti-OMe, F, Cl, and CN destabilize the ground and transition states to a similar degree. As a result, changes in the activation energy are minor. The net effect of these substituents on the cyclization rate is similar to that of the para substituents, and the corresponding computational data ﬁt well into the correlation in Fig. 12.49 Steric compression of enediyne moiety by ortho substituents is further illustrated in Fig. 14, which clearly shows that steric destabilization is accompanied by decrease in the distance between carbons C1 and C6. These computational predictions have been tested experimentally. Kinetic measurements conﬁrmed that both ortho-NO2 and ortho-CHO substituents substantially decrease activation energies for the Bergman cyclization supporting earlier computational predictions.51 Ortho substitution also allows one to control other steps in the cycloaromatization cascade. For example, intra-molecular hydrogen-atom (H-atom) abstraction from the ortho-OCH3 group effectively intercepts the p-benzyne intermediate in the Bergman cycloaromatization of 2,3-diethynyl-1-methoxybenzene before this 30 TS

2.2 2.8

25

Repulsion O N+

O

20

15

10

P

1.2 2.0 8.1 6.6

24.1 30.9 12.3 10.4

20.9 27.9

R

5 5.4 5.8 0

-5

Fig. 13 The steric assistance mechanism for the ortho-effect. Energy proﬁle for the para-isomer is given in dotted lines whereas data for ortho-isomer are shown in solid lines. Calculations were performed at the BLYP/6-31G** (in bold) and B3LYP/6-31G** levels. P stands for products, R stands for reactants. Adapted from reference49.

20

I. ALABUGIN ET AL. 4.25

H

C1-C6 distance in ortho-enediyne, Å

4.2

anti-CHO

4.15 syn-CHO

4.1 4.05

NO2 y = -0.0478x + 4.1961

4

R2 = 0.9778

3.95 3.9 0

1

2

3

4

5

6

δE(para-ortho)Reactant, kcal/mol

Fig. 14 The correlation between C1–C6 distance and relative energies of ortho and para isomers of substituted benzannelated enediynes.

intermediate undergoes either retro-Bergman ring opening or external H-atom abstraction. This process leads to the formation of a new diradical and renders the cyclization step essentially irreversible.52

CONTROL OF STERIC REPULSION WITH POSITIVELY CHARGED ORTHO SUBSTITUENTS

Since four-electron repulsion is the dominant factor in the reactant destabilization, any structural perturbation that either increases electron repulsion in the reactant or decreases the electron repulsion in the TS will decrease the activation energy for the cyclization. One way for placing an accelerating substituent in direct spatial proximity to the in-plane p-orbitals is to use appropriate ortho substituents in benzannelated enediynes. It was shown that ortho substituents exert a large effect on the cyclization rate and proposed that this observation can be used for the design of pH-sensitive enediynes.49 The acid-catalyzed Bergman cyclizations are interesting because cancer cells are more acidic (pH 5.5) 53 than normal cells (pH 7.5).54 Thus, the signiﬁcant increase in reactivity upon protonation can be used in the design of tumor-speciﬁc DNA cleaving agents.55 Table 1 provides examples of amino enediynes which become much more reactive toward the Bergman cyclization upon protonation on nitrogen because the presence of a positively charged ammonium moiety alleviates the p–p repulsion of the in-plane p-orbitals. The computational results given in Table 2 illustrate that acceleration of the Bergman cyclization by protonation of a spatially close amino group is a general phenomenon. The accelerating effect of ammonium groups is transmitted mainly through space, and there are several promising structural classes of amino enediynes with a wide range of basicity and reactivity which can be activated through protonation.

CYCLOAROMATIZATION REACTIONS

21

Table 2 Calculated energetics in kcal/mol for the Bergman cyclizations of protonated and unprotonated aminoenediynes Protonated Enediynesa

+

NH3

st ec

st ec

+

NH3

+

NH3

syn anti

Neutral Amines

DE6¼

DH6¼

DG6¼

DEr

DE6¼

29.2 28.4

28.6 27.1

29.7 29.1

10.6 11.7

32.3

30.3 30.2

29.6 29.3

31.3 31.1

9.6 9.2

31.5

25.5 29.9

24.6 28.7

26.2 30.4

10.0 10.2

27.2

a

st, staggered; ec, eclipsed. a The reaction barriers for the Bergman cyclization of corresponding neutral amino enediynes.

Interestingly, both donor (syn-OMe) and acceptor (NO2) substitution can increase the accelerating effect of an ortho ammonium group when the steric interference with the above substituents ‘‘pushes’’ the adjacent acetylene moiety toward the other acetylene group, increasing the C1–C6 bonding. This is an example of an interesting cooperating effect between two ortho-substituents further illustrated in Fig. 15.

HYBRIDIZATION EFFECTS

As a direct consequence of Bent’s rule,56 carbon compounds with a signiﬁcant amount of s-character in their bonds to an electronegative element, e.g., ﬂuoroalkynes, are unstable. Thus, reactions which involve change in hybridization at a carbon atom bearing such an acceptor can be promoted when the amount of s-character at the reactive atom decreases. The role of rehybridization in a variety of chemical and supramolecular processes has been extensively analyzed and reviewed recently.55,58 An illustrative example of how rehybridization can be used to control the Bergman cyclization is provided by substituent effects at the alkyne termini of enediynes. This effect in cycloaromatization chemistry was ﬁrst studied by Schreiner and coworkers, who found dramatic acceleration of the Bergman cyclization upon

22

I. ALABUGIN ET AL.

O

O

Me

+

NO2

NH3

NO2

NH3

+

NH3

NH3+

NO2

O

+

Me

Ea=25.6

Ea=25.5

Push-push

Ea=30.8 Pull-pull

Ea=22.6

Me

Ea=25.7

Push-pull

Fig. 15 Predicted cooperative effects on activation energies (in kcal/mol) at the B3LYP/ 6-31G** level for model enediynes (‘‘push’’ and ‘‘pull’’ denote through-space repulsive (steric) and attractive (H-bonding) interactions of ortho-substituents with in-plane p-orbitals of an adjacent acetylene moeity).

introduction of s acceptors at the terminal carbons of (Z)-1,5-hexadiyne-3-ene.59 In particular, cyclization of the enediynes with terminal ﬂuoro-substituents was predicted to have the lowest barrier and be signiﬁcantly exothermic. A subsequent experimental study found that the reverse reaction (retro-Bergman ring opening) in this system is endothermic, thus conﬁrming the earlier computational predictions.60 In contrast, the computational work of Jones and Warner found that acceptor substituents positioned at the ene part of the enediyne moiety decelerate the reaction61 in full accord with the earlier experimental results of Jones and coworkers. Rehybridization provides a uniﬁed explanation for these two sets of results.57 Since the transition state is reactant-like22 and TS and reactant have the same dominant Lewis structures, application of the NBO method for the analysis of the transition state is straightforward. The hybrid orbital ‘‘h’’ that connects the terminal acetylene carbon to the substituent undergoes the most dramatic rehybridization (sp-sp2 or 48%-33% s-character) in the Bergman cyclization of enediyne I in Fig. 16. According to Bent’s rule, terminal ﬂuorine substitution destabilizes the reagent by preventing this hybrid orbital from attaining its ‘‘natural’’ sp-hybridization (in other words, it is unfavorable to direct a hybrid orbital with 50% of s-character toward a strong acceptor). As a result, the hybrid h has only 36–37% of s-character in the reactant – a dramatic effect of F substitution! However, the differences in hybridization decrease in the TS, illustrating that the destabilizing effect of electron acceptor in the reactant is removed by rehybridization in the TS (Fig. 17).

6

p-Effects on reactant stability

As mentioned in the preceding section, p-effects on the stability of the reactants are going to be rather subtle in thermal cyclizations, since the determining factor in the activation barrier for this reaction is the formation of the bond between in-plane orbitals. One way to accelerate this reaction would be to destabilize the reactant p-system. The challenge is in designing a system where the reactant destabilization is not transferred to the transition state and product as well. An elegant approach to

CYCLOAROMATIZATION REACTIONS TSRBC

X=H X=F TSBC

41.0

36.59 21.0

35.6

17.1

27.9

31.2 3.3 3.9

0

3.3 X

X

X

X

X ED1

30.07

23.88

F

F

23.22 22.94 ED1

BZY

F

F

s-char of C, %

Energy, kcal/mol

23

TSBC

BZY

TSRBC

ED2

X ED2

Reaction coordinate

(a)

(b)

Reaction coordinate

Fig. 16 (a) Comparison of potential energy proﬁle for the formal Cope rearrangement of 3,4-diﬂuorohexa-1,5-diyne-3-ene with that of (Z)-hexa-1,5-diyne-3-ene, (b) Rehybridization in the C(F) bond along the reaction path. ED1 ¼ 3,4-diﬂuoro-hex- 3-ene-1,5-diyne; ED2 ¼ 1,6-diﬂuoro-hex-3-ene-1,5-diyne; BZY ¼ diﬂuoro-1,4-didehydrobenzezne; TSBC ¼ the transition state for the Bergman cyclization; TSRBC ¼ the transition state for the retro Bergman cyclization. 1.212 I

1.265

4.480 Ea = 31.2 1.210 F

I

4.452

1.978 ΔEr = 3.3 1.268 F 2.052

Ea = 24.0

F

F ΔEr = -17.1

F

F

Fig. 17 The Bergman cyclizations of parent and ﬂuoro-substituted enediynes with the triple bond and the incipient bond lengths and the activation energies calculated at the BS-UB3LYP/6-31G** level.

achieve such selective reactant destabilization is described in the next section which describes systems with selective destabilization of reactants and efﬁcient communication between the p and s arrays.

7 Transition state effects: communication of orthogonal orbitals in the transition state of radical-anionic cyclizations In order to activate remote substituent effects in cycloaromatization reactions, two orthogonal p-systems need to ﬁnd a way to communicate. At ﬁrst glance, this task

24

I. ALABUGIN ET AL.

seems impossible due to the lack of overlap between orthogonal orbitals. Even though small deviations from planarity remove strict p-orthogonality and mix the two p-systems, the mixing is insigniﬁcant. However, the mixing can be ampliﬁed due to the inverse proportionality of orbital interaction energy to the energy gap between the interacting orbitals. As a result, even a small spatial overlap can be ampliﬁed dramatically if the energy gap between ﬁlled and unﬁlled interacting orbitals approaches zero, or, in other words, when ﬁlled and unﬁlled MOs cross. Such MO crossings occur in cycloaromatization reactions as a direct result of the bondforming interaction that transforms the in-plane p-orbitals into an s-bond and a pair of non-bonding MOs (two radical centers) as shown in Fig. 18. Consequently, in the vicinity of the TS the highest occupied in-plane MO (HOMO-1) is destabilized to the extent that it becomes the new HOMO, whereas the unoccupied in-plane MO is stabilized to become the LUMO. In other words, in the reactant both HOMO and LUMO correspond to out-of-plane orbitals but, at shorter C1–C6 or C1–C5 distances, the frontier MOs are localized at in-plane non-bonding (radical) orbitals. LUMO+2 (reactant)

LUMO+1 (reactant)

H H

SOMO (reactant) SOMO (product)

Energy

Energy

SOMO (reactant)

SOMO (product)

H

H

H H

Reaction coordinate

Reaction coordinate

Fig. 18 Crossings of in-plane and out-of-plane frontier MOs in radical-anionic Bergman and C1C5 cyclizations (crossings for photochemical, dianionic and radical-cationic cyclizations involve the same MOs but differ in the number of electrons).

CYCLOAROMATIZATION REACTIONS

25

In this analysis, the activation barrier for both C1–C6 and C1–C5 cyclizations of enediyne radical-anions can be described as the avoided crossing between the out-ofplane and in-plane MOs (conﬁgurations). One-electron reduction populates the out-of-plane LUMO of the enediyne moiety. At the TS (the crossing), the electron is ‘‘transferred’’ between the orthogonal p-systems to the new (in-plane) LUMO. This effect leads to the accelerated cyclization of radical-anions of benzannelated enediynes, a large sensitivity of this reaction to p-conjugative effects of remote substituents and the fact that this selectivity is inverse compared to that of the Bergman cyclization. Similar electronic effects should apply to the other reductive cyclization reactions that were mentioned in the introduction. In the case of the Bergman cyclization and the C1–C5 cyclization of enediynes, both the activation barrier for cyclization as well as the thermodynamics of the reaction became more favorable upon one-electron reduction compared to the thermal counterparts. The cyclization barrier drops by up to 12 kcal/mol (in the C1–C5 cyclization) and the process becomes exothermic (as opposed to the endothermic cyclizations of the neutral counterparts) as illustrated in Fig. 19 and Fig. 20. Fig. 21 explains the accelerating effect of donor substituents on the radicalanionic Bergman cyclization. Donor substituents destabilize the radical-anion of the starting material while p-acceptors have an opposite effect. However, these effects (both the stabilization by acceptors and the destabilization by donors) are signiﬁcantly decreased in the product where the extra electron is transferred in the in-plane system and, thus, direct conjugation is not present anymore. Since the crossing point between the two states representing the starting material and the reactant is the cyclization TS, the energy difference between starting material and the crossing point is the activation energy. When the starting material is destabilized but the TS is not, the activation energy is low. When the starting material is stabilized, but . E, kcal/mol

46.1 40.1 .

H 34.0

31.2 29.5 . . .−

3 3.3

11.9 -

.

H .

-16.7

-

Fig. 19 The reaction energy proﬁles for thermal (on the left) and radical-anionic (on the right) C1–C6 and C1–C5 cyclizations of the parent enediyne computed at the B3LYP/ 6-31G** level.

26

I. ALABUGIN ET AL. E, kca l/mol 45.1 40.4

.

28.5 .

H

24.8

13.4 .

.

23.1

-.

6

0.3

H

-

.

. -18.8 -

Fig. 20 The reaction energy proﬁles for thermal (on the left) and radical-anionic (on the right) C1–C6 and C1–C5 cyclizations of the parent benzannelated enediyne computed at the B3LYP/6-31G** level.62 Ea, kcal mol-1

X X

X=NH2 LUMO (diradical)

X=H X=NO2 R LUMO (enediyne)

X

SOMO (radical-anion)

SOMO (radical-anion)

P

X Reaction coordinate

Fig. 21 The crossings of in-plane and out-of-plane frontier MOs in the radical-anionic Bergman cyclizations. A similar effect is observed in the C1–C5 cyclization.

stabilization is lost in the TS, the activation energy is high. Although new in cycloaroamatization chemistry, this approach is just another example of the classic ‘‘reactant stabilization’’ and ‘‘reactant destabilization’’ concepts often utilized in enzyme chemistry.

CYCLOAROMATIZATION REACTIONS

8

27

Effects on product stability

At this stage, both s- and p-effects are important. However, the inﬂuence of the p-effects dramatically increases. In fact, it is the formation of a p-aromatic system in the cycloaromatization reactions that makes these processes energetically feasible.

s-EFFECTS

Interaction of non-bonding electrons The non-bonding electrons can be coupled either directly through space (TS) or indirectly through antibonding (s*) bridge orbitals (through bond (TB) coupling). The most well-recognized of the s-effects is the TB interaction of the two radical centers in p-benzynes and related molecules.63 A well-recognized effect of this interaction is displayed in the lower reactivity of p-benzyne in H-abstraction reactions relative to that of the phenyl radical.55 TB interaction, which is absent in monoradicals,54 provides an additional 3–5 kcal/mol of stabilization energy to the p-benzyne-type diradicals. Since this stabilizing energy is lost with the ﬁrst H-atom abstraction, the p-benzyne diradicals are less reactive and more selective than simple phenyl radicals. Interestingly, coupling between the non-bonding orbitals is dramatically enhanced upon one-electron reduction of p-benzynes or didehydrofulvenes, possibly because the TS interaction adds to the TB coupling.12b

Zwitterionic products Full -polarization in diradicals can give rise to zwitterionic products. First examples were studied in detail by Carpenter and coworker who investigated solvent effects on rates and product distribution in Myers-Saito cyclizations.64 Polar solvents and substitution patterns that stabilize either positive or negative charges (or both) favor the zwitterionic products. For example, the presence of a dimethylamino group leads to stabilization of cations and isolation of pyrrolo-quinolines, rather than pyrido-indoles from eneyne-carbodiimides, as reported by Wang and coworkers (Scheme 14).65 Even though these products are not formed by a diradical mechanism, these polar species may still have relevance to DNA damage, because of the potential alkylating ability of their electrophilic (cationic) sites. Another example of a zwitterionic product of a cycloaromatization reaction was given by Kerwin and coworkers. Their ‘‘skipped’’ (aza)enediynes rearranged to yield (aza)eneyne-allenes that subsequently cyclized under addition of methanol (in a byproduct), which is consistent with a partitioning between a diradical and a zwitterionic reaction pathway (Scheme 15).66

28

I. ALABUGIN ET AL. Me2N

Me N

+ N

Me2N

NPh

NH Ph isolated in 84% yield N

N Ph N thermolysis in presence of 5 eq. trimethylsilyl chloride

Me2N

R = H, Me

Me2N

N

N

Ph

N N R not formed

Scheme 14 Zwitterionic Myers-Saito-type cyclization from reference65. H

X R

+

X

H H

R

-

X

methanol

OMe

R

R = Ph, H X = N, CH

Scheme 15 Formation of a methanol adduct in cyclization of eneyne-allenes, formed from ‘‘skipped’’ enediynes.

Strain increase: effect on the photochemical Bergman cyclization A very interesting experimental observation of Jones et al. illustrates a different effect of strain on the efﬁciency of photochemical Bergman cyclizations.43d Variations in the size of the cycle which does not incorporate the whole enediyne system, but only the vinyl part of the enediyne moiety (in contrast to the previously discussed data) affect the yield of the cycloaromatized product. Initially, an increase in the ring size leads to an increase in yield (Scheme 16). Increased ring strain in the C-4 (n ¼ 2) and C-5 (n ¼ 3) products may be a reason for lower yields (and longer irradiation times). The effect of strain on cyclizations is well-documented outside of cycloaromatization chemistry, as well. For example, annealing of a cyclopentene ring leads to a decrease in the cyclization rate and an inversion of the 5-exo/6-endo selectivity for all three patterns shown in Fig. 22. Interestingly, after reaching the maximum at the 6-membered cycle, the yields drop again. This decrease in efﬁciency occurs despite the appreciable reduction in the distance between the terminal acetylenic carbons relative to the 6-membered analogue. Here, the efﬁciency may simply be a function of how photochemical excitation is distributed in the reactive excited state. Calculated enediyne geometries suggest the cyclization is more efﬁcient for those enediynes where the terminal phenyl groups are rotated outside of the enediyne plane (Table 3). Rotation of the aromatic ring out of the enediyne plane forces the aromatic p-orbitals to overlap with the in-plane p-orbitals of the enediyne system as illustrated

CYCLOAROMATIZATION REACTIONS

29

H H H

n(H2C)

Ph

hν 2-propanol

n(H2C)

Ph H

yield n=2; 0% n=3; 13% n=4; 22% n=5; 16% n=6; 14%

Scheme 16 Model enediynes used in photochemical Bergman cyclization.

15

16

14

Fig. 22 Radicals used to study the effect of strain on the efﬁciency of 5-exo and 6-endo-dig cyclizations. Table 3 C1–C6 (‘‘c–d’’) distances in model enediynes Distance (A˚)

Ring Size n¼2 n¼3 n¼4 n¼5 n¼6

4.941 4.284 4.018 3.947 3.893

R

R R

R R

R Aryls are "in plane", the π-systems are orthogonal

R R Aryls are "out-of plane", the π-systems are in conjugation

Fig. 23 Effect of orientation of the aromatic ring on the conjugation with the enediyne system.

in Fig. 23. This involves the in-plane p-orbitals in a more extended conjugated system and decreases the energy gap between frontier in-plane orbitals. The key role of the in-plane p-system suggests that the excitation should be delivered to this system in order for the cyclization to proceed efﬁciently.

30

I. ALABUGIN ET AL.

p-EFFECTS

The most obvious effect on cycloaromatization, as the name implies, is the formation of an aromatic system. By delocalizing electrons in an aromatic ring, the product gains a high degree of stability, which is reﬂected in the small endothermicity of the Bergman cyclization and the exothermicity of the Myers-Saito cyclization. Since the Schmittel and Schreiner cyclizations are not true cycloaromatization reactions per se, they do not have the beneﬁcial effect of the formation of an aromatic system and are therefore much are more endothermic than their counterparts. Product stabilization is much more pronounced when the enediyne or ene-yneallene starting materials are not already part of an aromatic system, since forming an aromatic system constitutes a much higher degree of stabilization than expanding an aromatic system (Fig. 24). Conjugation of the radical center provides additional stabilization to the p-radical formed by the Myers-Saito and Schmittel cyclizations. Both the Schmittel and the Schreiner cyclization do not formally produce an aromatic system, Schleyer, Schreiner and co-workers compared these reactions to the Bergman and Myers-Saito cyclization using standard aromaticity criteria, such as magnetic susceptibility exaltations, aromatic stabilization energy (ASE), and nucleus independent chemical shift (NICS). Interestingly, the degree of cyclic electron delocalization found in these systems is comparable to that found in benzene. The effect of benzannelation however is smaller in magnitude, and for Schreiner cyclization it has an opposite effect leading to a slight decrease in reaction endothermicity.

Bergman

Erel.: 25.2 (Erel.: 24.6)

Erel.: 8.5 (Erel.: 14.4)

Erel.: 41.0 (Erel.: 37.2)

Erel.: 41.3 (Erel.: 37.4)

Erel.: 18.7 (Erel.: 19.5)

Erel.: -9.7 (Erel.: -4.5)

Erel.: 30.0 (Erel.: 25.2)

Erel.: 12.9 (Erel.: 11.9)

Schreiner

Myers-Saito

Schmittel

Fig. 24 Relative energies for the transition states and radical products for the different cyclization pathways (benzannelated systems in parentheses).

CYCLOAROMATIZATION REACTIONS

31

These results, of course, contrast the signiﬁcant effect of benzannelation in radical-anionic cycloaromatization reactions discussed above12 where not just the formation of a new aromatic cycle but also restoration of aromaticity in the previously existing cycle occurs at the same time in the cyclorearomatization process.

9

Conclusion

The great potential utility of cycloaromatization reactions is matched by the inherent complexity of electronic changes that accompany these processes. This complexity stems from the presence of two orthogonal p-arrays, both of which undergo signiﬁcant changes during the reactions. The asynchronous nature of s- and p-effects opens numerous opportunities to control these processes and will continue to provide challenges for experimentalists and theoreticians alike. Even though some of the general factors controlling the Bergman cyclization are applicable to newly discovered cycloaromatization reactions, new phenomena such as communication of orthogonal orbitals in radical-anionic, radical-cationic and dianionic cycloaromatizaton and cyclorearomatization will continue to stimulate future development of this interesting ﬁeld.

References 1. We are omitting reactions such as the trimerization of acetylene on transition metal catalysts, even though they could be considered cycloaromatization reactions 2. Bergman, R.G. (1973). Acc. Chem. Res. 6, 25 3. Roth, W.R., Hopf, H. and Horn, C. (1994). C. Chem. Ber. 127, 1765; Wenthold, P.G. and Squires, R.R. (1994). J. Am. Chem. Soc. 116, 6401; Davico, G.E., Bierbaum, V.M., De Puy, C.H., Ellison, G.B. and Squires, R.R. (1995). J. Am. Chem. Soc. 117, 2590 4. Rawat, D.S. and Zaleski, J.M. (2004). Synlett 393; Bhattacharyya, S. and Zaleski, J.M. (2004). Curr. Topics Med. Chem. 4, 1637 5. Basak, A., Mandal, S. and Bag, S.S. (2003). Chem. Rev. 103, 4077 6. Schreiner, P.R., Navarro-Vazquez, A. and Prall, M. (2005). Acc. Chem. Res. 38, 29 7. Klein, M., Walenzyk, T. and Ko¨nig, B. (2004). Collect. Czech. Chem. Commun. 69, 945 8. Galm, U., Hager, M.H., Van Lanen, S.G., Ju, J., Thorson, J.S. and Shen, B. (2005). Chem. Rev. 105, 739 9. Jones, G.B. and Fouad, F.S. (2002). Curr. Pharm. Design 8, 2415 10. Lewis, K.D., Wenzler, D.L. and Matzger, A.J. (2003). Org. Lett. 5, 2195; Lewis, K.D., Rowe, M.P. and Matzger, A.J. (2004). Tetrahedron 60, 7191 11. Kawatkar, S.P. and Schreiner, P.R. (2002). Org. Lett. 4, 3643 12. (a) Alabugin, I.V. and Kovalenko, S.V. (2002). J. Am. Chem. Soc. 124, 9052; (b) Alabugin, I.V. and Manoharan, M. (2003). J. Am. Chem. Soc. 125, 4495 13. Eshdat, L., Berger, H., Hopf, H. and Rabinovitz, M. (2002). J. Am. Chem. Soc. 124, 3822; Treitel, N., Eshdat, L., Sheradsky, T., Donovan, P.M., Tykwinski, R.R., Scott, L.T., Hopf, H. and Rabinovitz, M. (2006). J. Am. Chem. Soc. 128, 4703 14. Whitlock Jr., H.W., Sandvick, P.E., Overman, L.E. and Reichardt, P.B. (1969). J. Org. Chem. 34, 879 15. Tamao, K., Yamaguchi, S. and Shiro, M. (1994). J. Am. Chem. Soc. 116, 11715 16. Yamaguchi, S., Miyasato, M. and Tamao, K. (2003). Chem. Lett. 32, 1104

32

I. ALABUGIN ET AL.

17. Wenthold, P.G. (1999). J. Chem. Soc., Perkin Trans. 2, 2357; Hammad, L.A. and Wenthold, P.G. (2003). J. Am. Chem. Soc. 125, 10796 18. Navarro-Va´zquez, A., Prall, M. and Schreiner, P.R. (2004). Org. Lett. 6, 2981 19. Clark, A.E., Davidson, E.R. and Zaleski, J.M. (2001). Am. Chem. Soc. 123, 2650 20. Zimmerman, H.E. and Pincock, J.A. (1973). J. Am. Chem. Soc. 95, 3246 21. Schreiner, P.R. and Bui, B.H. (2006). Eur. J. Org. Chem. 5, 1162 22. Galbraith, J.M., Schreiner, P.R., Harris, N.R., Wei, W., Wittkopp, A. and Shaik, S. (2000). Chem. Eur. J. 6, 1446 23. Lefﬂer, J.E. (1953). Science 117, 340; Hammond, G.S. (1995). J. Am. Chem. Soc. 77, 334 24. Alabugin, I.V. and Manoharan, M. (2003). J. Phys. Chem. A 107, 3363 25. Prall, M., Wittkopp, A. and Schreiner, P.R. (2001). J. Phys. Chem. 105, 9265 26. Alabugin, I.V. and Manoharan, M. (2003). J. Am. Chem. Soc. 125, 4495 27. The only known exception is the fascinating class of radical-anionic and dianionic cyclizations where p-effects can be activated through MO crossings 28. Alabugin, I.V., Manoharan, M., Breiner, B. and Lewis, F. (2003). J. Am. Chem. Soc. 125, 9329 29. Alabugin, I.V. and Manoharan, M. (2005). J. Am. Chem. Soc. 127, 12583; Alabugin, I.V. and Manoharan, M. (2005). J. Am. Chem. Soc. 127, 9534 30. Koga, N. and Morokuma, K. (1991). J. Am. Chem. Soc. 113, 1907–1911 31. For theoretical analysis of the Myers-Saito cyclization, see: (a) Engel, B. and Hanrath, M. (1998). J. Am. Chem. Soc. 120, 6356–6361; (b) Schreiner, P.R. and Prall, M. (1999). J. Am. Chem. Soc. 121, 8615–8627 32. In very small macrocycles (7- and 8-membered rings), cyclization is not expected, despite the very short c–d distance of of 2.5 and 2.6 A˚, respectively. The ring strain in the cyclized product would simply be too high 33. Nicolaou, K.C., Smith, A.L. and Yue, E.W. (1993). Proc. Natl. Acad. Sci. USA 90, 5881 34. Snyder, J.P. (1989). J. Am. Chem. Soc. 111, 7630 35. Magnus, P., Carter, P., Elliott, J., Lewis, R., Harling, R., Pitterna, T., Bauta, W.E. and Fortt, S. (1992). J. Am. Chem. Soc. 114, 2544 36. Wandel, H. and Wiest, O. (2002). J. Org. Chem. 67, 388 37. Plourde, G.W., Warner, P.M., Parrish, D.A. and Jones, G.B. (2002). J. Org. Chem. 67, 5369 38. Schreiner, P.R. (1998). J. Am. Chem. Soc. 120, 4184 39. This notion has been the subject of a recent controversy, which was resolved 40. Kraka, E. and Cremer, D. (1994). J. Am. Chem. Soc. 116, 4929 41. Jafri, J.A. and Newton, M.D. (1978). J. Am. Chem. Soc. 100, 5012 42. One can argue that the presence of the in-plane C3-C4 bond and the 6-electron out-ofplane p-system renders the whole molecule non-antiaromatic. Therefore, we use the quotation marks for the term ‘‘antiaromatic region’’. However, use of this term is justiﬁed because the lack of the p–p* stabilization along the strong p–p repulsion provides an appealing analogy to antiaromatic molecules 43. (a) Funk, R.L., Young, E.R.R., Williams, R.M., Flanagan, M.F. and Cecil, T.L. (1996). J. Amer. Chem. Soc. 118, 3291; (b) Kaneko, T., Takanashi, M. and Hirama, M. (1999). Angew. Chem. Int. Ed. 38, 1267; (c) Choy, N., Blanco, B., Wen, J., Krishan, A. and Russell, K.C. (2000). Org. Lett. 2, 3761; (d) Jones, G.B., Wright, J.M., Plourde II, G., Purohit, A.D., Wyatt, J.K., Hynd, G. and Fouad, F. (2000). J. Amer. Chem. Soc. 122, 9872 44. Rawat, D.S. and Zaleski, J.M. (2004). Synlett 393; Bhattacharyya, S. and Zaleski, J.M. (2004). Curr. Topics Med. Chem. 4, 1637; Basak, A., Mandal, S. and Bag, S.S. (2003). Chem. Rev. 103, 4077; as wel l asKo¨nig, B. (2000). Eur. J. Org. Chem. 12, 381 45. Warner, B.P., Millar, S.P., Broene, R.D. and Buchwald, S.L. (1995). Science 269, 814

CYCLOAROMATIZATION REACTIONS

33

46. Bhattacharyya, S., Pink, M., Baik, M. and Zaleski, J.M. (2005). Angew. Chem. Int. Ed. 44, 592 47. Bhattacharyya, S., Dye, D.F., Pink, M. and Zaleski, J.M. (2005). Chem. Commun. 5295; For a review on metal-bound enediynes, see: Rawat, D.S. and Zaleski, J.M. (2004). Synlett 3, 393 48. Kim, C.-S. and Russell, K.C. (1998). J. Org. Chem. 63, 8229; Choy, N. and Russell, K.C. (1999). Heterocycles 51, 13; Kim, C.-S. and Russell, K.C. (1999). Tetrahedron Lett. 40, 3835; Kim, C.-S., Diez, C. and Russell, K.C. (2000). Chem. Eur. J. 6, 1555 49. Alabugin, I.V., Manoharan, M. and Kovalenko, S.V. (2002). Org. Lett. 4, 1119 50. Eliel, E.L., Wilen, S.H. and Doyle, M.P. (2001). Basic Organic Stereochemistry, p. 459, Wiley-Interscience, New York 51. Zeidan, T., Kovalenko, S.V., Manoharan, M. and Alabugin, I.V. (2006). J. Org. Chem. 71, 962 52. Zeidan, T., Manoharan, M. and Alabugin, I.V. (2006). J. Org. Chem. 71, 954 53. Osinsky, S.P., Levitin, I.Y., Bubnovskaya, L.N., Ganusevich II, Sigan, A.L., Tsykalova, M.V. and Zagorujko, L.I. (1999). Exp. Oncol. 21, 216; Tannock, I.F. and Rotin, D. (1989). Cancer Res. 49, 4373 54. Kraka, E. and Cremer, D. (2000). J. Am. Chem. Soc. 122, 8245 55. Hoffner, J., Schottelius, J., Feichtinger, D. and Chen, P. (1998). J. Am. Chem. Soc. 120, 376–385 56. Bent, H.A. (1961). Chem. Rev. 61, 275 57. (a) Alabugin, I.V. and Manoharan, M. (2007). J. Comp. Chem. 28, 373; (b) Additional examples of rehybridization are analyzed in Alabugin, I.V., Manoharan, M., Buck, M. and Clark, R.J. (2007) Theochem, 813, 21–27 58. For an analysis of rehybridization effects in supramolecular chemistry, see: Alabugin, I.V., Manoharan, M., Peabody, S. and Weinhold, F. (2003). J. Am. Chem. Soc. 125, 5973; Alabugin, I.V., Manoharan, M. and Weinhold, F. (2004). J. Phys. Chem. A 108, 4720 59. Prall, M., Wittkopp, A., Fokin, A.A. and Schreiner, P.R. (2001). J. Comp. Chem. 22, 1605 60. Wenk, H.H., Balster, A., Sander, W., Hrovat, D.A. and Borden, W.T. (2001). Angew. Chem. Int. Ed. Engl. 40, 2295 61. Jones, G.B. and Warner, P.M. (2001). J. Am. Chem Soc. 66, 2134 62. The starting points for both C1–C5 and C1–C6 cyclizations are characterized using C1–C6 distances to stress that starting point for both cyclizations is the same enediyne radical-anion. However, for transition states and products of C1–C5 and C1–C6 cyclizations, the respective incipient bond length (C1–C5 or C1–C6) was used as the reaction coordinate 63. (a) Hoffman, R., Imamura, A. and Hehre, W.J. (1968). J. Am. Chem. Soc. 90, 1499; Hoffman, R. (1971). Acc. Chem. Res. 4, 1; (b) Paddon-Row, M.N. (1982). Acc. Chem. Res. 15, 245; (c) Gleiter, R. and Schafer, W. (1990). Chem. Res. 23, 369–375; (d) Brodskaya, E.I., Ratovskii, G.V. and Voronkov, M.G. (1993). Russ. Chem. Rev. 62, 975 64. Hughes, T.S. and Carpenter, B.K. (1999). J. Chem. Soc., Perkin Trans. 2, 2291 65. Li, H., Petersen, J.L. and Wang, K.K. (2003). J. Org. Chem. 68, 5512 66. Feng, L., Kumar, D., Birney, D.M. and Kerwin, S.M. (2004). Org. Lett. 6, 2059

This page intentionally left blank

N-Acyloxy-N-alkoxyamides – structure, properties, reactivity and biological activity STEPHEN A. GLOVER School of Science and Technology, University of New England, Armidale, NSW 2350, Australia 1 Introduction 35 General 35 Background 37 2 Synthesis 39 3 Structure 43 General 43 Theoretical structures 44 X-ray structures 47 Spectroscopic properties 51 4 Chemical reactivity 59 Solvolysis studies – AAl1 reactivity 60 Nucleophilic substitution reactions – SN2 reactivity 70 Thermal decomposition reactions 90 5 Biological activity of N-acyloxy-N-alkoxyamides 97 Mutagenicity in the Ames Salmonella/microsome assay 97 Anticancer activity of N-acyloxy-N-alkoxyamides 115 6 Conclusions 115 Acknowledgements 116 References 117

1

Introduction

GENERAL

The properties of amides are determinants of the structure and characteristics of a wide range of molecules and particularly those of peptides and proteins.1 Universally, amides and amide linkages are characterised by a nitrogen that is largely sp2 hybridised and in which the lone pair resides in a 2pz orbital. As a consequence there is a strong interaction between the amide nitrogen and the carbonyl. While this has traditionally been described as a resonance delocalisation involving I and II in Fig. 1a, contemporary views strongly favour a third resonance contributor III in which there is s back donation from carbon to the sp2 hybridised nitrogen.2,3

35 ADVANCES IN PHYSICAL ORGANIC CHEMISTRY VOLUME 42 ISSN 0065-3160 DOI: 10.1016/S0065-3160(07)42002-0

r 2008 Elsevier Inc. All rights reserved

36

S.A. GLOVER

According to Wiberg, the p donation from nitrogen is best described as a Highest Occupied Molecular Orbital (HOMO)—Lowest Unoccupied Molecular Orbital (LUMO) interaction in which case there is little charge transfer to oxygen since this contributes weakly to the LUMO (Fig. 1b).2 Either way, the predominant interaction is between the nitrogen lone pair and the carbon of the carbonyl bond and the double bond character is the prime cause of both planarity at nitrogen and the restricted rotation about C–N bonds in amides. Properties of hydroxamic esters largely mirror those of amides. The attachment of one oxygen to the amide nitrogen does not dramatically alter the degree of pyramidality at nitrogen or the extent of lone pair delocalisation/interaction with the carbonyl.4–8 A good number of amides are not planar at nitrogen owing to twisting due to steric interactions (Fig. 2a),9–17 conﬁgurational properties that close the angles at nitrogen (Fig. 2b)11,18–20 or lactams that lock the nitrogen lone pair out of alignment with the carbonyl 2pz orbital (Fig. 2c).21–24 The result in all cases is a disconnection between the nitrogen lone pair and the amide carbonyl. In these cases, the amide nitrogen tends towards sp3 hybridisation and the C–N bonds assume single bond character. All these amides constitute the class of ‘‘twisted amides’’ although in Fig. 2b, twisting may be the result of lone pair disconnection rather than the cause thereof. However, they can all best be regarded as N-acylamines. Kirby’s conﬁgurationally rigid 1-azaadamantan-2-one 1 is the extreme of this class.21–23,25 The nitrogen properties of this lactam are clearly amine-like while the carbonyl is ketonic in all respects. Twisted amides undergo rapid hydrolysis or reduction22,25–30 and exhibit enhanced reactivity.31

O C

O C

N

O C

N

COLUMO O N

R C

NHOMO

N

I

II

III

(a)

(b)

Fig. 1 (a) Resonance and (b) HOMO–LUMO interaction in simple amides.

R3

O

O R3 C

C N

R N

N

R2

O

R1 (a)

(b)

(c)

Fig. 2 (a) A sterically twisted amide; (b) an angularly constrained amide; (c) a twisted lactam.

37

N-ACYLOXY-N-ALKOXYAMIDES

N

O

R3

O O

N H 2

1 a b c d

Y X

R2

O C

C O

N O

R1

3

X=Y=OR X=Cl,Y=OR X=OR,Y=NR2 X=Y=Halogen

The author has identiﬁed another class of amides that possess pyramidal nitrogens but are devoid of any steric or conﬁgurational imposition that results in enforced twisting about the amide linkage. ‘‘Anomeric amides’’, are deﬁned as amides that bear two heteroatoms at the amide nitrogen 2.32 There is ample evidence of pyramidality in N,N-dioxoamides 2a, N-halo-N-oxoamides 2b and N-aminoN-oxoamides 2c. N,N-dihaloamides 2d may well fall into the same class. In all of these, the amide nitrogen responds to the collective electronegativity of the substituents by rehybridising from sp2 to sp3. This enables an electron density distribution that better satisﬁes the electron demand of the nitrogen substituents, X and Y. The conﬁgurational change results in smaller angles at nitrogen and reduced p-character of the lone pair orbital with attendant disconnection from the amide carbonyl as evidenced by spectroscopic properties, radically reduced amide isomerisation barriers, reactivity patterns and theoretical attributes.6,32–36 This review will cover the structural, physical, spectroscopic, chemical and biological properties of the most widely studied class of these anomeric amides, the N-acyloxy-N-alkoxyamides 3. These not only exhibit all the structural and spectroscopic characteristics of anomeric amides 2, their reactivity patterns embrace a range of reactions at the amide nitrogen that in many respects parallel those at a saturated carbon.37–46 In addition, they are mutagenic as well as DNA damaging agents that have anticancer capability.37,38,40,46–50 To date nearly a hundred of these have been synthesised in our laboratory while, recently, a range of urea and urethane analogues have also been reported.51–53

BACKGROUND

Aromatic amines 4 are metabolised in vivo by cytochrome P450 mediated oxidation to phenolic and hydroxylamine derivatives 5 and 6. Phase II conjugation of the latter with PAPS or acyl transferase results in formation of the sulfuric or acetic acid esters 7. Nitrogen conjugation to give the N-acetyl analogues is also possible (Scheme 1).54–65 These metabolites have been shown to produce arylnitrenium ions 8 under solvolytic conditions and McClelland and Novak have demonstrated, respectively

38

S.A. GLOVER OH NH2

NH

NH2

P450

Ar

Ar

Ar

5

6

HO 4

Acetylase (PAPS) N Ar

H (Ac)

Ar

8

(SO3–) OAc N H (Ac)

7

Scheme 1.

by Laser Flash Photolysis (LFP) and azide clock methods, that the lifetimes of these nitrenium ions are discriminating; those aromatic amines that produce long-lived arylnitrenium ions are generally more carcinogenic and the rates of hydration of these ions in water are critical.66–85 Arylnitrenium ions are highly delocalised and steric hindrance to hydration at the 4-position seems to be critical in determining the lifetimes in solution. Nitrenium ions from carcinogenic 4-biphenylamine 9, N-acetyl-2-aminoﬂuorene 10, 2-naphthylamine 11 and benzidine 12 have been shown to have much longer lifetimes than other, less harmful arylamines and their reaction with guanine to afford C-8 (G-C8) adducts 13 is the primary cause of mutations of DNA.86–89 O

NH2

NHAc

9

Ar

N

(Ac)H

N

NH2 NH2 11

NH2 12

NH

N

10

N

N H2

Deoxyribose 13

In 1984, we demonstrated that N-alkoxy-N-acyl nitrenium ions 15 could be generated by the reaction of N-alkoxy-N-chloroamides 14 with Lewis acids such as Ag+ and Zn2+ and used these to form heterocycles by intramolecular aromatic substitution reactions (Scheme 2).90 In this manner, several novel N-acyl-3,4-dihydro-2,1benzoxazines 16a and N-acyl-4,5-dihydro-(1H,3H)-2,1-benzoxazepines 16b were made. Subsequent work91,92 and that of Kikugawa93–96 produced numerous syntheses involving alkoxynitrenium ions including formation of natural products.97–99 In a seminal theoretical paper, we showed that, at the Modiﬁed Neglect of Differential Overlap (MNDO) level, phenylnitrenium ions 17 as well as NH2, PH2,

39

N-ACYLOXY-N-ALKOXYAMIDES (CH2 )n O N Cl

(CH2 )n O N

Lewis acids R' O

(CH2 )n O N

R' O

14

R' O

15

16 a n=2 b n=3

Scheme 2.

SH and OH-substituted nitrenium ions 19a–d and their N-formyl analogues have singlet ground states with a similar degree of p-overlap with the electron-deﬁcient nitrogen (p-bond orders 40.9). Ab initio calculations (HF/6-31G*) on heteroatomsubstituted nitrenium ions showed that all the heteroatoms strongly stabilised the singlet state relative to their triplet state.100 Extensive double bond character in hydroxynitrenium ion was also determined by Schwarz and coworkers.101 N

H (CHO)

N

R X

N

H (CHO)

19 17

18

a X=NH2 b X=PH2 c X=SH d X=OH

For phenylnitrenium ions, we85 and others 72,78,102–104 have computed that there is extensive positive charge delocalisation into the aromatic ring and arylnitrenium ions are best described as 4-imino-2,5-cyclohexadienyl-1-yl carbenium ions 18. From appraisal of their respective resonance stabilisation, arylnitrenium ions and alkoxynitrenium ions should form with similar facility. On account of the fact that N-acetoxy-N-acetyl arylamines 7 are penultimate carcinogens in the metabolism of aromatic amines, N-acyloxy-N-alkoxyamides 3 were designed to test their potential as DNA-damaging agents.

2

Synthesis

N-Acyloxy-N-alkoxyamides 20 are synthesised from N-chlorohydroxamic esters 22 by replacement of chlorine by a carboxyl group (Scheme 3). Initially we employed silver acetate in anhydrous ether to make N-acetoxy derivatives.47 However, most have been made using sodium carboxylates in dry acetone by analogy with Finkelstein chemistry.5,38–40,42,43,46,48,49,105 The reactions can be monitored conveniently by thin layer chromatography and N-acyloxy-N-alkoxyamides generally

40

S.A. GLOVER

R1O

OAc (COR2) N O

H N 3

R 21

i ii

R1Br

Cl R1O

N

O

ButOCl 1

3

H N

O

R O

R 22

O

KO

R3 20

R1ONH2

3

R 23

Et3N

Cl

O 3

R 24

i AgOAc anhydrous diethyl ether ii NaOAc (NaOCOR2) anhydrous acetone

Scheme 3.

oxidise to brown spots on silica plates. Yields vary depending upon the stability of the N-acyloxy-N-alkoxyamide to hydrolysis but in most instances, conversion is clean, or minor impurities (such as the corresponding alkyl esters) can be removed using centrifugal chromatography. Scheme 3 outlines synthetic strategies for the introduction of a range of substituents on the amide, alkoxyl and acyloxyl side chains. Hydroxamic esters 23 are readily synthesised from potassium salts of hydroxamic acids 21 according to Cooley et al.7 or by condensation of the corresponding acid chloride 24 with an alkoxyamine. This group has synthesised a wide range of N-acyloxy-N-alkoxyamides that can be categorised as follows:

1. N-Acetoxy-N-alkoxybenzamides 25, with variation in the alkoxyl side chain 2. N-Acetoxy-N-butoxyarylamides 26, with variation on the benzamide ring 3. N-Acetoxy-N-arylmethyloxybenzamides 27, with variation on the benzyloxyl side chain 4. N-Aroyloxy-N-benzyloxybenzamides 28, with variation on the ester side chain 5. N-Acyloxy-N-butoxyamides 29, with variable amide and acyloxyl side chains 6. N-Acyloxy-N-alkoxyacetamides 30, with variable alkoxyl and acyloxyl side chains 7. N-Benzoyloxy- and N-acetoxy-N-benzyloxybenzamides 31–33, bearing one or more para-tert-butyl substituents 8. N-Acyloxy-N-alkoxyamides 34 and 35 where two amide groups are tethered by alkoxyl or acyl chains 9. N-Acyloxy-N-alkoxyamides 36 and 37 of a specialised nature

41

N-ACYLOXY-N-ALKOXYAMIDES

O

O R

Bu

N O AcO

O

X

N O AcO

AcO

25

26

27

a R=Et b R=Pr c R=Bu d R=Pentyl e R=Octyl f R=Pri g R=Bui h R=Penti i R=But j R=2-butyl

a X=H b X=MeO c X=Ph d X=Me e X=Cl f X=Br g X=NO2 h X=m-NO2 i X=But

a Y=H b Y=MeO c Y=PhO d Y=Ph e Y=Me f Y=Cl g Y=Br h Y=NO2 i Y=But

O

O R3

N O O O

Z

Y

N O

CH3

N O O O R2

28 a Z=H b Z=MeO c Z=Ph d Z=Me e Z=Cl f Z=CHO g Z=CN h Z=NO2 i Z=But j Z=CF3 k Z=m-NO2 l Z=m-MeO

O Bu

R1 N O O O R2

29 2

3

a R =Ph, R =Et b R2=Ph, R3=Pri c R2=Ph, R3=But d R2=Ph, R3=neopentyl e R2=Ph, R3=1-Adamantyl f R2=R3=Ph g R2=Pr, R3=Ph h R2=Pri, R3=Ph i R2=(S)-2-butyl, R3=Ph j R2=Neopentyl, R3=Ph k R2=1-Ad, R3=Ph l R2=But, R3=Ph m R2=Me, R3=2-naphthyl n R2=Me, R3=fluoren-1-yl o R2=Me, R3=9,10-anthraquinone-2-yl p R2=Me, R3=pyren-1-yl r R2=2-naphthyl, R3=Ph s R2=2,6-dimethylphenyl, R3=Ph t R2=3,5-dimethylphenyl, R3=Ph u R2=Me, R3=3,5-dimethylphenyl v R2=hexyl, R3=But w R2=hexyl, R3=Ph x R2=5-hexen-1-yl, R3=But y R2=5-hexen-1-yl, R3=Ph

30 1

2

a R =Bu, R =Me b R1=Bu, R2=Ph c R1=Bn, R2=Me d R1=Bn, R2=Ph e R1=Bu, R2=2-Np f R1=CH2(1-Np),R2=Me g R1=CH2(2-Np),R2=Me h R1=(CH2)2(2-Np),R2=Me i R1=(CH2)3(2-Np),R2=Me j R1=Bu, R2=fluoren-1-yl k R1=Bu, R2=pyren-1-yl l R1=Bu, R2=CH2(pyren-1-yl) m R1=Bu, R2=CH2(2-naphthyl) n R1=Bu, R2=(CH2)2(2-naphthyl)

42

S.A. GLOVER O

R3 N O O O

R1

Y

O

R1

O N O

O O

O

N O

R3

O CH3

But

R2 31 a R1=R2=H,

But 32 a Y=H; b Y=But,

R3=But

b R1=But , R2=R3=H c R1=R2=But, R3=H; d R1=H, R2=R3=But; e R1=R3=But, R2=H; f R1=R2=R3=But

O

O

O N O

Ph

(CH2)3

AcO 34

2

O

N O

a

O N O

O CH3

O

d O

b

CH3

36 a R =H b R1=Cl

Ph

c

O

1

O N OAc

(CH2)n N OAc 35 a n=7 b n=8

R1

O But

33 a R1=R2=But

37 a b c d

a,b,=Me, c,d=H a,b=H, c,d=Me a,b,d=H, c=Me a=Me, b,c,d=H

Recently, Shtamburg et al. synthesised a variety of N-acyloxy-N-alkoxyureas 38, N-acyloxy-N-alkoxycarbamates 39 as well as N-acetoxy-N-ethoxybenzamide 25a by an analogous procedure using appropriate sodium carboxylates in CH3CN.51–53 Compound 39h was synthesised from ethyl N-butoxy-N-chlorocarbamate using our NaOAc in acetone method.106

43

N-ACYLOXY-N-ALKOXYAMIDES OCOR2

OCOR2 R1O

N

O

R1O

O OR3

NR32 38

39

a R1=R2=R3=Me b R1=R3=Me, R2=Et c R1=R3=Me, R2=Pri d R1=R3=Me, R2=Ph e R1=Prn, R2=R3=Me f R1=Prn, R2=Me, R3=Me,H g R1=Prn, R2=Ph, R3=Me,H h R1=Et, R2=Me, R3=CH2(1-Np),H i R1=Et, R2=Me, R3=H j R1=Bu,R2=Me, R3=H k R1=n-dodecyl, R2=Me, R3=H

3

N

a R1=R2=R3=Me b R1=R3=Me, R2=Et c R1=n-octyl,R2=R3=Me d R1=R2=Me, R3=Et e R1=Me, R2=Ph, R3=Et f R1=Pri, R2=Me, R3=Et g R1=Me, R2=4-ClC6H4, R3=Me h R1=Bu, R2=C6H4, R3=Et

Structure

GENERAL

N-Acyloxy-N-alkoxyamides are archetypal anomeric amides. The combined electronegativity of the alkoxyl and acyloxyl oxygens strongly alters the geometry at the amide nitrogen and the attendant pyramidality manifests itself in radically reduced or negligible amide conjugation; the nitrogen lone pair is in an sp3 rather than a 2pz orbital (Fig. 3a). As a consequence, relative to amides and hydroxamic esters, they should have much longer N–C(O) bonds, low barriers to E–Z isomerisation and higher carbonyl stretch frequencies in their IR spectra. However the barrier to nitrogen inversion is likely to be low since, in the planar transition state for inversion, the lone pair can interact with the carbonyl (Fig. 3b). In contrast to N-acyloxyN-alkoxyamides, dialkoxyamines, which are also pyramidal at nitrogen, have abnormally high nitrogen inversion barriers. Here the planar inversion transition state is destabilised by sp2 hybridisation, which necessitates shorter N–O bonds, as well as lone pair repulsion.107–111 O R C

poor overlap N OAc

sp3N

O low barrier

R C AcO

OR

N

‡ O C

N

R

OR OAc

OR (a)

(b)

Fig. 3 (a) Poor p overlap in N-acyloxy-N-alkoxyamides; (b) stabilisation of the inversion transition state.

44

S.A. GLOVER σ*N— OR

O R

N

O

σ*N— OAc

nO

RCO

E

AcO σ*N—OAc

O

σN— OR

nO

σN— OAc

nOAc

(a)

R

90°

OAc

(b)

(c)

Fig. 4 (a) Anomeric overlap in N-acyloxy-N-alkoxyamides; (b) lone pair stabilisation through an nO–s*N–OAc interaction; (c) optimum conformation for anomeric overlap in Nacyloxy-N-alkoxyamides.

In addition to these unusual amide characteristics, the bisoxo substitution and the sp3 hybridisation at nitrogen results in a strong anomeric interaction between nO, the alkoxyl oxygen lone pair and the s*N–OAc bond (Fig. 4a). This is a common feature of all bisheteroatom-substituted amides that we have studied.32 In the case of N-acyloxy-N-alkoxyamides this anomeric interaction is likely to be strongest of the two possible anomeric effects; the s*N–OAc is lower in energy than s*N–OR while the alkoxyl oxygen p-type lone pair would be higher in energy than that of the acyloxyl ether oxygen (Fig. 4b). As illustrated in Fig. 4c, to maximise such interactions, the optimum geometry would require twist angles about the donor oxygen–nitrogen bond of around 901. Computed structures for model N-acyloxy-N-alkoxyamides, X-ray diffraction data for a number of congeners, as well as spectroscopic evidence fully support these qualitative arguments.

THEORETICAL STRUCTURES

Both AM1 calculations on N-acetoxy-N-methoxybenzamide38 and ab initio 6-31G* calculations on N-formyloxy-N-methoxyformamide 405,45 predict a strongly pyramidal nitrogen. Its lowest energy conformation at HF/6-31G* is depicted in Fig. 5a while structural data are provided in Table 1 together with that for N-methoxyformamide 41. O H

O N OCH3

H

O NHOCH3

H

NH2

CHO 40

41

42

The conformation in which the amide carbonyl and the formloxy groups are syn and both methyl and formyloxy carbonyl groups are exo to the nitrogen pyramid is lowest in energy. The average angle at nitrogen was 110.31 while dihedral angles give a Winkler–Dunitz amide distortion index of wN ¼ 58.5 (Table 1).112,113 The N–CO (1.405 A˚) and amide CQO (1.178 A˚) bond lengths are respectively longer and

45

N-ACYLOXY-N-ALKOXYAMIDES

Fig. 5 (a) HF/6-31G* lowest energy conformation of N-formyloxy-N-methoxyformamide; (b) Newman projection along the O1–N bond.

Table 1 Selected structural and spectroscopic properties of N-acyloxy-N-alkoxyamides 31b and 31f, -urea 38i, -carbamate 39g, theoretical models 40 and 41 and 1-aza-2-adamantanone 1 Parametera

31b

31f

38i

39g

40b

41b

1

rC ¼ O (A˚) rCN(A˚) rNOR(A˚) rNOAcyl(A˚)

1.2070 1.4414 1.4017 1.4396

1.2052 1.4394 1.4014 1.4414

1.222 1.426 1.398 1.426

1.198 1.424 1.396 1.474

1.178 1.405 1.360 1.375

1.188 1.373 1.373

1.210 1.455 – –

b1 (1) b2 (1) b3 (1) Sb (1) d (1)c ob4 (1)d

110.6 109.0 104.5 324.1 35.9 108.0

109.4 108.6 105.5 323.5 36.5 107.8

113.5 111.6 108.5 333.5 26.5 111.2

113.4 111.4 109.3 334.1 25.9 111.4

110.8 110.6 109.4 330.8 29.2 110.3

115.0 114.1 111.3 340.4 19.6 113.5

109.0 109.0 110.0 328.0 32.0 109.3

49.3 21.4 135.8 163.6

50.4 19.3 134.0 165.0

25.6 39.1 162.0 148.6

34.9 29.2 153.0 158.7

33.8 28.8 150.3 155.3

24.0 28.0 158.4 154.4

60.0 120.0 120.0 60.0

13.9 5.0 65.6 96.2 137.6 1730 174.2

15.5 4.4 65.3 96.7 141.6 1726 174.2

6.8 7.7 57.1 104.0 69.0 1720 158.5

2.9 7.9 56.3 95.5 67.1 1785

2.5 4.10 58.5 102.3 112.6

2.0 2.4 49.6

90.0 0.0 60.0

o1 o2 o3 o4

(1) (1) (1) (1)

t(1)e wC (1)e wN (1)e C–O–N–O(CO) (1) C(O)–O–N–OR(1) nmax (cm1) d13C amide CQO a

See Fig. 9 for deﬁnition of angles. Geometry calculated at HF/6-31G* level of theory. c d ¼ 360Sb. d ob4 ¼ Sb/3. e Amide distortion parameters deﬁned in accordance with Winkler–Dunitz.112,113 b

1731 200

46

S.A. GLOVER H

N

O

O H

O2CH

MeO

O

MeO N

H

H

CO 1852(s)/1818(a) (1.171) CN 1182 (1.462) t = 90°; = 105.8° -472.824551

(a)

H H

N

O

CO 1861(s)/1824(a) (1.172) CN 1190 (1.452) t = 90°; = 104.2° -472.833170

H

O H

CO 1784 (1.193) CN 1231 (1348) t=0˚; =120° -169.887003

H N

HCO2

HCO2

H OMe

CO 1796 (1.188) CN 1229 (1.373) t=2.0°; = 113.5° -284.349123

CO1838(s)/1819(a) (1.178) CN 1232 (1.405) t = 0.5°; =110.3° -472.845200

MeO

N

H N

O

H

N

MeO

O H

CO 1843 (1.178) CN 1020 (1.436) t = 90°; =105.2° -284.324215

(b)

CO 1813 (1.183) CN 1083 (1.427) t = 90°; =107.5° -169.858270

(c)

Fig. 6 Amide CQO and C–N vibrational frequencies (cm1) and bond lengths (A˚; in parentheses), twist angles (t), average angles at nitrogen ob4 computed at HF/6-31G* level for ground state and orthogonal conformations of (a) N-formyloxy-N-methoxyformamide 40, (b) N-methoxyformamide 41 and (c) formamide 42. Energies (Hartrees) at B3LYP/6-31G*// HF/6-31G* level.

shorter than those computed for N-methoxyformamide (1.373 A˚ and 1.188 A˚) reﬂecting signiﬁcantly decreased interaction with the amide carbonyl. The CH3–O–N–OCHO dihedral angle was 102.31 at the HF/6-31G* level reﬂecting a strong nO–s*N–OCHO anomeric overlap in this direction (Fig. 5b). The dihedral angle HCO–O–N–OMe of 112.61 reﬂects a poorer alignment for the alternative nOCHO–s*N–O interaction. The anomeric overlap in the preferred conformation also results in a shorter CH3O–N bond in 40 (1.360 A˚) than that found for methoxyformamide 41 (1.373 A˚) despite the reduced sp3 character in the latter (ob4 ¼ 113.5, wN ¼ 49.61). However, a more meaningful comparison would be with O-methylhydroxylamine, which like 40 at the HF/6-31G* level, is sp3 hybridised at nitrogen with an N–O bond length of 1.399 A˚. Gas-phase vibrational frequencies can be computed using molecular orbital methods,114,115 which can also provide structural data about transition states or unstable conformations. Fig. 6 gives HF/6-31G* gas-phase carbonyl and C–N stretch frequencies for ground-state and fully twisted conformations of N-formyloxy-N-methoxyformamide (Fig. 6a), N-methoxyformamide (Fig. 6b) and formamide (Fig. 6c). Fully twisted formamide and N-methoxyformamide have carbonyl stretch frequencies that are 29 and 47 cm1, respectively, higher than the planar or near-planar ground state structures. The C–N stretch frequencies are however reduced concomitantly by 147 and 208 cm1, respectively, indicating a signiﬁcant loss of lone pair overlap in the twisted forms. The carbonyl vibrational frequencies for The frequency change is small when compared to the difference between CQO (1700 cm1) and C–O

(1100 cm1) stretches in line with the prevailing view that in planar amides relatively little charge is transferred to oxygen. The changes in frequency represent a small stiffening of the carbonyl bond at best.

N-ACYLOXY-N-ALKOXYAMIDES

47

both twisted forms of N-formyloxy-N-methoxyformamide 40 are strongly coupled but the averages of both the symmetrical and asymmetrical stretch frequencies (1835 and 1842 cm1) are comparatively similar to that of the ground state conformation (1828 cm1) (Fig. 6a). In this case, the C–N stretch frequency is reduced by only 50 or 42 cm1. Thus, when compared with formamide or N-methoxyformamide, both the carbonyl stretch and the C–N bond stretch frequencies are much less sensitive to the orientation of the sp3 lone pair. The small changes in the frequencies and bond lengths upon twisting the lone pair of N-formyloxy-N-methoxyformamide out of conjugation (Fig. 6a) would suggest that, at HF/6-31G* level in the gas phase at least, some residual interaction is lost. However, it should be borne in mind that changes in the degree of pyramidality will also affect bond lengths. While the C–N bond length in the ground state of 40 (1.405 A˚) is of the same order as twisted formamide 42 (1.427 A˚), the bond lengthens by 0.005–0.006 A˚ in the twisted forms of 40. To a large degree this must be due to the tighter angles at nitrogen, which result in orbitals with greater p character. The extent of nitrogen lone pair-carbonyl overlap is also reﬂected in the barriers to E–Z isomerisation. B3LYP/6-31G*//HF/6-31G* barriers to rotation in formamide (18.0 kcal mol1), methoxyformamide (15.6 kcal mol1) are similar in concurrence with our results in an earlier study using B3LYP/6-31G(D) optimised geometries.6 However both barriers are signiﬁcantly higher than the smallest barrier for isomerisation of formyloxymethoxyformamide (7.5 kcal mol1; nitrogen lone pair and carbonyl syn). Though the nitrogen of the hydroxamic ester is computed to be signiﬁcantly pyramidal (wN ¼ 49.6), the similarity in IR stretch frequencies and isomerisation barriers to those of the formamide conﬁrm that the impact of one oxygen substituent upon amide characteristics is far smaller than that brought about by bisoxo-substitution. The calculated variations in CQO and C–N vibrational frequencies for formamide (+29 and 147 cm1, respectively) upon twisting the lone pair into the OCN plane are in accordance with changes in bond lengths (0.001 and+0.008 A˚ respectively) and reﬂect the contemporary view that p-donation to the carbonyl carbon in the planar form does not result in an equivalent loss in CQO pi-bond character.2 Similar results are found for N-methoxyformamide. The small differences in CQO bond lengths between planar (sp2 N) and twisted (sp3 N) formamide and methoxyformamide, as well as the similarity in CQO bond length of all three amides in the ground state (sp2 N in 42, sp3 N in 40) also lend support to Wiberg’s HOMO–LUMO theory of amide resonance.

X-RAY STRUCTURES

X-ray data are available for several N-acyloxy-N-alkoxyamides 31b and 31f,5 an N-acyloxy-N-alkoxyurea 38i and an N-acyloxy-N-alkoxycarbamate 39g.51 The structure of all four is dominated by pyramidality at the amide nitrogen. While amides are typically planar or close to planar with average angles at nitrogen close

48

S.A. GLOVER

to 1201 in the overwhelming majority of structures studied to date,5 these bisoxosubstituted species have nitrogen atoms that are highly sp3 hybridised. Fig. 7 depicts crystal structures of N-acyloxy-N-alkoxyamides 31b and 31f and those of urea 38i and carbamate 39g are given in Fig. 8. Principal bond lengths, angles at nitrogen and Dunitz amide distortion parameters (Fig. 9) are listed in Table 1 together with those for Kirby’s most twisted amide 1.

Fig. 7 ORTEP116 depiction of (a) N-benzoyloxy-N-(4-tert-butylbenzyloxy)benzamide 31b and (b) N-(4-tert-butylbenzoyloxy-N-(4-tert-butylbenzyloxy)-4-tert-butylbenzamide 31f with displacement ellipsoids shown at the 20% level.

49

N-ACYLOXY-N-ALKOXYAMIDES

Fig. 8 ORTEP116 depiction of (a) N-acetoxy-N-ethoxyurea 38i and (b) methyl N-4chlorobenzoyloxy-N-methoxycarbamate 39g with displacement ellipsoids shown at the 20% level.

τ

O ω3

ω2 AcO1

R3

C

ω1 O2R ω4

τ=1/2(ω1+ω2) χC=(ω1−ω3+π) = (ω4−ω2+π) χN=(ω2−ω3+π) = (ω4−ω1+π) (a)

β3 AcO1 O2R β2 N β1 C O

R3 (b)

Fig. 9 (a) Winkler–Dunitz parameters for deﬁning twist (t) and pyramidality (wN, wC) in amides; (b) angles at nitrogen.

Crystal data for all four amides are uniformly similar. While amides 31b and 31f (|wN| ¼ 65.61 and 65.31, respectively) are more pyramidal than the urea and carbamate (|wN| ¼ 57.11 and 56.31, respectively) all four possess sp3-hybridised nitrogens. Average angles at nitrogen ob4 for the two amides are smaller than that required by pure tetrahedral geometry as exempliﬁed by the tetrahedral nitrogen in 1-aza-2-adamantanone 1. In all cases, the high degree of pyramidality attributed to the presence of two electronegative oxygen atoms at nitrogen, conﬁrms the predictions of HF/6-31G* calculations.5 On the basis of geometries for all acyclic amides in the Cambridge Structural Database in 2002,117 amides 31b and 31f are the most pyramidal amides of this type.5 Sp3 hybridisation at nitrogen results in decreased lone pair 2pz character, which manifests itself in three ways: (i) Diminished co-linearity between nitrogen lone pair and carbonyl carbon 2pz orbitals

50

S.A. GLOVER

(ii) Long N–C bond lengths (iii) Reduced E–Z isomerisation barriers All four structures bear testimony to these facts. Both 31b and 31f exhibit a signiﬁcant degree of twisting around the N–C(O) bond (t ¼ 13.9 and 15.51, respectively). The experimental barrier to E–Z isomerisation in N-acyloxy-N-alkoxyamides has been estimated to be below 8–9 kcal mol1 32 in line with the theoretical barrier for 40 (7.5 kcal mol1). Since the lone pair on nitrogen resides in an sp3 hybrid orbital, twisting angles are likely to arise from steric interactions or crystal packing rather than any residual p-orbital overlap considerations. Nitrogen lone pairs in urea 38i and carbamate 39g are closer to the carbonyl carbon 2pz plane (Table 1). The N–C(O) bonds (rC–N, Table 1) are all long when compared with normal, acyclic amides (average bond length of 1.359 A˚, median 1.353 A˚5). Although N–C(O) bonds of N-acyloxy-N-alkoxyamides cannot be compared directly with tertiary N,N-dialkylamides on account of the different hybridisation at nitrogen, they are only slightly shorter than the corresponding bond in the fully twisted lactam 1-aza-2-adamantanone 1 even though the twist angle in 31b and 31f is relatively small (Table 1). The urea and carbamate carbonyl bond lengths cannot be compared directly with the amides. However, the crystallographic carbonyl bond lengths in the amides, 31b and 31f, are similar to that in 1-aza-2-adamantanone 1 (Table 1). In general, when compared with C(O)–N bonds, there is much less variation in carbonyl bond lengths in planar and twisted amides24 and the same is true of anomeric amides. In general, carbonyl bond lengths are relatively insensitive to the degree of lone pair interaction. Kirby has attributed this to a negative hyperconjugative effect involving an nO–s*C–N interaction impacting in an opposite sense upon the CQO bond length21 but, as has been proposed by Wiberg, transfer of nitrogen p-density to the carbonyl p* orbital is the main process and oxygen gains relatively little electron density from this.2,3 Similar arguments can account for the relative insensitivity of CQO bonds to loss of p-overlap through pyramidalisation at nitrogen in N-acyloxyN-alkoxyamides. Geometries of all four amides indicate a preference in the solid state for a conformation in which the alkoxyl oxygen lone pair, nOR, is largely collinear with the s*N–OAc bond. Dihedral angles for C–O–N–OAc in Table 1 are all close to 901. This anomeric interaction is clearly predominant as the alternative, C(O)–O–N–OR dihedral angles indicate poor nOAc–s*N–OR interactions. As a consequence, N–OR bonds should be shorter than normal. As with the theoretical models, a comparison between experimental N–OR bond lengths and those of the parent hydroxamic esters is inappropriate on account of differences in hybridisation at nitrogen; in hydroxamic esters nitrogen is much closer to sp2 hybridised leading naturally to shorter N–O bonds. However, a comparison can be made to the experimental bond length in hydroxylamines since the nitrogen is fully tetrahedral in these, as well as in N-acyloxy-N-alkoxyamides. Typically, the N–O bond in hydroxylamines is of the order of 1.44–1.46 A˚ and values in this range have also been computed for the fully twisted hydroxamic esters (Fig. 6b) where the nitrogen also assumes a pyramidal

N-ACYLOXY-N-ALKOXYAMIDES

51

geometry.6 The rNOR distances of 1.402, 1.401, 1.398 and 1.396 A˚ in 31b, 31f, 38i, 39g respectively represent very signiﬁcant shortening, presumably on account of an effective nO–s*N–OAc anomeric interaction. SPECTROSCOPIC PROPERTIES

Carbonyl stretch frequencies, carbonyl 13C and amide 15N chemical shifts for a wide range of N-acyloxy-N-alkoxyamides are listed in Table 2 together with those of the precursor hydroxamic esters. Spectroscopically, mutagens can be categorised into six types: (i) (ii) (iii) (iv) (v) (vi)

N-Alkanoyloxy-N-alkoxyarylamides (entries 1–41) N-Alkoxy-N-aroyloxyarylmides (entries 42–52) N-Alkanoyloxy-N-alkoxyalkylamides (entries 53–58) N-Alkoxy-N-aroyloxyalkylamides (entries 59–68) N-Acyloxy-N-alkoxyureas (entries 69–72) N-Acyloxy-N-alkoxycarbamates (entries 73–77)

Infrared spectroscopy N-Acyloxy-N-alkoxyamides are characterised by two high frequency double bond absorptions corresponding to the stretching modes of the ester and amide carbonyls. The acyloxyl substituents are ester groups but the bonding to nitrogen results in higher than normal CQO stretch frequencies. For aliphatic groups (R2 ¼ alkyl, Table 2, entries 1–41 and 53–58), these are all in the range of 1767 cm1 through to 1800 cm1 whereas carbonyls of saturated esters are generally found between 1735 and 1750 cm1.124,125 Acetoxy compounds have carbonyls that are uniformly close to or 41790 cm1 while at the lower end of the range are higher alkylamides (entries 12–19). Bulky pivaloyloxy and adamantane-1-carboxy esters in 29k and 29l (entries 18 and 19) exhibit distinctly lower carbonyl absorptions of 1767 and 1771 cm1 in line with the known impact of sterically induced angular deformation at carbonyls,124,126 although, as discussed below, steric effects upon solvation cannot be discounted.127 Carbonyls of aroyloxyl groups (R2 ¼ Aryl, Table 2, entries 42–52 and 59–68) range from 1750 to 1767 cm1 and are similarly increased relative to normal aromatic esters (1715–1730 cm1). Most benzoyloxy compounds give carbonyls between 1760 and 1765 cm1 but those that are para–substituted with donor groups, as well as those with polycyclic aroyloxyl groups, have carbonyls in the lower range of 1750–1760 cm1. Thus the impact of hydroxamic ester substitution through nitrogen at the ester oxygen is to raise the acyloxyl carbonyls into the range for acid chlorides and acid anhydrides.128

R1

Entry/ structure

13

C and selected

R2

R3

15

N chemical shiftsa (CDCl3) for N-acyloxy-N-alkoxyamides

Amide n (cm1) (d13C) [d15N]

Ester n (cm1) (d13C) [d15N]

Hydroxamic ester n (cm1) (d13C) [d15N]

26c38 26f38 26e38 26d38 26g38 26b38 26i38 26h118 29m48 29n119 29p119

Bu Bu Bu Bu Bu Bu Bu Bu Bu Bu Bu

Me Me Me Me Me Me Me Me Me Me Me

4-Biphenylyl p-Brphenyl p-Clphenyl p-Mephenyl p-NO2phenyl p-MeOphenyl p-Butphenyl m-NO2phenyl 2-Naphthyl Fluoren-1-yl Anthraquinone2-yl

1721 (173.01) 1731 (173.12) 1728 (172.85) 1730 (173.93) 1730 (172.15) 1728 (173.18) 1721 (173.70) 1727 (171.39) 1724 (174.03) 1742b 1726b

1794 (166.24) 1791 (167.94) 1792 (167.82) 1790 (168.01) 1790 (167.83) 1790 (168.02) 1794 (168.08) 1794 (167.71) 1792 (167.98) 1785b 1792b

1684 1696 1695 1695 1695 1692 1679 1693 1684 1681 1672

(166.38) (165.45) (165.29) (166.44) (164.14) (166.12) (166.42) (164.00) (166.70) (167.10) (182.40)

12 13 14 15 16 17 18 19

29g119 29w120 29y120 29h119 29i119 29j119 29k119 29l119

Bu Bu Bu Bu Bu Bu Bu Bu

Pr 1-Hexyl 5-Hexen-1yl Pri (S)-2-Butyl Neopentyl 1-Adamantyl But

Phenyl Phenyl Phenyl Phenyl Phenyl Phenyl Phenyl Phenyl

1721 (174.4) 1723 (174.2) [124] 1730 (174.4) 1721 (174.3) 1725 (174.5) 1721 (174.4) 1723 (174.4) 1722 (174.7)

1774 1782 1783 1775 1774 1778 1767 1771

1654 1654 1654 1654 1654 1654 1654 1654

(165.65) (165.65) [197] (165.65) (165.65) (165.65) (165.65) (165.65) (165.65)

20

25a38

Et

Me

Phenyl

1789 (166.18)

1679 (166.50) [200]

21 22

25b48 25f38

n-Pr Pri

Me Me

Phenyl Phenyl

1789 (168.18) 1788 (166.21)

1678 (166.42) 1684 (166.77)

23 24 25

25c47 25g38 25j121

Bu Bui 2-Bu

Me Me Me

Phenyl Phenyl Phenyl

1795 (167.80) 1790 (166.05) 1790 (168.30)

1654 (165.65) 1684 (166.54) 1685 (166.90) [200]

26 27 28

25i121 25e47,48 27a47

But Octyl Benzyl

Me Me Me

Phenyl Phenyl Phenyl

1724 (174.25) [124] 1724 (174.30) 1724 (174.68) [127] 1732 (173.90) 1724 (174.15) 1719 (174.90) [127] 1707 (174.9) 1728 (174.29) 1728 (174.12)

1786 (168.40) 1798 (168.21) 1798 (168.08)

1686 (167.90) 1684 (166.51) 1678 (166.26)

S.A. GLOVER

1 2 3 4 5 6 7 8 9 10 11

(170.49) (171.10)) (171.0) (174.5) (173.9) (169.5) (174.4) (175.7)

52

Table 2 Infrared carbonyl absorption frequencies (CHCl3), (R1ON(OCOR2)COR3) and precursor hydroxamic esters

R1

Entry/ structure

R2

R3

29 30 31 32 33 34 35

27g38 27f38 27h38 27d38 27i38 27e38 27c38

p-Brbenzyl p-Clbenzyl p-NO2benzyl 4-Biphenylyl p-Butbenzyl p-Mebenzyl p-PhObenzyl

Me Me Me Me Me Me Me

Phenyl Phenyl Phenyl Phenyl Phenyl Phenyl Phenyl

36

27b38

Me

Phenyl

37

37a

2,6-diMebenzyl

Me

Phenyl

38 39

37b121 37c121

3,5-diMebenzyl 3-Mebenzyl

Me Me

Phenyl Phenyl

40 41

37d121 34105

2-Mebenzyl – (CH2)6–

Me Me

Phenyl di-Phenyl

42 43 44 45 46

29f49 28b40 28i40 28d40 28a40

Bu Benzyl Benzyl Benzyl Benzyl

Phenyl p-MeOphenyl p-Butphenyl p-Mephenyl Phenyl

Phenyl Phenyl Phenyl Phenyl Phenyl

47 48 49 50 51 52

28e40 28f40 28j40 28g40 28h40 28c119

Benzyl Benzyl Benzyl Benzyl Benzyl Benzyl

p-Clphenyl p-CHOphenyl p-CF3phenyl p-CNphenyl p-NO2phenyl 4-Biphenylyl

Phenyl Phenyl Phenyl Phenyl Phenyl Phenyl

53 54 55

29v120 29x120 30a105

Bu Bu Bu

1-Hexyl 5-Hexen-1-yl Me

But But Me

1731 (174.04) 1731 (174.08) 1732 (174.20) 1729 (174.14) 1725 (173.94) 1729 (174.14) 1725 (174.17) [124] 1725 (174.17) [124] 1727 (174.20) [124] 1728 (174.10) 1729 (174.20) [124] 1726 (174.20) 1724 (174.3)

Ester n (cm1) (d13C) [d15N] 1791 1793 1793 1791 1790 1791 1795

(168.02) (168.02) (168.19) (166.10) (167.89) (168.10) (168.15)

Hydroxamic ester n (cm1) (d13C) [d15N] 1680 1687 1691 1674 1685 1681 1690

(166.50) (166.68) [203] (166.59) (164.91) (167.89) (166.24) (157.97) [196]

1795 (168.15)

1684 (159.93) [203]

1791 (168.20)

1684 (166.70)202]

1791 (168.10) 1792 (168.10)

1686 (166.00) 1684 (165.60) [196]

1792 (168.10) 1789 (168.20)

1650 (166.20) 1662 (166.00)

1728 (174.50) 1718 (174.25) 1738 (174.36) 1733 (174.36) 1731 (174.28) [123] 1734 (174.20) 1735 (174.07) 1734 (174.12) 1732 (174.04) 1733 (173.93) 1729 (174.40) [123]

1758 1750 1756 1756 1758

(164.00) (164.04) (164.21) (164.21) (164.13)

1654 1678 1678 1678 1678

(165.50) (166.26) (166.26) (166.26) (166.26) [197]

1759 1761 1765 1763 1764 1757

(163.25) (163.21) (163.07) (162.70) (162.37) (164.20)

1678 1678 1678 1678 1678 1678

(166.26) (166.26) (166.26) (166.26) (166.26) (166.26) [197]

1718 (182.50) 1719 (182.50) 1746 (176.20)

1779 (170.70) 1785 (170.50) 1797 (167.70)

1683 (176.00) 1683 (176.00) 1678 (167.90)

53

p-MeObenzyl

121

Amide n (cm1) (d13C) [d15N]

N-ACYLOXY-N-ALKOXYAMIDES

Table 2 (continued )

54

Table 2 (continued ) R1

Entry/ structure

R2

R3

Amide n (cm1) (d13C) [d15N]

Ester n (cm1) (d13C) [d15N]

Hydroxamic ester n (cm1) (d13C) [d15N]

56 57 58

30c105 35a122 35b122

Benzyl di-benzyl di-benzyl

Me Me Me

Me –(CH2)7 –(CH2)8

1736 (176.30) 1742b 1741 (178.48)

1793 (167.90) 1794b 1791 (167.43)

1690 (168.00) [196] 1689 (154.50) 1690b

59 60 61 62 63 64 65 66 67 68

30d105 30b105 29a46 29b46 29c46 29d46 29e46 30e105 30j123 30k123

Benzyl Bu Bu Bu Bu Bu Bu Bu Bu Bu

Ph Ph Ph Ph Ph Ph Ph 2-Naphthyl 1-Fluorenyl 1-Pyrenyl

Me Me Et Pri But Neopentyl 1-Adamantyl Me Me Me

1738 1738 1745 1736 1728 1735 1721 1737 1746 1743

1762 1765 1766 1767 1763 1763 1755 1757 1756 1751

1690 1678 1664 1657 1683 1653 1681 1678 1678 1678

69 70 71 72

38e52 38f52 38k52 38l51

n-Pr Pr Bu Et

Me Me Me Me

N(Me)2 NHMe NH2 NH2

73 74 75 76 77

39a52 39c52 39d52 39e51 39h106

Me Octyl Me Me Bu

Me Me Me 4-Clphenyl Ph

OMe OMe OEt OEt OEt

(176.50) (176.50) (180.10) (182.60) (182.80) (177.20) (182.20) (176.5) (176.5) (176.70)

1732c 1730c 1730c 1720c 1780c 1780c 1755c 1785c (158.5) 1762d (158.2)

(164.10) (164.10) (164.10) (164.10) (164.00) (163.80) (164.10) (164.30) (164.10) (165.20)

1784c 1795c 1790c 1798c 1805c 1805c 1780c 1780 (163.7) 1772d (164.3)

(168.00) (167.90) (171.80) (174.90) (176.20) (169.50) (175.40) (167.90) (167.90) (167.90)

1695c 1685c 1680c 1680c 1765c 1745c 1765 1724d (157.7)

a15

N shift relative to nitromethane 0 ppm. Not recorded. c Solid state or thin ﬁlm IR. d Thin ﬁlm IR. b

S.A. GLOVER

55

N-ACYLOXY-N-ALKOXYAMIDES

In contrast to CQO bond lengths, CQO vibrational frequencies are very much more sensitive to the extent of lone pair overlap in amides. In particular, increasing combined electronegativity of substituents at nitrogen and the attendant pyramidalisation and electronic effects result in a marked increase in the nC ¼ O values although the changes reﬂect only small variations in force constants.32 With the exception of 25i (entry 26) arylamides (R3 ¼ aryl, Table 2, entries 1–52) exhibit carbonyls in the range of 1718–1742 cm1, on average some 51713 cm1 higher than their precursor hydroxamic esters. N-Acyloxy-N-alkoxyalkylamides (R3 ¼ alkyl, Table 2, entries 53–68) exhibit carbonyls on average 57715 cm1 higher than the hydroxamic esters from which they are derived and their amide absorptions appear on average about seven wave numbers higher than their arylamide counterparts. Clearly, most, if not all sp2 character at nitrogen is lost upon bisoxo substitution and the extent of nitrogen lone pair interaction with the amide carbonyl is radically altered. As a consequence, the carbonyl stiffens measurably. N-AcetoxyN-tert-butoxybenzamide 25i has an anomalously low carbonyl (1707 cm1) that may suggest a smaller degree of pyramidality in this substrate due to steric hindrance. The vibrational frequencies of N-acyloxy-N-alkoxyamides correspond to that observed for the twisted 1-aza-2-adamantanone 1 (1731 cm1).21,23 Where an amide nitrogen lone pair (either through twisting/pyramidalisation or pyramidalisation alone) loses conjugation with the carbonyl, the conﬁguration is analogous to an ester rather than a ketone. As with esters, acid halides and anhydrides, or peresters,129 the carbonyl stretch frequency is higher than that of ketones and aldehydes on account of destabilisation of the polar form of the carbonyl by the strongly –I effect of the nitrogen or in these compounds, bisoxo-substituted nitrogen (Fig. 10 (II)). Steric effects on both the amide and the acyloxyl side chain are similar. Tert-butyl and adamantyl groups on the amide side chain in 29v, 29x, 29c, and 29e (Table 2 entries 53 and 54, 63 and 65) result in lower stretch frequencies that, on average, are only 40 cm1 higher than their precurser hydroxamic esters. Streck and coworkers have suggested that such changes in dialkyl ketones can be ascribed to destabilisation of resonance form II through steric hindrance to solvation which, in the case of tert-butyl counteracts the inductive stabilisation.127 When comparing the series of (para-substituted benzoyloxy)benzyloxybenzamides (28a–j, Table 2, entries 43–52), one notable exception is the 4-methoxy analogue 28b, which exhibits a much lower carbonyl stretch at 1718 cm1. The through resonance of the methoxyl lone pair reduces the electron-withdrawing effect of the acyloxyl substituent possibly resulting in a lower degree of pyramidality at the amide nitrogen and better resonance into the amide carbonyl.

O C

O C R

N I

R

N II

Fig. 10 Resonance forms in twisted or pyramidal amides.

56

S.A. GLOVER

Three other series of related anomeric amides have been studied in these laboratories. The closely related N,N-dialkoxyamides exhibit carbonyls in their solution infrared spectra in the range of 1705–1715 cm1, some 20–30 cm1 higher than hydroxamic esters.32 The signiﬁcantly higher range for N-acyloxy-N-alkoxyamides can clearly be ascribed to the increased electron-withdrawing effect of the ester functionality as opposed to an alkoxyl group. N-Chlorohydroxamic esters are also anomeric and their carbonyl stretch frequencies are in the range of 1720–1745 cm1, on average 40 cm1 higher than their hydroxamic esters precursors in line with the lower electronegativity of chlorine.32 In addition, related N-alkoxy-N-aminoamides in the form of N,N0 -dialkoxy-N,N0 -diacylhydrazines are less pyramidal at nitrogen and have solution state carbonyls in the range of 1700–1722 cm1 for arylamides and 1733–1744 and 1707–1723 cm1 for non-degenerate carbonyls of acetamides.35,130 N-Acyloxy-N-alkoxyamides are clearly the most pyramidal of all anomeric amides reported to date. The situation with N-acyloxy-N-alkoxyureas and carbamates is similar although infrared data were mostly determined by liquid ﬁlm or condensed phase (KBr/nujol mull).52,131 However, the limited data for N-acyloxy-N-alkoxyureas (Table 2, entries 69–72) give amide carbonyl frequencies ca. 1730 cm1 that are raised by some 37–40 cm1 by acyloxylation. Values for carbamates (Table 2, entries 73–77) are higher (mostly 1780 cm1) but are raised to a lesser extent (10–20 cm1) relative to their parent carbamates. Clearly, carbonyl vibrational frequencies will be inﬂuenced strongly by the adjacent amino or alkoxyl group in both analogues. Nuclear magnetic resonance spectroscopy 13

C NMR spectroscopy. 13C NMR resonances for both carbonyls of the N-acyloxyN-alkoxyamides and that of the parent hydroxamic ester are given for most compounds in Table 2. With the exception of para–substituted benzyloxy-N-acyloxy-N-alkoxyamides, amide carbonyl 13C NMR values of 56 congeners differ from those of their precursor hydroxamic esters by on average +8.0(70.6) ppm. Steric and electronic effects inﬂuence hydroxamic esters and N-acyloxy-N-alkoxyamides similarly. This includes substrates with branching a to the amide carbonyl. Two classes of N-acyloxy-N-alkoxyamides are worthy of mention: a. Mutagens with benzyloxy and para–substituted benzyloxy side chains 27a–i (Table 2, entries 28–36) are an exception. Their carbonyl shifts differ by between 6–16 ppm relative to their parent alkoxyamides. In this subset, this variation is almost entirely due to changes in the carbonyl chemical shift of the hydroxamic ester (d168–158) since N-acyloxy-N-alkoxyamides are all very close to d174. This proves to be another manifestation of pyramidality at nitrogen in N-acyloxyN-alkoxyamides. Hydroxamic esters are largely sp2 hybridised at nitrogen and the similarity in carbonyl stretch frequencies in hydroxamic esters and simple amides conﬁrms their strong resonance interaction. This results in some positive polarisation at the amide nitrogen but overall, lone pair overlap with the carbonyl

N-ACYLOXY-N-ALKOXYAMIDES

57

carbon would be facilitated by donor groups and disfavoured by electronwithdrawing substituents on the alkoxyl group. However, amide carbonyl chemical shifts in the analogous N-acyloxy-N-alkoxyamides, where nitrogen is strongly sp3 hybridised and which exhibit little or no amide resonance, are insensitive to the electronic effect of alkoxyl substituents at nitrogen. Amide carbonyl vibrational frequencies for the series behave similarly; while those of the hydroxamic esters in this subset vary between 1678 and 1691 cm1, the N-acyloxy-N-alkoxyamide carbonyls span only six wave numbers. b. Branching a to the amide carbonyl (Table 2, entries 53, 54, 62, 63 and 65) affects the mutagen and hydroxamic esters similarly and causes a marked downﬁeld shift of up to 6 ppm relative to the acetamide substrate (Table 2, entry 60). Branching a to the ester carbonyls results in a downﬁeld shift of 8 ppm relative to acetoxyl esters (Table 2, entries 15, 16 18 19 versus 23). Associated CQO infrared stretches are also lowered in these substrates and both trends support stabilisation of the polar resonance form II in Fig. 10. A simple model of inductive stabilisation is indicated but steric effects must also be implicated. Pri, But and 1-adamantyl produce similar downﬁeld shifts of around 8 ppm on both the amide and ester side chains, despite the greater inductive effect of the tertiary systems. Similarly, neopentyl, where branching is removed one carbon from the carbonyl produces only a small shift of o1 ppm, which is smaller than the 3 ppm produced by unbranched ethyl (on R3) or propyl (on R2), which have a similar or weaker inductive effect. Streck and coworkers showed that in a range of solvents, the 13C carbonyl shifts in dialkyl ketones were affected similarly by branching at the a-position.127 In chloroform, the carbonyls of di-tert-butylketone and diisopropylketone were 11–12 ppm downﬁeld of that of acetone, which they attributed to a mixture of inductive and steric effects. With tertiary systems, particularly in dipolar solvents, hindrance to solvent stabilisation of the polar, basic form of the carbonyl offsets the inductive stabilisation of the branched alkyl. 13C NMR data presented here support this. While amide carbonyl chemical shifts in N-acetoxy-N-alkoxyamides are uniformly shifted to higher frequency by about 8 ppm relative to their precursor hydroxamic esters, they are not ketonic and appear some 20 ppm upﬁeld of aryl/alkyl ketones. The shift in the ‘‘most twisted’’ amide 1 is at a signiﬁcantly higher frequency (200 ppm) and has been labelled as aldehydic by Kirby and coworkers.22 However, rigid bicyclic ketones exhibit much higher carbonyl 13C NMR shifts and 200 ppm represents a shift to higher ﬁeld of similar magnitude relative to 2-adamantanone, which has a carbonyl 13C NMR resonance at 218 ppm. The origin of this upﬁeld shift from that of ketones, parallels that of esters, anhydrides and acid chlorides. Contrary to conventional reactivity arguments, which imply that substitution at carbonyls by electronegative atoms reduces electron density at the carbonyl carbon and hence promotes addition to carbonyls, a systematic study of 13C NMR shift data for ester carbonyls shows that electron density is actually greater at such carbons (reactivity enhancement is actually due to destabilisation of the ground states of the esters by the electron-withdrawing substituents).132,133 Our observations are in line with those of Neovonen et al. Electron-withdrawing nitrogen in

58

S.A. GLOVER

twisted/pyramidal amides destabilises resonance form II (Fig. 10) relative to ketones. These carbonyls have greater double bond character resulting in higher electron density at carbon, higher ﬁeld carbonyl chemical shifts and higher nCQO. 15

N NMR spectroscopy. 15N NMR data have been obtained for a limited number of mutagens and hydroxamic esters and chemical shifts are presented in Table 2. Secondary and tertiary amide nitrogens generally resonate in the region of 200 to 250 ppm relative to nitromethane.134 Formamide resonates at 267.5 ppm and N,N-dimethyl-p-toluamide resonates at 282.6 ppm.135 15N in N-butoxybenzamide (Table 2, entry 13) resonates much further downﬁeld at 195.7 ppm relative to nitromethane and that of other hydroxamic esters in this study all resonate between 196 and 203 ppm. The large shift to higher frequency of ca. 87 ppm relative to N,N-dimethyl-p-toluamide represents a signiﬁcant impact of electronegative oxygen on the electron density at an amide nitrogen. An acyloxyl group would be expected to be more strongly electron-withdrawing than an alkoxyl group but bisoxo substitution in N-heptanoyloxy-N-butoxybenzamide 29w (Table 2 entry 13) results in a smaller downﬁeld shift of about 73 ppm. Downﬁeld shifts for other N-acyloxyN-alkoxyamides were similar. Successive alkoxylation of isobutoxyamine to give N-ethoxy-N-methoxyisobutoxyamine results in 15N downﬁeld shifts of 96 and 72 ppm.136 Thus acyloxylation would be expected to produce a downﬁeld shift at least of the order of 100 ppm. Computed properties, IR carbonyl stretch frequencies as well as isomerisation barriers in hydroxamic esters support a largely intact interaction between the nitrogen lone pair and the carbonyl in these amides resulting in a signiﬁcant donation of the lone pair to the carbonyl carbon despite the electronegativity of the alkoxyl oxygen. However, bisoxo substitution, as is generally found for anomeric amides, results in strong pyramidalisation at nitrogen and attendant disconnection of the nitrogen and carbonyl. The complete localisation of the lone pair on nitrogen in N-acyloxy-N-alkoxyamides must, to a degree, offset the electron-withdrawing effect of the acyloxyl group. Yamada has demonstrated that twisting amides (though without complete pyramidalisation) produces relatively small upﬁeld shifts of only 12 ppm.137 However, when compared with N,N-dimethyl-p-toluamide, the amide nitrogen in N-p-toluoylaziridine, which like N-acyloxy-N-alkoxyamides has a completely pyramidal nitrogen, is upﬁeld by some 34 ppm.135 On the strength of this, it would appear that in N-acyloxy-N-alkoxyamides, pyramidalisation and loss of amide resonance probably accounts to a large degree for the smaller than expected shift. The role of steric effects is unclear but the anomeric effect could also contribute to an increase in electron density at nitrogen. X-ray data for the two N-acyloxyN-alkoxyamides, a urea and a carbamate outlined above show clear evidence, both from bond lengths and conformations, of an anomeric interaction; RO–N bonds are short when compared to alkoxyamines. This interaction is responsible for SN1, SN2, homolytic and rearrangement reactions of N-acyloxy-N-alkoxyamides (vide infra) and has also been supported computationally. Acyloxylation of the hydroxamic esters results in both pyramidalisation as well as anomeric donation from the

59

N-ACYLOXY-N-ALKOXYAMIDES

alkoxyl oxygen to the s*NOAc orbital according to Fig. 4a. This could also result in increased electron density at nitrogen. The smaller shift of 72 ppm upon a second alkoxylation of isobutoxyamine might also reﬂect increased anomeric interactions in N-ethoxy-N-methoxyisobutoxyamine.136 Dynamic 1H NMR spectroscopy. In stark contrast to parent hydroxamic esters, which usually exhibit line broadened alkoxyl group resonances in their 1H NMR spectra, at or even signiﬁcantly above room temperature,138,139 in CDCl3 the alkoxyl and acetoxyl protons in N-acetoxy-N-alkoxybenzamides give rise to sharp signals well below room temperature. In d8-toluene, the benzylic and acetoxyl methyl resonances of N-acetoxy-N-benzyloxybenzamide 43 showed signiﬁcant line broadening below 250 K but remained isochronous down to 190 K. O

O

C Ph

C N

OCH2Ph

OAc 43

Ph

N

OCH2Ph

Cl 44

These results clearly indicate that barriers to all isomerisation processes are at least less than about 8 kcal mol1. In N-benzyloxy-N-chlorobenzamide 44 the amide isomerisation was not observable but the anomeric overlap resulted in diastereotopic benzylic hydrogens, which at coalescence afforded a barrier for rotation about the N–OBn bond of around 10.3 kcal mol1.32 Like its N-chloro analogue, the amide isomerisation barrier in 43 is too low to be observed by 1H NMR and even though there is deﬁnitive X-ray and theoretical evidence for anomeric effects in N-acyloxyN-alkoxyamides, the barrier to isomerisation about the N–OBn bond must be lower than 10.3 kcal mol1. The nO–s*N–Cl anomeric interaction in 44 is predicted to be stronger than the nO–s*N–OAc interaction in 43 on purturbation arguments.32

4

Chemical reactivity

N-Acyloxy-N-alkoxyamides are intrinsically reactive at the amide nitrogen. Three factors contribute to this: (i) Sp3 character of the nitrogen itself, which results in reactivity akin to sp3 carbon. SN1 and SN2 reactivity, both of which are improbable at sp2 carbon and planar amide nitrogens, become feasible (ii) The substitution pattern strongly favours an nO–s*N–OAc anomeric effect resulting in weakening of the N–OAc bond (iii) Alkoxyl oxygen lone pair involvement promotes heterolytic SN1 and SN2 reactions at nitrogen, homolysis reactions as well as molecular rearrangements (vide infra)

60

S.A. GLOVER

As part of our programme aimed at understanding the source of their chemical mutagenicity, we have fully investigated the reactivity of N-acyloxy-N-alkoxyamides under a range of conditions. They have been shown to undergo acid-catalysed solvoysis reactions producing alkoxynitrenium ions, SN2 reactions with a range of organic and inorganic nucleophiles, homolysis reactions that produce alkoxyamidyl radicals as well as the heteroatom rearrangements on nitrogen (HERON) rearrangement, an intramolecular reaction discovered in these laboratories. Urea and carbamate analogues undergo alcoholysis to give N,N-dialkoxyamides.

SOLVOLYSIS STUDIES – AAL1 REACTIVITY37– 40

The ﬁrst indication that N-acyloxy-N-alkoxyamides reacted by an acid-catalysed process came from preliminary 1H NMR investigations in a homogeneous D2O/ CD3CN mixture, which indicated that N-acetoxy-N-butoxybenzamide 25c reacted slowly in aqueous acetonitrile by an autocatalytic process according to Scheme 4 (k is the unimolecular or pseudo unimolecular rate constant, K0 the dissociation constant of acetic acid and K the pre-equilibrium constant for protonation of 25c).38 Upon addition of a solution of sulfuric acid in D2O the reaction of N-acetoxyN-alkoxyamides obeys pseudo-unimolecular kinetics consistent with a rapid reversible protonation of the substrate followed by a slow decomposition to acetic acid and products according to Scheme 5. Here k is the unimolecular or pseudo unimolecular rate constant and K the pre-equilibrium constant for protonation of 25c. Since under these conditions water (D2O) was in a relatively small excess compared with dilute aqueous solutions, the rate expression could be represented by the following equation:

d½S d½AcOH ½S½H3 Oþ ¼ ¼ k½SHþ ¼ k:K ¼ k0 ½S dt dt ½H2 O

½H3 Oþ ¼ kH ½H3 Oþ . where the pseudo-unimolecular rate constant k0 ¼ ½Hk:K 2 O K' AcOH + H2O K

H3O+ + S SH+

k

AcO– + H3O +

SH+ + H2O

products + AcOH

Scheme 4.

H3O+ + S SH+

k

K

SH+ + H2O

products + AcOH

Scheme 5.

(1)

61

N-ACYLOXY-N-ALKOXYAMIDES

Pseudo unimolecular rate constants k0 for sulfuric acid-catalysed solvolysis of 25c in CD3CN/D2O (adjusted to a constant ratio of 3.8:1) were found to be linearly dependent upon the acid concentration (Fig. 11) and the gradient afforded a composite rate constant kH of (2.4170.10) 102 l mol1 s1 at 308 K. From the intercept, ko, the rate constant for uncatalysed solvolysis, was at least three orders smaller and zero within experimental error. A similar linear dependence and near-zero uncatalysed rate constant was demonstrated for other N-acetoxyN-alkoxybenzamides given in Table 3. Scheme 6 depicts three possible hydrolysis mechanisms.140 The ﬁrst (pathway (i)) is normal acid-catalysed ester hydrolysis in which attack of solvent (H2O) upon the protonated intermediate is rate determining. The second (pathway (ii)) is the 6 5

104k' (s-1)

4 3 2 1 0 0.010

0.020

[H3O+](mol.l-1)

Fig. 11 308 K.

Dependence of k0 on acid concentration in the solvolysis of 25c in CD3CN–D2O at

Table 3 Rate constants for acid independent and uncatalysed solvolysis of N-acetoxyN-alkoxybenzamides in CD3CN/D2Oa at 308 K Substrate (R, X)

1 1 102k308 s ) H (l mol

25a (R ¼ Et) 25c (R ¼ Bu) 25f (R ¼ Pri) 25g (R ¼ Bui) 26b (X ¼ MeO) 26d (X ¼ Me) 26e (X ¼ Cl) 26f (X ¼ Br) 26g (X ¼ NO2)

3.82 (0.22) 2.41 (0.10) 30.19 (1.63) 1.91 (0.15) 29.56 (3.04) 8.96 (0.62) 1.02 (0.13) 1.19 (0.07) 0.245 (0.02)

a

CD3CN/D2O 3.8:1.

105k308 (s1) o 1.76 2.31 5.40 0.02 1.47 13.40 4.24 0.43 2.40

(2.90) (1.30) (9.15) (1.96) (14.30) (8.50) (2.58) (1.50) (0.70)

r 0.996 0.995 0.996 0.994 0.984 0.999 0.975 0.997 0.990

62

S.A. GLOVER O

+

O PhCON

H3O+ fast

CH3

[0.48]

+

O H

PhCON

OBu

+

O

H2 O Path (i) AAc2

CH3 OBu

slow

H2 O

Path (ii)AAl2 slow

Path (iii)AAl1 slow [0.33]

+

PhCO N OBu

+ +

+

O

H OH

O PhCON

fast

H2 O

H

AcOH H2 O

CH3 OBu + H O 2

[0.69]

H

+

O H [0.69]

PhCON

+ +

AcOH H2 O

+

AcOH H3O+

OBu

fast

fast [1.00]

OH PhCON OBu

+

Scheme 6.

disfavoured AAl2 process which would involve displacement of acetic acid through nucleophilic attack by a water molecule at nitrogen. Pathway (iii), the AAl1 mechanism is typically found for the acid-catalysed hydrolysis of tertiary alkyl or benzyl esters, which generate stabilised carbenium ions.141 Rate equation (1) indicates that kH should be inversely proportional to the activity of water for solvolysis by the AAl1 mechanism and independent of it if the bimolecular processes (pathways (i) and (ii)) pertain. Fig. 12 illustrates that acid independent rate constants at different volume fractions of D2O in CD3CN, kH, were linearly dependent upon the inverse of aD2 O in CD3CN as determined from the corresponding activities of H2O in CH3CN.142 This is in accord with the AAl1 mechanism (pathway (iii), Scheme 6). Further evidence for the AAl1 mechanism was obtained from a solvent kinetic isotope study. The theoretical kinetic isotope effects for intermediates in the three reaction pathways as derived from fractionation factors are indicated in parentheses in Scheme 6.143,144 For the AAl1 mechanism (pathway (iii)) a solvent KIE (kH2 O =kD2 O ) between 0.48 and 0.33 is predicted while both bimolecular processes (pathways (i) and (ii)) would have greater values of between 0.48 and 0.69. Acidcatalysed hydrolysis of ethylene oxide derivatives and acetals, which follow an A1 mechanism, display KIEs in the region of 0.5 or less while normal acid-catalysed ester hydrolyses (AAc2 mechanism) have values between 0.6 and 0.7.145,146 From rates of the solvolysis of 25c at different sulfuric acid concentrations in D2O/CD3CN and H2O/CD3CN (Fig. 13), the observed solvent KIE was found to be 0.44 (70.02) conﬁrming, therefore, that the transition state for solvolysis lies along

63

N-ACYLOXY-N-ALKOXYAMIDES 0.225

KH(l mol s-1)

0.175

0.125

0.075

0.025 1.125

1.175

1.225

1.275 1/aH O

1.325

1.375

1.425

2

Fig. 12 Dependence of kH for 25c on activity of water in CD3CN–D2O at 308 K.

6 5

kH(H2O)/kH(D2O) =0.44 CD3CN/D2O

104k' (s-1)

4 3 2

CD3CN/H2O

1 0 0.004

0.014

0.024

[H3O+](mol l-1)

Fig. 13 Dependence of k0 on acid concentration for solvolysis of 25c in CD3CN–D2O and CD3CN–H2O at 308 K.

pathway (iii), between the protonated ester and nitrenium ion. Fractionation factors and solvent kinetic isotope effects are not expected to differ greatly in aqueousorganic mixtures.147 Thus N-acyloxy-N-alkoxyamides undergo acid-catalysed solvolysis forming N-acyloxy-N-alkoxynitrenium ions. However, the rate of uncatalysed reaction was negligible under the same conditions. Anomeric weakening of the N–O bond in the neutral species is insufﬁcient to promote heterolysis. However, protonation of the acyloxyl group, renders this substituent at nitrogen more electronegative (Fig. 14a)

64

S.A. GLOVER

C H

O

C O

O sp3 N R O

R'

C

H nO — σ*NO (a)

δ+ O

C O

R'

O δ+ R N O

O C N O sp2 HO

R

C O R'

(b)

(c)

Fig. 14 (a) Anomerically weakened NO bond in the protonated intermediate; (b) transition state; (c) products from AAl1 reaction of N-acyloxy-N-alkoxyamides.

Population of the s*NOAc orbital weakens the bond rendering it unstable, resulting in formation of a resonance-stabilised nitrenium ion (Fig. 14c). Arrhenius data and acid independent rate constants at 308 K, k308 H , for a wide range of N-acyloxy-N-alkoxybenzamides are given in Table 4. With the exception of the benzoyloxylated series, 28, the activation energies are in the region of those cited for the acid-catalysed hydrolysis of tertiary alkyl, diphenylmethyl and a-methylallyl esters namely ca. 29 kcal mol1 141 Entropies of activation are positive in keeping with a dissociative transition state but for the most part are much more positive than the ester values (ca. 9.5 cal K1 mol1), which accords with much looser and later transition states with substantial alkoxynitrenium ion character (Fig. 14b). AAc2 hydrolysis of esters have strongly negative DSz between 24 and 36 cal K1 mol1.148 The isopropoxy compound 25f reacts about an order of magnitude faster than 25a, 25c and 25g. A measurably larger DSz is consistent with additional relief of steric compression in the transition state; the protonated intermediate (Fig. 14a) would be sp3 hybridised at nitrogen while the alkoxynitrenium ion (Fig. 14c) would be sp2 hybridised at nitrogen. The isobutoxy compound 25g, in which the branching is one methylene removed from the oxygen atom has similar parameters to straight chain substrates 25a and 25c. Para-substituents on the benzamide side chain in 26, the benzyloxy side chain in 27 and the benzoyloxy side chain in 28 would be expected to affect the rate determining step differently. The inﬂuence of para-substituents on the benzamide and benzyloxyl side chains upon the pre-equilibrium protonation step is likely to be negligible considering their remoteness from the site of protonation and their electronic inﬂuence must rather impact upon the rate determining N–O bond heterolysis step. Parasubstituents on the leaving group should impact upon both the protonation and bond heterolysis steps. HF/6-31G* calculations on N-methoxy-N-benzoylnitrenium ion gave a geometry in which N2pz–O2pz are coplanar yielding a N–O bond order of 1.6 and a bond length of 1.191 A˚, which is typical of pure N–O double bonds (Fig. 15a).149,150 The LUMO (Fig. 15b) was localised on the C(O)NO atoms with some contribution into the aromatic ring and the positive charge resides on the nitrogen, the alkoxy substituent and the acyl group. Group charges for these were N ¼ +0.1, OCH3 ¼ +0.47

65

N-ACYLOXY-N-ALKOXYAMIDES

Table 4 Arrhenius and rate data for acid-catalysed solvolysis of N-acetoxy-Nalkoxybenzamides (25, 26, 2738,39 and 2840)a,b ln A

EA (kcal mol1)

DSz (cal K1mol1)

102 k308 H (l mol1 s1)c

r

25a (R ¼ Et) 25c (R ¼ Bu) 25f (R ¼ Pri) 25g (R ¼ Bui)

41.46 42.02 45.73 41.27

(0.43) (1.15) (2.02) (0.29)

27.37 27.92 28.64 27.66

(0.29) (0.69) (1.27) (0.19)

21.85 22.95 30.33 21.47

(0.24) (2.17) (1.34) (0.17)

3.72 2.60 33.50 1.92

0.996 0.998 0.995 0.994

26b (X ¼ MeO) 26c (X ¼ Ph) 26d (X ¼ Me) 26e (X ¼ Cl) 26f (X ¼ Br) 26g (X ¼ NO2) 26i (X ¼ But)

43.05 44.92 44.01 39.05 39.96 31.25 44.56

(1.29) (1.06) (0.13) (2.37) (0.63) (0.78) (1.23)

26.99 29.45 28.66 26.46 27.04 22.76 29.04

(0.81) (0.67) (0.17) (1.50) (0.41) (0.50) (0.76)

25.01 28.80 26.92 17.05 18.87 1.58 27.99

(2.56) (2.10) (0.53) (4.71) (1.24) (1.55) (2.44)

33.860 3.899 5.875 1.452 1.455 0.252 5.337

1.000 0.999 1.000 0.995 1.000 0.999 0.999

27a (Y ¼ H) 27b (Y ¼ MeO) 27c (Y ¼ PhO) 27d (Y ¼ Ph) 27e (Y ¼ Me) 27f (Y ¼ Cl) 27g (Y ¼ Br) 27h (Y ¼ NO2) 27i (Y ¼ But)

45.01 44.50 38.19 42.99 47.06 41.89 40.01 34.43 43.82

(1.32) (2.60) (2.80) (0.78) (0.71) (1.26) (0.45) (1.09) (0.94)

30.57 28.37 25.46 29.19 31.14 29.26 28.16 25.63 29.26

(0.84) (1.65) (1.72) (0.50) (0.45) (0.79) (0.29) (0.69) (0.60)

28.90 27.87 15.36 24.89 32.96 22.69 18.96 7.88 26.53

(2.63) (5.18) (5.56) (1.55) (1.41) (2.51) (0.88) (2.17) (1.86)

0.499 14.870 3.160 0.876 2.062 0.265 0.239 0.057 1.821

0.999 0.993 0.999 1.000 1.000 0.999 1.000 0.998 1.000

28a (Z ¼ H) 28b (Z ¼ MeO) 28d (Z ¼ Me) 28e (Z ¼ Cl) 28f (Z ¼ CHO) 28j (Z ¼ CF3) 28g (Z ¼ CN) 28h (Z ¼ NO2)

39.44 41.20 40.50 38.83 39.10 40.10 36.20 37.00

(0.58) (1.30) (1.20) (0.84) (1.30) (1.20) (1.30) (1.70)

17.84 21.30 20.04 16.67 17.05 19.15 11.32 12.99

(1.15) (2.65) (2.39) (1.67) (2.58) (2.39) (2.58) (3.37)

27.49 28.56 28.23 27.08 27.04 27.56 25.22 25.53

(0.38) (0.84) (0.72) (0.53) (0.84) (0.79) (0.79) (1.05)

0.411 0.395 0.369 0.424 0.578 0.689 0.630 0.863

0.999 0.997 0.998 0.999 0.997 0.998 0.997 0.995

a

CD3CN:D2O 3.8:1. Errors in parentheses. c From Arrhenius data at 308 K. b

and PhCO ¼ +0.42. Substituents at nitrogen that can stabilise this developing charge would be expected to accelerate the heterolysis step. Using Arrhenius derived rate constants at 308 K for substrates 25c, 26b–g and 26i (Table 4) an excellent ﬁt was obtained with Hammett s+ substituent constants yielding a r-value of 1.4 (r ¼ 0.995). While s+ correlations are expected where there is direct conjugative interaction between the aryl substituent and the positively charged centre,151,152 the ﬁt in this case is consistent with a signiﬁcant build up of localised positive charge on nitrogen in the rate determining step. The small negative Hammett r-value is appropriate for nitrenium ion formation beta to the aromatic ring and rate enhancement by electron-releasing substituents can best be ascribed to a diminution of positive charge at the amide carbonyl carbon in 45 thereby facilitating the development of positive charge at nitrogen. A similar correlation

66

S.A. GLOVER

Fig. 15 (a) HF/6-31G* optimised geometry and (b) LUMO of N-methoxy-Nbenzoylnitrenium ion.

(s+, r ¼ 0.74) has been reported for the acid-catalysed decomposition of o-diazoacetophenones in which carbenium ion character is also developed alpha to the carbonyl.153,154 Y δO δ+ OBu N δ+ X

O

CH3

O

O N

δ+ MeO

O

CH3

O δ+ H

δ+O H 45

OBu

46

Ph

N

δ+ O C H2

O

CH3

δ+O H 47

The Arrhenius parameters for this series of compounds show the usual trend expected for a unimolecular heterolysis in the rate-determining step. Electronreleasing groups (negative s+) have much more positive DSz indicative of a looser transition state together with a higher EA as a consequence of the greater degree of N–O bond stretching. With the exception of 26b all the substrates in this series exhibit an isokinetic relationship (Fig. 16, ln A ¼ 0.5EA16.5, r2 ¼ 0.998). The deviation (of 3 ln A units) from this relationship by the para-methoxy substrate would suggest a different mechanism of activation. Para-methoxylation in 26b results in a more organised transition state than those found with other activating substituents and the reaction is also driven by a slightly lower EA. The tighter transition state could in part be due to the planarity requirements for conjugative interaction with the carbonyl. However, by analogy with the decomposition of the o-diazoacetophenones, a resonance interaction in the transition state such as that in 46 would also account for these activation parameters.153,154 The products resulting from acid-catalysed solvolysis of N-acetoxy-N-butoxybenzamides 26 in acetonitrile–water mixtures, as illustrated in Scheme 7, are

67

N-ACYLOXY-N-ALKOXYAMIDES 48 46 27b

44

26b

42 R2 = 0.999

lnA

40

27c

38 R2 = 0.997

36

R2 = 0.974

34

26 27 28

32 30 5.00

10.00

15.00

20.00 25.00 E A(Kcalmol-1)

30.00

35.00

40.00

Fig. 16 Isokinetic relationships for acid-catalysed solvolysis of 26, 27 and 28 in CD3CN/D2O at 308 K.

acetic acid, butanol 50, benzoic acids 52, butanal 54, benzohydroxamic acids 55 and butyl benzoates 56. All of these products except acetic acid are derived from the N-hydroxy-N-butoxybenzamide intermediate 49 which is formed by capture of the nitrenium ion 48 by water (Scheme 7).38,39 Aldehyde 54 and the hydroxamic acids 55 were generated together in an acidcatalysed elimination reaction (Scheme 7 pathway (ii)). A crossover experiment indicated that esters are formed in a concerted rearrangement concomitant with the likely formation of the hydroxynitrene 57 (Scheme 7 pathway (iii)); while there is no evidence to date for the formation of hydroxynitrene, joint solvolysis of equimolar quantities of N-acetoxy-N-butoxy-p-chlorobenzamide 26e and N-acetoxyN-benzyloxybenzamide 27a afforded signiﬁcant quantities of butyl p-chlorobenzoate (36%) and benzyl benzoate (54%) as the only esters. This is an example of a HERON reaction, which has been identiﬁed in these laboratories as a characteristic rearrangement of bisheteroatom-substituted amides.32,33,42,43,155–158 Since ester formation was shown to prevail in neutral or low acid concentrations, it could involve the conjugate anion of the hydroxamic acid (vide infra).158 The source of alcohol 50 is most probably acid-catalysed hydrolysis of 49 to the nitrosocarbonylbenzene intermediates 51, which, like acid chlorides, react with water to give benzoic acids 52 (Scheme 1 pathway (i)).159 Acylnitroso intermediates 51 were trapped as the Diels–Alder adducts 53 in reactions in CH3CN/H2O and in 18 the presence of cyclopentadiene. In CH3CN/10% H18 O 2 O, 53 was enriched in

68

S.A. GLOVER O Ar

CH2Pr N O

O

H+

O Ar

N O + 48

– AcOH

O

O

H2O

CH2Pr

Ar

H+

CH3

CH2Pr

N O HO 49 O

O Ar H

CH2Pr N O O + H+

H2O

O

Path (i) HOCH2Pr

Ar OH 52

Ar N O

50

O

51

N 53

H

+

O

HH

Ar

CPr N HO

Path (ii)

O

Ph O N O HO

CH2Pr

O

OH Ar

PrCHO + Ar N

54

Ph

O

OH

55

NH OH

O

Path (iii)

+

Ar

N OH

OBu 56

57

Scheme 7.

providing unequivocal evidence for both the trapping of the nitrenium ion intermediate 48 by solvent water molecules and subsequent hemiacetal-like hydrolysis to the nitrosocarbonylbenzene 51.39 Studies at different acid strengths indicated that pathways (i) and (ii) (Scheme 7) were favoured over the non-acid-catalysed HERON reaction pathway (iii) at higher acid concentrations. Substituents on the benzyloxy side chain in 27a–i would be expected to affect ease of alkoxynitrenium ion formation through their inductive inﬂuence at the alkoxyl oxygen, which, because alkoxynitrenium ions are strongly delocalised, gains signiﬁcant positive charge in the transition state 47. An isokinetic plot using Arrhenius data for this series indicated that 27b (X ¼ MeO) and 27c (X ¼ PhO) deviated markedly from the rest of the series, which gave a good linear relationship (Fig. 16, ln A ¼ 0.54EA23.0, r2 ¼ 0.974).39 Substrates in this series were found to react by two mechanisms; those bearing +M substituents, 27b–d, reacted by an E1 process (Scheme 8, pathway (i)) while the remainder, like N-acetoxy-N-butoxybenzamides 26, reacted through the AAl1 process (Scheme 8, pathway (ii)). Evidence for this dichotomy was obtained from kinetic isotope effects and 18O incorporation studies 39 in CH3CN/10% H18 2 O. Whereas the normal AAl1 reaction (pathway (ii)) in CH3CN/10% H18 2 O would yield labeled Diels–Alder product 60, reaction of 27b under these conditions gave 18O labeled benzyl alcohol 59 and unlabelled 53. Benzyl cation 58 is lost in the

69

N-ACYLOXY-N-ALKOXYAMIDES

C5H6

O O CL2

Ph N O

X E1

O

CL2

Ph O CH3

N O O

–AcOH

O+ H

H218O X 18

L L 58

Path (ii) AAl1 Slow

–AcOH

O N

Ph

X

N O

Slow Path (i)

O+ H

CH3

‡

O

O H

Ph 53 X

CL2 59

Path (iii) O

O Ph N +

O

H218O

CL2 O

X

Ph

N O

Ph

CL2

18

18

O18

C5H6

N

X

N O H+ O H

O 60

CL2

Ph

HO

Scheme 8.

Ph

H+

N O O Z O

O

O

O Ph

K

Ph

Ph

Ph

Ph N O

k

N O O +O H

Z

O Z HO

Scheme 9.

rate-determining step rather than through a fast elimination from the benzyloxynitrenium ion (Scheme 8, pathways (ii) and (iii)). Acid-independent rate constants for 27b and 27d where L ¼ H and D gave signiﬁcant second order isotope 3 2 D effects, kH H/kH ¼ 1.32 and 1.18, respectively, indicative of sp to sp rehybridisation of the benzylic carbon in the transition state. In contrast, substrates with +I (27e, Scheme 8, X ¼ Me) and M (27h, Scheme 8, X ¼ NO2) afforded kH/kD ¼ 1.02 and 0.99. Reaction product studies were consistent with this change of mechanism. 27b gave benzyl alcohol, benzoic acid and benzoic anhydride and no products from solvolysis intermediate N-benzyloxy-N-hydroxybenzamide.39 For series 27, two Hammett correlations are therefore operative. Substrates 27b–d bearing +M substituents correlate with s+ with r ¼ 2.0, while the remainder (27a, 27e–i) correlate with s with r ¼ 1.6.39 No change in mechanism was found for the benzoyloxy series 28. Fig. 16 illustrates a uniform isokinetic relationship. Here substituents Z impact on both protonation and heterolysis steps (Scheme 9). Electron donating substituents increase the basicity of the benzoyl carbonyl, which promotes protonation and hence shifts the

70

S.A. GLOVER

equilibrium constant K, to the right. However, electron donating substituents, while increasing the concentration of the protonated intermediate, decrease the partial positive charge at the carbonyl carbon thereby reducing the ease of N–O bond heterolysis (and magnitude of k). These opposing effects manifest themselves in low overall sensitivity to the electronic effects of the para substituents. Rate constants at 308 K (Table 4) correlate with Hammett s substituent constants with positive slope but with a low gradient (r ¼ 0.32, r2 ¼ 0.87).40 A similar sensitivity has been reported for normal acid-catalysed solvolysis of alkyl benzoates where substituents also inﬂuence protonation and solvolysis steps in the opposite sense.160 Reversible protonation of benzoic acids has a reported r+ value of 1161 thus a similar sensitivity in the protonation of 28 would yield a r-value for the heterolysis step in the region of+1.5 and entirely consistent with build up of negative charge at a centre indirectly conjugated to the substituent.y

NUCLEOPHILIC SUBSTITUTION REACTIONS – SN2 REACTIVITY

N-Acyloxy-N-alkoxyamides have been found to undergo SN2 reactions with a number of organic and inorganic nucleophiles including anilines, thiols, hydroxide and azide anions. Reaction products from all of these processes are themselves reactive anomeric amides and outcomes have uncovered novel chemistry of this unusual class of compounds. In SN2 reactions, substituents at the central atom that can stabilise cationic character will also stabilise the SN2 transition state leading to products. As well as lowering the energy, such stabilisation moves SN2 saddle points in the direction of cationic species resulting in incipient cationic character in the transition structure and longer bonds to both the nucleophile and the leaving group.162–165 Work in these laboratories and elsewhere has established that nitrenium ions are strongly stabilised by neighboring heteroatoms (see background section and Fig. 15).38,91–95,97–100,166,167 This, together with the fact that N-acyloxyN-alkoxyamides possess a tetrahedral nitrogen bearing a good leaving group, should promote SN2 reactions leading to displacement of carboxylate. The anomeric substitution results in negative hyperconjugation and anchimeric assisted weakening of the N–OAc bond (Scheme 10). Reactions with aromatic amines – the HERON reaction N-Acyloxy-N-alkoxyamides are mutagenic towards TA100 and TA98 strains of Salmonella typhimurium without metabolic activation. They are direct-acting mutagens and react as electrophiles directly with nucleophilic centres in DNA.38–40,46–49,50 The principle DNA target of many electrophiles at physiological pH is known to be N7 of Guanine (G-N7), the most nucleophilic centre in the y logðk0 X =k0 H Þ ¼ logðK X =K H Þ þ logðkX =kH Þ ¼ 1:0sþ þ rs and a plot of ðlogðk0 X =k0 H Þ þ sþ Þ versus s yields a better correlation (r2 ¼ 0.94) and a r-value for the heterolysis step in the region of +1.5.

71

N-ACYLOXY-N-ALKOXYAMIDES Nu R

O C

O δ+ OR

N O δ–

R

C

N

–OAc

Nu

OR

O CH3

Scheme 10.

CH3

O N

O

NH Ph OR

CH3

61

N

AcOH

N Ph

RO

AcO

62

25

Ph

O

CH3 N N

OR 63

H3C

N

N

Ph 64

N

N

CH3

Ph 65

Scheme 11.

major groove.60,168–171 N-Acyloxy-N-alkoxyamides have been shown to damage DNA at this position.49,105 To evaluate their relative potential for reaction towards G-N7 and other nucleophiles, the rates of reaction of various N-acyloxyN-alkoxyamides with the aromatic base N-methylaniline, have been determined routinely.41–43 Initially, a reaction of N-acetoxy-N-butoxybenzamide 25c with N-methyl aniline 61 in d4-methanol or aqueous acetonitrile, afforded an excellent yield of butyl benzoate 63(R ¼ Bu) and acetic acid. Close examination of these highly coloured reaction mixtures indicated the presence of crystals of N,N0 -dimethyl-N,N0 diphenyltetrazene 65 (Scheme 11, R ¼ Bu). The reaction is promoted by polar solvents as reactants are unchanged in pure acetonitrile. A crossover experiment using a mixture of N-acetoxy-N-butoxy-p-toluamide 26d and N-acetoxy-N-ethoxybenzamide 25a afforded clean yields of butyl p-toluate and ethyl benzoate thus pointing to an intramolecular rearrangement.41 The process, which is depicted in Scheme 11 and which appears to be common to all classes of N-acyloxy-N-alkoxyamides, involves an SN2 reaction leading to displacement of the acyloxyl group and formation of the intermediate N-alkoxy-N(N0 -methylanilino)benzamide 62. These are strongly anomeric on account of a highenergy lone pair on nitrogen and the electronegative oxygen of the alkoxyl group. Anomeric weakening of the N–O bond in 62 results not in heterolysis to alkoxide

72

S.A. GLOVER O

NR2

O

N H

Ar O R (a)

1.49Å

c aN 1.77Å b 1.79Å

O

NH2 H

1.50Å c a

1.74Å b

O H

O H

(b)

(c)

N

NH2

1.81Å

Fig. 17 (a) Transition state for a HERON reaction of an N-alkoxy-N-aminoamide; computed transition states for migration of hydroxyl in N-amino-N-hydroxyformamide at (b) AM1 (a ¼ 651, b ¼ 451 and c ¼ 651) and (c) HF/6-31G* (a ¼ 631, b ¼ 481 and c ¼ 691) levels.

ion, which would be energetically unfavourable, but rather migration of the alkoxyl group from the amide nitrogen to the carbonyl carbon and heterolysis of the N–C bond. This results in the formation of 1-methyl-1-phenyldiazene 64, which under the reaction conditions dimerizes to the tetrazene 65. These rearrangement processes are characterised by a transition state in which the alkoxyl group migrates from the amide nitrogen to the carbonyl carbon (Fig. 17a) and therefore involves HEteroatom Rearrangements On Nitrogen the HERON reaction.z155,157,158 This reaction, though published in the early 1990s, was named in the literature in a 1995 paper in which we presented a semiempirical molecular orbital treatment of the HERON reactions of a range of anomeric amides.155 AM1 predicted, adequately, the pyramidal ground state properties of various anomeric amides, as well as the relative propensity for HERON migration of a variety of geminal heteroatom substituents on nitrogen. In general, the activation energy was lowered with an amino substituent and more electronegative migrating atoms or electron deﬁcient groups. In the case of N-hydroxy-N-aminoformamide, a transition state was predicted that indicated partial bonding between the migrating oxygen and both the amide nitrogen and carbon atoms in a plane perpendicular to that of N–CQO (Fig. 17b). The amide bond was largely unaltered. A similar result was obtained from HF/6-31G* calculations (Fig. 17c). Subsequently, a DFT evaluation of the HERON reaction of N-methoxyN-dimethylaminoformamide to give methyl formate and 1,1-dimethyldiazene conﬁrmed the earlier predictions at the AM1 and HF/6-31G* level on the simpler system (Fig. 18). The activation energy in the gas phase is modest at 21.4 kcal mol1 and the rearrangement is predicted to be exothermic by some 5.5 kcal mol1.33 The transition state geometry shows that the N–C bond is intact and the carbonyl bond is virtually unaltered relative to the starting structure. An intrinsic reaction coordinate study indicated that the N–C bond cleaves in concert with O–C bond formation after the transition state. In essence, this reaction represents an intramolecular SN2 reaction on the amide carbonyl. Further analysis of the computed transition state indicates the charge redistribution (Fig. 18). Predictably N2 becomes z First presented to the 2nd Heron Island Conference on Reactive Intermediates and Unusual Molecules, Heron Island, Australia, 1994.

73

N-ACYLOXY-N-ALKOXYAMIDES

1.51Å

1.22Å

Group charges in the ground and transition state.

N2 N1 1.27Å 2.00Å

— N2(CH3)2 — N1 — OCH3 — CHO

1.71Å

GS

TS

0.00 0.00 -0.02 +0.02

+ 0.51 -0.19 -0.34 -0.01

EA = 21.4 kcalmol-1

H=5.5 kcalmol -1 1.21Å

N1 N2

O

1.39Å

1.39Å O

Fig. 18 B3LYP/6-31G*-derived energetics, ground and transition state properties for the HERON reaction of N-methoxy-N-dimethylaminoformamide to 1,1-dimethyldiazene and methyl formate.

more positive at the transition state while N1 and the migrating group develop negative charge. Electron-releasing groups on the donor nitrogen and electron-withdrawing groups on the migrating oxygen should lower the energy of the transition state, and the rearrangement should also proceed better in polar solvents that can stabilise this charge separation. Interestingly, there is no appreciable change in charge at the carbonyl and neither electron-donor groups nor electron-withdrawing groups at this position should exert a major inﬂuence upon the energetics of the reaction. Experimental evidence in support of these computational results has accrued from a number of reactivity studies. The HERON rearrangement of the intermediate N-alkoxy-N-(N-methylanilino)benzamides proceeds with a low activation energy in methanol as it has never been detected under the reaction conditions even at temperatures well below ambient.42,43 This is not surprising as the B3LYP/6-31G* activation energy for the model reaction in the gas phase is an upper limit and it would probably be signiﬁcantly lower in solution. However, the related dimeric N,N0 -diacyl-N,N0 -dialkoxyhydrazines 66 are stable, but rearrange above room temperature to esters 67 and nitrogen in stepwise HERON rearrangements (Scheme 12).131,156 These hydrazines are unsymmetrical and exhibit all the characteristics of anomeric amides including pyramidality at nitrogen and anomeric effects where one nitrogen is a lone pair donor to the adjacent s*NOAc bond.35 Here the product from the ﬁrst HERON reaction, N-acyl-N-alkoxydiazene 68, ultimately undergoes a second HERON process with very low EA.36,172

74

S.A. GLOVER O

O OR

R' N

HERON

N

R' OR

R'

RO

OR

67

+ N2

R' O

O

66

HERON

N N

67

68

a R=Et, R'=4-C6H4X b R=CH2(4-C6H4X), R'=CH3

Scheme 12.

Using these hydrazines we have veriﬁed the electronic effects predicted from the theoretical studies.33 A series of 4-substituted N,N0 -dibenzoyl-N,N0 diethoxyhydrazines, 66a, correlated with Hammett s+ constants with r ¼ 0.35, (r2 ¼ 0.98). Here electron-releasing groups stabilise developing positive charge at the donor nitrogen. N,N0 -diacetyl-N,N0 -dibenzyloxyhydrazines, 66b, correlated with Hammett s constants with r ¼ +1.02 (r2 ¼ 0.91) in keeping with stabilisation of developing negative charge at the migrating oxygen.33 Kinetic studies. HERON reactions of N-acyloxy-N-alkoxyamides, which can be conveniently followed by NMR in d4-methanol by monitoring the disappearance of the mutagen or aniline and formation of ester and tetrazene, conform to classical bimolecular kinetics being ﬁrst order in both mutagen and N-methylaniline.41–43,46,105 Arrhenius parameters and bimolecular rate constants (308 K) for a range of N-acyloxy-N-alkoxyamides 25–30 are collated in Table 5. Typically, SN2 reactions of alkyl halides by anionic displacement involving charge delocalisation in the transition state exhibit entropy values in the range –4.5 to –9.5 cal K1 mol1. The entropies of activation for SN2 reactions at the amide nitrogen are however in the region of –24 cal K1 mol1 as would be expected for an associative transition state in which there is increased charge separation relative to the ground state (Fig. 19).174 Solvation will be more important and a contribution to the negative DSz must also be the more ordered arrangement of solvent molecules about the transition state.174 This effect is offset to a small extent by relief of steric compression. Anomerically substituted amides have pyramidal nitrogens as a consequence of dual substitution at nitrogen with electronegative atoms.5,6,32–36 Consequently there is a sharper angle at nitrogen in the ground state than in the transition state, where nitrogen approximates sp2 hybridisation. The activation energies are modest and reﬂect partial bond formation/breaking in a normal SN2 reaction. These reactions have an analogy in the SN2 reactions of a-haloketones such as phenacyl bromides.175 The rate-enhancing effect of a-carbonyl groups on SN2 processes at carbon is well known, and has been attributed to conjugation of the p-orbital on carbon in the SN2 transition state with the carbonyl p-bond,164,175–177 stabilisation of ionic character at the central carbon as outlined by Pross,164,178 as well as electrostatic attraction to the carbonyl carbon.176 Although there

75

N-ACYLOXY-N-ALKOXYAMIDES

Table 5 Arrhenius activation energies, entropies of activation and rate constants at 308 K for reaction of N-acetoxy-N-alkoxybenzamides 25, N-acetoxy-N-butoxybenzamides 26, N-benzoyloxy-N-benzyloxybenzamides 2842,43 and other N-acyloxy-N-alkoxyamides with N-methylaniline in d4-methanol Substrate (R, X, Y, Z)

ln A

EA (kcal mol1)

DSz (cal K1 mol1)

104 k308 2 (l mol1 s1)

r

25a (R ¼ Et) 25b (R ¼ Pr) 25c (R ¼ Bu) 25d (R ¼ Pent) 25e (R ¼ Oct) 25f (R ¼ Pri) 25g (R ¼ Bui) 25h (R ¼ Penti)

15.86 19.49 17.78 19.89 13.94 19.48 17.53 18.90

(1.7) (1.1) (0.5) (0.1) (1.4) (1.0) (1.1) (0.7)

11.90 13.61 12.62 13.84 10.55 14.75 13.00 13.30

(0.93) (1.00) (0.39) (0.16) (0.69) (0.87) (0.75) (0.62)

30.98 23.76 27.16 22.98 34.80 23.79 27.66 24.93

(3.3) (2.1) (1.0) (0.2) (2.9) (1.9) (2.1) (1.4)

276.5 628.4 581.5 643.3 363.0 96.7 239.3 581.5

0.994 0.992 0.999 1.000 0.991 0.993 0.993 0.995

26b (X ¼ OMe) 26c (X ¼ Ph) 26d (X ¼ Me) 26e (X ¼ Cl) 26f (X ¼ Br) 26g (X ¼ NO2) 26i (X ¼ But)

15.91 15.56 13.35 12.11 21.15 15.11 18.54

(1.1) (1.2) (0.1) (1.0) (0.9) (2.3) (0.9)

11.59 11.28 10.49 9.05 14.64 10.87 13.07

(0.67) (0.70) (0.14) (0.38) (1.04) (1.23) (0.71)

30.88 31.57 35.97 38.43 20.47 32.46 25.65

(2.2) (2.5) (0.2) (2.0) (1.7) (4.6) (1.7)

481.5 559.3 222.3 678.3 609.4 693.9 585.4

0.993 0.992 1.000 0.996 0.987 0.983 0.994

28a (Z ¼ H) 28b (Z ¼ OMe) 28d (Z ¼ Me) 28e (Z ¼ Cl) 28i (Z ¼ But)

14.60 14.30 5.91 19.61 21.13

(2.6) (1.8) (0.0) (1.3) (0.4)

11.16 11.64 6.37 13.48 15.39

(1.50) (1.05) (0.00) (0.79) (0.19)

33.48 34.08 50.73 23.53 20.52

(5.3) (3.6) (0.0) (2.6) (0.7)

260.8 87.7 109.8 881.2 177.0

0.991 0.996 1.000 0.998 1.000

27a (Y ¼ H) 29a (R2 ¼ Ph, R3 ¼ Et) 29f (R2 ¼ R3 ¼ Ph) 29l R2 ¼ But, R3 ¼ Ph) 30a (R1 ¼ Bu, R2 ¼ Me) 30b (R1 ¼ Bu, R2 ¼ Ph) 30c (R1 ¼ Bn, R2 ¼ Me) 30d (R1 ¼ Bn, R2 ¼ Ph)

14.73 (1.73) 17.90 (1.7)

11.53 (0.95) 15.56 (0.95)

33.18 (5.5) 26.97 (3.3)

155.6 5.32

0.993105 0.996173

18.6 (3.0)

12.20 (1.67)

25.06 (7.9)

2550.0

0.983105

13.50 (1.0)

11.54 (0.57)

35.56 (2.0)

45.8

0.993119

14.28 (1.48)

12.63 (0.95)

34.14 (5.0)

16.9

0.995105

11.89 (1.27)

10.17 (0.30)

38.91 (6.0)

87.5

0.986105

17.48 (1.96)

15.03 (1.67)

27.69 (6.9)

8.2

0.989105

19.98 (1.27)

15.92 (0.72)

23.39 (4.5)

23.4

0.998105

is no comparative rate data for reactions on amines or alkoxyamine, these arguments could also apply to substitution at the amide nitrogen in N-acyloxyN-alkoxyamides. These transition state properties have been modelled computationally at the HF/6-31G* level.45 The transition state geometry and properties for the reaction of N-formyloxy-N-methoxyformamide 40 with ammonia giving 69 (Scheme 13), is illustrated in Fig. 20. In the transition state the nucleophile and leaving group,

76

S.A. GLOVER Ph Me O N R

Ph NH

O

Me N H δ+

OR' OAc

N

O

OR'

N

N(Me)Ph

R

R

OR' δ– OAc

Fig. 19 Ground state and transition state geometries for the bimolecular reaction of N-methylaniline and N-acetoxy-N-alkoxyamides. δ+ NH3

NH3 O H

O CH3 N H O O 40

O H

O N O CH3 δ- O

H

NH3

N

O CH3

O

O H

O H

69

Scheme 13.

Fig. 20 HF/6-31G* optimised geometry and changes in group charges (relative to reactants) for the transition state for SN2 displacement of formate from N-formyloxy-N-methoxyformamide 40 by ammonia.

though displaced in the direction of the amide nitrogen lone pair, are approaching linearity and the amide nitrogen lone pair, the methoxyl group and the amide formyl substituent are approximately coplanar. Deviation from the classical linear transition state for SN2 displacement at carbon is not unexpected considering the complex interplay of lone pairs on N2, O3 and O5. The methoxyl group is oriented to maximise alignment of the p-type lone pair on O5 with the N2–O3 bond and

77

N-ACYLOXY-N-ALKOXYAMIDES

therefore the s*N–O orbital, suggesting it plays a negative hyperconjugative role even at the transition state. The reaction is clearly concerted, the acetoxyl group departing as N–N bond formation occurs. At the 6-31G* level of theory the amide carbonyl is close to the C4–N2–O5 plane and largely disconnected with the amide nitrogen lone pair which lies in the same plane. Analysis of the group electrostatic charges in the ground state of the reactants and the transition state supports the experimental solvation effect described above and which leads to more strongly negative DSz values. Fig. 20 gives the changes in total charge on the nucleophile (Dq ¼ +0.34) and the formyloxyl leaving group (0.67). The greater build-up of negative charge in the formyloxyl group compared to positive charge accrual on the nucleophile is indicative of a non-synchronous process and of partial nitrenium ion character in the transition state (Dq(N) ¼ 0.23). A detailed analysis of the charges on each fragment along the reaction coordinate showed that even at the N1–N2 bond distance of 2.5 A˚, where the charge on ammonia is only +0.04 and the geometry at nitrogen is virtually unchanged and tetrahedral, the formate has well developed anionic character (0.54). The bulk of the positive charge (+0.54) resides mainly on the amide nitrogen and the methoxyl group and represents partial alkoxynitrenium ion character. This is transferred in part to the ammonia nitrogen at the transition state and the bulk of the difference in charge is computed to reside on the amide nitrogen.45 B3LYP//HF/6-31G* energies for reactants, transition state and products for the reaction in Scheme 13 are presented in Table 6. In accordance with the increased charge separation in the transition state and products, the gas-phase EA of 22.6 kcal mol1 for the reaction reduces to just 4.4 kcal mol1 with inclusion of semiempirical solvation energies in water while the overall reaction, which is very endothermic in the gas phase becomes exothermic by 5.1 kcal mol1 with solvation.179 The calculated EA is lower than the experimental values for substitution by N-methylaniline in methanol which fall in the range of 6–15 kcal mol1 (Table 5).42,43 However, in aqueous solution these barriers would be lower than in methanol. Analogous calculations for the reaction of ammonia with N-formyloxy-N-methyl formamide predict a much higher activation barrier, which emphasises the role of the alkoxyl substituent in SN2 reactions on N-acyloxy-N-alkoxyamides.45 Table 6 B3LYP/6-31G*//HF/6-31G* energies of reactants, transition states and products for the reaction of N-formyloxy-N-methoxyformamide with ammonia (Scheme 13)a Structure NH3 CHON(OCHO)OCH3 40 [CHON(NH3)OCH3]+ 69 HCO–2 T.s. (Fig. 20) a

E (Hartree)

Eaq (Hartree)b

56.5472467 472.841903 340.008802 189.175640 529.353172

56.5579596 472.842752 340.116406 189.292434 529.393668

Lowest energy conformers; ZPE not included. Incorporates solvation energies calculated by the SM5.4 method of Cramer.179

b

78

S.A. GLOVER

Electronic effects. Nucleophilic attack is favoured by electron-withdrawing groups on the amide and the acyloxyl side chains. Interpolated bimolecular rate constants at 308 K for the series of para-substituted N-acetoxy-N-butoxybenzamides 25c, 26b–g and 26i (Table 5) gave a weak but positive Hammett correlation with s constants (r ¼ 0.13, r ¼ 0.86).42,43 These SN2 reactions are analogous to those of aniline and substituted pyridines with phenacyl bromides, which have similar Arrhenius activation energies and entropies of activation in methanol (EA ¼ 14–16 kcal mol1, DSz ¼ 27 to 31 cal K1 mol1) and 4-substituted phenacyl halides afforded a similar Hammett correlation with pyridine in methanol (s, r ¼ 0.25).175 The bimolecular rate constants for the limited series of N-benzoyloxyN-benzyloxybenzamides in Table 5 correlated strongly with Hammett s constants (r ¼ 1.7, r ¼ 0.97).42,43 Stabilisation of developing carboxylate character supports the computed charge redistribution in the transition state.45 Rate constants for the reaction of a series of substituted anilines 70–72 with N-acetoxy-N-butoxybenzamide 25c are presented in Table 7. NH 2 R

R

R'

R' 70

a R=Me, R'=H b R=Et, R'=H c R=H, R'=Me

NHR

NH 2

71

X 72 a X=MeO b X=Me c X=H d X=Cl e X=Br f X=NO 2

a R=Me b R=Ph

Table 7 Bimolecular rate constants for the reaction of anilines 70–72 with N-acetoxyN-butoxybenzamide 25c in d4-methanol42, 43 Aniline 71a N-Mea 70c 3,5-diMe 70a 2,6-diMe 70b 2,6-diEt 71b N-Phb 72f X ¼ NO2 72d X ¼ Cl 72c X ¼ H 72e X ¼ Br 72b X ¼ Me 72a X ¼ OMe a

104 k308 (l mol1 s1) 2

104 k298 (l mol1 s1) 2

104 k278 (l mol1 s1) 2

581.5 193.4 8.4 15.4 –

290.9

62.8 22.9 0.9 1.0 – 36.7 68.6 89.6 106.0 273.3 929.0

From Table 5. Included for comparison only.

b

–

1.6

79

N-ACYLOXY-N-ALKOXYAMIDES

Data for 72 correlated with Hammett s+ with r ¼ 0.91, r ¼ 0.95, which accords with the developing charge separation in the transition state. However, the Hammett reaction constant is smaller than that determined for the reaction of substituted anilines at alkyl and acyl carbon centres (r ¼ 2.0 to 3.0)180 and together with the strong sensitivity to substituents on the leaving group, reﬂects an early transition state in which positive charge is not transferred directly to the nucleophile. Rather, because of incipient stabilisation by the alkoxyl group, there is signiﬁcant nitrenium ion character. Steric effects. As expected, bulky groups on the mutagen as well as on the nucleophile have a signiﬁcant impact on ease of SN2 reaction at nitrogen. Rate constants at 308 K reﬂect the trends that would be expected for the relative steric inﬂuences of the alkoxy side chains in 73 (Table 5). Substrates with unbranched propyl, butyl and pentyl chains all exhibit similar rates at 308 K as would be expected. However the substrates 25f (R ¼ Pri) and 25g (R ¼ Bui) exhibit signiﬁcantly lower rates than 25a (R ¼ Et) and 25b (R ¼ Pr), respectively, in accordance with their higher degree of steric hindrance brought about by branching. 25h with isopentyloxy and 25c with butyloxy side chains react with similar rates indicating that branching remote from the reactive site has little inﬂuence. Analysis of relative activation parameters provides more insights into the variations in these rate constants. An isokinetic plot of EA versus ln A for straight chain and isopentyl substrates is linear and all substrates, where branching of the alkyl chain is at least d to the reactive centre, nitrogen, behave similarly (Fig. 21, ﬁlled squares).43,181,182 The isopentyloxy and butyloxy substrates 25h and 25c have very similar transition state 21 20

y = 1.8588x - 5.8462 R2 = 0.9914

19

25f

LnA

18 25g

17 16 15 14 13 10

11

12 13 E A(kcalmol-1)

14

15

Fig. 21 Isokinetic relationship for SN2 reaction of N-methylaniline with N-acetoxyN-alkoxybenzamides 73 in d4-methanol.

80

S.A. GLOVER

properties so branching d to the nitrogen has little effect. In the associative step, factors that facilitate overlap in the classical SN2 transition state complex will lower the energy of the transition state. However, more effective overlap must involve greater reorganisation resulting in a reduction in DSz, which becomes more negative. Structure 25f bearing isopropyloxyl and, to a lesser extent, 25g with isobutyloxyl deviated from the isokinetic relationship (Fig. 21, open squares) reﬂecting different transition state characteristics brought about by the steric crowding closer to the reaction site. It is likely that in these cases, to attain the same degree of reorganisation as needed in the straight-chain substrates, a higher activation energy is needed. Alternatively, straight chain substrates with the same activation energy as the isopropyl and isobutyl compounds would need an earlier transition state requiring less reorganisation. O O Ph

O R

AcO R Et, Pr, Bu, pentyl, octyl Pri, Bui, pentyli

Bu

N O

N O O O R

R=Me, Et, Pri, But, CH2But, 1-adamantyl

R=Me, Pr, (S)-2-butyl, Pri, But, CH2But, 1-adamantyl, Ph

R

Bu

N O

73

Ph

BzO

74

75

Steric effects on the amide side chain in 74 were marked. Out of a series of N-benzoyloxy-N-butoxyalkylamides 30b, 29a–e, only acetamide, 30b, and propanamide, 29a, afforded Arrhenius data in the normal temperature range 260–290 K. Rate constants at 308 K (Table 5) indicated that the propanamide (5.3 l mol1 s1) reacted an order of magnitude slower than the acetamide (87.4 l mol1 s1) and further branching at the a carbon through isopropyl substitution in 29b, tert-butyl in 29c and adamantyl in 29e completely inhibited SN2 attack by N-methylaniline at the amide nitrogen, even at higher temperatures.46,173 Removing branching to the b carbon in the neohexamide 29d returned reactivity, which was slow at the only temperature studied (2.2 l mol1 s1 at 313 K).46 SN2 reactivity with neutral nucleophiles like N-methylaniline parallels that of a-haloketones with amines, which are also strongly affected by steric effects on the a0 -carbon.183 In general, the rates of SN2 reactions are strongly adversely inﬂuenced by steric effects and branching b to the reactive centre and the same appears to be true for N-acyloxy-N-alkoxyamides. Rates of reaction of a-haloketones are also radically different with ionic nucleophiles relative to neutral amino nucleophiles, which react much more slowly. This has been attributed to a tighter transition state with ionic nucleophiles like acetate and azide leading to a classical SN2 transition in which there is additional stability through orbital overlap with the carbonyl carbon 2pz orbital.183,184 Azide has been found to react with N-acyloxy-N-alkoxyamides and N-chloro-N-alkoxyamides, even those with branching a to the carbonyl (vide infra).

81

N-ACYLOXY-N-ALKOXYAMIDES

Table 8 Bimolecular rate constants for the reaction of N-methylaniline with N-acyloxy-Nbutoxybenzamides 75 (25c, 29f–l) in d4-methanol at 308 K Substrate (R, R2) 25c (R ¼ H) 29g (R2 ¼ Pr) 29h (R2 ¼ Pri) 29i (R2 ¼ (S)-2-butyl) 29j (R2 ¼ CH2But) 29k (R2 ¼ Ad) 29l (R2 ¼ But) 29f (R2 ¼ Ph) a

104 k303 (l mol1 s1)

pKAa

22.9 122.0 91.0 97.0 78.0 61.0 33.8 1844.0

4.76 4.82 4.85 4.80 4.79b 4.86b 5.03 4.20

pKA of departing carboxylic acid. Calculated value.

b

Steric effects of branching on the acyloxyl side chain in 75 are insigniﬁcant. Table 8 gives relative rate constants for the reaction of N-methylaniline with a range of N-butoxy-N-alkanoyloxybenzamides 25c, 29a–l in d4-methanol at 303 K and the corresponding value for N-butoxy-N-benzoyloxybenzamide 29f together with pKAs of the leaving group carboxylic acid.119 Comparison of rate constants for straight chain and branched substrates clearly indicates that steric inﬂuences are minimal. Log of the reaction rate constant yields a reasonable linear correlation with pKA (gradient ¼ 4.8, r ¼ 0.969), which supports the theoretical transition state properties in which substantial charge transfer to the carboxyl group takes place. Steric effects on the nucleophile, aniline, were clearly evident. Rate constants for bimolecular attack of 2,6-dimethyl- 70a, 2,6-diethyl- 70b, and 3,5-dimethylaniline 70c at 308 K indicate that the ortho-substituted anilines react more than an order of magnitude slower at the same temperature (Table 7). Structure 70c must be able approach the reactive nitrogen more closely.42,43 A comparison of the rate constants for reaction of aniline 72c, N-methyl- 71a and N-phenylaniline 71b provides further evidence of steric effects although the very small rate constant for the diphenylamine could also be accounted for by reduced nucleophilicity on account of lone pair resonance into the additional phenyl ring. Bimolecular reactions of aniline with N-acyloxy-N-alkoxyamides are model SN2 processes in which reactivity is dictated by a transition state that resembles normal SN2 processes at carbon. Electronic inﬂuences of substituents support a nonsynchronous process which has strong charge separation at the transition state and which is subject to steric effects around the reactive centre, at the nucleophile but not on the leaving group. The sp3 character of nitrogen and disconnection between the amino group and the amide carbonyl renders these reactions analogous to the displacement of halides in a-haloketones. The mutagenic activity of N-acyloxy-N-alkoxyamides is believed to involve SN2 reactions of guanine N7 at the amide nitrogen with displacement of carboxylate.49 Remarkably, their biological activity is similarly affected by substituents at the amide nitrogen (vide infra).

82

S.A. GLOVER

Reactions with hydroxide ions Treatment of N-acyloxy-N-alkoxybenzamides 76a with dilute aqueous sodium hydroxide, at room temperature, resulted in the rapid formation of alkyl benzoates 67.40 A crossover experiment using N-acetoxy-N-butoxy-p-chlorobenzamide 26e and N-acetoxy-N-benzyloxybenzamide 27a resulted in the exclusive formation of butyl p-chlorobenzoate (46%) and benzyl benzoate (43%) esters along with the hydrolysis products, p-chlorobenzoic acid and benzoic acid indicating that ester formation involves an intramolecular process. Using a series of N-(p-substituted benzoyloxy) mutagens 28a, 28d–h, it was shown to occur by SN2 reaction of hydroxide at nitrogen, which gives as intermediate the hydroxamic acid 77 (Scheme 14).40 The presence of excess base ensured conversion of the hydroxamic acid intermediate into the conjugate anion 78 resulting in a HERON reaction leading to formation of ester 67 and presumably NO. The formation of non-crossover esters in the acid-catalysed solvolyses of N-acetoxyN-alkoxyamides at low acid concentrations has also been observed (Scheme 7, pathway (iii)) and presumably involves the conjugate anion of the hydroxamic acid 49.39 While the lone pair on the hydroxyl oxygen of hydroxamic acid 77 would be tightly bound resulting in a weak nOH–s*N–OR interaction, the anion 78 would possess a high energy pair of electrons that would enhance the anomeric effect and drive the rearrangement. HF/6-31G* optimised structures and B3LYP/6-31G*//HF/6-31G* energies of model ground state 78b, transition state 79b, and rearranged product 80b of the HERON step are illustrated in Fig. 22. The transition state 79b for rearrangement of N-methoxy-N-methanohydroxamate anion is computed to be early; the N1–O2 bond is appreciably longer at the transition state, the O2–N1–C1 angle is large and the carbonyl is essentially intact (Fig. 22b). Unlike the HERON reaction of

O

R'

O O OH R' HO– R' S 2 N OR N N OH N O O –R''CO2– RO RO R'' O 76 77 78

O R'

R'

67 a R=Bn,R'=Ph, R''=Ar b R=Me, R=R''= H

N O δ– 79

NO OR

Scheme 14.

O

R δ–O

O

OR

R'

NO

80

83

N-ACYLOXY-N-ALKOXYAMIDES

Fig. 22 HF/6-31G* optimised geometries and B3LYP/6-31G*//HF/6-31G* energies (Hartree) of (a) ground state 78b, (b) transition state 79b, and (c) HERON rearranged product 80b. Table 9 Bimolecular rate constants for the reaction of dilute hydroxide with 28a, 28d–h at 275.4 K Substrate (Z) 28d (Z ¼ Me) 28a (Z ¼ H) 28e (Z ¼ Cl) 28f (Z ¼ CHO) 28g (Z ¼ CN) 28h (Z ¼ NO2)

(l mol1 s1) k275 2 2.27 3.16 3.82 4.91 6.97 8.14

(0.39) (0.65) (0.69) (0.59) (0.09) (1.26)

N-alkoxy-N-aminoamides, where the amide bond breaks in concert with the migration, the transition state in this case leads to the tetrahedral alkoxide intermediate 80b (Fig. 22c), which decomposes to methyl formate by elimination of NO. B3LYP/6-31G*//HF/6-31G* gave an activation energy for the HERON step of only 5.3 kcal mol1 (10.0 kcal mol1 including solvation effects). Formation of the tetrahedral intermediate was exothermic by 10.7 kcal mol1 and the overall reaction producing methyl formate and NO is endothermic by 21 kcal mol1 with solvation, which suggests reversibility of the HERON step in this case.158 Rate constants for the reaction of base with N-aroyloxy-N-benzyloxybenzamides 76a (using 28a, 28d–h) could only be obtained at low temperature (275.4 K) and low base concentration and are presented in Table 9. Rates correlated with Hammett s constants with positive slope (r ¼ +0.55, r ¼ 0.992) in keeping with the BAl2 rather than the BAc2 mechanism of base hydrolysis of benzoate esters, which correlate with Hammett s but with a much larger r value in the range 2.0–2.4. The observed sensitivity to substituents is in keeping with stabilisation by electron-withdrawing groups of partial carboxylate character in the transition state. Reaction with azide ions N-Acyloxy-N-alkoxyamides also react rapidly with azide to give quantitative yields of non-crossover esters and two equivalents of nitrogen. As with hydroxide, the

84

S.A. GLOVER N R'

O N

N

N OR O R'' O

O

δ– N

SN2

N R'

N

N

N N

OR

O δ–

N N OR

O

R''

R'

O

R'' O

O

76

81

83

82 -N2 O

N2 +

OR

HERON

R'

N N

OR

R' O

67

68

a R=alkyl, R'=alkyl,aryl, R''=CH3 b R=Me, R'=R''=H

Scheme 15.

reaction rates are fast, but could be monitored dilatometrically in aqueous acetonitrile at 294 K. At this temperature, a reaction between N-acetoxy-N-benzyloxybenzamide 27a showed ﬁrst order dependence upon concentrations of both mutagen and azide and yielded a bimolecular rate constant of ca. 1.9 l mol1 s1.44 The product formation is outlined in Scheme 15. The intermediate N-alkoxy-N-azidoamide 82a, which is also formed rapidly from N-alkoxy-N-chloroamides, reacts by loss of nitrogen to give 1-acyl-1-alkoxydiazene 68a. The same intermediate was formed in the stepwise decomposition of N,N0 -diacyl-N,N0 -dialkoxyhydrazines (Scheme 12)131,156 and its HERON decomposition to ester 67a and nitrogen, a very low EA process, has been veriﬁed computationally.36,172 In our detailed theoretical study of reaction pathways of the model N-azidoN-methoxyformamide 82b we showed that decomposition by loss of nitrogen was the energetically most favourable process with an EA of only 5.3 kcal mol1 at B3LYP/6-31G*.36 In addition this step is exothermic by 42–44 kcal mol1. Thermal decomposition of 68b to methyl formate 67b and nitrogen has an EA of only 2.9 kcal mol1 and is exothermic by 95 kcal mol1. Overall, the conversion of 82b to methyl formate 67b and two molecules of nitrogen is thus predicted to be exothermic by 137–139 kcal mol1. The nucleophilic attack of azide on 76b was also modeled at the pBP/DN*//HF/ 6-31G* level (Table 10, Fig. 23).44 In the gas phase the transition state is preceded by an ion–molecule complex (Fig. 23a) which is at a minimum on the reaction coordinate and in which the azide anion is 3.7 A˚ from the amide nitrogen, and approaching the axis of the formyloxy–nitrogen (N1–O9) bond but in which the azide possesses all the anionic charge. The reaction is computed to be exothermic in the gas phase (DE ¼ 13.3 kcal mol1) and more so with solvation (DE ¼ 17.0 kcal mol1) but

85

N-ACYLOXY-N-ALKOXYAMIDES

Table 10 pBP/DN*//HF/6-31G* energies of reactants, ground state complex, transition state and products for reaction of azide with N-formyloxy-N-methoxyformamide 76ba Structure

E (Hartree)

Eaq (Hartree)b

HCON(OCHO)OMe 76b N 3 [76b–N3] complex HCON(N3)OMe 82b HCO 2 83b T.s. 81b

473.05455 164.31803 637.39671 448.09735 189.29648 637.40085

473.06748 164.43783 637.50844 448.11401 189.41833 637.49775

a

ZPE not included. Energies include AM1 solvation energies.

b

N4

N2

N3

N3

N4 N2

O8

C7 C6

O11 N1 O9 C10

(a)

O5

N1–N2=3.722Å N1–O9=1.383Å O9–N1–N2=166.5° N1–N2–N3=135.6° C7-N1–O9=109.9° O9-N1–O5=107.6° C7–N1–O5=111.5° O8–C7–N1–N2=162.9°

O5 C6 C7

N1

O8 O9 C10

O11

N1–N2=2.403Å N1–O9=2.053Å N2–N3=1.170Å N3–N4=1.140Å N1–C6=1.479Å N1–O5=1.251Å C7–N1–N2=74.3° C7–N1–O9=84.9° O5–N1–N2=92.9° O5–N1–O9=98.9° N2–N1–O9=158.6° C7–N1–O5=109.1° O8–C7–N1–O5=175.4° O9–N1–O5–C6=97.4° O8–C7–N1–O5=174.9°

(b)

Fig. 23 HF/6-31G* geometries for (a) the ground state complex and (b) the transition state in the reaction of azide anion with N-formyloxy-N-methoxyformamide 76b.

the transition state, while at a saddle-point, is lower in energy than the ion–molecule complex (EA ¼ 2.6 kcal mol1) in accord with the expected stabilisation due to charge delocalisation; the charge on azide (1) is shared by both the azide (0.76) and the formate group (0.62) in the transition state. Decreased solvent stabilisation of the transition state relative to the intermediate complex is important and inclusion of semiempirical solvation energies resulted in a small positive EA of 6.7 kcal mol1. A similar EA was determined for the reaction of ammonia but there solvation, which exerts greater stabilisation in the transition state than on the ground state, reduced the gas phase EA from 22.6 to 4.4 kcal mol1. The computed energetics suggests a facile SN2 reaction in agreement with the experimental ﬁndings. Attack of hydroxide, though not modelled computationally, has similar, large rate constants in aqueous organic medium. Rate constants for the reaction of N-acyloxy-N-alkoxyamides with aromatic amines or glutathione are slower by two to three orders of magnitude.42,43,105,185 In the computed transition state (Fig. 23b) the azide and formyloxyl groups are approaching colinearity (N2–N1–O9 ¼ 158.61) and the SN2 displacement is

86

S.A. GLOVER

predicted to be concerted but non-synchronous. The increase in negative charge of the formate group [Dqformate ¼ 0.65] is larger than the decrease in negative charge over the azide group [Dqazide ¼ +0.24], which is indicative of more N1–O9 bond stretching than N2–N1 bond formation. In this respect the transition state is similar to that reported for the corresponding reactions of ammonia with the same mutagen model.45 However, the transition states for those reactions, involving neutral nucleophiles, exhibited signiﬁcant charge separation as opposed to the charge delocalisation in the azide reaction transition state. Near coplanarity of O8–C7–N1–O5–C6 indicates that the lone pair on N1 is completely out of conjugation with the carbonyl at the transition state and both the C7 and O5 p-orbitals are aligned with the partial bonds between N1 and the incoming azide and departing formate. The EA for azide reaction on N-formyloxy-N-methylformamide was 22.2 kcal mol1 higher than reaction with the N-methoxy model. A similar result was obtained to the reactions of ammonia and reinforces the activating inﬂuence of N-alkoxy groups upon SN2 reactions in anomeric N-acyloxy-N-alkoxyamides. Finally, the corresponding reactions of azide with N-chloro-N-alkoxyamides has been used as an excellent source of highly hindered esters.44 The rearrangement of 68–67 (Scheme 15), apart from being highly favourable energetically, is characterised by an early transition state with little disruption to the carbonyl. In these HERON reactions, the N–C(O) bond breaks in concert with the MeO–C bond formation after the transition state along the reaction coordinate and the process is akin to an SN2 reaction on an acyl carbon. Accordingly ester formation is not restricted by the constraints in normal Fischer esteriﬁcation where formation of the tetrahedral intermediate is destabilised by bulky alcohol and acyl groups.186–188 Reaction with thiols185 Glutathione is a biological reducing and conjugating agent with a manifold of physiological functions.189 It is present in cells in concentrations up to 5 mmol and reacts with a range of xenobiotics, in particular molecular electrophiles or metabolites. Its role in this regard is to render such molecules soluble and excretable.190,191 N-Acyloxy-N-alkoxyamides are molecular electrophiles and react with DNA inducing mutagenic changes. Preliminary work has indicated that they interact with glutathione in cells and an NMR study of the behaviour of N-acetoxy-N-butoxybenzamide 25c and glutathione 84 in d6-DMSO/D2O indicated rapid consumption of both reactants. The products were the oxidised form of glutathione 86, the hydroxamic ester butoxybenzamide 85, and acetic acid (Scheme 16). By monitoring NMR integrals the reaction was shown to be bimolecular (k303 ca. 2 102 l mol1 s1) but 2 thiol was consumed at twice the rate of 25c. The reaction that best accounts for the products invokes initial attack upon N-acetoxy-N-butoxybenzamide by glutathione giving N-butoxy-N-(S-glutathiyl)benzamide intermediate 88a (Scheme 17). Rapid, non-rate-determining reaction of 88a with a second molecule of glutathione yields the disulﬁde 89a and hydroxamic ester 85. N-Acetylcysteine was unreactive but L-cysteine

87

N-ACYLOXY-N-ALKOXYAMIDES O

O O

N O

N H

CH3 D2 O/d6-DMSO

O

O

85

25c

AcOH HOOC

H2N O HOOC

H N

H2N

O

O

HN

N H

O

COOH

N H

COOH

S 2

HS 86

84

Scheme 16.

H S R R

87

O

OBu N OAc

O –AcOH

N S

OBu

R 25c

: SH

S

R

O

R

88

S 89

Ph

N H 85

OBu

a RS=glutathiyl b RS=(L)-cysteinyl methyl ester c RS=(L)-cysteinyl ethyl ester

Scheme 17.

methyl and ethyl esters 87b and 87c reacted in a similar fashion to glutathione. In particular, the L-cysteine ethyl ester reaction with several N-benzoyloxyN-benzyloxybenzamides 28 could be monitored nicely by 1H NMR in d4-methanol under pseudo unimolecular conditions and over a range of temperatures. Arrhenius studies afforded data presented in Table 11. Activation energies and entropies of activation are in the same region as those for SN2 reactions of neutral N-methylaniline. Large negative DSz values are again in accord with increased charge separation with attendant solvation in the transition state. The rate constants at 298 K correlate with Hammett s values with r ¼ 1.13 (r ¼ 0.985) in keeping with a development of benzoate character in the transition state. The reactions of glutathione and cysteine esters have been modeled by the reaction of methanethiol with N-formyloxy-N-methoxyformamide 40 at the BP/ DN*//6-31G* level.45 The transition state (Fig. 24) possesses similar characteristics

88

S.A. GLOVER

Table 11 Arrhenius activation energies, entropies of activation and rate constants at 298 K for reaction of N-benzoyloxy-N-benzyloxybenzamides 28 with L-cysteine ethyl ester in d4methanol EA (kcal mol1) DSz (cal K1 mol1) 104 k298 (l mol1 s1) 2

ln A

Substrate (Z)

28a (Z ¼ H) 9.46 28b (Z ¼ OMe) 21.89 28d (Z ¼ Me) 10.43 28e (Z ¼ Cl) 14.36 28h (Z ¼ NO2) 15.30

(1.17) (3.84) (1.89) (2.49) (3.02)

8.71 16.36 9.22 10.80 10.79

(0.7) (2.2) (1.1) (1.3) (1.6)

S1

43.66 18.99 41.73 33.94 32.07

51.3 31.3 57.0 200.0 523.2

0.9862 0.9819 0.9738 0.9716 0.9786

S1-N2 = 2.062Å N2-O3 = 2.408Å N2-C4 = 1.501Å N2-O5 = 1.296Å S1-N2-O3 = 147.3° C4-N2-O5 = 109.0° C4-N2-O3 = 69.5° C4-N2-S1 = 100.3° O5-N2-O3 = 103.9° O5-N2-S1 = 108.7° N2-S1-C10 = 95.3°

C10

C6

N2

O5

(2.3) (7.6) (3.8) (4.9) (6.0)

r

O3-N2-O5-C6 = -80.9° O7-C4-N2-O5 = -14.4°

O7 C4 O9 O3

C8

Δq (OCHO) = -0.72 Δq (CH3SH) = +0.40 Δq (CHO) = +0.06 Δq (N) = +0.25 Δq (OCH3) = +0.02

Fig. 24 HF/6-31G* optimised geometry and changes in group charges (relative to reactants) for the transition state for SN2 displacement of formate from N-formyloxy-N-methoxyformamide 40 by methanethiol.

to that for the reaction of ammonia. The nucleophile and leaving groups are approaching colinearity (147.31) and the methoxyl group is nearly orthogonal to the departing NO(CHO) bond (80.91) which is marginally longer (2.408 A˚) than where ammonia (2.316 A˚) was concerned. Nitrenium ion character is likewise developed in an asynchronous transition state. Charge separation (Dq (OCHO) ¼ 0.72; Dq (CH3SH) ¼ +0.4)) is greater for the thiol reaction which can be accounted for by the lower electronegativity of sulfur although the rates of L-cysteine ethyl ester correlate with Hammett s with slightly lower sensitivity (r ¼ 1.1) than the reaction with anilines (r ¼ 1.7). The model reaction has a low EA of 4.2 kcal mol1 and is exothermic by 5.9 kcal mol1 (Table 12). Clearly the experimental energies are higher than that predicted for the model reaction but the bulkier cysteine residues can be expected to inﬂuence the ease of reaction relative to methyl in methanethiol.

89

N-ACYLOXY-N-ALKOXYAMIDES

Table 12 pBP/DN*//HF/6-31G* energies of reactants, transition states and products for the reaction of N-formyloxy-N-methoxyformamide 40 with methanethiola Structure

E (Hartree)

Eaq (Hartree)

CH3SH CHON(OCHO)OCH3 40 [CHON(SHCH3)OCH3]+ HCO 2 T.s. (Fig. 24)

438.77211 473.05454 722.34182 189.29648 911.79233

438.77458 473.05540 722.42609 189.41328 911.82333

a

Lowest energy conformers; ZPE not included.

O X

N O 90

OR

O

R'OH X

Me

N

OR

+ AcOH

OR'

O a X=NR''2 b X=OR''

91

Scheme 18.

Alcoholysis reactions Kostyanovsky and coworkers have demonstrated that prolonged reaction of N-acyloxy-N-alkoxyureas 90a in neat alcohols leads to displacement of acetate and formation of symmetrical and mixed N,N-dialkoxyureas 91a (Scheme 18).52 The reactions work with primary alcohols (MeOH, EtOH and PrOH) but 2-propanol required longer reaction times. Tert-butanol was unreactive, presumably for steric reasons. The corresponding carbamates 90b reacted similarly giving 91b, but in certain cases the major product was, instead, the dialkylcarbonate. Methanolysis of ethyl N-acetoxy-N-methoxycarbamate 92 afforded ethyl methylcarbonate 96 and the dimethoxy product 95 in 44% and 34% yields, respectively, which was accounted for by SN1 reactions to give the alkoxynitrenium ions 93 that undergo loss of alkoxynitrene to form the alkoxycarbonyl acyl cation 94 (Scheme 19, pathway (i)). It is possible and energetically more feasible that this reaction proceeds rather by a HERON rearrangement (Scheme 19, pathway (ii)). While this reaction is likely to be slow, the alcoholysis reactions were carried out over long reaction times. These authors also reported that N-acyloxy-N-alkoxyamides did not undergo methanolysis under the same conditions but treatment of N-acetoxy-N-ethoxybenzamide 25a in more forcing conditions with NaOMe in DME afforded a mixture of ethyl and methylbenzoate, 97 and 98 (Scheme 20). They attributed the formation of the former to a HERON reaction and methyl benzoate to the direct attack of methoxide at the amide carbonyl. It is seems plausible that the HERON reaction to give ethyl benzoate 97 involves the anion of N-ethoxybenzohydroxamic acid 99, formed by methoxide reaction at

90

S.A. GLOVER O

O

MeOH

OMe

EtO

N OAc

EtO

-AcOH Path (i)

92

N

OMe

O

-MeON EtO

93

94 MeOH

Path (ii) O N OMe

EtO

O

HERON

OMe

-MeON

OMe

EtO

95

96

Scheme 19.

MeO –/DME

O Ph

N

OEt

O

O Ph

OEt

Ph

OMe

OAc 25a

97 O

-MeOAc X

98

HERON N

OEt

O 99

Scheme 20.

the ester carbonyl, by analogy with the reaction of N-acyloxy-N-alkoxyamides and hydroxide discussed above.

THERMAL DECOMPOSITION REACTIONS

N-Acyloxy-N-alkoxyamides are thermally unstable at modest temperatures. The pyramidality at nitrogen and anomeric weakening of N–O bonds results in decomposition by several pathways including both heterolytic and homolytic processes. Homolytic cleavage of weakened N–OAcyl bonds results in formation of a longlived alkoxyamidyl radical and a reactive acyloxyl radical. Preliminary studies have indicated that the alkoxyamidyls, so produced can intercept the products of acyloxyl radical reactions. However, the N-acyloxy-N-alkoxyamides are also suitably disposed to HERON rearrangements in which the acyloxyl substituent migrates to give anhydrides. Recently, examples of both types of processes have been demonstrated.192

91

N-ACYLOXY-N-ALKOXYAMIDES

Free radical decomposition A number of N-acyloxy-N-alkoxybenzamides 100 have been shown to decompose in mesitylene at temperatures above 1001C giving a complex mixture of products including the N-alkyl adduct 106, the corresponding alkyl ester 108, anhydride 101 and a product (up to 60% yield in cases), which was characterised as 1,4,2-dioxazoline 104 (Scheme 21). Minor products included hydroxamic ester 105 and mesitylene adduct 109. Formation of both adduct 106, mesitylene adduct 109 and heterocycle 104 can be ascribed to homolysis (Scheme 21, pathway (i)). Decarboxylation of acyloxyl radical gives radicals, R0 d, which either combine with the persistent free radical 102 giving 106 or, by abstracting hydrogen from solvent, give 3,5-dimethylbenzyl radicals, which combine with 102 to give 109. Formation of the dioxazoline 104 would appear to require hydrogen abstraction from the oxymethylene position of radical 102 to give the diradical 103, which undergoes cyclisation. B3LYP/6-31G* calculations predict the all cis singlet state of 103e is more stable than the lowest triplet state by some 26 kcal mol1 and, as indicated in Scheme 21, is strongly dipolar (Table 13). Intramolecular hydrogen transfer to the carbonyl oxygen followed by electron transfer is less likely due to the absence of an obvious oxidant. In addition, hydrogen transfer from carbon to oxygen in the methoxyformamidyl 102e is computed to be endothermic by 13 kcal mol1 using B3LYP/6-31G*. These unusual products as well as adducts 109 and 106 may be attributed to the persistence of alkoxyamidyls; they are isoelectronic with nitroxyl radicals and have signiﬁcant delocalisation onto oxygen and hence, thermodynamic stability. The ESR

O R'' –NOCH2R

N O 100 –NOCOR'

O 101

O

OCH2R R' O

O

R''

N

O

O

R''

103

N

O

104

R' Δ

Mesitylene Path (i)

O R''

H H

O

R

N O

OCH2R NH

O R''

R''

102

– N2

105

O

O

OCH2R

N R''

108

N

R''

R

R

O

OCH2R N R' 106

HERON Path (iii)

O R''

R

H

O

HERON Path (ii)

O R''

R'

107

OCH2R

OCH2R N

R'' 2

109

Scheme 21.

a R =4-ClPh; R'=Me, R''=Ph b R= 4-ClPh, R'=Me, R''=But c R= Pr, R'=Hexyl, R''=Ph d R= Pr, R'=Hexyl, R''= But e R=R'=R''=H

92

S.A. GLOVER

Table 13 B3LYP/6-31G* energies of optimized geometries for reaction of methyl radical with methoxyamidyl S0 102e Structure 103e S0 (cis– cis) 103e T1 (trans– trans) CH4 102e (trans– trans) CH3d T.s. (Fig. 25b) DEb EAc

Energy (au) 283.076675 283.034466 40.518428 283.713092 39.838234 323.528342 0.0437771 0.0229837

Energy (kcal mol1) 0.0 26.4a

27.4 14.4

a

Relative to So. E (CH4+103e S0) – E (CH3d+102e). c E (T.s.) – E (CH3d+102e). b

hyperﬁne coupling constant to nitrogen in amidyls (AN) is typically 1.48 mT while that of N-alkoxyamidyls is 1.04 mT, indicating a signiﬁcant delocalisation away from the nitrogen atom.193,194 Unlike electrophilic amidyls, which add to alkenes and arenes, alkoxyamidyls are more nucleophilic and persistent.194 Their lifetimes in solution appear to enable capture of less stable free radicals formed in their presence either in solvent cage reactions or after escape from the solvent cage. In the case of formation of diradicals 103 leading to the dioxazoline 104, the nature of the abstracting species is as yet unclear although alkyl radical is more likely than acyloxyl radical, which would decarboxylate rapidly at these temperatures. N-Alkyl adduct 106 is most probably formed in a cage reaction. Formation of ester 108 could involve dimerisation of alkoxyamidyls 102 and thermal rearrangement of the hydrazines 107 to esters and nitrogen according to Scheme 12, although a HERON reaction (Scheme 21, pathway (iii)) cannot be discounted under these conditions (vide infra). Anhydrides 101 are almost certainly formed by a HERON reaction (Scheme 21, pathway (ii)). Yields of products are extremely variable and highly dependent upon reaction conditions (temperature, concentration, atmosphere) and the nature of the side chains. For instance, N-acetoxy-4-chlorobenzyloxybenzamide 100a afforded mostly ester 108a and little cyclic product whereas the corresponding pivalamide 100b gave dioxazoline 104b, ester 108b and anhydride 101b in proportions of 6:1:9.6. Similarly, N-butoxy-N-heptanoyloxybenzamide 100c gave dioxazoline 104c, ester 108c and anhydride 101c in proportions of 2.2:1:5.1 but the corresponding pivalamide 100d afforded mostly anhydride 101d. The hydrogen abstraction from alkoxyamidyl 102 has been modeled at B3LYP/ 6-31G* level by the reaction of methyl radical (R0 ¼ CH3 d), with methoxyformamidyl 102e giving 103e and methane. Energies are presented in Table 13. Formation of methane and 103e from alkoxyamidyl and methyl radical has an activation barrier of only 14.4 kcal mol1 but this would be expected to be lower

93

N-ACYLOXY-N-ALKOXYAMIDES

1.093Å

1.236Å

1.354Å 1.396Å O1

N1 C2 O2 C1 1.438Å

1.536Å

1.218Å

1.237Å H2

(a)

N1 1.385 154.3° O2 C1 1.369 1.223

C2 H1 1.364

1.357Å 1.419Å

N1 C1 O2

1.269Å

O1

O1C1N1O2=0.2° C1N1O2C2=0.2° N1O2C2H1=-0.1° N1O2C2H2=179.9°

H1 C2 H2

O1

(b)

(c)

Fig. 25 B3LYP/6-31G* structures for (a) methoxyformamidyl 102e, (b) transition state and (c) 103e from methyl radical reaction with N-methoxyformamidyl. Table 14 Arrhenius activation energies, entropies of activation and rate constants at 373 K for thermal decomposition of N-acetoxy-N-butoxybenzamides 26a, 26b, 26d, 26e and N-acetoxy-N-(4-nitrobenzyloxy)benzamide 27h in mesitylene192 ln A

Substrate (X, Y) 26a (X ¼ H) 26b (X ¼ MeO) 26d (X ¼ Me) 26e (X ¼ Cl) 27h (Y ¼ NO2)

27.27 29.99 31.28 26.03 30.05

(2.02) (3.21) (0.56) (1.59) (5.28)

EA(kcal mol1) 27.65 29.45 31.17 26.04 30.08

(1.55) (2.45) (0.44) (1.20) (4.05)

DSz (cal K1 mol1) 6.32 0.92 1.63 8.78 0.81

(4.0) (6.4) (1.1) (3.2) (10.5)

105 k373

r

4.20 5.62 1.99 10.7 2.53

0.9938 0.9931 0.9996 0.9958 0.9823

with larger alkoxy or benzyloxy groups. The transition state is computed to be early along the reaction coordinate (Fig. 25b), resembling methyl and amidyl radicals. The methyl radical centre is largely planar at the transition state and transfers about 0.12 electrons to the alkoxyamidyl. In keeping with this, hydrogen abstraction followed by heterocycle formation should be favoured by electron-withdrawing groups at the oxymethylenic position (C2 in Fig. 25b), and yields of heterocycle 104a with 4-nitrobenzyloxy- and 4-chlorobenzyloxy groups were greater than with benzyloxy or 4-methylbenzyloxy substituents. Reaction leading to the triplet diradical is exothermic by 1.0 kcal mol1 but singlet formation would be exothermic by 27.4 kcal mol1. Arrhenius data for thermal decomposition of several N-acyloxy-N-alkoxyamides is presented in Table 14. While the reactions, when monitored by HPLC obeyed ﬁrst order kinetics, rate constants must represent the summation of rate constants of parallel processes, which were not followed individually, and EA and DSz represent an envelope of energies and entropies for several processes.

HERON reactions Thermal decomposition of N-acyloxy-N-alkoxyamides in mesitylene yielded, in certain cases, signiﬁcant quantities of anhydrides formed by intramolecular processes. HERON reactions of N-acyloxy-N-alkoxyamides would be expected to favour migration of acyloxyl groups since anomeric effects are strongest in this direction

94

S.A. GLOVER O R3

Na+

O

Path (i) R

3

Na+

O2CR2 N OR1 N OR1 O 111 112 C R2 O O O Na+ NaH+ R3 2 Path (ii) R3 R N OR1 N OR1 O C O O 114 113 C R2 O Na+ O R2 Na+ O R3 C O Path(iii) R3 N O2CR2 N O O R1 O 116 115 R1

Scheme 22.

(Figs. 4 and 5b). However, such reactions ought only to manifest themselves in gaseous or non-polar media; in aqueous or polar organic solutions (H2O/CH3CN, CH3OH) heterolysis to solvated alkoxynitrenium and carboxylate ions is favoured as evidenced by the AAl1 solvolysis reactions described earlier in this chapter. SN2 reactions too are favoured in polar solvents on account of polar separation in the transition states. Anhydrides have never been observed as side-products from the AAl1 or SN2 reactions of N-acyloxy-N-alkoxyamides by our group37–43 or in Shtamburg’s recent studies.52 Mesitylene, on the other hand, is a high boiling, nonpolar medium in which it is reasonable to expect HERON reactions, provided these are energetically possible. Such reactions ought also to be feasible in the gas phase at higher temperatures. The ﬁrst unequivocal evidence for HERON reactions of N-acyloxy-N-alkoxyamides emerged from ESI-tandem mass spectrometric studies on a number of N-acyloxy-N-alkoxyamides.158 Under these conditions, free from solvent, the sodiated parent ion 110 can be detected, which fragments under collision induced dissociative conditions into three sodiated product ions (Scheme 22, Table 15). These products exhibited masses corresponding to the sodiated N-alkoxyamidyl radical 111, which in most cases was the major pathway (Scheme 22, pathway (i)) but the second most intense product ion was the sodiated anhydride 113 formed through HERON rearrangement of the acyl group (Scheme 22, pathway (ii)). The ester 115 that would be formed through an alternative HERON migration of the alkoxyl group (Scheme 22, pathway (iii)) was a weak product ion in all cases where it was present, and absent in the fragmentations of all aliphatic amides 110h–l. The fragments from these reaction processes are presumed to be acyloxyl radical 112, alkoxynitrene 114 and acyloxynitrene 116.

95

N-ACYLOXY-N-ALKOXYAMIDES Table 15

a b c d e f g h i j k l

M/z for sodiated ions from ESI-MS of N-acyloxy-N-alkoxyamides 110a–la 110

111

113

115

274 302 362 404 370 426 336 288 302 394 316 330

215 243 303 345 249 305 215 167 181 273 195 209

187 215 275 317 249 249 249 201 215 307 229 243

201 229 – – 235 291 – Trace – – – –

a

ESI conditions in positive ion mode: H2O/MeOH; drying gas 350 1C; capillary ¼ 30 V; CID ¼ 5 to 30 eV.

R3

O

Na+

N OR1 O C R2 O

110 a R1=Bu, R2=Me, R3=Ph b R1=Bu, R2=Me, R3=3,5-diMeC6H3 c R1=Bu, R2=Me, R3=fluoren-1-yl d R1=Bu, R2=Me, R3=2-anthraquinon-2-yl e R1=R2= R3=Ph f R1=4-But, R2=R3=Ph g R1=Bu, R2= R3=P h

h R1=Bu, R2= Ph R3=n-Pr i R1=Bu, R2= Ph R3=Bui j R1=Bu, R2= Ph R3=1-Ad k R1=Bu, R2= Ph R3=But l R1=Bu, R2= Ph R3=neopentyl

According to B3LYP/6-31G* calculations on N-formyloxy-N-methoxyformamide 40 and its HERON rearrangement transition states, in the gas phase, migration of acyloxyl group is favoured over the alkoxyl group by about 4 kcal mol1 (Table 16). Two conformations of the methoxyl migration (Fig. 26c and d) were found with very similar energies. The three transition states (Fig. 26b–d) are relatively similar with the migrating formyloxyl oxygen O1 in Fig. 26b or methoxyl oxygen O3 in Fig. 26c and d perpendicular to the O3–N1–C1 or O1–N1–C1 planes, respectively. The HERON transition state for formyloxyl group migration (Fig. 26b) appears looser and later, which would be expected on account of the polarities involved. The N1–O1 and C1–O1 bonds in Fig. 26b are longer than those to O3 in either Fig. 26c or d. The magnitude of the activation energies indicates that such migrations are unlikely to compete with heterolysis of the N–OAc in polar solvents. While the inﬂuence of the spectator sodium ion on these reactions is unknown, clearly the ESI-MS-MS data reﬂects the expected migration tendencies of acyloxyl versus

96

S.A. GLOVER

Table 16 B3LYP/6-31G* computed energies for ground state and HERON migration transition states for N-formyloxy-N-methoxyformamide 40 Structure

E (Hartree)

Relative E (kcal mol1)

Frequency (cm1)

Fig. Fig. Fig. Fig.

472.851865 472.790221 472.782035 472.783513

0.0 38.7 43.4 42.9

271i 210i 282i

26a 26b 26c 26d

O4 1.416Å 1.203Å

N1

O2

C3

1.382Å

C1

C2

1.446Å

<110.4°>

O3 O1 2.071Å

O2

O1

76°

N1

C2

1.559Å-

C1 O4

C3

1.799Å-

(a)

O3

(b)

C3

C3

O2 O3

O3 1.723Å

N1

C2

2.041Å

C1 (c)

1.798Å

N1

O4

C2 O1

O2

1.467Å 56°

61°

1.496Å

O4

1.939Å

O1

(d)

C1

Fig. 26 B3LYP/6-31G* geometries for (a) ground state of N-formyloxy-N-methoxyformamide 40 and transition states for migration of (b) formyloxyl and (c,d) methoxyl groups.

alkoxyl moieties in these substrates. The alkoxyl oxygen, though more electronegative than nitrogen, is a better donor atom than the acyloxyl oxygen, which is bonded to an electron withdrawing carbonyl carbon. Hence, the acyloxyl group migration would be preferred. The formation of anhydrides 101 in the thermal decomposition reactions of N-acyloxy-N-alkoxyamides in non-polar mesitylene (Scheme 21, pathway (ii)) can thus be attributed to HERON migration of acyloxyl groups. However, the similarity of EAs for the two isomeric transition states raises the possibility that in these reactions some, or all of the ester might be generated by HERON migration of the alkoxyl substituent (Scheme 21, pathway (iii)).

N-ACYLOXY-N-ALKOXYAMIDES

5

97

Biological activity of N-acyloxy-N-alkoxyamides

N-Acyloxy-N-alkoxyamides constitute a class of mutagenic compounds and the premise that, as N-alkoxy analogues of metabolites from aromatic amines, they would act as electrophiles towards DNA, proved to be correct. Almost all those N-acyloxy-N-alkoxyamides tested to date are mutagenic in the Ames test.195,196 Moreover, unlike the aromatic amines that spawned the research into this class of amides, they are direct-acting mutagens that damage DNA without the need for metabolic activation, i.e. they are in an ultimate mutagenic form.47–49 It is widely understood that metabolic activation is intrinsic to the mode of action of many substances that are known to be carcinogenic60,65,170,197–199 and metabolites so produced are more active than their precursors. Usually such metabolites are electrophilic towards DNA and react with the DNA bases at nucleophilic positions such as N7, N2 and O6 of guanine (G-N7, G-N2 and G-O6) or N3 of adenine (A-N3).169,170,200 Pullman and Pullman illustrated that the electrostatic potential in the region of N7-G is not only the most negative, but that the magnitude of the negative potential increases signiﬁcantly at guanines in single-stranded and duplex DNA.169 It is this site in the major groove that is most often modiﬁed by alkylating agents such as mustards, alkyl sulfonates or dialkylsulfates.170,200 These form primary alkyl cations, or, as in the case of aziridinium ions from nitrogen mustards, masked alkyl cations which are hard Lewis acids and would be expected to react at N7-G according to Pearson’s hard–soft acid/base theory. N-Acyloxy-N-alkoxyamides are no exception and have been shown to react with plasmid DNA at G-N7 in the major groove at physiological pH as well as at A-N3 in the minor groove in slightly acidic solution.49,105 However, the active form is not the alkoxynitrenium ion as originally advocated, rather N-acyloxy-N-alkoxyamides behave as molecular electrophiles towards DNA. Their direct mutagenicity makes them excellent vehicles for investigating mechanisms of mutagenesis and structural inﬂuences upon the mutagenic process and we have utilised many analogues as probes for drug–DNA interactions.46,49,50

MUTAGENICITY IN THE AMES SALMONELLA/MICROSOME ASSAY195,196

The Ames test is a bacterial assay for identifying substances that can produce genetic damage that leads to mutations and it utilises various mutant forms of S. typhimurium that have a histidine requirement that is reverted by mutagenic agents. In the presence of these, they produce this amino acid, which is essential for their growth. Assays are performed on agar plates and mutations are registered through colonies proliferated from reverted bacterial cells. The test is quantitative and in the right dose range usually affords linear dose response curves; numbers of revertants (colonies) are proportional to the exposure to mutagenic material. One of the most widely used strains, TA100, is reverted by point mutations, that is, by basepair substitutions upon replication after DNA damage, which restores the genetic functionality. Such damage can be inﬂicted by electrophilic species.

S.A. GLOVER

1400 1200 1000 800 600 400 200 0

-S9

0

0.03

0.06 0.12 (μmol/plate)

Revertants

Revertants

98

0.25

(a)

1000 900 800 700 600 500 400 300 200 100 0

y = 3056.8x + 157.48 R2 = 0.9934

y = 2716.3x + 161.3 R2 = 0.9953

0

0.1 0.2 (μmol/plate)

0.3

(b)

Fig. 27 Dose–response plots for N-benzoyloxy-N-benzyloxybenzamide 28a in S. typhimurium TA100: (a) comparison of activity with and without S9 and (b) linear plots without S9.

The assay can be performed using mutagenic substances that react directly with DNA or, where metabolic activation is necessary, with pre-mutagen in the presence of rat liver homogenate that is enriched in mixed function oxidases (termed S9). Metabolic oxidation (if that is what is required) results in ultimate or penultimate mutagenic forms, which act as electrophiles towards S. typhimurium. N-Acyloxy-N-alkoxyamides have been shown to mutate S. typhimurium TA100 to the same extent both with, and without the need for metabolic activation.47,48 Fig. 27a illustrates activities of N-benzoyloxy-N-benzyloxybenzamide 28a at various doses both in the presence and absence of S9. Activity is largely the same at all but the maximum dose where the activity in the absence of S9 is reduced owing to toxicity of mutagen at this dose level (such toxicity is removed in presence of S9). Fig. 27b shows duplicate linear dose responses for this mutagen and typiﬁes both the linearity and reliability of the assay with N-acyloxy-N-alkoxyamides. In these assays, there is a background spontaneous reversion (value at zero dose), which is subtracted from reversion rates at all doses. Gradients yield mutagenic activity at 1 mmol per plate, which are used for comparative purposes. O

O

O

N O

N

N O

117

118

O

O

O

O 28a

The need for electrophilicity at the amide nitrogen was demonstrated by comparison of the mutagenic activity of 28a with that of two analogues with similar structure, N-benzoyloxy-N-benzylbenzamide 117, which possesses a leaving group but without anomeric destabilisation and N-benzyl-N-benzyloxybenzamide 118,

99

N-ACYLOXY-N-ALKOXYAMIDES 1800.0 1600.0 y = 2572.3x + 282.58 R2 = 0.9914

Revertants

1400.0 1200.0

28a

1000.0 800.0 600.0 400.0

117/118

200.0 0.0 0

0.1

0.2

0.3 0.4 (μmol/plate)

0.5

0.6

Fig. 28 Mutagenic activities N-benzoyloxy-N-benzyloxybenzamide 28a, N-benzoyloxyN-benzylbenzamide 117 and N-benzyl-N-benzyloxybenzamide 118 in S. typhimurium TA100 without S9. KM

kA

[Mutagen]

[Mutagen][DNA]

[Adduct]

+ [Carboxlate]

Path (i) Path(ii)

kN

– Carboxylate

Path(iii)

KN [Nitrenium ion] + [Carboxylate]

kN [Nitrenium ion][DNA]

Scheme 23.

which is a hydroxamic ester without a leaving group at nitrogen.49,105 While the latter is not an electrophile and would not be expected to undergo either nitrenium ion formation nor SN2 reactions at nitrogen, 117 could in principle react by either mechanism. However calculations of SN2 activation energies for replacement of formate from N-formyloxy-N-methylformamide by ammonia, methanethiol and azide all indicated a much higher barrier for this process when compared to the N-formyloxy-N-methoxyformamide.44,45 In a direct comparison, N-benzoyloxyN-benzyloxybenzamide 28a gave an excellent dose response over the range 0–0.5 mmol per plate while 117 and 118 were completely inactive over the same dose range and returned only background reversion rates (Fig. 28). The experiment showed that only a nitrogen bearing a leaving group anomerically destabilised by the geminal alkoxyl oxygen (an anomeric amide), is reactive. The activity of N-acyloxy-N-alkoxyamides can be attributed to direct attack of the mutagens upon DNA (Scheme 23, pathway (i)) or, as was originally proposed, reaction to form N-acyl-N-alkoxynitrenium ions (Scheme 23, pathway (ii)). In either

100

S.A. GLOVER

case, two processes are involved, which are described by equilibrium constants KM and kA along pathway (i) and by KN and kN along pathway (ii). Equilibrium constants KM and KN represent respectively binding of intact mutagen or nitrenium ion into the grooves of DNA. Rate constants kA and kN are respective rates of reaction of bound mutagen or nitrenium ion at G-N7 of DNA. O

O R

X

N O AcO

25

O Bu

N O AcO

26

O Y

N O

N O O O

AcO

27 Z

28

Table 17 presents mutagenic activities in S. typhimurium TA100 at 1 mmol per plate without S9 for a wide range of mutagens investigated to date. A number of qualitative observations can be made:

A comparison of the activity levels of N-acetoxy-N-butoxybenzamides 26, N-acetoxy-N-benzyloxybenzamides 27 and N-benzoyloxy-N-benzyloxybenzamides 28 with common para substituents showed increasing mutagenicity with size or hydrophobicity. Activities increased with increasing aromatic substitution. Activity increases with larger alkoxyl substituents in N-acetoxy-N-alkoxybenzamides 25 indicating a hydrophobic effect upon mutagenicity (entries 1–4). The greater relative activity of all tested members of series 28 (entries 22–32) when compared with the N-acetoxy analogue, 27a (entry 14), as well as variation in activity of various 28 congeners and other mutagens with different leaving groups showed clearly that the mutagens must bind intact to DNA rather than generate N-acyl-N-alkoxynitrenium ions intracellularly, since 27a and all members of 28 would generate the same ionic intermediate. If this were the case, mutagenic activities would be similar. Mutagenic activities across series 28 (entries 22–32) showed a distinct dependence upon the electronic effect of the substituent; activity is greatest with electron donor and weakest with electron-withdrawing substituents at Z. Steric effects of substituents on various side chains are important as exempliﬁed by tert-butylated mutagens (entries 13, 21, 30, 41–47). While substrates bearing one tert-butyl group on the para position of either the benzyloxy, benzoyloxy or benzamide rings have slightly lower activities within each series, the presence of two or more such groups leads to strongly reduced activities.

Activity is strongly inﬂuenced by mutagen functionality and structure, its reactivity, hydrophobicity and the steric effects of substituents.

X,Y,Z,Ra 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

log Pc

pKAd

E 2s e

E 2s f

E 3s g

Ih

Calc.i

2.49 2.53 2.88 2.40

1.54 2.02 4.11 1.86

4.76 4.76 4.76 4.76

0 0 0 0

0 0 0 0

0 0 0 0

0 0 0 0

2.35 2.48 3.06 2.44

0.14 0.05 0.18 0.04

2.50 2.73 2.92 2.48 2.54 2.69 2.16 2.45 2.37

2.44 2.32 4.12 2.93 3.0 3.27 2.48 2.48 4.15

4.76 4.76 4.76 4.76 4.76 4.76 4.76 4.76 4.76

0 0.55 2.41 1.24 0.97 1.16 1.76 0 2.78

0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0

2.60 2.49 2.73 2.56 2.62 2.67 2.37 2.61 2.69

0.10 0.24 0.19 0.08 0.08 0.02 0.21 0.16 0.32

2.63 2.70 3.06 3.09 2.78 2.76 2.74 2.60

2.93 2.81 3.96 4.61 3.42 3.49 3.76 4.64

4.76 4.76 4.76 4.76 4.76 4.76 4.76 4.76

0 0 0 0 0 0 0 0

0 0.55 0.55 2.41 1.24 0.97 1.16 2.78

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

2.73 2.62 2.94 2.83 2.68 2.74 2.79 2.78

0.10 0.08 0.12 0.26 0.10 0.02 0.05 0.18

3.22 3.06 3.70 3.29 3.12

4.83 4.70 6.51 5.32 5.39

4.20 4.31 4.21 4.37 3.99

0 0 0 0 0

0 0 0 0 0

0 0.55 2.41 1.24 0.97

0 0 0 0 0

3.15 3.08 3.39 3.20 3.17

0.07 0.02 0.31 0.09 0.05

Diff.j

101

22 23 24 25 26

PhCON(OAc)OR (R) 25a Et 25b Pr 25e Octyl 25f Pri X-C6H4CON(OAc)OBu (X) 26a H 26b MeO 26c Ph 26d Me 26e Cl 26f Br 26g p-NO2 26h m-NO2 26i But PhCON(OAc)OCH2C6H4-Y (Y) 27a H 27b MeO 27c PhO 27d Ph 27e Me 27f Cl 27g Br 27i But PhCON(O2CC6H4-Z)OBn (Z) 28a H 28b MeO 28c Ph 28d Me 28e Cl

log TA100b

N-ACYLOXY-N-ALKOXYAMIDES

Table 17 Ames mutagenicity data for Salmonella typhimurium TA100 and physicochemical parameters for N-acyloxy-N-alkoxyamides together with predicted activities according to QSAR (Equation 4)

102

Table 17 (continued ) log TA100b

log Pc

pKAd

E 2s e

E 2s f

E 3s g

Ih

Calc.i

2.87 2.96 2.69 3.20 2.94 3.23

4.58 4.86 4.86 6.54 4.86 4.70

3.73 3.55 3.42 4.35 3.45 4.09

0 0 0 0 0 0

0 0 0 0 0 0

1.01 0.51 1.76 2.78 0 0

0 0 0 0 0 0

2.88 2.97 2.83 3.39 3.00 3.09

0.01 0.01 0.14 0.19 0.06 0.14

33 34 35 36 37 38 39 40

28f CHO 28g CN 28h NO2 28i But 28k m-NO2 (bzyl) 28l m-MeO (bzyl) R1CON(O2CR2)OR3(R1,R2,R3) 34 Ph,Me,-(CH2)629f Ph,Ph,Bu 30a Me,Me,Bu 30b Me,Ph,Bu 30c Me,Me,Bn 30d Me,Ph,Bn 37a Ph,Me,2,6-diMeBn 29u 3,5-diMePh, Me,Bu

2.96 2.70 1.94 2.65 2.22 2.90 3.04 2.74

3.55 4.34 0.54 2.44 1.04 2.93 3.91 3.42

4.76 4.20 4.76 4.20 4.76 4.20 4.76 4.76

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

2.91 3.01 2.07 2.48 2.21 2.62 3.01 2.87

0.05 0.31 0.13 0.17 0.01 0.28 0.03 0.13

41 42 43

31a 31b 33a

4-ButPh,Ph,Bn Ph,Ph, 4-ButBn 4-ButPh,Me, 4-ButBn

3.50 3.05 2.89

6.54 6.54 6.31

4.20 4.20 4.76

2.78 0 2.78

0 2.78 2.78

0 0 0

0 0 0

3.24 3.19 2.86

0.26 0.14 0.03

44 45 46 47

31c 31d 31e 32a

Ph, 4-ButPh , 4-ButBn 4-ButPh, 4-ButPh,Bn 4-ButPh,Ph, 4-ButBn Ph,3,5-diButPh,Bn

1.86 2.61 1.64 2.07

8.24 8.24 8.24 8.24

4.35 4.35 4.20 4.38

0 2.78 2.78 0

2.78 0 2.78 0

2.78 2.78 0 0

0 0 0 0

3.44 3.48 3.28 4.13

1.58 0.87 1.64 2.06

48 49 50 51 52 53 54

29m 30g 29r 30f 30h 30i 30m

2-Naphthyl,Me,Bu Me,Me, CH2-2-Np Me,2-Np,Bu Me,Me, CH2-1-Np Me,Me,(CH2)2-2-Np Me,Me, (CH2)3-2-Np Me, CH2-1-Np, Bu

3.59 3.57 3.64 3.53 3.41 3.47 3.41

3.44 2.03 3.44 2.03 2.31 2.73 3.38

4.76 4.76 4.16 4.76 4.76 4.76 4.30

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

1 1 1 1 1 1 1

3.74 3.34 3.61 3.34 3.42 3.54 3.62

0.15 0.23 0.03 0.19 0.01 0.07 0.21

27 28 29 30 31 32

Diff.j

(0.71) (1.09) (0.89) (1.05) (0.85) (0.79) (0.65)

S.A. GLOVER

X,Y,Z,Ra

X,Y,Z,Ra 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71

30k Me,1-Pyr,Bu 30l Me, CH2-1-Pyr,Bu 29n 2-Fl,Me,Bu 29o 2-Aq,Me,Bu RCON(OBz)OBu (R) 29a Et 29b Pri 29c But 29d CH2But 29e 1-Ad PhCON(O2CR)OBu (R) 29g Pr 29h Pri 29i (S)-2-Bu 29j CH2But 29k 1-Ad 29l But 29s 2,6-diMePh 29t 3,5-diMePh

log TA100b

log Pc

pKAd

E 2s e

E 2s f

E 3s g

Ih

Calc.i

Diff.j

3.92 3.70 2.93 2.39

4.59 4.54 4.18 2.83

3.68 4.30 4.76 4.76

0 0 0 0

0 0 0 0

0 0 0 0

1 1 1 1

3.83 3.95 3.94 3.57

0.09 0.25 1.01 1.18

2.39 2.41 1.96 2.38 2.51

3.10 3.66 4.37 4.31 5.12

4.20 4.20 4.20 4.20 4.20

0 0 0 0 0

0 0 0 0 0

0 0 0 0 0

0 0 0 0 0

2.66 2.82 3.02 3.00 3.23

0.27 0.41 1.06 0.62 0.72

2.47 2.05 2.23 2.42 2.73 2.30 2.43 2.73

3.51 3.66 4.08 4.31 5.12 4.37 5.31 5.30

4.82 4.85 4.80 4.80 4.86 5.03 3.56 4.34

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

2.91 2.96 3.06 3.13 3.36 3.19 3.14 3.30

0.44 0.91 0.83 0.71 0.63 0.89 0.71 0.57

(0.95) (0.61) (0.15) (0.32)

N-ACYLOXY-N-ALKOXYAMIDES

Table 17 (continued )

a

103

X, Y, Z, and R’s are substituents speciﬁed for each series; Np ¼ naphthyl, Pyr ¼ pyrenyl, Fl ¼ ﬂuorenyl, Aq ¼ 9,10-anthraquinone-2-yl and Ad ¼ 1-adamantyl. b Log of the gradient of the dose-response plot in the linear region; activity at 1 mmol per plate. c Log P is the hydrophobicity (log octanol–water partition coefﬁcient) calculated according to the Ghose–Crippen algorithm.201 d pKA of the carboxylic acid corresponding to the N-acyloxyl substituent. e Taft steric parameter for para substituent on benzamide side chains. f Taft steric parameter for para substituent on benzyloxy side chains. g Taft steric parameter for para substituent on benzoyloxy side chains. h Indicator variable for polycyclic aromatics; I ¼ 1 if present, I ¼ 0 if absent. i Calculate log TA100 according to QSAR (Equation 4). j Difference between predicted and observed log TA100; values in parentheses are computed with I ¼ 0.

104

S.A. GLOVER

Activity and structure The observation that different leaving groups impart quite different mutagenic activity is good support for binding of the drugs to DNA before reacting. Activity must follow pathway (i) of Scheme 23 where the overall structure controls the equilibrium constant KM. Further support for this came from in vitro DNA damage studies from which we showed that the DNA damage proﬁles of mutagens with different substitution patterns vary markedly and, as well, have different pH dependence.49,105 Since mutagens with identical backbones and different leaving groups would produce common nitrenium ions, which would give rise to identical damage patterns, these results are inconsistent with nitrenium ion formation (Scheme 23, pathway (ii)). However, they do not distinguish between intact binding followed by nitrenium ion formation (Scheme 23, pathway (iii)) as opposed to direct SN2 attack upon the bound mutagen. Activity and stability Series 28 contains a diverse range of electronic effects on the benzoyloxy ring. A plot of log TA100 against Hammett s constants for all but the 4-methoxy 28b (toxic) and 4-phenyl 28c (signiﬁcantly more hydrophobic) was linear with negative gradient (Fig. 29).48 In reactivity, the same series of mutagens exhibited positive Hammett correlations for AAl1 acid catalysed solvolysis (r ¼ +0.32) and SN2 reactivity with N-methylaniline (r ¼ +1.7), hydroxide (r ¼ +0.55) and L-cysteine ethyl ester (r ¼ +1.1), all reactions in which the benzoyloxyl group leaves with electrons and which are therefore accelerated by electron-withdrawing groups. The negative Hammet r-value for 3.7 ρ = -0.49 R = 0.913

3.5

Log TA100

3.3

3.1

Me

H

m-MeO Cl

But

CN 3-NO2

2.9

CHO

2.7

2.5 -0.2

4-NO2

0

0.2

σ

0.4

0.6

0.8

Fig. 29 Hammett correlation of mutagenicities of 28 at 1 mmol per plate without S9.

N-ACYLOXY-N-ALKOXYAMIDES

105

mutagenicity therefore indicates that activity correlates not with their reactivity (which would require a positive r-value) but with their stability to chemical reactions. Furthermore, activity should correlate positively with pKA of the departing carboxylic acid although this dependency is likely to be greater for Series 28. Experimental or calculated pKAs for all substrates are found in Table 17. This inverse correlation has been noted in other instances of biological activity. The cytotoxic activity of a range of duocarmycin analogues behaved similarly.202,203 In these situations, biological activity will be reduced for reactive substrates that are consumed intracellularly or in a biological assay. In the case of N-acyloxyN-alkoxyamides, those mutagens that are less prone to solvolysis or reaction with adventitious nucleophiles such as glutathione, will be present at higher concentrations at DNA, the intracellular target. Activity and hydrophobic effects The dependence of drug activity and most other biological interactions on hydrophobicity is well established.204 Qualitatively, the mutagenicity of N-acyloxyN-alkoxyamides shows a clear dependence upon this property, which can be represented by the log of the octanol–water partition coefﬁcient, P. Log P can be evaluated additively by a variety of algorithms and using the Ghose–Crippen method,201 mutagenic activity of mutagens of entries 1–43 in Table 17 correlated with log P according to the following equation: log TA100 ¼ 0:20ð0:02Þ log P þ 2:06; ð0:09Þ ðn ¼ 43; r ¼ 0:819; s ¼ 0:2; F ¼ 84Þ

(2)

Relatively little is known about the role of speciﬁc interactions, including the role of hydrophobicity, in targeting small molecules to DNA.205,206 A number of studies have pointed to hydrophobicity being a factor in determining the activity of mutagenic intercalating agents207,208 and of a number of classes of indirect-acting mutagens.204,209 The latter must undergo metabolic activation to an electrophilic form and the impact of structural changes upon binding to DNA is unclear as other hydrophobic interactions, predominantly those with activating enzymes, play a role. There is reasonable evidence from a range of studies that mutagens requiring metabolic activation have a log P dependence in the region of 1.0. To date, with the exception of one study on sulfonate esters, the activities of direct-acting mutagens have been thought to be independent of hydrophobicity and this has even been seen as a characteristic of direct mutagenicity.204–206 Equation 2 indicates that the activity of N-acyloxy-N-alkoxyamides, unquestionably a class of direct-acting mutagens, is dependent upon their hydrophobicity.48 The lower coefﬁcient in these studies must relate to their association with bacterial DNA. Activity and steric effects The inﬂuence of steric effects on mutagenicity of N-acyloxy-N-alkoxyamides is clear from the data presented in Table 17. Despite their much higher hydrophobicities, all

106

S.A. GLOVER

mono tert-butylated mutagens, exhibit lower or similar activity to that of their unsubstituted variant (Table 17, entries 13/5, 21/14, 30/22, 41/22, 42/22 and 43/14). The N-benzoyloxy-N-benzyloxybenzamides bearing two or more tert-butyl groups exhibit radically reduced activities considering their increased hydrophobicity (Table 17, entries 44–47). Steric effects are also apparent in the activities of several mutagens in series 28. In Fig. 29 the 4-tert-butyl group in 28i, the 4-nitro substituent in 28h and perhaps the CHO in 28f result in lower activity than would be expected on electronic grounds alone. Realistic prediction of the mutagenic activities of this class of mutagens requires a physicochemical evaluation of steric effects of substituents and the most widely utilised measures of such are Taft steric parameters, Es (Table 17).204,210 Quantitative structure activity relationships (QSAR) governing activity of N-acyloxyN-alkoxyamides Entries 1–43 in Table 17 constitute an excellent training set for QSAR. Substrates from classes 25–28 have representative electronic, and steric requirements as well as wide ranging hydrophobicity (over ﬁve log P units). Multivariate analysis incorporating log P, pKA and steric effects E1s , E2s and E3s for para substituents on the benzamide, benzyloxy and benzoyloxy side chains afforded a QSAR (Equation 3): log TA100 ¼ 0:29ð0:03Þ log P þ 0:21ð0:09ÞpK A þ 0:15ð0:04ÞE 1s þ 0:16ð0:04Þ E 2s þ 0:11ð0:05Þ E 3s þ 0:86ð0:46Þ ðn ¼ 43; r ¼ 0:897; s ¼ 0:16; F ¼ 30:3Þ

ð3Þ

Relative to QSAR (2), there was a considerable improvement in the correlation, which showed a strong dependence upon log P and, as expected from the full set, a moderate pKA dependence. Based on the errors, steric impact is real for the benzamide and the benzyloxy side chains while that for the benzoyloxy is less reliable.50 Observed versus predicted mutagenicities from QSAR (3) are plotted in Fig. 30 (ﬁlled squares). The mutagenic activity of N-acyloxy-N-alkoxyamides reﬂects their interaction with the primary target, which in this case is bacterial DNA. The predictive model (Equation 3) allows discovery of structural factors that either increase or diminish DNA damage. Such effects can operate either upon binding to DNA or reactivity with DNA. Both types of structural impacts have been observed. Intercalation effects50. The mutagenicities of seven naphthalene-containing mutagens (119, Table 17, entries 48–54) are plotted in Fig. 30 (circles). All six show mutagenic activity on average about 0.9(70.1) log TA100 unit higher than predicted from QSAR (3), which would indicate a greater propensity to damage DNA over and above that determined by hydrophobicity and pKA alone. Activity enhancement

107

N-ACYLOXY-N-ALKOXYAMIDES 4.0

Observed TA100

3.5

3.0

2.5

2.0

1.5 1.5

2.0

2.5 3.0 Predicted TA100

3.5

4.0

Fig. 30 Predicted versus observed mutagenic activities (log TA100) for (a) mutagens in Table 17 (entries 1–43, ﬁlled squares) and (b) naphthalene substrates (entries 48–54, circles) based upon QSAR (3).

is most probably greater than that considering that steric impact of such groups is not accounted for in QSAR (3). Mutagenicity data is supported by DNA damage studies, which showed signiﬁcantly greater damage for mutagens bearing naphthalene.49,105 O R3

O R1

N O O C R2 O R1= 2-Np, 1-Np, CH2-2-Np, (CH2)2-2-Np and R2=R3=Me R2= 2-Np, CH2-2-Np and R1=Bu, R3=Me R3=2-Np and R1=Bu, R2=Me

119

Me

O Bu

N O O C R O R=1-pyrenyl, CH2(1-pyrenyl)

120

R

Bu N O O C Me O

R=2-Fluorenyl, 9,10-anthraquinon-2-yl

121

Such activity enhancement can be ascribed to planarity of the naphthalene nucleus and its disposition for p–p stacking between the DNA base pairs. There is no literature precedence for involvement of naphthalene in such processes unless it is p electron deﬁcient.211,212 However, our results suggest that both alkylated and acylated naphthalenes are intercalators with bacterial DNA.

108

S.A. GLOVER

Reconstitution of the QSAR to incorporate an indicator variable I for the presence (I ¼ 1) or absence (I ¼ 0) of a naphthalene nucleus yields QSAR (4). log TA100 ¼ 0:28ð0:03Þ log P þ 0:21ð0:09ÞpK A þ 0:14ð0:04Þ E 1s þ 0:15ð0:04Þ E 2s þ 0:09ð0:05Þ E 3s þ 0:86ð0:07ÞI þ 0:93ð0:44Þ ðn ¼ 50; r ¼ 0:927; s ¼ 0:16; F ¼ 45:6Þ

ð4Þ

Observed versus predicted mutagenicities from QSAR (4) and incorporating all seven naphthalene containing mutagens are plotted in Fig. 31 in which naphthalene data are represented by open circles. Predicted activities of all mutagens by this equation are presented in Table 17. For entries 48–58, values in parentheses are differences between predicted (I ¼ 0) and observed activities. Remarkably, comparison of the dependence upon the indicator variable I, and log P gives an estimate of the value of including naphthalene anywhere in a mutagen’s structure. Naphthalene is worth 3 log P units of hydrophobicity to binding to DNA. It is presumed that the intercalative binding increases residence time upon DNA resulting in greater probability of the DNA damage leading to a mutation event. The parent hydroxamic acid of 30g, N-(2-naphthylmethyloxy)acetamide was completely inactive in the Ames test so that the activity of the mutagen, is not due to naphthalene intercalation itself.213 The fact that presence of naphthalene on the leaving group is equally effective lends further support for pathway (i) in Scheme 23, which invokes pre-association of intact mutagens with DNA. QSAR (4) would appear to be a useful measure of whether other similar polycyclic aromatics can intercalate with DNA. Two pyrene-containing mutagens N-butoxyN-(1-pyrenecarboxoyloxy)acetamide and N-butoxy-N-(1-pyreneacetoxy)acetamide 29k

4.0

29l

Observed TA100

3.5

3.0

29n

2.5

29o

2.0

1.5 1.5

2.0

2.5

3.0

3.5

4.0

Predicted TA100

Fig. 31 Predicted versus observed mutagenic activities (log TA100) for (a) mutagens in Table 17 (entries 1–43 (squares), 48–54 (circles)) and (b) other polycyclic substrates (entries 55–58, triangles, diamonds) based upon QSAR (4).

109

N-ACYLOXY-N-ALKOXYAMIDES

120 ﬁt QSAR (4) with indicator value I ¼ 1 and deviate signiﬁcantly with I ¼ 0 (Fig. 31, triangles; Table 17, entries 55 and 56). Presumably the pyrene nucleus has a similar intercalative binding ability to naphthalene. However, data for a ﬂuorene analogue and a mutagen bearing an anthraquinone-2-yl structure 121 best ﬁt with I ¼ 0, (Fig. 31, diamonds; Table 17, entries 57 and 58). Both these may not intercalate in the same manner. To generalise these ﬁndings more mutagens bearing pyrene, ﬂuorene and anthraquinone moieties must be tested. Other polycyclic aromatic systems can be treated likewise. Thus, the mutagenicities of N-acyloxy-N-alkoxyamides may provide an intracellular means of quantifying structural elements of molecules in terms of their ability to target drugs to DNA; they can be used as probes for drug–DNA interactions.50 Steric effects46. The mutagenicities of all mono 4-tert-butylated mutagens 26i, 27i, 28i, 31a,b and di-tert-butylated 33a (Table 17, entries 13, 21, 30, 41–43) are predicted well according to the QSAR (4) (Fig. 32, ﬁlled squares). In particular, the activities of the triarylated mutagens 28i, 31a and 31b bearing a single tert-butyl group are adequately predicted. Binding and reaction of these mutagens with DNA would appear to be unimpeded by the presence, on one aryl ring, of the bulky tert-butyl group. However mutagenicities for 122 and 123 with more than one tert-butyl group were markedly lower. The four di-tert-butylated N-benzoyloxy-N-benzyloxybenzamides 31c–e and 32a (Fig. 32, solid circles, Table 17, entries 44–47) deviate very signiﬁcantly from the predicted values and the triarylated mutagens 31f and 32b were completely inactive. 4.5

Observed TA100

4.0 3.5 3.0 30b 29a

29d

29e 32a

2.0 29c 1.5 1.5

31d

29b

2.5

2.0

2.5

31c 31c

3.5 3.0 Predicted TA100

4.0

4.5

Fig. 32 Predicted versus observed mutagenic activities (log TA100) for mutagens in Table 17 (entries 1–43, 48–58 squares), di-tert-butylated analogues 31c–e, 32a (122, 123, entries 44–47, closed circles) and alkylamides 30b, 29a–e (124, entries 36, 59–63, triangles) based upon QSAR (4).

110

S.A. GLOVER O

O

R3

R1 N O O C O

R N O O C O

R2 But

R1= R2=But, R3=H R1=R3=But, R2=H R2=R3=But, R1=H R1= R2=R3=But

But R=But, H 123

122

Lack of reactivity towards SN2 reactions at nitrogen would in itself account for low mutagenic activity. However, the tert-butyl groups on 31c–f and 32a,b are well removed from the reactive nitrogen and, as well, SN2 reactions of mutagens 31c–e with N-methylaniline in methanol at 303 1C occur with relatively similar rate constants to that of unsubstituted 28a, and of mono tert-butylated systems 28i and 31a,b (Table 18). The steric impediment of para aryl substituents to damage of DNA and which is accounted for in the QSAR’s (3) and (4) through Taft steric parameters E1s , E2s and E3s , presupposes that the mutagen can enter the major groove where bulkiness of substituents presumably impacts upon alignment of the mutagen for reaction with guanine. The radical decrease in activity of mutagens 31c–f and 32a,b suggests that, in these cases, the drugs cannot react with DNA since they are either too large to enter the major groove or, once in the groove, are unable to achieve the transition Table 18 Bimolecular rate constants for the reaction of N-acyloxy-N-alkoxyamides bearing bulky substituents with N-methylaniline in d4-methanol Substrate (Z, R1, R2, R3)

104k303(l mol1 s1)

Substrate (R1, R2, R3)

104k313(l mol1 s1)

193.3a 117.2a 798.5c

30b (R1 ¼ Bu,R2 ¼ Ph) 29a (R2 ¼ Ph, R3 ¼ Et) 29d (R2 ¼ Ph, R3 ¼ neopentyl) 29b (R2 ¼ Ph, R3 ¼ Pri) 29c (R2 ¼ Ph, R3 ¼ But) 29e (R2 ¼ Ph, R3 ¼ 1adamantyl)

97.2b 8.1 2.2

28a, Z ¼ H 28i Z ¼ But 31a R1 ¼ R2 ¼ H,R3 ¼ But 31b R1 ¼ But, R2 ¼ R3 ¼ H 31c R1 ¼ R2 ¼ But, R3 ¼ H 31d R1 ¼ H R2 ¼ R3 ¼ But, 31e R1 ¼ R3 ¼ But, R2 ¼ H a

From reference 43. From reference.105. c From reference 46. d No reaction at this temperature. b

578.5c 353.0c 156.6c 369.8c

d d d

111

N-ACYLOXY-N-ALKOXYAMIDES

state for reaction with G–N7, the putative site of DNA attack. A comparison of the activity of 33a with that of 31e, or the mono tert-butylated N-benzoyloxyN-benzyloxybenzamides 28i, 31a,b with the di-tert-butylated substrates 31c–e suggests that there is a critical bulk beyond which, even with conformational ﬂexibility, inclusion in the major groove or reaction of the nucleotide at the electrophilic amide nitrogen becomes difﬁcult. The computed transition state for SN2 reaction at nitrogen requires the ‘‘backside’’ attack of the nucleophile on the nitrogen and approximately on line with the bond to the acyloxyl leaving group.45 With such constraints the transition state may be reached with one tert-butyl group on one of the three benzene rings but is unattainable with two such substituents. With respect to the generalised Scheme 23, it is likely that the di-tert-butylayted mutagens have a very low KM. Fig. 33a gives the AM1-optimised transition state for the reaction of N-formyloxy-N-methoxyformamide with 9-methylguanine. Using the angular and bond length constraints in this transition state (at G-N7, mutagen N1 and formyloxy oxygen O1), mutagen 31b models into the major groove as shown in Fig. 33b. It is clear, even form this crude representation that, while one tert-butyl group can be accomodated along the direction of the groove, para positions on the other two aromatic rings could not include a second tert-butyl substituent without imposition of very signiﬁcant steric constraints. Notably, X-ray data5 or AM1 calculations on tri-tert-butylated mutagen 31f provide an indication of the maximum dimensions between any two tert-butyl groups as between 14 and 17 A˚, which greatly exceed the width of the major groove of B-DNA (12 A˚).

N7-N1 = 1.643 Å N1-O1 = 1.976 Å

12Å

O4

C3 N1

N7

6Å O1

C1 O3

O2

(a)

C2

(b)

Fig. 33 (a) AM1-optimised transition state for reaction of N-formyloxy-N-methoxyformamide with G-N7; (b) conformation of 31b in the major groove of DNA constrained to transition state in (a) with side chains minimised by molecular mechanics.

112

S.A. GLOVER

While all di-tert-butylated substrates exhibit no or low activity, 31d is less affected. While this may be attributed to the particular conﬁguration, it is also the least reactive towards SN2 reactions (Table 18), which, on account of the inverse relationship between mutagenicity and reactivity, would generate higher activity. O R

Bu N O O C Ph O

R=Me, Et, Pri, But, CH2But, 1-adamantyl 124

Mutagenicity of a limited number of alkylamides 124 have been investigated.46,50,105 In particular, the series 29a–e together with 30b (Fig. 32, triangles, Table 17, entries 36, 59–63) constitute a series with increasing branching adjacent to the amide carbonyl. Remarkably, the deviations from the predicted mutagenicities increase through the series Me(30b)oEt(29a)oPri(29b)oBut(29c), the latter being virtually inactive. The deviation of the bulky 1-adamantamide derivative 29e is similar to that of the pivalamide 29c. Clearly branching at the a-carbon is an important factor controlling activity. Comparison of the activity of the pivalamide 29c with that of the neohexamide 29d also indicates a diminishing steric inﬂuence as the tert-butyl group is moved one methylene away from the carbonyl. Unlike the di and tri-tert-butylated mutagens 31c–f and 32a,b the overall dimensions of substrates 29a–b would not be expected to impede their access to the major groove. We attribute the systematic reduction in activities to an increase in the activation barrier for reaction at G-N7 as a consequence of branching at the a-carbon. The ease of SN2 reaction of these substrates with N-methylaniline broadly supports this (Table 18). The rate constant at 313 K for reaction of propanamide 29a is an order of magnitude smaller than that of the acetamide 30b,105 while mutagens with the isopropyl and the tert-butyl amide side chains, 29b,c, are unreactive with N-methylaniline at this temperature. In addition, while the pivalamide 29c is unreactive, the neohexamide 29d, in which the branching is one methylene removed from the reactive nitrogen is susceptible to SN2 reaction with N-methylaniline, albeit with a smaller rate constant than 29a. The adamantane carboxamide mutagen 29e is also resistant to SN2 reactions at nitrogen. Clearly there are some differences between reactivity and mutagenic activity, in that the isobutyramide substrate 29b, which exhibits impaired mutagenicity, is unreactive towards SN2 reactions with N-methylaniline and the adamantamide 29e, though resistant to SN2 reaction with N-methylaniline at matching temperatures, is still marginally mutagenic. However, such reactions are also dependent upon the nature of the nucleophile and the solvent environment. For instance, while the tert-butyl-containing amide side chain in 29c completely inhibits attack by

113

N-ACYLOXY-N-ALKOXYAMIDES

N-methylaniline at nitrogen, similarly substituted N-chloroamides react readily with azide, which is known to be a stronger nucleophile.44 Differences between experimental SN2 reaction rates and the mutagenicity data may also be related to the inﬂuence of the hydrophobic environment within the major groove; solvent organisation about a polar SN2 transition state44,45 accounts for substantially negative entropies of activation in the reactions of these mutagens with N-methylaniline (Table 5). Such a contribution would presumably be different or absent in reactions at G-N7. Overall, though, relative mutagenicities in this series follow SN2 reactivities and would appear to be controlled by steric inhibition of the substitution reaction at the amide nitrogen rather than any steric inhibition to binding in the major groove. In these substrates the magnitude of kA in Scheme 23 would appear to be limiting. As pyramidal amides5,32 their SN2 reactivity with neutral nucleophiles like N-methylaniline parallels that of a-haloketones with amines, which, as described in an earlier section, are also strongly affected by steric effects on the a0 -carbon.183 SN2 reactions are in general strongly and adversely inﬂuenced by steric effects and branching b to the reactive centre and the same appears to be true for N-acyloxy-Nalkoxyamides 30b and 29a–e. Broadly speaking, their mutagenic activity is affected similarly. The arguments presented herein lend the strongest support to SN2 attack by G-N7 at the amide nitrogen of N-acyloxy-N-alkoxyamides and for pathway (i) in Scheme 23 rather than pathway (iii) in which, once bound to DNA, the mutagens undergo SN1 formation of reactive nitrenium ion. Though beyond the scope of this review, steric inhibition of mutagenicity by remote214,215 and proximal alkyl substituents 216,217 has recently been established by Boche and coworkers in studying the Ames mutagenicity of 4-nitro- and 4-aminobiphenyls. The effects with these are more difﬁcult to interpret than our mutagens since in both cases enzymatic activation is involved (nitroreductase [–NO2] and cytochrome P450 oxidases [–NH2]) and intercalative binding is also a possibility. Inhibition to reaction with G-C8 was also proposed to explain reduced mutagenicities. In our system, metabolic activation is not required. Furthermore, there is no suggestion that simpler mutagens without polycyclic aromatics intercalate with DNA.50 It is clear from this study that bulky distal groups can have a signiﬁcant inﬂuence upon the actual groove binding with bacterial DNA whereas proximal branching inhibits reaction with DNA. O Ph

Bu N O O C R O

R=Me, Bu, Pri,(S)-2-Bu But, CH2But, 1-adamantyl, 2,6-diMePh, 3,5-diMePh 125

114

S.A. GLOVER

Activities of another group of mutagens 125 bearing increasingly bulky leaving groups showed yet a third type of steric effect. All alkyl groups adjacent to the benzoyloxy carbonyl in 29g–l (Fig. 34, ﬁlled circles, Table 17, entries 64–69) reduced activity but branching resulted in an average deviation from the predicted activity of 0.8(70.1) log TA100. In this series, rates of SN2 reaction with aniline were correlated negatively with pKA of the departing carboxylic acid and were independent of steric effects (Table 8). Like alkylamides 29a–e, this series would gain access to the major groove but substituents in this case present no barrier to SN2 reactivity. Rather, the bulkiness must impact upon ease of reaction with DNA (upon kA in Scheme 23) by imposing constraints on the transition state for reaction at G-N7. The activity difference between dimethylated benzoyloxy systems 29s and 29t (Fig. 34, triangles, Table 17, entries 70 and 71) supports this. While the 3,5dimethylbenzoyl system in 29t would subtend the normal small angle between the aromatic ring and the carbonyl, the 2,6-dimethyl substitution would result in a high degree of twisting resulting in greater steric demand (Fig. 35). Notably, N-(2,6-dimethylbenzyloxy)-N-acetoxybenzamide 37a and N-acetoxyN-butoxy-3,5-dimethylbenzamide 29u (Fig. 34, black squares, Table 17, entries 39 and 40) were well modelled by QSAR (4). The difference between predicted and observed activities for 37a (0.03) and N-acetoxy-N-isopropoxybenzamide 25f (0.04) (Table 17, entry 4) suggests that steric effects proximate to the reactive nitrogen but on the alkoxyl side chain are less important in terms of reactivity with DNA. This is most probably a consequence of the greater ﬂexibility in this side chain relative to the acyloxyl or amide groups.

4.5

Observed TA100

4.0 3.5 3.0 29t 29g 29j 29s 29I 29h 29i

2.5 2.0 1.5 1.5

2.0

2.5

29k

3.0 3.5 Predicted TA100

4.0

4.5

Fig. 34 Predicted versus observed mutagenic activities (log TA100) for mutagens in Table 17 (entries 1–43, 48–58 squares), 29g–l(entries 64–69, circles) and 29t and 29s (entries 70 and 71, triangles) based upon QSAR (4).

115

N-ACYLOXY-N-ALKOXYAMIDES O

O Ph

N O

OBu

Ph

O

H3 C

N O

OBu O

CH3 H3C 29s

CH3 29t

Fig. 35 Conformations in N-(2,6-dimethylbenzoyloxy)- and N-(3,5-dimethylbenzoyloxy)N-butoxybenzamides. ANTICANCER ACTIVITY OF N-ACYLOXY-N-ALKOXYAMIDES

Preliminary investigations highlighted the cytotoxic and anticancer activity of N-acyloxy-N-alkoxyamides. Four mutagens, N-acetoxy-N-butoxybenzamide 26a, N-acetoxy-N-benzyloxybenzamide 27a, N-(4-methylbenzoyloxy)-N-benzyloxybenzamide 28d and N-acetoxy-N-butoxy-2-naphthamide 29m were found to be reproducibly cytotoxic towards P388 mouse leukemia but with IC50 values in the range of 15–19 mmol. QSAR (4) provides a means of selecting side chains for optimal mutagenicity and DNA damage. N-Butoxy-N-(pyrene-1-carboxoyloxy)acetamide 30k with a log TA100 ¼ 3.92, is the most potent mutagen we have discovered to date. This together with three other mutagens N-benzoyloxy-N-butoxybenzamide 29f (log TA100 ¼ 2.7), N-acetoxy-N-(2-naphthylmethyloxy)-acetamide 30g (log TA100 ¼ 3.57) and N-butoxy-N-(2-naphthoyloxy)benzamide 29r, (predicted log TA100 ¼ 4.14) were tested in the NCI 60 cell line in vitro screen.218 The standardised screen evaluates activity in 60 human cancer cell lines. Overall quality of activity is measured by a mean over all strains of the growth inhibition index (average of the doses causing 50% growth inhibition, GI50, across all 60 cell lines), and can be used for comparative purposes. Of the four mutagens, the pyrene analogue was most active with a mean GI50 of 4.63. Data from the other three mutagens gave GI50 values close to the cut-off concentration of 1 104 M. The mean GI50 of 30k rivals that of some anticancer agents widely used in cancer chemotherapy (busulfan 3.7; cyclophosphamide 3.67; melphalan 4.8; uracil nitrogen mustard 4.63).219 However, the establishment of QSAR (4) and further reﬁnement in its predictive accuracy should lead to identiﬁcation of better anticancer drug candidates based upon the N-acyloxy-N-alkoxyamides.

6

Conclusions

N-Acyloxy-N-alkoxyamides are a class of biologically active amides, which are members of the recently identiﬁed class we have termed anomeric amides. These are

116

S.A. GLOVER

amides with two heteroatoms bonded at the nitrogen. Structurally, spectroscopically and chemically, this functional group is best described as an N-acylamine since the nitrogens are sp3 hybridised resulting in little or no amide nitrogen overlap with the carbonyl. In N-acyloxy-N-alkoxyamides the heteroatoms are oxygens and this review has dealt with the unique manifestations of this substitution pattern where substituents are acyloxyl and alkoxyl groups. The physical and spectroscopic properties of N-acyloxy-N-alkoxyamides conﬁrm pyramidality at nitrogen and the disconnection of the nitrogen lone pair from the amide carbonyl. The presence of an acyloxyl and an alkoxyl group at nitrogen also results in an anomeric interaction between the oxygens, which is facilitated by the sp3-hybridised nitrogen. Experimental observations, including X-ray analysis are fully supported by results from computational chemistry. Anomeric weakening of the N–OAc bond results in four principle types of reactions:

AAl1 (or SN1)-type leading to alkoxynitrenium ions with acid-catalysis SN2-type with various organic and inorganic nucleophiles Homolytic dissociation reactions leading to stabilised alkoxyamidyl radicals HERON reactions either directly, or of products of solvolysis or SN2 displacement

They are direct-acting mutagens towards S. typhimurium TA100 with demonstrable ability to react with plasmid DNA at G-N7 or A-N3. The versatile synthetic protocol has enabled us to synthesise and test, reliably, many representatives and establish a working QSAR that has enabled us to understand intimate details of their interaction with bacterial DNA. Their biological activity can be sheeted home to binding to DNA followed, most likely, by an SN2 reaction with G-N7. Electronic and steric effects of substituents are in accord with this mechanism. The QSAR (4) cannot only highlight structural factors most likely to impact positively or negatively upon their biological activity, it can be used to probe drug–DNA interactions. QSAR (4) has already enabled identiﬁcation of candidates with reasonable anticancer activity and can direct future synthesis and design to this end. This study probably represents the most systematic and detailed analysis of any set of direct-acting chemical mutagens to date.204

Acknowledgements The author acknowledges the many research students who have made signiﬁcant contributions to this work as well as invaluable collaborations with Dr. Antonio Bonin, Prof Jeanne Buccigross, Prof Mike Novak, Prof Arvi Rauk and Dr. David Tucker.

N-ACYLOXY-N-ALKOXYAMIDES

117

References 1. Greenberg, A., Breneman, C. and Liebman, J.F. (2000). The Amide Linkage: Structural Significance in Chemistry, Biochemistry and Material Science. Wiley & Sons, Inc, New York 2. Wiberg, K.B. (2000). Origin of the amide rotational barrier. In The Amide Linkage. Structural Significance in Chemistry, Biochemistry and Materials Science, Greenberg, A., Breneman, C.M. and Liebman, J.F. (eds), p. 33. John Wiley & Sons, Inc., New York 3. Wiberg, K.B. and Breneman, C. (1992). J. Am. Chem. Soc. 114, 831 4. Challis, B. and Challis, J. (1979). Amides and related compounds. In Comprehensive Organic Chemistry, Nitrogen Compounds, Carboxylic Acids, Phosphorous Compounds, Sutherland, I.O. (ed.), Vol. 2, p. 957. Pergamon Press, Oxford 5. Gillson, A.-M.E., Glover, S.A., Tucker, D.J. and Turner, P. (2003). Org. Biomol. Chem. 1, 3430 6. Glover, S.A. and Rauk, A. (1996). J. Org. Chem. 61, 2337 7. Cooley, J.H., Bills, W.D. and Throckmorton, J.R. (1960). J. Org. Chem. 25, 1734 8. Johnson, J.E., Nalley, E.A., Kunz, Y.K. and Springﬁeld, J.R. (1976). J. Org. Chem. 41, 252 9. Yamada, S. (1999). Rev. Heteroatom Chem. 19, 203 10. Yamada, S. (2000). Sterically hindered twisted amides. In The Amide Linkage. Structural Significance in Chemistry, Biochemistry and Materials Science, Greenberg, A., Breneman, C.M. and Liebman, J.F. (eds), p. 215. John Wiley & Sons, Inc, New York 11. Yamada, S. and Homma, A. (2002). Chem. Commun. 2656 12. Yamada, S. and Noguchi, E. (2001). Tetrahedron Lett. 42, 3621 13. Cow, C.N., Britten, J.F. and Harrison, P.H.M. (1998). Chem. Commun. 1147 14. Duspara, P.A., Matta, C.F., Jenkins, S.I. and Harrison, P.H.M. (2001). Org. Lett. 3, 495 15. Matta, C.F., Cow, C.C., Sun, S., Britten, J.F. and Harrison, P.H.M. (2000). J. Mol. Struct. 523, 241 16. Bain, A.D., Chen, H. and Harrison, P.H. (2006). Can. J. Chem. 84, 421 17. Matta, C.F., Cow, C.C. and Harrison, P.H.M. (2003). J.Mol. Struct. – Theochem. 660, 81 18. Otani, Y., Nagae, O., Naruse, Y., Inagaki, S., Ohno, M., Yamaguchi, K., Yamamoto, G., Uchiyama, M. and Ohwada, T. (2003). J. Am. Chem. Soc. 125, 15191 19. Ohwada, T., Achiwa, T., Okamoto, I., Shudo, K. and Yamaguchi, K. (1998). Tetrahedron Lett. 39, 865 20. Alkorta, I., Cativiela, C.C., Elguero, J., Gil, A.M. and Jime´nez, A.I. (2005). New J. Chem. 29, 1450 21. Kirby, A.J., Komarov, I.V., Kowski, K. and Rademacher, P. (1999). J. Chem. Soc. Perkin 2, 1313 22. Kirby, A.J., Komarov, I.V. and Feeder, N. (2001). J. Chem. Soc. Perkin 2, 522 23. Kirby, A.J., Komarov, I.V., Wothers, P.D. and Feeder, N. (1998). Angew. Chem. – Int. Edit. Engl. 37, 785 24. Greenberg, A. (2000). The amide linkage as a ligand: its properties and the role of distortion. In The Amide Linkage. Structural Significance in Chemistry, Biochemistry and Materials Science, Greenberg, A., Breneman, C.M. and Liebman, J.F. (eds), p. 47. John Wiley & Sons, Inc., New York 25. Kirby, A.J., Komarov, I.V. and Feeder, N. (1998). J. Am. Chem. Soc. 120, 7101 26. Bashore, C.G., Samardjiev, I.J., Bordner, J. and Coe, J.W. (2003). J. Am. Chem. Soc. 125, 3268 27. Brown, R.S. (2000). Studies in amide hydrolysis: the acid, base and water reactions. In The Amide Linkage. Structural Significance in Chemistry, Biochemistry and Materials Science, Greenberg, A., Breneman, C.M. and Liebman, J.F. (eds), p. 85. John Wiley & Sons, Inc., New York

118

S.A. GLOVER

28. Lopez, X., Mujika, J.I., Blackburn, G.M. and Karplus, M. (2003). J. Phys. Chem. A 107, 2304 29. Morgan, K.L., Rawlins, M.L.M. and Montgomery, M.N. (2005). J. Phys. Org. Chem 18, 310 30. Mujika, J.I., Mercero, J.M. and Lopez, X. (2005). J. Am. Chem. Soc. 127, 4445 31. Yamada, S. and Katsumata, H. (1999). J. Org. Chem. 64, 9365 32. Glover, S.A. (1998). Tetrahedron 54, 7229 33. Glover, S.A., Mo, G. and Rauk, A. (1999). Tetrahedron 55, 3413 34. Glover, S.A. and Rauk, A. (1999). J. Org. Chem. 64, 2340 35. Glover, S.A., Mo, G., Rauk, A., Tucker, D. and Turner, P. (1999). J. Chem. Soc. Perkin Trans. 2, 2053 36. Glover, S.A. and Rauk, A. (2002). J. Chem. Soc. Perkin Trans. 2, 1740 37. Campbell, J.J., Glover, S.A. and Rowbottom, C.A. (1990). Tetrahedron Lett. 31, 5377 38. Campbell, J.J., Glover, S.A., Hammond, G.P. and Rowbottom, C.A. (1991). J. Chem. Soc. Perkin Trans. 2, 2067 39. Bonin, A.M., Glover, S.A. and Hammond, G.P. (1994). J. Chem. Soc. Perkin Trans. 2, 1173 40. Bonin, A.M., Glover, S.A. and Hammond, G.P. (1998). J. Org. Chem. 63, 9684 41. Campbell, J.J. and Glover, S.A. (1992). J. Chem. Soc. Perkin Trans. 2, 1661 42. Campbell, J.J. and Glover, S.A. (1999). J. Chem. Res. (S) 8, 474 43. Campbell, J.J. and Glover, S.A. (1999). J. Chem. Res. (M) 8, 2075 44. Glover, S.A. and Mo, G. (2002). J. Chem. Soc. Perkin Trans. 2, 1728 45. Glover, S.A. (2001). Arkivok xii, 143 46. Andrews, L.E., Bonin, A.M., Fransson, L., Gillson, A.-M.E. and Glover, S.A. (2006). Mutat. Res. 605, 51 47. Gerdes, R.G., Glover, S.A., Ten Have, J.F. and Rowbottom, C.A. (1989). Tetrahedron Lett. 30, 2649 48. Bonin, A.M., Banks, T.M., Campbell, J.J., Glover, S.A., Hammond, G.P., Prakash, A.S. and Rowbottom, C.A. (2001). Mutat. Res. 494, 115 49. Banks, T.M., Bonin, A.M., Glover, S.A. and Prakash, A.S. (2003). Org. Biomol. Chem. 1, 2238 50. Andrews, L.E., Banks, T.M., Bonin, A.M., Clay, S.F., Gillson, A.-M.E. and Glover, S.A. (2004). Aust. J. Chem. 57, 377 51. Shishkin, O.V., Zubatyuk, R.I., Shtamburg, V.G., Tsygankov, A.V., Klots, E.A., Mazepa, A.V. and Kostyanovsky, R.G. (2006). Mendeleev Commun. 222 52. Shtamburg, V.G., Klots, E.A., Pleshkova, A.P., Avramenko, V.I., Ivonin, S.P., Tsygankov, A.V. and Kostyanovsky, R.G. (2003). Russ. Chem. Bull. 52, 2251 53. Shtamburg, V.G., Tsygankov, A.V., Klots, E.A. and Kostyanovsky, R.G. (2004). Mendeleev Commun. 208 54. Kadlubar, F., Fu, P.P., Jung, H., Shaikh, A.U. and Beland, F.A. (1990). Environ. Health Perspect. 87, 233 55. Beland, F.A. and Kadlubar, F.F. (1990). Metabolic activation and DNA adducts of aromatic amines and nitroaromatic hydrocarbons. In Chemical Carcinogenesis and Mutagenesis, Cooper, C.S. and Grover, P.L. (eds), p. 267. Springer Verlag, Secaucus 56. Lin, D.X., Bryant, M.S., Maleveille, C., Friesen, M., Bartsch, H., Lang, N.P. and Kadlubar, F.F. (1994). Environ. Health Perspect. 102, 11 57. Kadlubar, F.F. (1985). Environ. Health Perspect. 62, 19 58. Kadlubar, F.F., Miller, J.A. and Miller, E.C. (1978). Cancer Res. 38, 3628 59. Humphreys, W.G., Kadlubar, F.F. and Guengerich, F.P. (1992). Proc. Natl. Acad. Sci. USA 89, 8278 60. Blackburn, G.M. and Kellard, B. (1986). Chem. Ind. 607, 687, 770 61. Boche, G., Bosold, F. and Schroder, S. (1988). Angew. Chem. Int. Ed. Engl. 27, 973 62. Famulok, M., Bosold, F. and Boche, G. (1989). Angew. Chem. Int. Ed. Engl. 28, 337

N-ACYLOXY-N-ALKOXYAMIDES 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103.

119

Bosold, F. and Boche, G. (1990). Angew. Chem. Int. Ed. Engl. 29, 63 Famulok, M. and Boche, G. (1989). Angew. Chem. Int. Ed. Engl. 28, 468 Colvin, M.E., Hatch, F.T. and Felton, J.S. (1998). Mutat. Res. 400, 479 Panda, M., Novak, M. and Magonski, J. (1989). J. Am. Chem. Soc. 111, 4524 Helmick, J.S., Martin, K.A., Heinrich, J.L. and Novak, M. (1991). J. Am. Chem. Soc. 113, 3459 Helmick, J.S. and Novak, M. (1991). J. Org. Chem. 56, 2925 Novak, M., Kahley, M.J., Eiger, E., Helmick, J.S. and Peters, H.E. (1993). J. Am. Chem. Soc. 115, 9453 Davidse, P.A., Kahley, M.J., McClelland, R.A. and Novak, M. (1994). J. Am. Chem. Soc. 116, 4513 Fishbein, J.C. and McClelland, R.A. (1995). J. Chem. Soc. Perkin Trans. 2, 663 Novak, M. and Lin, J. (1999). J. Org. Chem. 64, 6032 McClelland, R.A., Davidse, P.A. and Hadzialic, G. (1995). J. Am. Chem. Soc. 117, 4173 Novak, M., Kahley, M.J., Lin, J., Kennedy, S.A. and James, T.G. (1995). J. Org. Chem. 60, 8294 Novak, M. and Kennedy, S.A. (1995). J. Am. Chem. Soc. 117, 574 Fishbein, J.C. and McClelland, R.A. (1996). Can. J. Chem. 74, 1321 McClelland, R.A., Kahley, M.J. and Davidse, P.A. (1996). J. Phys. Org. Chem. 9, 355 McClelland, R.A. (1996). Tetrahedron 52, 6823 McClelland, R.A., Kahley, M.J., Davidse, P.A. and Hadzialic, G. (1996). J. Am. Chem. Soc. 118, 4794 Sukhai, P. and McClelland, R.A. (1996). J. Chem. Soc. Perkin Trans. 2, 1529 Novak, M., Kayser, K.J. and Brooks, M.E. (1998). J. Org. Chem. 63, 5489 Novak, M. and Rajagopal, S. (2001). N-Arylnitrenium ions. In Advances in Physical Organic Chemistry, Tidwell, T.T. and Richard, J.P. (eds), p. 167. Academic Press Ltd., London Ramlall, P. and McClelland, R.A. (1999). J. Chem. Soc. Perkin Trans. 2, 225 Novak, M. and Rajagopal, S. (2002). Chem. Res. Tox. 15, 1495 Glover, S.A. and Novak, M. (2005). Can. J. Chem. 83, 1372 Kennedy, S.A., Novak, M. and Kolb, B.A. (1997). J. Am. Chem. Soc. 119, 7654 McClelland, R.A., Gadosy, T.A. and Ren, D. (1998). Can. J. Chem. 76, 1327 McClelland, R.A., Ahmad, A., Dicks, A.P. and Licence, V.E. (1999). J. Am. Chem. Soc. 121, 3303 Cheng, B. and McClelland, R.A. (2001). Can. J. Chem. 79, 1881 Glover, S.A., Goosen, A., McCleland, C.W. and Schoonraad, J.L. (1984). J. Chem. Soc. Perkin Trans. 1, 2255 Glover, S.A., Goosen, A., McCleland, C.W. and Schoonraad, J.L. (1987). Tetrahedron 43, 2577 Glover, S.A., Rowbottom, C.A., Scott, A.P. and Schoonraad, J.L. (1990). Tetrahedron 46, 7247 Kikugawa, Y. and Kawase, M. (1984). J. Am. Chem. Soc. 106, 5728 Kikugawa, Y. and Shimada, M. (1987). Chem. Lett. 1771 Kawase, M., Kitamura, T. and Kikugawa, Y. (1989). J. Org. Chem. 54, 3394 Kikugawa, Y. (1996). Rev. Heteroatom. Chem. 15, 263 Wardrop, D.J. and Basak, A. (2001). Org. Lett. 3, 1053 Wardrop, D.J., Landrie, C.L. and Ortiz, J.A. (2003). Synlett 9, 1352 Miyazawa, E., Sakamoto, T. and Kikugawa, Y. (2003). J. Org. Chem. 68, 5429 Glover, S.A. and Scott, A.P. (1989). Tetrahedron 45, 1763 Schroder, D., Grandinetti, F., Hrusak, J. and Schwarz, H. (1992). J. Phys. Chem. 96, 4841 Li, Y., Abramovitch, R.A. and Houk, K.N. (1989). J. Org. Chem. 54, 2911 Srivastava, S., Ruane, P.H., Toscano, J.P., Sullivan, M.B., Cramer, C.J., Chiapperino, D., Reed, E.C. and Falvey, D.E. (2000). J. Am. Chem. Soc. 122, 8271

120

S.A. GLOVER

104. Sullivan, M.B., Brown, K., Cramer, C.J. and Truhlar, D.G. (1998). J. Am. Chem. Soc. 120, 11778 105. Banks, T.M. (2003). Reactivity, Mutagenicity and DNA Damage of N-acyloxy-N-Alkoxyamides. Ph.D., University of New England 106. Clements, B. and Glover, S.A. Unpublished data 107. Mu¨ller, K. and Eschenmoser, A. (1969). Helv. Chim. Acta 52, 1823 108. Rudchenko, V.F. (1993). Chem. Rev. 93, 725 109. Epiotis, N.D., Cherry, W.R., Shaik, S., Yates, Y.L. and Bernardi, F. (1977). Structural Theory of Organic Chemistry. Part III, p. 131, Springer-Verlag, Berlin 110. Mu¨ller, K. (1970). Helv. Chim. Acta 53, 1112 111. Shustov, G.V., Zolotoi, A.B. and Kostyanovsky, R.G. (1982). Tetrahedron 38, 2319 112. Dunitz, J.D. (1979). X-Ray Analysis and Structure of Organic Molecules. Cornell University Press, London 113. Winkler, F.K. and Dunitz, J.D. (1971). J. Mol. Biol. 59, 169 114. Hehre, W.J., Radom, L., v.R. Schleyer, P. and Pople, J.A. (1986). Ab Initio Molecular Orbital Theory. John Wiley & Sons, Inc, New York 115. Scott, A.P. and Radom, L. (1996). J. Phys. Chem. 100, 16502 116. Johnson, C.K. (1976). ORTEP-II: A FORTRAN Thermal Elipsoid Plot Program for Crystal Structure Illustrations. Report ORNL-5138. Oak Ridge National Laboratory, Oak Ridge, TN, USA 117. Allen, F.H. (2002). Acta Cryst. B58, 380 118. Glover, S.A. and Rowbottom, C.A. Unpublished data 119. Glover, S.A. and Schumacher, R.R. Unpublished data 120. Glover, S.A. and Mary, C. Unpublished data 121. Glover, S.A. and Price, H. Unpublished data 122. Collis, M. and Glover, S.A. Unpublished data 123. Glover, S.A. and Clay, S.F. Unpublished data 124. Bellamy, L.J. (1975). The Infra-red Spectra of Complex Molecules, Chapter 11, p. 203, Chapman and Hall, London 125. Silverstein, R.M. and Webster, F.X. (1998). Spectrometric Identification of Organic Compounds. John Wiley & Sons, Inc, New York 126. Adelfang, J.L., Hess, P.H. and Cromwell, N.H. (1961). J. Org. Chem. 26, 1402 127. Nyquist, R.A., Streck, R. and Jeschek, G. (1996). J. Mol. Str. 377, 113 128. Bellamy, L.J. (1975). The Infra-Red Spectra of Complex Molecules, Chapter 8, p. 141, Chapman and Hall, London 129. Fisher, J.J. and McMahon, T.B. (1988). J. Am. Chem. Soc. 110, 7599 130. De Almeida, M.V., Barton, D.H.R., Bytheway, I., Ferriera, J.A., Hall, M.B., Liu, W., Taylor, D.K. and Thomson, L. (1995). J. Am. Chem. Soc. 117, 4870 131. Shtamburg, V.G. Private communication 132. Neuvonen, H. and Neuvonen, K. (1999). J. Chem. Soc. Perkin Trans. 2, 1497 133. Neuvonen, H., Neuvonen, K., Koch, A., Kleinpeter, E. and Pasanen, P. (2002). J. Org. Chem. 67, 6995 134. Levy, G.C. and Lichter, R.L. (1979). Nitrogen-15 Nuclear Magnetic Resonance Spectroscopy. John Wiley & Sons, Inc, New York 135. Bennet, A.J., Somayaji, V., Brown, R.S. and Santarsiero, B.D. (1991). J. Am. Chem. Soc. 113, 7563 136. Rudchenko, V.F., Ignatov, S.M., Chervin, I.I. and Kostyanovsky, R.G. (1988). Tetrahedron 44, 2239 137. Yamada, S. (1996). J. Org. Chem. 61, 941 138. Bauer, L. and Exner, O. (1974). Angew. Chem., Int. Ed. Engl. 13, 376 139. Brown, D.A., Glass, W.K., Mageswaran, R. and Mohammed, S.A. (1991). Magn. Reson. Chem. 29, 40

N-ACYLOXY-N-ALKOXYAMIDES

121

140. Isaacs, N.S. (1995). Physical Organic Chemistry (2nd edn). p. 519, Longman Scientiﬁc and Technical, New York 141. Adam, K.R., Lauder, I. and Stimson, V.R. (1962). Aust. J. Chem. 15, 467 142. French, H.T. (1987). J. Chem. Thermodyn. 19, 1158 143. Gold, V. (1969). Adv. Phys. Org. Chem. 7, 259 144. Isaacs, N.S. (1995). Physical Organic Chemistry (2nd edn). p. 308, Longman Scientiﬁc and Technical, New York 145. Pritchard, J.G. and Long, F.A. (1956). J. Am. Chem. Soc. 78, 6008 146. Isaacs, N.S. (1995). Physical Organic Chemistry (2nd edn). p. 389, Longman Scientiﬁc and Technical, New York 147. Gold, V. (1969). Adv. Phys. Org. Chem. 7, 322 148. Isaacs, N.S. (1995). Physical Organic Chemistry (2nd edn). p. 514, Longman Scientiﬁc and Technical, New York 149. Schmidt, M.W. and Gordon, M.S. (1986). Inorg. Chem. 25, 248 150. Schmidt, M.W., Truong, P.N. and Gordon, M.S. (1987). J. Am. Chem. Soc. 109, 5217 151. Okamoto, Y. and Brown, H.C. (1957). J. Org. Chem. 22, 485 152. Brown, H.C. and Okamoto, Y. (1958). J. Am. Chem. Soc. 80, 4979 153. Tsuno, Y., Ibata, T. and Yukawa, Y. (1959). Bull. Chem. Soc. Jpn. 32, 960 154. Yukawa, Y. and Tsuno, Y. (1959). Bull. Chem. Soc. Jpn. 32, 965 155. Buccigross, J.M. and Glover, S.A. (1995). J. Chem. Soc. Perkin Trans. 2, 595 156. Buccigross, J.M., Glover, S.A. and Hammond, G.P. (1995). Aust. J. Chem. 48, 353 157. Glover, S.A. (2006). HERON rearrangement. In: Merck Index (14th edn), Organic Name Reactions ONR-43, O’Neil, M.J. (ed.), Merck & Co., Inc., Whitehouse Station, NJ. 158. Glover, S.A., Rauk, A., Buccigross, J.M., Campbell, J.J., Hammond, G.A., Mo, G., Andrews, L.E. and Gillson, A.-M.E. (2005). Can. J. Chem. 83, 1492 159. Rowe, J.E. and Ward, A.D. (1968). Aust. J. Chem. 21, 2761 160. Isaacs, N.S. (1995). Physical Organic Chemistry (2nd edn). p. 162, Longman Scientiﬁc and Technical, New York 161. Ritchie, C.D. and Virtanen, P.O.I. (1972). J. Am. Chem. Soc. 94, 1589 162. Thornton, E.R. (1967). J. Am. Chem. Soc. 89, 2915 163. Jencks, W.P. (1972). Chem. Rev. 72, 705 164. Pross, A. (1995). Theoretical and Physical principles of Organic Chemistry. John Wiley & Sons, Inc, New York 165. Harris, J.M., Shafer, S.G., Moffatt, J.R. and Becker, A.R. (1979). J. Am. Chem. Soc. 101, 3295 166. Glover, S.A., Goosen, A., McCleland, C.W. and Schoonraad, J.L. (1984). J. Chem. Soc. Perkin Trans. 2, 2255 167. Glover, S.A., Jones, K.M., McNee, I.R. and Rowbottom, C.A. (1996). J. Chem. Soc. Perkin Trans. 2, 1367 168. Harvey, R.G. and Geacintov, N.E. (1988). Acc. Chem. Res. 21, 66 169. Pullman, A. and Pullman, B. (1980). Int. J. Quantum Chem., Symp. 7, 245 170. Gniazdowski, M. and Cera, C. (1996). Chem. Rev. 96, 619 171. Ford, G.P. and Smith, C.T. (1989). J. Comp. Chem. 10, 568 172. Thomson, L.M. and Hall, M.B. (2000). J. Phys. Chem. A 104, 6247 173. Andrews, L.E. and Glover, S.A. Unpublished results 174. Isaacs, N.S. (1995). Physical Organic Chemistry (2nd edn). p. 422, Longman Scientiﬁc and Technical, New York 175. Forster, W. and Laird, R.M. (1982). J. Chem. Soc. Perkin Trans. 2, 135 176. Carrol, F.A. (1998). Perspectives Structure and Mechanism in Organic Chemistry, p. 489. Brooks/Cole Publishing Company, New York 177. Halvorsen, A. and Songstad, J. (1978). J. Chem. Soc. Chem. Commun. 327 178. McLennan, D.J. and Pross, A. (1984). J. Chem. Soc. Perkin Trans. 2, 981

122

S.A. GLOVER

179. Chambers, C.C., Hawkins, G.D., Cramer, C.J. and Truhlar, D.G. (1996). J. Phys. Chem. 100, 16385 180. Jaffe´, H.H. (1953). Chem. Rev. 53, 191 181. Lefﬂer, J.E. and Grunewald, E. (1963). Rates and Equilibria of Organic Reactions. Wiley, New York 182. Isaacs, N.S. (1995). Physical Organic Chemistry (2nd edn). p. 116, Longman Scientiﬁc and Technical, New York 183. De Kimpe, N. and Verhe´, R. (1988). The Chemistry of a-Haloketones, a-Haloaldehydes and a-Haloimines. John Wiley and Sons, Chichester 184. Thorpe, J.W. and Warkentin, J. (1973). Can. J. Chem. 51, 927 185. Adams, M., Glover, S.A. and Tucker, D.J. Unpublished results 186. Sutherland, I.O. (1979). Esters. In Comprehensive Organic Chemistry, Nitrogen Compounds, Carboxylic Acids, Phosphorous Compounds, Sutherland, I.O. (ed.), Vol. 2, p. 871. Pergamon Press, Oxford 187. Mulzer, J. (1991). In Comprehensive Organic Synthesis Vol 6, Heteroatom Manipulation. Winterfeldt, E. (ed.), p. 323. Pergamon Press, Oxford 188. Mulzer, J. (1995). In Comprehensive Organic Functional Group Transformations. Moody, C.J. (ed.), Vol. 5, p. 121. Pergamon, Oxford 189. Vina, J. (1990). Glutathione: Metabolism and Physiological Functions. CRC Press, Inc, Boca Raton 190. Mitchell, S.C. (1996). Biological Interactions of Sulfur Compounds. Taylor & Francis, London 191. Jakokby, W.B. (1981). Detoxiﬁcation and drug metabolism: conjugation and related systems. In Methods in Enzymology, Vol. 77. Academic Press, New York 192. Glover, S.A., Johns, J., Van Losenoord, A., Mary, C. and Garcia, P. Unpublished results 193. Ingold, K.U. and Walton, J.C. (1987). Landolt Bo¨rnstein II 17c. Springer-Verlag, Berlin 194. Glover, S.A., Goosen, A., McCleland, C.W. and Schoonraad, J.L. (1986). J. Chem. Soc. Perkin Trans. 2, 645 195. Mortelmans, K. and Zeiger, E. (2000). Mutat. Res. 455, 29 196. Maron, D.M. and Ames, B.N. (1983). Mutat. Res. 113, 173 197. Miller, E.C. and Miller, J.A. (1981). Cancer 47, 2327 198. Miller, J.A. (1970). Cancer Res. 30, 559 199. Singer, B. and Kusmierek, J.T. (1982). Ann. Rev. Biochem. 52, 655 200. Rajski, S.R. and Williams, R.M. (1998). Chem. Rev. 98, 2723 201. Ghose, A.K., Pritchett, A. and Crippen, G.M. (1988). J. Comp. Chem. 9, 80 202. Baraldi, P.G., Cacciari, B., Romagnoli, R., Spalluto, G., Boyce, C.W. and Boger, D.L. (1999). Bioorg. Med. Chem Lett. 9, 3087 203. Boger, D.L. (1995). Acc. Chem. Res. 28, 20 204. Hansch, C. and Leo, A.J. (1995). Exploring QSAR, Fundamentals and Applications in Chemistry and Biology, Part I. American Chemical Society, Washington, DC 205. Debnath, A.K., Shusterman, A.J., Lopez de Compadre, R.L. and Hansch, C. (1994). Mutat. Res. 305, 63 206. Tuppurainen, K. (1999). Chemosphere 38, 3015 207. Ferguson, L.R. and Denny, W.A. (1979). J. Med. Chem. 22, 251 208. Ferguson, L.R. and Denny, W.A. (1980). J. Med. Chem. 23, 269 209. Leo, A., Panthananickal, A., Hansch, C., Theiss, J., Shimkin, M. and Andrews, A.W. (1981). J. Med. Chem. 24, 859 210. Hansch, C., Leo, A.J. and Hoekman, D. (1995). Exploring QSAR, Fundamentals and Applications in Chemistry and Biology, Part II. American Chemical Society, Washington, DC 211. Sartorius, J. and Schneider, H.-J. (1997). J. Chem. Soc., Perkin Trans. 2, 2319

N-ACYLOXY-N-ALKOXYAMIDES

123

212. Qian, X., Huang, T.-B., Wei, D.-Z., Fan, M.-C. and Yao, W. (2000). J. Chem. Soc., Perkin Trans. 2, 715 213. Bonin, A.M., Clay, S.F. and Glover, S.A. Unpublished results 214. Klein, M., Erdinger, L. and Boche, G. (2000). Mutat. Res. 467, 69 215. Glende, C., Klein, M., Schmitt, H., Erdinger, L. and Boche, G. (2002). Mutat. Res. 515, 15 216. Klein, M., Voigtmann, U., Haack, T., Erdinger, L. and Boche, G. (2000). Mutat. Res. 467, 55 217. Glende, C., Schmitt, H., Erdinger, L., Engelhardt, G. and Boche, G. (2001). Mutat. Res. 498, 19 218. Boyd, M.R. and Paull, K.D. (1995). Drug Development Research 34, 91 219. COMPARE Program. National Cancer Institute/National Institute of Health Developmental Therapeutics Program, http://dtp.nci.nih.gov/index.html

This page intentionally left blank

The interplay between experiment and theory: computational NMR spectroscopy of carbocations HANS-ULLRICH SIEHL Institute of Organic Chemistry I, Ulm University, Albert-Einstein-Allee 11, D-89069, Ulm, Germany

1 Introduction 125 2 Alkyl and cycloalkylmethyl cations 126 3 Vinyl cations 133 4 Cycloalkyl cations 142 5 m-Hydrido-bridged carbocations 144 6 Bicyclic and polycyclic carbocations 145 7 p-Stabilized carbocations 150 8 Heteroatom stabilized carbocations 156 9 Final remarks 158 10 Conclusions 160 Acknowledgments 160 References 160

1

Introduction

Basic questions of physical organic chemistry have always triggered the development for better tools and methods in chemistry. A fruitful interplay of experimental and computational methods has guided the further development of the ﬁeld. Exploring the chemistry of carbocations by NMR spectroscopy in super acid solution combined with today’s state of the art of quantum chemical calculations has been particularly successful. NMR spectroscopy has evolved as the important experimental method for the direct study of structure and dynamics of carbocations in solution and recently also in the solid state. For general structure elucidation contemporary NMR techniques rival X-ray crystallography, which is of particular difﬁcult for carbocations. Quantum chemical methods have developed as indispensable tools to complement experimental results. Despite great interest in ab initio calculation of experimentally observable molecular properties, reliable calculations of NMR chemical shift and coupling constant have become routine only recently.1 This chapter will review in a non-comprehensive attempt recent applications of quantum chemical calculations of NMR chemical shifts and spin–spin coupling constants in carbocation chemistry. The IGLO method, an important breakthrough for the calculation of magnetic properties, was developed in the mid-1980s by Kutzelnigg et al.2 After some 125 ADVANCES IN PHYSICAL ORGANIC CHEMISTRY VOLUME 42 ISSN 0065-3160 DOI: 10.1016/S0065-3160(07)42003-2

r 2008 Elsevier Inc. All rights reserved

126

H-U. SIEHL

applications to carbocations were published in a landmark paper by Schindler3 in 1987 the calculation of carbocations and related species rapidly developed into a ﬂourishing ﬁeld4 exploited by many research groups, especially by Schleyer and collaborators.5,6 These early IGLO applications to carbocations were an important contribution to the recognition of computation of NMR chemical shifts as a potential structural tool.7 Further historical, methodological and practical aspects of calculation of NMR chemical shifts and spin–spin coupling constants not particular related to carbocations have been summarized elsewhere and will not be considered here.8,9 Explanations for the acronyms and abbreviations used in theoretical chemistry and in this review are published by the IUPAC organization.10 Recently the GIAO method (Gauge Independent Atomic Orbitals)11 for calculation of NMR chemical shifts has become the de facto standard and has been implemented in major quantum chemistry packages. The GIAO approach facilitates accurate NMR shift calculations via electron-correlated methods.12 GIAO-DFT methods have evolved as a standard tool in particular for the calculation of NMR chemical shifts for larger molecules and transition metal complexes. Quantum chemical calculations of NMR chemical shifts using DFT methods, however, for some types of carbocations are less satisfactory. DFT calculations lack possibilities for systematic improvements compared to calculations with traditional methods for treating electron correlation. Calculations using the GIAO method in conjunction with second-order Møller–Plesset perturbation theory (GIAO-MP2) have convincingly demonstrated the importance of electron correlation effects in NMR chemical shift calculations and could resolve a number of problems concerning the interpretation of experimental NMR spectra of carbocations.13–17

2

Alkyl and cycloalkylmethyl cations

The CH+ 5 cation 1, protonated methane, is the parent of hypercoordinated carbocations containing a ﬁve coordinated carbon atom. It is elusive in solution and has not been observed by NMR spectroscopy but gas-phase infrared investigations have shown its ﬂuxional structure which has been proven by ab initio molecular dynamic simulation.18 H H C H

H H

1 13 For the CH+ C NMR chemical shift of 5 with a Cs symmetrical structure 1 a –11.5 ppm is calculated. This is 7.1 ppm more shielded than that calculated for the hydrocarbon CH4 at the GIAO-MP2 level.19 The shielding effect in 1 is as expected for a hypercoordinated carbocation carbon.20 The calculated minimum energy structure for diprotonated methane (CH2+ 6 ) 2 has C2v symmetry with two 3c–2e

COMPUTATIONAL NMR SPECTROSCOPY OF CARBOCATIONS

127

bonding interactions. The 13C NMR chemical shift of this ﬂeeing dication is calculated –25.7 ppm, 14.2 ppm more shielded than that of 1.21 Similar structures of type 3 (X ¼ NH3, PH3, H2O, H2S, HF, HCl, CO, N2, CO2 and CS2) have also been calculated.22 H H H

H C

2+

X H

H

H

2+

H C H

H

2

3

Quantum chemical calculations predict the primary ethylcation, C2H+ 5 to have m-hydrido-bridged structure 5 which is 6–8 kcal/mol more stable than the Kekule linebond structure for the primary cation 4. The ethylcation is not stable enough to be observable directly in superacid media. The NMR chemical shifts were calculated for both isomers 4 and 5 using the GIAO-MP2 method for CCSD optimized structures.23 H

H

H H H 1 2 H

Hb

H H

H

5

4

The calculated 13C NMR chemical shifts of C1 and C2 in 4 are 361.17 and 73.96 ppm, respectively, while in the symmetrically bridged form 5 the calculated shift for the carbon atoms is 163.97 ppm. The shielded chemical shift (–0.02 ppm) of the bridging hydrogen (Hb) in 5 is in accord with a 3c–2e bond. The small value of the NMR spin–spin coupling constant in 5 (1J(CHb) ¼ 13 Hz) calculated with the EOM-CCSD method24 as compared to 1J(CH) ¼ 125–250 Hz in hydrocarbons is diagnostic for such 1,2-hydrogen-bridged structures. The C2 symmetric structure for the 2-propyl cation 6 is the only open-chain mini25 mum on the C3H+ 7 potential energy surface calculated at MP2/6-311G(d,p) level. H H 3C

CH 3

6

IGLO/II chemical shift calculations for this geometry agree reasonably well (Dd ¼ +4 ppm) for the C+ carbon (exp. 320.6 ppm), and slightly worse for the methyl carbons (exp. 51.5 ppm) Dd ¼ –7 ppm.26 GIAO-MP2/TZP/DZ NMR calculations result in somewhat better agreement for the chemical shift of the methyl groups.6 As expected for C2 symmetry, the 2-propyl cation 6 is chiral and racemizes by methyl group rotation, passing through a Cs transition state with a calculated barrier of ca. 0.7 kcal/mol. This process is rapid on the NMR time scale, even at temperatures well below 77 K.

128

H-U. SIEHL

The 13C chemical shift tensors of the isopropyl cation were calculated at GIAOMP2/TZP/DZ and GIAO-B3LYP/TZP/DZ levels.27 The principal component d11 (coincident with the C–H bond along the C2 axis of symmetry) was calculated to be 545 ppm, the experimental value measured by CP MAS-NMR was 497 ppm. Only modest improvements of the calculated chemical shifts tensors result when counterions (FHF and SbF 6 ) were added to the carbocation structure to model the effect of ion pairing and environment. The chemical shift tensors of the isopropyl cation in model ion pairs have also been calculated using GIAO-B3LYP MP2/ 6-311G(d,p) and various other basis sets.28 The tert-butyl cation structure (7) with Cs symmetry is better suited for hyperconjugation than the C3h form and is thus energetically slightly favored.29 The energy surface for methyl-group rotation is however very ﬂat. CH 3 H 3C

CH 3

7

The IGLO/DZ//MP2/6-31G(d,p) calculated 13C chemical shifts deviate from experiment by +11 ppm for the central carbon (exp. 335.2 ppm) and –1.3 ppm (averaged) for the CH3 groups (exp. 45.6 ppm). The 13C chemical shift for the central carbon of the protio-tert-butyl dication (8) has been calculated as 327.5 ppm (IGLO/II//MP2/6-31G(d)).30 CH3

H H C H

H

2+

CH 3

8

This value is shielded compared to the experimental and calculated shift of the tert-butyl cation 7. This was taken as evidence that the dication 8 even in strong superacids is only in a very limited equilibrium with the t-butyl cation. The 2-methyl-1-triisopropylsilylpropyl-2-cation (9) is the ﬁrst experimentally accessible b-silyl substituted alkyl cation in superacid solution.31,32 β Sii Pr3 H 2C α CH 3 H 3C

9

The GIAO-HF/TZP-DZ//MP2/TZP calculated shifts show fortuitously small differences to the experimental shifts even for the formally charged carbon (–7.3 ppm)

COMPUTATIONAL NMR SPECTROSCOPY OF CARBOCATIONS

129

and for silicon (+8.1 ppm). The large deviations of the GIAO-MP2 calculated shifts, in particular for those atoms involved in b-Si-hyperconjugation (Ca –28 ppm, Cb +15 ppm, 29Si +35 ppm) can be rationalized assuming an overestimation of s-delocalization effects. The 2-butyl cation is the smallest secondary cation that can be stabilized either by C–C or C–H hyperconjugation. Experimental results give evidence for two equilibrating isomers.33 MP2/6-311G(d,p) calculations show that the symmetrically hydrido-bridged structure 11 is marginally more stable than the partially methylbridged structure 10.34,35 CH3

CH3 H

H

H

H

H

H3 C

10

CH 3 H

11

NMR chemical shifts for 10 and 11 have been calculated using IGLO-HF, GIAO-HF and GIAO-MP2 methods and various basis sets.6 As expected the chemical shifts for the C–C-hyperconjugatively stabilized structure is very sensitive toward the degree of bridging, which depends on the computational method used for the geometry optimization. The 2-methyl-2-butyl cation (12) is the smallest tertiary carbocation structurally suitable for stabilization through C–C hyperconjugation. H3 C

CH3 H

H3C

H

12

A comparison of IGLO/DZ//MP2/6-31G(d) calculated and measured 13C NMR chemical shifts demonstrates that the partially methyl bridged isomer is the preferred species.36 It was demonstrated that the calculated 13C chemical shifts are highly sensitive toward hyperconjugational distortion, i.e. the degree of bridging (ca. 6 ppm/deg from 681 to 981). Electron density methods such as GIAO-DFT methods require much less computational resources in terms of cpu time, memory and disk space compared to wave-function methods such as GIAO-MP2. A systematic study of a set of 16 alkyl- and cycloalkyl cation (Scheme 1) was performed to investigate the performance of GIAO-B3LYP methods for prediction of 13C NMR chemical shifts for these types of carbocations.37 The structures were optimized at the B3LYP/6-31G(d) level and, for comparison, also at the MP2/6-31G(d) level. NMR chemical shifts calculations were done at the GIAO-B3LYP/6-311G(d,p) level. No signiﬁcant changes for the calculated shifts were observed with a diffuse function added to the basis set. Generally, the GIAODFT calculated chemical shifts are more downﬁeld than the experimental shifts. This is due to deﬁciencies of the applied method. GIAO-DFT calculations are

130

H-U. SIEHL

H

H

H

Scheme 1 A systematic GIAO-B3LYP NMR study of 16 alkyl- and cycloalkyl cations.

known to overestimate the paramagnetic contributions to the chemical shielding; this results in overly deshielded chemical shifts. The effect is, however, not uniform for all carbons in the carbocations studied. The chemical shifts of the carbons at positions most sensitive to the positive charge, i.e. the formally positive charged carbon C1 and the a- and b-carbon atoms were analyzed in detail. The shifts of these carbons are expected to show the largest effects from the positive charge. The more remote carbons follow a regular pattern much as in neutral compounds carrying an electron withdrawing substituent unless special effects such as transannular interactions are operative. In general a very good correlation of predicted and experimental chemical shifts was obtained. The slope and intercept for a correlation equation were determined separately for the C+ carbon C1 and the a- and b-carbon positions and were found to be different for cation structures preferentially stabilized by b-C–H and b-C–C hyperconjugation. The chemical shifts calculated using the GIAO-B3LYP/6-311G(d,p) method for MP2/6-31G(d) optimized structures are generally somewhat closer to the experimental values. Scaling of GIAO-B3LYP/ 6-311G(d,p) calculated chemical shifts for structures optimized at the B3LYP/ 6-31G(d) level generally, however, gives reasonable agreement with experiment at reduced computational costs. The largest deviations from the experimental values are found for the predicted shift for the b-carbons of carbocations stabilized preferentially by b-C–C hyperconjugation. This deviation is not so much due to an inadequate level for the calculation of the geometry but a consequence of the limitations in the quantum chemical model used for the calculation of NMR chemical shifts. Chemical shift calculations using the GIAO-MP2 method give results which are closer to the experimental data but still show some deviations in particular for the shift of the b-carbon. It has been demonstrated that b-C–C–s-bond hyperconjugatively stabilized carbocations may require methods for electron correlation which are more demanding than DFT and MP2 procedures, such as CCSD and CCSD(T) methods for the adequate calculation of chemical shifts. NMR chemical shift calculations using these higher correlation methods have been limited to benchmark calculations of model carbocations because of prohibitive computational costs. Experimental NMR spectra of only a few carbocations have been compared with coupled cluster calculated NMR chemical shifts.38

COMPUTATIONAL NMR SPECTROSCOPY OF CARBOCATIONS

131

The approach of scaling GIAO-B3LYP/6-311G(d,p)//B3LYP/6-31G(d) chemical shift calculations for carbocations is limited to alkyl and cycloalkyl carbocations which are predominantly stabilized by either b-C–C- or b-C–H hyperconjugation. Calculation of 13C NMR chemical shifts for some p–p resonance stabilized carbocations, such as benzyl type carbocations, indicate that the same scaling approach is not applicable to carbocations stabilized by p–p conjugation. 39 Cyclopropylethyl cations 13 and 14 have been investigated by 1H and 13C NMR spectroscopy.40–42 H

β

H CH3

H

13

α

β H

α

H H

H γ

H α

H CH3

14

H

15

16

The E-conformation 13 is lower in energy than the Z-isomer 14. These are the smallest cyclopropyl substituted carbocations which can be investigated in solution by high resolution NMR. The corresponding primary cyclopropylmethyl cation 15 cannot directly be observed by high resolution NMR in solution because it is energetically less favorable than the bicyclobutonium ion 16 and thus only a minor 13 isomer in the fast dynamic equilibrium of the C4H+ C- and 1H 7 cations 15 and 16. NMR chemical shifts have been calculated for MP2/6-31G(d) optimized structure of 13 using different approaches to model the n-electron space (GIAO-SCF, GIAODFT(B3LYP), IGLO-DFT(PW91), GIAO-MP2) and various basis sets.42 The best overall correspondence between experimental and calculated chemical shifts of 13 for all carbons or hydrogens was obtained with the GIAO-MP2/TZP/DZ method. The 13C chemical shifts are calculated 7–3 ppm too deshielded, the 1H NMR chemical shifts are calculated also too deshielded with a maximum deviation of –0.6 ppm for the CH3 group. Larger relative differences between the magnetically different protons are calculated as compared to experiment. Nuclear spin–spin coupling constants were calculated for the MP2/6-31G(d) optimized geometry of 13 with SOS-DFT(Perdew)/IGLO-III using a ﬁnite perturbation level (FPT level).43 The experimentally determined vicinal coupling 3 J(Ha,Hb) ¼ 12.5 Hz was calculated as 3J(Ha,Hb) ¼ 13.6 Hz. The calculated 3 J(Ha,CH3) coupling of 6.1 Hz is also in good agreement with the measured value of 6.6 Hz. The C,H-coupling constant of the sp2-hybridized carbon to the C+–H atom 1J(Ca,Ha) ¼ 165 Hz (exp.) is calculated in fair agreement as 154 Hz. For the corresponding Z-isomer 14 the experimental value of 3J(Ha,Hb) ¼ 8.0 Hz corresponds to 8.2 Hz calculated and 3J(Ha,CH3) ¼ 6.2 Hz (exp.) corresponds to 6.5 Hz (calc.). The comparative study of the experimental NMR spectra of E-1-cyclopropyl2-(triisopropylsilyl)ethyl cation (17) and the computational model structure E-1cyclopropyl-2-(trimethylsilyl)ethyl cation (18) demonstrates another application of calculations of 1H and 13C NMR chemical shifts and nuclear spin–spin coupling constants. In particular vicinal 3J(H,H) spin–spin coupling constants are useful for

132

H-U. SIEHL

Fig. 1 400 MHz 1H NMR spectrum of E-1-cyclopropyl-2-(triisopropylsilyl)ethyl cation (17) in SO2ClF/SO2F2 at –105 1C, reference TMA (N+(CH3)4) d ¼ 3.00 ppm. Insets: expansions for CH- and CH2-signals.

the assignment of stereochemistry of this type of carbocations.44 γ ´´ γ´

β´ H

H anti β

α H

17

Si( iPr) 3 Hs yn

anti γ ´´ β´ H Hβ

γ´

α

H

Si(CH3 ) 3 Hs yn

18

The 400 MHz 1H NMR spectrum of 17 (Fig. 1) shows characteristic 3J(H,H) coupling constants for the non-equivalent syn and anti oriented b-CH2-hydrogens to the a-methine-hydrogen which are useful for the assignment of the stereochemistry. The connectivity and the coupling pattern, in particular the diagnostically valuable AMM0 X-spin system formed by Hb0 , Hb(anti), Hb(syn), Ha were analyzed by H,C-COSY and H,H-COSY45 NMR spectra. The spin–spin coupling constants (Perdew/IGLO-III (FPT level))43 at MP2/ 6-31G(d) geometry were calculated for both, the Z- and the E-isomeric structures E-18 and Z-18. The quite satisfactory agreement (DJ ¼ 0.1–1 Hz) of the calculated 3 J(H,H) coupling constants for 3J(Ha,Hb-anti) ¼ 14.2 Hz, 3J(Ha,Hb-syn) ¼ 6.9 Hz, 3 J(Ha,Hb0 ) ¼ 12.3 Hz for the trimethylsilyl-substituted E-isomer model cation 18 with

COMPUTATIONAL NMR SPECTROSCOPY OF CARBOCATIONS

133

the observed values for the experimentally investigated triisopropyl-substituted cation 17 3J(Ha,Hb-anti) ¼ 14.1 Hz, 3J(Ha,Hb-syn) ¼ 7.0 Hz, 3J(Ha,Hb0 ) ¼ 13.3 Hz, conﬁrm the trans arrangement of Ha and Hb, the syn/anti assignment and the E-conformation for the observed carbocation. The decisive vicinal coupling constant for the Z-isomer Z-18 is calculated to be 3J(Ha,Hb0 ) ¼ 7.9 Hz. The better agreement between experimental and GIAO-B3LYP/6-311G(d,p)//MP2/6-31G(d) calculated chemical shifts for the E-isomer (Ha Dd(exp.–calc.) ¼ –0.5 ppm) than calculated for a Z-isomer (HaDd ¼ +1.05 ppm) is in accord with the assignment to the E-conﬁguration to 17. In particular, the large difference calculated for the Cg0 and Cg00 carbons in the Z-isomer (DdCg0 /dCg00 ¼ 21.76), as compared to a calculated difference of DdCg0 /dCg00 ¼ 2.79 for the E-isomer, is not in accord with the experimental shifts (DdCg0 /dCg00 ¼ 1.92). Fig. 2 shows a visual comparison between the experimentally observed multiplet structure for the a-hydrogen of 17 (the X part of the AMM0 X-spin system) and the calculated patterns for the E-18 and Z-18 model structures, respectively. These examples show that the quantum chemical calculation of nuclear spin–spin coupling constants which has been available for the general scientiﬁc community only quite recently is an additional valuable and reliable tool for the interpretation and assignment of NMR spectra of carbocations. Taken together quantum chemical ab initio calculation of chemical shifts and nuclear spin–spin coupling constants opens the possibility for a complete simulation of NMR spectra solely based on calculated carbocation structures. An experimental and calculational NMR investigation of dicyclopropyl substituted cyclobutylmethyl cation (19)45 has shown that IGLO/DZ//B3LYP/6-31G(d) calculated 13C NMR chemical shifts facilitate the assignment of the spectra.

19

3

Vinyl cations

The parent vinyl cation is elusive in solution. It has a m-hydrido-bridged structure 20 which is favored over the linear structure 21 both experimentally in the gas phase (6 kcal/mol) and computationally (3.8 kcal/mol at CCSD(T)).46 The chemical shifts were originally calculated using IGLO/DZ method.3 The hydrido-bridged form 20 has a characteristic upﬁeld shift (–2.6 ppm) for the bridging hydrogen, while the 13C NMR chemical shift for symmetrically bridged carbons is 104.3 ppm. H H C C H

H C C H H

20

21

134

H-U. SIEHL

9.20

9.15

9.10

9.05

9.00

(a)

8.70

8.65

8.60

8.55

8.50

(b)

10.20

10.15

10.10

(c)

Fig. 2 400 MHz 1H NMR signal of the a-hydrogen in 1-cyclopropyl-2-(trialkylsilyl)ethyl cations. (a) Experimental spectrum of a-hydrogen of E-1-cyclopropyl-2-(triisopropylsilyl) ethyl cation 17, 400 MHz, –105 1C, SO2ClF/SO2F2. (b) Simulated spectrum of calculated E-1-cyclopropyl-2-(trimethylsilyl)ethyl cation (E-18) (MP2/6-31G(d) geometry), using GIAODFT (B3LYP) 6-31G(d,p) calculated chemical shifts and SOS-DFT (Perdew) IGLO-III calculated H,H-coupling constants and a line width adapted from the experimental spectrum. (c) Simulated spectrum of calculated Z-1-cyclopropyl-2-(trimethylsilyl)-ethyl cation (Z-18) (MP2/6-31G(d) geometry), using GIAO-DFT (B3LYP) 6-31G(d,p) calculated chemical shifts and SOS-DFT (Perdew) IGLO-III calculated H,H-coupling constants and a line width adapted from the experimental spectrum.

Despite the successful prediction of chemical shifts for a great structural variety of carbocations some difﬁculties have been encountered for vinyl cations.47 The effect of electron correlation, basis sets and geometry on calculated NMR spectra of vinyl cations has been studied in some detail also for the parent vinyl cation in its linear form.48 Comparative experimental and computational NMR studies, however, have

COMPUTATIONAL NMR SPECTROSCOPY OF CARBOCATIONS

135

been reported only for a few vinyl cation structures.47,49–53 H

H

H

C

H

H C

H H

H

H

CH3

H

22

H

H

CH3

H C

C H

H

H H 3C

H3 C

23

The allyl-resonance stabilized E- and Z-pent-1,3-dienyl-2-cations (22 and 23) are the smallest member of vinyl cations observed as persistent species in superacid solution.49 These are difﬁcult to generate experimentally50 but structures with only ﬁve heavy atoms are suitable candidates for coupled cluster model calculations. A challenging task of quantum chemistry was to assign the 13C NMR spectrum of the mixture of isomers (Fig. 3), which exhibits pairs of signals of 22 and 23 which differ only by a few ppm, to the chemical shifts for the speciﬁc carbon atoms of the E- and Z-isomers, respectively. The minimum structures of 22 and 23 were calculated including electron correlation at the MP2 level of theory using a polarized triple-zeta (TZP) basis. Quantum chemical shift calculations for 22 and 23 have been performed with a TZP basis for carbon and a DZ basis for hydrogen using gauge-including atomic orbitals (GIAOs). The GIAO-HF-SCF calculations deviate signiﬁcantly for the positively charged carbon atoms of the allyl-type resonance system showing up to 40 ppm too deshielded values compared to the experimentally observed chemical shifts. The basic HF-SCF method, as well as DFT approaches,51 are not sufﬁcient in predicting satisfactorily the shielding tensors in this type of carbocations. Inclusion of electron correlation using MP2, CCSD and CCSD(T) calculations are in good agreement with experiment and allow an unequivocal assignment of the Z- and E-isomers 22 and 23, respectively. The mean deviation between experimental and calculated NMR chemical shifts at the CCSD(T) level is 2.0 ppm. It is noteworthy but most likely fortuitous, that the lower level MP2 calculations gave marginally better results. These examples demonstrates the degree of accuracy provided by state-of-the-art coupled-cluster calculations of NMR chemical shifts, which allow unequivocal assignment of NMR signals of isomeric carbocations, that differ by only a few ppm.

Fig. 3 100 MHz 13C NMR spectrum of a mixture of E-1,3-dienyl-2-cation (22) and Z-1,3dienyl-2-cation (23) in SO2ClF/SO2F2 at –120 1C; reference d (CD2Cl2) ¼ 52.8 ppm.

136

H-U. SIEHL

13

C NMR chemical shifts of a series of higher substituted a-vinyl substituted vinyl cations 24–27 were calculated to explore the sensitivity of the predicted isotropic shifts to electron correlation, basis set and geometry effects in differently substituted 1,3-dienyl-2-cations.51 H H

H

H CH3

H3C

24 H3C H3C

H3 C

H

25 H

H3 C

26

CH3

H 3C

H CH3

H3 C

CH3

27

Owing to the size of these molecules, the high computational cost and the somewhat limited availability of Coupled-Cluster-NMR-methods only the less expensive second-order Møller–Plesset perturbation theory approach for electron correlation (GIAO-MP2) was considered. MP2/6-31G(d,p) optimized geometries were used for the NMR calculations. The sufﬁcient convergence of the geometries with respect to the wave-function model was conﬁrmed by GIAO-MP2 NMR calculation using these geometries and MP2 geometries obtained with larger basis sets. The absolute 13 C NMR shieldings for all carbons vary little with geometries obtained with larger basis sets, generally by o1 ppm. This indicates that MP2/6-31G(d,p) geometries are in general sufﬁciently accurate for the comparative NMR study of these type of carbocations. The choice of the basis set used for the NMR calculation is however quite important. A TZP basis at carbon is essential for the accurate determination of 13C chemical shifts. If a TZP basis is used also for hydrogen, the GIAO-MP2 calculated 13C NMR chemical shifts show an overall better agreement between experimentally observed and calculated 13C NMR chemical shifts (D3–4 ppm), about 2 ppm smaller deviations from the experimentally observed chemical shifts compared to the data obtained with the previously often used MP2/TZP/DZ combination of basis sets for carbon and hydrogen. Similar as for 22 and 23 the GIAO-HF approximation does not provide an adequate treatment of the electronic effects responsible for the chemical shift in carbocations 25–27. The correlation error, i.e. the difference between the HF-SCF and MP2 calculated NMR chemical shifts, is very large, up to 50 ppm for the sp-hybridized carbon of these vinyl cations. HF-SCF/TZP calculated NMR chemical shifts for some of these cations predict a reverse relative assignment of the terminal carbons of the allyl resonance moiety. GIAO-DFT methods with the B3LYP hybrid functional as well give unsatisfactory results. The DFT calculated NMR chemical shifts (B3LYP/TZP//MP2/6-31G(d,p)) for these carbocations show deshielding deviations from the experimental values which are smaller than with the basic HF-SCF approach but still very signiﬁcant (up to 23 ppm). The B3LYP/TZP calculated chemical shifts also predict for some of these cation structures the wrong order for the

COMPUTATIONAL NMR SPECTROSCOPY OF CARBOCATIONS

137

chemical shift of the terminal carbons of the allyl resonance moiety. Chemical shifts predicted by GIAO-HF-SCF and GIAO-DFT calculations thus cannot be relied upon for unequivocal assignment of NMR signals of dienyl cations. The mean deviation (3–4 ppm) between experimental and GIAO-MP2/TZP calculated chemical shifts for all carbons in these cations is satisfactorily small, only somewhat larger than that for the GIAO-CCSD(T) calculations for 22 and 23 (2 ppm) and only a small fraction of the cost of the very demanding GIAO-CCSD(T) calculations is required. The GIAO-MP2/TZP approach is therefore a suitable method to describe the chemical shift tensor of vinyl cations, which are simultaneously stabilized by p-resonance and s–C–C hyperconjugation of b-methyl groups. The GIAO-MP2/TZP calculated 13C NMR chemical shifts of the cyclopropylidene substituted dienyl cation 27 show for almost all carbon positions larger deviations from the experimental shifts than the other cations 22–26. The GIAO-MP2/TZP method overestimates the inﬂuence of s-delocalization of the positive charge into the cyclopropane subunit on the chemical shifts. Electron correlation corrections for cyclopropylidenemethyl cations such as 27 and 28 are too large to be adequately described by the GIAO-MP2 perturbation theory method and higher hierarchies of approximations such as coupled cluster models are required to rectify the problem. The a-cyclopropylcyclopropylidenmethyl cation 28 is the only vinyl cation experimentally accessible50 in superacid solution which enjoys neither p-bond nor heteroatom stabilization of the positive charge. The 13C NMR spectrum measured at –138 1C is shown in Fig. 4. γ´

β

β´

γ

α

28

The positive charge in 28 is stabilized by b-s–C–C-hyperconjugation with the C–C-ring bonds of the two cyclopropyl moieties. In the parlance of VB theory this is described by resonance of 28 with non-bonding resonance limiting structures, the homoallenyl cation type structure 28a, the homopropargyl cation type structure 28b and the Dewar-type limiting resonance structure 28c. C

C

28c

C

C

28b

C C

28

C

C

28a

In terms of MO theory the stabilization is described by a linear combination of a two-electron ﬁlled antisymmetric 3c–2e-Walsh orbital of the cyclopropane ring with the vacant p-orbital at the C+–carbon atom (Fig. 5). GIAO-HF calculated chemical shifts using a TZP basis for carbon and a DZ basis for hydrogen for the MP2/6-31G(d) optimized geometry of the vinyl cation 28 show large errors of up to 45 ppm because electron correlation is not taken into account (Scheme 2). GIAO-MP2 calculations consider interactions among electrons

138

H-U. SIEHL

Fig. 4 100 MHz 13C NMR spectrum of cyclopropylcyclopropylidenemethyl cation 28 in SO2ClF/SO2F2 at –138 1C, reference: N(CH3)+ 4 d ¼ 55.65 ppm, ~ CH3-signal of tert-butyl cation, d ¼ 48.24 ppm (C+-signal, d ¼ 335.43 ppm omitted). Inset (15–60 ppm) CH- and CH2-signals of 28 only.

(electron correlation) and yield reasonable results for all positions except for the lowﬁeld signal of the sp-hybridized positively charged Ca–carbon, which is calculated 22 ppm too shielded compared to the experimental shift.52 The conspicuously selective poor performance of the NMR chemical shift calculation for Ca of 28 at the MP2 level disappears when electron correlation effects are treated adequately (Scheme 2). High-level coupled-cluster calculations for 28, which were among the ﬁrst real chemical application of these techniques, gave excellent agreement.53 Using the GIAO-CCSD(T) approach, the deviations between experiment and theory for the a-carbon are reduced to 1 ppm. The congruence of calculated and observed 13C shifts for cation 28 suggests that the geometry adopted in superacid solution is similar to the gas-phase prediction. These results support the assumption that the 13C NMR chemical shifts of carbocations in strong superacids are more or less unperturbed by interaction with the medium. This is in accord with

COMPUTATIONAL NMR SPECTROSCOPY OF CARBOCATIONS

139

E

p pz-orbital

antisymmetric Walsh orbital

Fig. 5 Schematic representation of the linear combination of the ﬁlled antisymmetric Walsh orbital of a a-cyclopropyl ring and the vacant pz-orbital of the C+–carbon of a carbocation leading to delocalization of the positive charge in a 4c–2e MO.

the fact that the weakly coordinating anions54 such as SbF5X, Sb2F10X (XQCl) and higher oligomeric anions, with delocalized negative charge, which are present in superacid solutions containing a large excess of SbF5 exhibit a very low nucleophilicity. If any site-speciﬁc solvation or structured ion pairing of the carbocations and the oligomeric anions takes place in these media, the inﬂuence on the 13C NMR chemical shift is negligible. Model studies of basis set dependence and correlation effects have been extended to other, experimentally elusive vinyl cations, the cyclopropylidene methyl cation (29), the a-cyclopropyl vinyl cation (30) and the a-methyl vinyl cation (31).

H

H

H

H

H

29

CH3

30

31

A b-silyl stabilized vinyl cation, the 1-bis(trimethylsilyl)methyl-2-bis(trimethylsilyl) ethenyl cation (32) was investigated by dynamic 13C NMR spectroscopy (Fig. 6).55 (H3 C) 3 Si

Si(CH3) 3 C C C

H Si(CH3 ) 3

(H3C)3 Si

32

140

H-U. SIEHL Cβ Cγ Cγ Cβ´ Cα exp. spectrum

GIAO-CCSD(T)

GIAO-CCSD

GIAO-MP2

GIAO-DFT (B3LYP)

GIAO-HF

Scheme 2 13C NMR chemical shifts of cyclopropylcyclopropylidenmethyl cation 28. bottom to top: calculated GIAO-HF, GIAO-DFT (B3LYP), GIAO-MP2, GIAO-CCSD, GIAO-CCSD(T) with TZP/DZ basis for MP2/6-31G(d,p) geometry and (top) experimental chemical shifts.

This cation is the ﬁrst example of an experimentally accessible carbocation which is stabilized neither by p–p-conjugation nor by b-C–H or b-C–C–s-bond hyperconjugation but only by the superior ability of b-C–Si–s-bonds for hyperconjugative stabilization which is the origin of the b-silyl effect. Cation 32 is stabilized by four silyl groups in b-position to the sp-hybridized C+–carbon Ca. The two silyl groups bond to the sp2-hybridized b-carbon of the vinyl cation moiety are ﬁxed in plane with the formally vacant p-orbital at the Ca–carbon, thus optimal oriented for b-Si–C–s-bond hyperconjugation with the pzC+ orbital. The two silyl groups bond to the methine carbon contribute also to the hyperconjugative stabilization of the positive charge. The rotation around the C+-CH(SiMe3)2 bond is controlled

COMPUTATIONAL NMR SPECTROSCOPY OF CARBOCATIONS

141

Fig. 6 100 MHz 13C NMR spectrum of 1-bis(trimethylsilyl)methyl-2-bis(trimethylsilyl) ethenyl cation (32) in SO2ClF/SO2F2, –100 1C, reference CD2Cl2 d ¼ 53.80 ppm. Inset: expanded region (between 2 and –0.5 ppm) showing the temperature dependence of the methine Si(CH3)3 signals between –130 and –100 1C.

by strong b-C–Si–s-bonds hyperconjugation of these silyl groups with the formally vacant p-orbital at the Ca–carbon. This causes temperature dependent kinetic line broadening of the signals of the CH(SiMe3)2-methyl groups (Fig. 6, inset). The experimentally determined energy barrier of 7.5 kcal/mol1 is in fair agreement with extrapolations from ab initio calculations of model cations 33–37 at different levels. IGLO/II NMR calculations for the model cations also account for the strong shielding effect. The calculations for model cations 33–36 with varied torsion angles for the b-silyl groups conﬁrmed nicely the dihedral dependence of the b-silyl effect. SiH3

H C C C H

SiH 3

33

SiH3

H

H

C C C H

SiH3

34

SiH3

H

H

C C C H

SiH 3

35

H

H C C C

H H

SiH3

SiH3

36

H3 Si C C C H 3Si

SiH3 H

SiH3

37

An unusual zwitterionic structure 38 formed by reaction of 1,4-di(tert-butyl) butadiyne with two equivalents of di(tert-butyl)aluminum hydride was investigated

142

H-U. SIEHL

by X-ray crystal structure determination, 1H, methods.56

13

C NMR and IR spectroscopic

(H3C) 3 C (H3C)3 C C1 (H3 C) 3 C

C C 2

Al

C4 H

3

C(CH3 )3

Al H

C(CH3 ) 3

C(CH3 ) 3

38

The array of unsaturated carbons C1–C4 resembles a butadien-2-yl cation but has a heavily distorted geometry. The formally dicoordinated carbon C2 has an unusual non-linear bond arrangement (1631). The Al–C1-bond has a favorable geometric arrangement for hyperconjugation with a formally vacant p-orbital at C2. Quantum chemical calculations of geometry, NBO charges and NMR chemical shifts were performed at B3LYP/6-31G(d) level of theory. The charges for the carbons C2 and C4, the formally allyl-resonance positions, are given as +0.15 and –0.14, respectively. The experimental and GIAO-B3LYP/6-31G(d) calculated 13C NMR chemical shifts for the carbons of the butadienyl moiety (d, exp./calc. ppm), 116.7/108 (C1); 161.3/ 160 (C2); 129.9/120 (C3); 156.0/140 (C4), deviate by 9, 1, 10 and 16 ppm, respectively. The good agreement for the formally positively charged C2 is probably fortuitous. A fast exchange process which could not be explained by the authors renders the 13C-signals for the t-butyl groups at the two vastly different Al-substituents equivalent at room temperature and can be slowed down at –30 1C.

4

Cycloalkyl cations

The cyclopentyl cation (39) undergoes a rapid degenerate rearrangement which can be frozen out at cryogenic temperatures as shown by solid state CPMAS 13C NMR spectra.57 MP2/6-31G(d,p) calculations show that cyclopentyl cation has a twisted conformation 4058 in which the axial hydrogens are bend toward the carbocation center. This is due to the pronounced geometrical distortion caused by the hyperconjugative interaction of the b-s–C–H-bond with the formally vacant 2pz-orbital at the C+ carbon of this secondary carbocation. H H H

39

40

41

COMPUTATIONAL NMR SPECTROSCOPY OF CARBOCATIONS

143

IGLO/DZ calculations of the 13C NMR chemical shifts reveal that the chemical shifts are very sensitive to small changes in geometry and that an MP2/6-31G(d,p) optimized geometry is required to obtain chemical shifts which are in agreement with the experimental data.6 Earlier IGLO/DZ calculations assuming a planar (C2v) structure3 or a symmetrically hydrido-bridged (Cs) structure 41 lead to larger differences. CH3

H3 C H3C

CH3 H3C

H3 C

CH3 CH3

CH3

42

The nonamethylcyclopentyl cation (42) also undergoes a fast dynamic equilibrium process caused by circumambulatory trans-migrating methyl groups which lead to fast averaging of all quaternary carbons and dynamic line broadening of the a-, b- and g-methyl signals.59,60 B3LYP/6-31G(d) calculations show the structure to be a trivalent tertiary carbocation with onset of bridging of the axial-oriented b-methyl groups toward the cationic carbon because of hyperconjugative b-C–C–s-bond–2pz–C+-orbital interactions. GIAO-DFT calculated 13C NMR chemical shifts empirically scaled using a linear correlation37 gave good agreement between the calculated averaged chemical shift for the ring carbons (108.5 ppm) and the experimental value (110.1 ppm). The two chair conformations of tertiary methylcyclohexyl cation 43 and 44 are in rapid equilibrium. CH 3

H H CH 3

H H 43

44

The two conformations are two stereoisomers where the formally vacant p-orbital is either equatorially (43) or axially (44) oriented, in hyperconjugative interaction in 43 with the b-C–C bonds or in 44 with the b-C–H bonds, respectively. Experimental dynamic 13C NMR measurements showed the C–C hyperconjugative isomer 43 to be marginally (1–3 kJ/mol) more stable.61 The interpretation of the experimental data was conﬁrmed by calculation of 13C NMR chemical shifts of 43 and 44.62 The calculated shifts in particular for those carbons involved in the hyperconjugative stabilization of the positive charge are critically dependent on the level of theory used. For example, the b-carbon shift for the C–C hyperconjugatively stabilized isomer 43 calculated using IGLO-HF or GIAO-HF methods deviate from experiment by –20 ppm whereas GIAO-MP2 methods with DZP/DZ (59.8 ppm) or TZP/DZ basis sets (62.2 ppm) show good agreement with the experimental estimate

144

H-U. SIEHL

of 60 ppm. Similar for the (2-propyl)-cyclohexyl cation b-C–C- and b-C–H–s-bond hyperconjugatively stabilized isomers 45 and 46 have been characterized by experimental and computational 13C NMR spectroscopic investigations.63 CH3 CH

H3C

H

H H

H

45

5

CH3 CH CH3

46

l-Hydrido-bridged carbocations

The experimentally observed 1H and 13C NMR chemical shifts of the m-hydridobridged64 cyclooctyl cation (47) were reproduced reasonably well by IGLO/DZ and IGLO/II NMR chemical shift calculations for MP2/6-31G(d) geometries (Cs symmetry) except for a large deviation of the m-hydrido-bridged hydrogen (calc. (IGLO/ II//MP2/6-31G(d)): –10.5 ppm, exp.: –7.7 ppm).65

H

H

H

47

48

49

CSGT-DFT NMR calculations on HF-SCF optimized geometries have been performed for cyclooctyl (47), cyclononyl (48) and cyclodecyl (48) cations and overall somewhat better agreement compared to previous IGLO-HF calculations was obtained.66 The chemical shift for the highly shielded m-hydrogen in 47 was calculated to be –8.9 ppm. The shift for the bridgehead carbons deviate, however, by about +10 ppm, demonstrating again that HF and DFT levels are not appropriate to model geometries and chemical shifts for hydrido-bridged carbocations.6,67 If electron correlation is taken into account using GIAO-MP2 method much better results were obtained including the shifts for the bridgehead carbons.68

H

H

50

51

A distinct upﬁeld shift of the bridging hydrogen was also calculated for the in-bicyclo[4.4.4]tetradecyl cation (50)69 using GIAO-HF methods, but the large discrepancy of about 6.5 ppm between theoretical and experimental d 1H values could

COMPUTATIONAL NMR SPECTROSCOPY OF CARBOCATIONS

145

not satisfactorily be explained at that time. A series of NMR calculations have been performed for 50, where the m-hydrogen was displaced in-plane and out-of-plane relative to the symmetric minimum on the Born–Oppenheimer energy surface.70 The chemical shift of the m-hydrogen did not change much. It was shown, however, that HF methods are not sufﬁcient for both the geometry optimization and the NMR calculations. GIAO-B3LYP/6-31G(d)//B3LYP/6-31G(d) could reproduce the large upﬁeld chemical shift of the bridging hydrogen. A large upﬁeld shift (–12 ppm) for the m-hydrogen was predicted for the experimentally elusive m-hydrogen-bridged bicyclo[3.3.3]undecyl cation (51). 2

H

52

CSGT-DFT NMR calculations were also performed for some other types of cyclic and acyclic hydrido-bridged cations and dications.71 Again, some discrepancies between experimental and theoretical results are apparent. For the triply-m-hydridobridged carbodication 52, a model for 4c–2e bonding unit, which however is not a minimum on the energy surface, a high-ﬁeld bridging hydrogen (–9 ppm) and a low-ﬁeld shift for the bridgehead carbons (223 ppm) was calculated.72

C H C H C

53

The unusual carbocation structure 53, a linear 5-center 4-electron C-H-C-H-C array can be derived from three anthracenes ‘joined up’ around a C-H-C-H-C core.73 The GIAO-DFT computed 1H and 13C NMR chemical shifts for the bridging hydrogens (2.9 ppm) and carbons (112 and 182 ppm) of the 5-center array differ considerably from those found in typical 3-center 2-electron systems.

6

Bicyclic and polycyclic carbocations

The cyclobutyl/cyclopropylmethyl cation system (C4H+ 7 ) has probably been the focus of more studies than any other carbocation system except the 2-norbornyl cation. Bridged cyclobutyl cations 16 are called bicyclobutonium ions. Bicyclobutonium

146

H-U. SIEHL

ions have a pentacoordinated g-carbon. The formally vacant p-orbital at the a-carbon of a formal secondary cyclobutyl cation interacts with the back-lobe of the endo-H at the g-carbon to form a transannular bond to the g-carbon as depicted in structure 54. exo

H

H

H

C endo

α

γ

H

H H

54

55

Another isomer on the energy hyper-surface of isomeric [C4H7]+ cations is the cyclopropylmethyl cation 15. The positive charge in a-cyclopropyl-substituted carbocations is stabilized by interaction of the ﬁlled antisymmetric Walsh orbital of the cyclopropyl ring (see Fig. 5) with the formally vacant p-orbital at the carbocation carbon. This overlap is maximal for the preferred bisected conformation of cyclopropylmethyl cations, as depicted for the parent cation 15 in structure 55. Computational investigations, including IGLO/DZ calculations, of isomers of the [C4H7]+ cation show that the bicyclobutonium structure 54 is preferred and the corresponding cyclopropylmethyl cation structure 55 is a minor isomer marginally higher in energy. Experimental NMR spectra of the parent [C4H7]+ cation74 are complicated because a low barrier rearrangement processes between 54 and 55 give raise to averaged peaks. The two structures, the bicyclobutonium ion 54 and the cyclopropylmethyl cation 55 are in fast equilibrium. A set of three symmetric bicyclobutonium ions interconverting via a set of three cyclopropylmethyl cations accounts for the rapid averaging of the methylene carbons observed in solution-state NMR studies (Scheme 3, RQH). Analysis of the temperature dependence of solution 13C NMR spectra75 comparison of the observed averaged chemical shifts with IGLO/DZ and GIAO-MP2 computed chemical shifts of the two isomers 54 and 55, solid state CPMAS 13C NMR spectra76 and equilibrium isotope effects77 conﬁrm the evidence for the presence of the two isomeric cations 54 and 55.6 R

R

H H R

R

R

R

H H

H H

Scheme 3 Rearrangement of [C4H6R]+ cations, for RQH a threefold degenerate interconversion of bicyclobutonium ion (54) and cyclopropylmethyl cations (55).

COMPUTATIONAL NMR SPECTROSCOPY OF CARBOCATIONS

147

The 1-methylbicyclobutonium ion (56) and the 1-(trimethylsilyl)bicyclobutonium ion (57) also undergo fast threefold methylene rearrangements (Scheme 3, RQCH3 or Si(CH3)3). H γ

CH 3

H

α

H

γ

Si(CH3) 3

α

H

57

56

CH3 H

Si(CH3 ) 3 H

H

H

59

58

However, contrary to the parent cation, the corresponding isomeric cyclopropylmethyl cation structures 58 and 59 are no minimum structures (MP2/6-31G(d)) and do not contribute to the averaged chemical shifts. H γ

SiH3 H

SiH3 α

H

H

60

61

The 1-silylcyclobutyl cation [1-SiH3-C4H6]+ (60), which serves as a model compound for [1-Si(CH3)3-C4H6]+ 57, has a hypercoordinated puckered 1-silylbicyclobutonium structure which is about 2.8 kcal/mol lower in energy than the (10 -silylcyclopropyl) methyl cation (61) which is a transition state (MP2/6-31G(d)).16,78 13 C NMR chemical shifts calculated (GIAO-MP2/TZP/DZ) for the MP2/6-31G(d) optimized geometry of the SiH3-substituted model cation 60 (Cb/b0 : 79.6 ppm; Cg: –14.5 ppm; ¼ 4 av. Cb/Cb0 /Cg chemical shift: 48.2 ppm; Ca: 133.9 ppm) are in good agreement with the experimental values for the 1-trimethylsilyl substituted bicyclobutonium ion (57) av. (Cb/Cb0 /Cg: 48.9 ppm; Ca: 137.4 ppm), whereas the chemical shifts calculated for the model silyl-substituted cyclopropylmethyl cation 61 are not in accord with experiment. 3-endo-trialkylsilylbicyclobutonium ions 62 (RQCH3 or Met2Bu) could be characterized by 1D 1H and 13C NMR spectroscopy including 2D NMR methods (H,H-COSY and H,C-COSY) as the ﬁrst static bicyclobutonium ions. H γ

H α

R3Si

62

R 3Si

γ

H α

H

63

The endo conformation was conﬁrmed by comparison with calculated 13C NMR chemical shifts (GIAO-MP2/TZP/DZ) of MP2/6-31G(d,p) optimized model

148

H-U. SIEHL

structures 3-endo- (62, RQH) and 3-exo-SiH3 (63, RQH) substituted bicyclobutonium ions. The assignment was also conﬁrmed by SOS-DFT (Perdew/ IGLO-III) calculation of the transannular 3J(Ha,Hg) spin–spin coupling constant which is 5.5 Hz measured experimentally and 5.9 Hz calculated for the endo-silyl isomer 62 (RQH) but is only 1.2 Hz calculated for the exo-silyl isomer 63 (RQH).78 The structure of the 2-norbornyl cation has been a focal point of controversy in physical organic chemistry. Experimental NMR spectroscopy and computational methods have been the decisive tools, favoring the hypercoordinated symmetric bridged structure 30, a protonated nortricyclane.79 The tricoordinated 2-norbornyl cation 31 is not a local minimum (MP2/6-31G(d)) on the energy surface.80

7

H

6

1

4

H 2

64

5

5

4

3

7

3

6 1

2

65

GIAO-MP2/TZP//MP2/6-31G(d) calculation of 13C NMR chemical shifts (C1, C2: 127.0 ppm; C3, C7: 40.5 ppm; C4: 41.5 ppm; C5: 23.5 ppm; C6: 24.0 ppm) agree within +4 ppm with the experimental chemical shifts measured at –158 1C (C1, C2: 124.8 ppm; C3, C7: 36.6 ppm; C4: 37.9 ppm; C5: 20.2 ppm; C6: 21.3 ppm).6 Similar 13C NMR and 1H NMR chemical shifts have also been calculated using GIAO-MP2/DZP//MP2/DZP. The spin–spin coupling constants calculated for the same geometry using an EOM-CCSD method24 are in accord with the experimentally measured ones. The 1J(C6,C1) and the 1J(C6,C2) coupling constants of –5.6 Hz (which are not experimentally available) are particular diagnostic for the nortricyclane framework of the hypercoordinated 2-norbornyl cation 64. 1J(C,C) coupling constants in most acyclic and cyclic organic compounds usually have positive sign and are in the range of 10–40 Hz. The small negative coupling constants calculated for the bridged structure 64 are characteristic for bicyclic molecules including a small-ring and are often attributed to interference between one-bond and two-bond couplings in this type of small-ring systems.81 NMR chemical shifts of 2-norbornyl cation related model structures, some with partly restricted geometries, have been calculated using the CSGT-B3LYP method.82 As expected for s-electron delocalized systems the chemical shift for carbons C1, C2 and C6 involved in the hypercoordinative bonding was found to be very sensitive toward small geometrical changes and vary greatly, whereas the shifts of carbons C3, C4, C5 and C7 are similar compared to the results of more appropriate electron correlation methods such as GIAO-MP2.

COMPUTATIONAL NMR SPECTROSCOPY OF CARBOCATIONS

149

2-Methyl-2-norbornyl (66) and 1,2-dimethyl-2-norbornyl (67) cations have partially bridged structures. 5 4 3

7

6

2

1

67

66

68

The IGLO/DZ and the SOS-DFPT-IGLO/BII calculated NMR chemical shifts using B3LYP/6-31G(d) optimized geometries agree well with the experimental values for 66 and 67.6 A strong dependence of calculated chemical shifts on the extent of bridging was found for the camphenyl cation (68) using GIAO-DFT calculations.83 For a B3LYP/6-31G(d) optimized geometry the 13C NMR chemical shift for the C2 (C+) carbon was 23 ppm too deshielded. Using the MP2/6-31G(d) optimized structure the shift was 44 ppm too shielded, whereas the GIAO-DFT(B3LYP)/ 6-31G(d)//MP4/6-31G(d) data gave the best agreement (266.8 ppm) with the experimental shift (261.9 ppm). R2S

69

Silanorbornylcations 69 were characterized by experimental and computational C and 29Si NMR chemical shifts. 29Si NMR chemical shift calculations at the GIAO-DFT(B3LYP)/6-311G(3d,p)//MP2/6-311G(d,p) level are in good agreement with the experimental data (Dd ¼ –3.8 to 2.7 ppm). The calculated 13C NMR chemical shifts deviate between +8.3 and 14.6 ppm.84

13

H

H

70

71

Similarly to the 2-norbornyl cation, comparison of calculated (IGLO/DZ//MP2/ 6-31G(d)) and experimental 13C NMR chemical shifts allowed to differentiate between the hypercoordinated 70 and the trivalent form 71 of the bicyclo[2.1.1.]hexyl cation.85 The experimental (157.8 ppm) and calculated (158.5 ppm) values for C1 and C2 (averaged signal) are reported to be nearly identical for the symmetrically bridged

150

H-U. SIEHL

structure 70, but do not correspond to the IGLO/DZ calculated chemical shift for the C+-carbon C2 in the classical structure 71 (207.4 ppm). On the basis of the calculated energies and the IGLO/DZ results, it was concluded that the bicyclo[2.1.1.]hexyl cation prefers the symmetrically bridged 70 over the trivalent structure 71. A series of hypercoordinated square-pyramidal carbocations were optimized at the MP2/6-31G(d) level and the 13C NMR chemical shifts of the carbocations were calculated using IGLO-HF and GIAO-MP2 methods.86

72

73

74

75

76

The parent square-pyramidal ion 72, the monomethyl-substituted analogs 73 and 74 with apical (73) and basal (74) substitution have not been observed experimentally. Good agreement between the calculated and experimental 13C NMR chemical shifts was obtained for the experimentally accessible 1,2-dimethyl-substituted analog 75 and the trimethyl-substituted analog 76. The GIAO-MP2/DZP-DZ calculated shift of the apical carbon (–17.3) for the bishomo square-pyramidal cations 77 agrees well with the experimental value of –17.2 ppm. 2+

78

77

The ﬁrst chemical application of the IGLO-HF method has been the correct description of the unexpectedly shielded 13C NMR chemical shifts of the 1,3-dehydroadamantane-5,7-diyl-dication 78.87

7

p-Stabilized carbocations

The cyclobutadienyl dication 79 (RQH) and the tetramethylcyclobutadienyl dication 79 (RQCH3) according to quantum chemical calculations at the MP4/ 6-31G(d)//HF/6-31G(d) level are predicted to have puckered structures which are more stable than the planar forms.88 R

R

2+

R R

79

The calculated chemical shifts for the ring carbons and the methyl carbons for the puckered geometry of the tetramethylcyclobutadienyl dication 79 (RQCH3)

COMPUTATIONAL NMR SPECTROSCOPY OF CARBOCATIONS

151

(209 and 18.7 ppm, respectively, IGLO/DZ) are nearly identical with the experimental chemical shifts (209.7 and 18.8 ppm). Arenium cations were the ﬁrst carbocations investigated by GIAO-MP2 NMR chemical shift calculations. H

H

80

81

82

A good agreement between experimental 13C NMR chemical shifts of the 2,4cyclohexadien-1-yl cation 80 and the calculated values at the GIAO-MP2/TZP-DZ// MP2/6-31G(d) level was obtained13 whereas earlier IGLO/DZ calculations for 80 and 81 deviated signiﬁcantly from experiment.89 On the basis of the observed 13C NMR chemical shifts, the total chemical shift difference criteria,90 IGLO/DZ chemical shift calculations and comparison with the model spiro[2.5]oct-4-yl cation (82), it was concluded that the phenonium ion 81 is a spirocyclopropylbenzenium ion.91 A linear correlation between calculated NPA charges and 13C NMR chemical shifts was found for 80 and some methyl- and triﬂuoromethyl substituted homologs.92 IGLO-HF 29Si NMR studies of silylated arenium ions 83 and 84 and comparison with experimental data have been reported.93,94

H

SiR3

H

SiR3

CH 3

83

84

Me2 Si

SiMe2

R2

R3

R4

R5 R1

85

The results indicate that the formation of long-lived trimethyl substituted silyl cations, in the presence of aromatic solvents, as claimed by Lambert et al.95 is not feasible under these conditions. Persistent silicenium ions require sterically more shielding substituents at silicon or hypercoordinative stabilization.96–98 13C and 29Si NMR chemical shifts were calculated for a series of disilylated arenium ions 85 using density functional theory (DFT). The calculations predict consistently the unsaturated carbon atoms to be too deshielded by 8–15 ppm. Applying an empirical correction, the deviation between experiment and theory was reduced to –0.4 to 9 ppm, and the 13C NMR chemical shift of the highly diagnostic Cipso is reproduced by the calculations (Dd ¼ –3.8 to 2.7 ppm).99 b-Silyl substituted benzyl cations have been prepared under superacid conditions in solution. Below –120 1C the interconversion of the anti- (86) and syn- (87)

152

H-U. SIEHL

Fig. 7 100 MHz 13C NMR of anti- (86) and syn- (87) 1-p-anisyl-2-triisopropyl-silyl-ethylcation, SO2ClF/SO2F2, –126 1C, ext. capill. Ref.: TMS (d ¼ 0.0 ppm) in CD2Cl2/SO2ClF.

1-p-anisyl-2-triisopropylsilyl-ethylcation is slow on the NMR time scale (Fig. 7).100 Two sets of four lines are observed, one for the C2 and C6 ortho and C3 and C5 meta carbons in the anti isomer 86 and another set with lower intensity for the C20 and C60 ortho and C30 and C50 meta carbons in the syn isomer 87. At higher temperatures kinetic line broadening and coalescence is observed (Fig. 8). Upon warming above the coalescence temperature of about –110 1C four lines for the aromatic methine carbons with decreasing line width are observable until decomposition takes place at about –70 1C (Fig. 8). From line shape analysis the energy barrier for the isomerization process is obtained as DG# ¼ 7.5 kcal/mol.101 H

β

α

H

SiiPr3

β

α

1

SiiPr 3

1´

6 5

2

6´

3

5´

2´ 3´ 4´

4

O

O

H3C

CH3

86

87

NMR chemical shift calculations were performed for the B3LYP/6-31G(d) optimized anti and syn isomeric structures of the analogous 1-p-anisyl-2-SiH3 substituted

COMPUTATIONAL NMR SPECTROSCOPY OF CARBOCATIONS

153

Fig. 8 Temperature dependence –141 1C (bottom trace) to –71 1C (top trace) of 100 MHz 13C NMR signals of the ortho and meta carbons of anti- (86) and syn- (87) 1-p-anisyl-2triisopropylsilyl-ethylcation in SO2ClF/SO2F2.

ethylcation which serve as computational models for the experimentally observed 2-triisopropylsilyl-substituted cations 86 and 87. B3LYP/6-31G(d) calculated NMR chemical shifts for both the anti and syn isomer generally show deviations of more than 10 ppm and up to about 20 ppm for the SiR3 substituted b-carbon (calc. RQH, exp. RQCH(Me2)) compared to the experimental values. The deviations from experiment are smaller (o8 ppm) when the 6-31G(d,p) basis set is used. SOS-DFT IGLO/III using the Perdew–Wang100 functional yields the smallest deviations o6 ppm (excluding the b-carbon chemical shift). The calculated relative differences between the chemical shifts for the two isomers allow, however, unequivocal assignment of the experimental signals even at the lowest level of the calculations. Substituted benzylic mono- and dications (88 and 89) were investigated by 1H and 13 C NMR spectroscopy and IGLO-DFT calculations.102 CH2 H3C

CH 2 CH3

H3 C

R1

R3

R1

R3

R2

CH3 CH3

H2C

CH2

H3C

R

R2

88

89

154

H-U. SIEHL

The results suggest that chinoid type structures are the predominant resonance contributors for 88. The IGLO/DZ//3-21G calculated 13C NMR chemical shifts of benzylic monocations 88 correlate reasonably well with the experimentally obtained data. The 13C NMR chemical shifts of the carbocation centers (CH2 carbon) are calculated 10.6–12.5 ppm too deshielded. Similar results were obtained for benzylic dications 89. NMR chemical shifts of arenium ions derived from various classes of polycyclic aromatic hydrocarbons have been calculated using GIAO-DFT methods.103 In conjunction with the evaluation of ring current effects and NICS values (Nuclei Independent Chemical Shift), calculations of 1H and 13C NMR chemical shifts for a series of ﬂuorenylidene dications were performed.104 For some homoaromatic carbocations the NICS values and chemical shifts have been calculated.105,106 IGLO-HF and GIAO-MP2 calculated 13C NMR chemical shifts for bishomoaromatic 7-norbornenyl 90 and 7-norbornadienyl cation 91 have also been reported.107

90

91

The cyclobutenyl cation (92) and the homotropylium cation, C8H+ 9 93 are both prototypes of homoaromatic systems. CH2 CH2 CH2

92

93

94

The cyclobutenyl cation 92 is one of the ﬁrst examples demonstrating that electron correlation is required both for geometry optimization and NMR chemical shift calculations.14 The IGLO/6-31G(d,p) calculated 13C NMR chemical shifts of the planar form of a homotropylium cation 94 clearly deviate from the experimental values (mean deviation D ¼ 45.6 ppm) alternating in the seven-membered ring between 122 and 194 ppm, whereas those of the non-planar structure for the homotropylium cation 93 are in good agreement with experiment (mean deviation D ¼ 6.2 ppm).108 DFT-calculated 13C NMR chemical shifts for the pentamethylcyclopentadienyl cation and the (CH3)5-cyclopentenyl cation 97 provide conclusive evidence that the

COMPUTATIONAL NMR SPECTROSCOPY OF CARBOCATIONS

155

alleged synthesis and isolation of the potentially antiaromatic (CH3)5-cyclopentadienyl cation was not achieved.109

15 5 ppm

282 p pm 144 ppm

1 44 ppm 191 ppm

204 ppm

283 ppm

258 ppm

16 5 pp m

177 ppm

15 6 ppm

251 ppm

2 53 pp m

H H 6 3 ppm

63 ppm

95

96

97

Instead the allyl cation 97 was obtained. The reported experimental 13C NMR data (d13C(exp): d ¼ 250/243, 153, 60 ppm) are in agreement with the cyclopentenyl cation structure 97, but differ signiﬁcantly from the calculated 13C NMR chemical shift for both Jahn-Teller distorted valence isomeric cyclopentadienyl cation structures 95 and 96. Carbocations have been generated in zeolites and were characterized by CPMAS 13 C NMR spectroscopy and accompanying quantum chemical calculations.110 13 C NMR isotropic shifts calculated at the GIAO-MP2/TZP/DZ level for the 1, 3-dimethylcyclopentenyl (98) and 1,2,3-trimethylcyclopentenyl cations (99)111 are in agreement with the observed experimental spectra of the carbocations in the zeolite.112

98

99

100

In both cases the calculated values are downﬁeld of the experimental shifts (by 5 ppm). It was suggested that the presence of the zeolite and motional averaging has some effect on the chemical shifts. The heptamethylcyclopentenyl cation 100 was also characterized on zeolite.113 13C NMR chemical shifts were calculated using GIAO-MP2 for DFT optimized geometries. Discrepancies between experimental and calculated chemical shift of up to 17 ppm were suggested to result from geometrical distortions of the carbocation structure in the zeolite cage. Other carbocations on zeolites, such as arenium114 and 1,3-arylsubstituted allyl cations115 have also been investigated by experimental and computational NMR methods.

156

8

H-U. SIEHL

Heteroatom stabilized carbocations

The experimental 13C NMR chemical shift of the simplest member of the acylium ions, the formyl cation [HCO]+ was reported as d ¼ 139.5 (measured under CO pressure of 85 atm) which compares well with the GIAO-MP2 calculated shift of d ¼ 136 ppm.116,117 The analogous ﬂuoroformyl cation [FCO]+ and protonated ﬂuoroformic acid [FC(OH)2]+ were characterized by 13C NMR spectroscopy, experimentally118 (d ¼ 117.5 and 157.8 ppm, respectively) as well as computationally (d ¼ 118.6 and 170.6 ppm, respectively).119 A systematic study of experimental and calculated 13C and 17O NMR chemical shifts of oxonium, carboxonium120 triﬂuoromethyl-121 and trimethylsilyl-122 substituted carboxonium ions and corresponding acyl cations123,124 showed GIAO-MP2 to be superior to IGLO-HF and GIAO-HF chemical shift calculations. Whereas for 13C NMR shifts TZP/DZ basis sets are sufﬁcient, larger basis set such as QZ2P/QZ2P are required for an accurate prediction of 17O NMR chemical shifts. The principal components of the 13C chemical shift tensors for some acylium cations were determined by slow speed MAS 13C NMR spectroscopy and quantum chemical methods.125 A good agreement between the theoretical prediction and the experimental tensor components of the acylium ions was found. Computational studies of 13C and 19F NMR chemical shifts of a number of ﬂuorocarbocations126 and other halomethyl cations and their protonated forms have been studied experimentally and computationally.127 It was shown by spin–orbit corrected IGLO-DFT calculations that for halomethyl cation CX3 (X ¼ F, Cl, Br, I) the experimentally observed decrease in the 13C NMR chemical shifts with increasing atomic number of the halogen is due to spin–orbit coupling.128 Previous explanations for this example of a ‘normal halogen dependence’ via electronegativity arguments were based on fortuitous empirical correlations. Comparison between IGLO-DFT, GIAO-HF and GIAO-MP2 results for these systems also demonstrates an increasing importance of electron correlation along the series. The 13C, 15N and 17O NMR chemical shifts of some substituted methyl cations and the corresponding protonated dications were calculated by the GIAO-MP2 method for MP2/6-31G(d) optimized geometries.129 The o-, m- and p-phenylenebis(1,3-dioxolanium) dications 101 and the related tris(1,3-dioxolanium) trication have been prepared and the calculated 13C and 17O NMR chemical shifts (GIAO-DFT) closely match the experimental values.130 H O

O

H

O C

H

H O

O C

H

C C

O

C

H O

101

102

103

104

O H

COMPUTATIONAL NMR SPECTROSCOPY OF CARBOCATIONS

157

The calculated 13C NMR chemical shift of the carbonyl carbon of monoprotonated benzaldehyde131,132 for the E-form 102 (205.5 ppm) and that for the Z-form 103 (207.4 ppm) agree well with the experimental shifts of 203.5 and 205.9 ppm, respectively. Protonation of a-substituted cinnamic acids such as 104 was studied by 13C NMR spectroscopy and IGLO-HF calculations.133 Protonated deltic acid (105) and related compounds,134,135 as well as protonated urea 106 (X ¼ O)136 and thiourea 106 (X ¼ S)137 have been investigated by 13C NMR spectroscopy and quantum chemical calculations.138 OH

HO

H2 N

OH

105

XH

OH

C

S

NH2

H3 C

106

H H3C

CH3

107

O S CH3

108

GIAO-MP2 calculated NMR chemical shifts for DFT optimized geometries and comparison with experimental data were used to study the site of protonation of dimethyl sulfoxide.139 The calculated 13C NMR chemical shift of O-protonated DMSO 107 (40.0 ppm) matches with the experimental value of 34.3 ppm. The calculated 13C NMR chemical shift of S-protonated DMSO 108 is 3 ppm deshielded compared to that calculated for 107. NMR chemical shifts of protonated carbonic acid 109 and 110 were investigated experimentally and computationally using IGLO/II chemical shift calculations for MP2/6-31G(d) optimized geometries.140 H O C

H

O

H

O

C

O

H O

H

109

2

H O H

110

The data indicate that in strong acidic solution H2CO3 may be in equilibrium with protonated carbonic acid. The structures of carbamic acid and its O- and N-protonated forms 111 and 112 were calculated at the MP2/6-31G(d) level.140 OH

O

C H2N

OH

111

C H3 N

OH

112

The calculated 13C and 15N NMR chemical shifts, using the IGLO-HF and GIAO-MP2 methods, however, deviate substantially from the experimental results. This is in contrast with the previously reported results on protonated guanidines 113

158

H-U. SIEHL

and 114141 where GIAO-MP2 calculated 13C and found to be in agreement with the experiment. NH2 H2N

C

15

2

NH3

NH2

H2 N

113

C

N NMR chemical shifts were

NH2

114

13

C and 15N NMR chemical shifts were calculated for protonated diazomethane [CH3NN]+ and the cyanodiazonium ion [NCNN]+.142 For the C-protonated diazomethane, which is more stable than N-protonated species, the IGLO-HF calculated 13C NMR chemical shift of 54.1 ppm agrees with the experimental value of 44.5 ppm. 15N NMR chemical shifts were calculated as 237.9 and 379.9 ppm. The 1H, 6Li and 13C NMR chemical shifts of the cyclopropenium cation and its lithium derivatives 115–118 were calculated both at GIAO-DFT and GIAO-HF level using B3LYP optimized geometries.143 Li

H

Li

Li Li

Li

115

H

H

116

Li

Li

117

Li Li

118

It was shown that lithium is even more effective than the amino group in stabilizing the ethyl, vinyl, allenyl and cyclopropenium cations.144

9

Final remarks

Calculation of NMR parameters such as chemical shift and spin–spin coupling constants has evolved into an important tool in carbocation chemistry as well as in related ﬁelds such as silylenium ion and borane chemistry which are not considered in this chapter. The GIAO method for calculation of magnetic parameters gives in many cases better results and has nowadays replaced the earlier IGLO method. The approach of combined quantum chemical calculation of structures and NMR chemical shifts has proven to be more than only a substitute for unavailable X-ray structures in carbocation chemistry. The performance of various theoretical models and computational methods has been extensively evaluated. Applications of quantum chemical calculation of NMR spin–spin coupling constants145 in carbocations are nowadays still somewhat limited but will become more important in the future. Adequately calculated geometries are essential for chemical shift calculations of carbocations. Carbocations that are stabilized by s-C–H- or s-C–C-hyperconjugation or by hypercoordination (‘bridging’) have bond lengths and bond angles, which differ signiﬁcantly from conventional bonding schemes. Geometry optimization of these

COMPUTATIONAL NMR SPECTROSCOPY OF CARBOCATIONS

159

type of carbocations generally requires 6-31G(d) or higher basis sets for describing the one-electron space. MP2-pertubation theory is the recommended method in the wave-function approach for modeling the n-electron space taking into account electron correlation in carbocations, which enjoy signiﬁcant hyperconjugation or hypercoordination. The disregard or the non-biased treatment of electron correlation in ab initio NMR chemical shift calculations leads in difﬁcult cases to unsatisfying or even erroneous results. Some reports of alleged perfect congruence between experimental and computed NMR chemical shifts might well be serendipitous results due to fortuitous error cancellation: The right answer for the wrong reason. Calculated relative NMR shifts are magnetic shielding differences and are prone to error cancellation effects. DFT methods using for example the Becke-3-parameter-Lee-YangParr hybrid functional (B3LYP) with 6-31G(d) or higher basis set for GIAO NMR chemical shift calculations of carbocations are cheap compared to wave-function methods, which consider electron correlation such as GIAO MP2 or GIAO coupled cluster methods. DFT quite often give fair and satisfactory results for a great structural variety of carbocation but are not always reliable. DFT methods lack the strong hierarchy of levels of increasing complexity and accuracy present in wave-function methods. Extrapolation and systematic improvement of GIAO DFT results is difﬁcult. Density functional methods show frequently large deviations of calculated NMR chemical shifts for atoms at particular positions in a carbocation. The fact that DFT methods work well for most carbon atoms in a carbocation while do poorly for a few others is confusing only at the ﬁrst sight. A systematic investigation often provides more detailed information on particular electronic and structural features due to hyperconjugative and hypercoordinative stabilization modes of carbocations. Care should be taken, if the object under study might have ‘unusual’ electron distribution. Those atoms in carbocations which are involved in hyperconjugation and hypercoordination are part of 3c–2e or higher multiple center two electron bonds. These bonding situations generally are not adequately described by DFT methods. The GIAO-MP2 method has convincingly demonstrated the importance of electron correlation effects in NMR chemical shift calculations of carbocations and generally leads to good or excellent agreement with experiment. However, GIAO-MP2 in certain cases tends to overestimate electron correlation effects on the absolute shielding constants.146 The GIAO coupled cluster schemes developed by Gauss and coworkers provide magnetic shielding data with quantitative accuracy but at very high computationally costs. CCSD(T) NMR calculations have been used extensively for benchmark calculations147 and very successfully for some challenging 13C NMR chemical shift problems in carbocation chemistry.49,53 It should be kept in mind that quantum chemical calculations of structures and magnetic properties generally are done for the isolated carbocation without taking into account its environment and media effects such as solvent, site-speciﬁc solvation or counterion effects. This is a critical question since NMR spectra of carbocations with a few exceptions are studied in superacid solutions and properties calculated for the gas-phase species are of little relevance if the electronic structure of carbocations is strongly perturbed by solvent effects. Provided that appropriate methods are used,

160

H-U. SIEHL

and that the results for the geometry optimization and shielding tensors are sufﬁciently converged with respect to wave-function model and basis set, the congruence of calculated and observed NMR chemical shifts suggests that the geometry adopted by a carbocation in superacid solution is similar to the gas-phase prediction and unperturbed by interaction with the environment.

10 Conclusions The interplay between theory and experiment is especially successful in the ﬁeld of reactive intermediates such as carbocations. The accurate predictions of the spectroscopic properties of carbocations by modern quantum theory facilitate their experimental characterization and identiﬁcation. Quantum chemical calculations of NMR parameters have provided detailed information on structures and stabilization modes of carbocations and have provided deeper insights into intriguing questions and long-standing controversies in carbocation chemistry. The conundrum of the structure of the 2-norbornyl cation148 has been solved with the help of ‘computational’ NMR spectroscopy. Contemporary experimental NMR spectroscopy and quantum chemical methodology in conjunction with the application of the b-silyl effect ﬁnally could solve the cyclopropylmethyl/bicyclobutonium cation (C4H+ 7 ) problem which has been enigmatic for about half a century.148 The everlasting need for better experimental tools and theoretical methods to explore the ﬁeld of carbocation chemistry had also a signiﬁcant inﬂuence on the further development of ab initio methods for the calculation of NMR shielding and indirect spin–spin coupling constants. Carbocation chemistry thus serves as a forerunner for a close integration of experimental and computational approaches in all areas of chemistry.

Acknowledgments The author acknowledges the contributions of research collaborators mentioned in the references and ﬁnancial support by the Deutsche Forschungsgemeinschaft (DFG), the Fonds der Chemischen Industrie, the Alexander von Humboldt Foundation (AvH-Stiftung) and the Deutsche Akademische Austauschdienst (DAAD). Discussions and cooperation with Dr. Ju¨rgen Gauss (University of Mainz) and Olga Malkina and Vladimir G. Malkin (Bratislava) are also gratefully acknowledged.

References 1. Helgaker, T., Jaszunski, M. and Ruud, K. (1999). Ab initio methods for the calculation of NMR shielding and indirect spin-spin coupling constants. Chem. Rev. 99, 293–352 2. Kutzelnigg, W., Fleischer, U. and Schindler, M. (1990). In NMR Basic Principles and Progress, Vol. 23, pp. 165. Springer-Verlag, Berlin Heidelberg 3. Schindler, M. (1987). J. Am. Chem. Soc. 109, 1020

COMPUTATIONAL NMR SPECTROSCOPY OF CARBOCATIONS

161

4. Siehl, H.-U. and Vrcek, V. (2004). Calculation of NMR parameters in carbocation chemistry. In Calculation of NMR and EPR Parameters, Theory and Applications, Kaupp, M., Bu¨hl, M. and Malkin, V.G. (eds), Chapter 23, pp. 373–397. Wiley-VCH, Weinheim, Germany 5. Buzek, P., Sieber, S. and Schleyer, P.v.R. (1992). Chem. Uns. Zeit 26, 116 6. Schleyer, P.v.R., Maerker, C., Buzek, P. and Sieber, S. (1997). In Stable Carbocation Chemistry, Prakash, G.K.S. and Schleyer, P.v.R. (eds), Chapter 2 and references therein. John Wiley & Sons Inc., New York 7. Bu¨hl, M. (1998). In Encyclopedia of Computational Chemistry, Schleyer, P.v.R., Schreiner, P.R., Allinger, N.L., Clark, T., Gasteiger, J., Kollmann, P. and Schaefer, H.F. III (eds), p. 1835. Wiley, Chichester 8. Schleyer, P.v.R., Schreiner, P.R., Allinger, N.L., Clark, T., Gasteiger, J., Kollmann, P. and Schaefer, H.F.III (eds) (1998). Encyclopedia of Computational Chemistry. Wiley, Chichester 9. Kaupp, M., Bu¨hl, M. and Malkin, V.G. (eds), (2004). Calculation of NMR and EPR parameters, theory and applications. Part A: Introductory Chapters 1–5; Part B NMR Parameters, Methodological Aspects, Chapters 6–20. Wiley-VCH, Weinheim, Germany 10. (a) Brown, R.D., Boggs, J.E., Hilderbrandt, R., Lim, K., Mills, I.M., Nikitin, E. and Palmer, M.H. (1996). Pure Appl. Chem. 68, 387; (b) For an electronic version see the IUPAC web site: www.iupac.org/reports/1996/6802brown/index.html 11. Wolinski, K., Hinton, J.F. and Pulay, P. (1990). J. Am. Chem. Soc. 112, 8251 12. Gauss, J. and Stanton, J.F. (2004). In Kaupp, M., Bu¨hl, M. and Malkin, V.G. (eds), Calculation of NMR and EPR Parameters, Theory and Applications. Electron-Correlated Methods for the Calculation of NMR Chemical Shifts. Chapters 8, pp. 123–140. WileyVCH, Weinheim, Germany 13. Sieber, S., Schleyer, P.v.R. and Gauss, J. (1993). J. Am. Chem. Soc. 115, 6987 14. Sieber, S., Schleyer, P.v.R., Otto, A.H., Gauss, J., Reichel, F. and Cremer, D. (1993). J. Phys. Org. Chem. 6, 445 15. Buzek, P., Schleyer, P.v.R., Vancik, H., Mihalic, Z. and Gauss, J. (1994). Angew. Chem. 104, 470 16. Siehl, H.-U., FuX, M. and Gauss, J. (1995). J. Am. Chem. Soc. 117, 5983 17. (a) Schleyer, P.v.R., Gauss, J., Bu¨hl, M., Greatrex, R.A. and Fox, M.A. (1993). J. Chem. Soc. Chem. Commun. 1766; (b) Bu¨hl, M., Gauss, J., Hofmann, M. and Schleyer, P.v.R. (1993). J. Am. Chem. Soc. 115, 12385 18. (a) Marx, D. and Savin, A. (1997). Angew. Chem. Int. Ed. 36, 2077; (b) Marx, D. and Parrinello, M. (1999). Science 284, 59 (and references therein); (c) Schreiner, P.R. (2000). Angew. Chem. Int. Ed. 39, 3239 (and references therein); (d) Klopper, W. and Kutzelnigg, W. (1990). J. Chem. Phys. 94, 5256 19. Rasul, G., Prakash, G.K.S. and Olah, G.A. (1998). Proc. Natl. Acad. Sci. USA 95, 7257 20. Schleyer, P.v.R., Lenoir, D., Mison, P., Liang, G., Prakash, G.K.S. and Olah, G.A. (1980). J. Am. Chem. Soc. 102, 683 21. Olah, G.A. and Rasul, G. (1996). J. Am. Chem. Soc. 118, 8503 22. Rasul, G., Prakash, G.K.S. and Olah, G.A. (2002). Proc. Natl. Acad. Sci. USA 99, 9635 23. Perera, S.A., Bartlett, R.J. and Schleyer, P.v.R. (1995). J. Am. Chem. Soc. 117, 8476 24. Perera, S.A., Sekino, H. and Bartlett, R.J. (1994). J. Chem. Phys. 101, 2186 25. Koch, W., Schleyer, P.v.R., Buzek, P. and Liu, B. (1992). Croat. Chim. Acta 65, 655 26. Schleyer, P.v.R., Koch, W., Liu, B. and Fleischer, U. (1989). J. Chem. Soc. Chem. Commun. 1098 27. Nicholas, J.B., Xu, T., Barich, D.H., Torres, P.D. and Haw, J.F. (1996). J. Am. Chem. Soc. 118, 4202 28. Farcasiu, D., Haˆncu, D. and Haw, J.F. (1998). J. Phys. Chem. A 102, 2493

162

H-U. SIEHL

29. Sieber, S., Buzek, P., Schleyer, P.v.R., Koch, W. and Carneiro, J.W. de M. (1993). J. Am. Chem. Soc. 115, 259 30. Olah, G.A., Hartz, N., Rasul, G. and Prakash, G.K.S. (1993). J. Am. Chem. Soc. 115, 6985 31. H.-U. Siehl, unpublished results; presented in a part at the 6th WATOC, Lugano, 2002, August 4–9 and the International Conference on Reaction Intermediates and Reaction Mechanism, Ascona, 2002, July 7–12 32. Freudenberger, C. (2003). Ph.D. Thesis, University of Ulm 33. Saunders, M., Hagen, E.L. and Rosenfeld, J. (1968). J. Am. Chem. Soc. 90, 6882 34. Carneiro, J.W. de M., Schleyer, P.v.R., Koch, W. and Raghavachari, K. (1990). J. Am. Chem. Soc. 112, 4064 35. Sieber, S., Buzek, P., Schleyer, P.v.R., Koch, W. and Carneiro, J.W. de M. (1993). J. Am. Chem. Soc. 115, 259 36. Schleyer, P.v.R., Carneiro, J.W. de M., Koch, W. and Forsyth, D.A. (1991). J. Am. Chem. Soc. 113, 3990 37. Vrcek, V., Kronja, O. and Siehl, H.-U. (1999). J. Chem. Soc. Perkin Trans. 2, 1317 38. Gauss, J. (2002). J. Chem. Phys. 116, 4773 39. Vrcek, V. and Siehl, H.-U. Unpublished results 40. Olah, G.A., Kelly, D.P. and Jeuell, C.L. (1970). J. Am. Chem. Soc. 92, 2544 41. Falkenberg-Andersen, C., Ranganayakulu, K., Schmitz, L.R. and Sorensen, T.S. (1984). J. Am. Chem. Soc. 106, 178 42. Backes, A.C. (1999). Ph.D. Thesis, University of Ulm 43. Malkin, V.G., Malkina, O.L., Casida, M.E. and Salahub, D.R. (1994). J. Am. Chem. Soc. 116, 5898 44. Siehl, H.-U., Backes, A.C. and Malkina, O. (1999). Organosilicon Chem. IV, 150–158 45. Prakash, G.K.S., Reddy, V.P., Rasul, G., Casanova, J. and Olah, G.A. (1998). J. Am. Chem. Soc. 120, 13362 46. (a) Crofton, M.W., Jagod, M.-F., Rehfuss, B.D. and Oka, T. (1989). J. Chem. Phys. 91, 5139; (b) Curtiss, L.A. and Pople, J.A. (1988). J. Chem. Phys. 88, 7405; (c) Liang, C., Hamilton, T.P. and Schaefer III, H.F. (1990). J. Chem. Phys. 92, 3653 47. Stanton, J.F., Gauss, J. and Siehl, H.-U. (1996). Chem. Phys. Lett. 262, 183–186 48. Gauss, J. and Stanton, J.F. (1997). J. Mol. Struct. (Theochem.) 398–399, 73–80 49. Siehl, H.-U., Mu¨ller, T. and Gauss, J. (2003). J. Phys. Org. Chem. 16, 577 50. Mu¨ller, T. (1993). Ph.D. Thesis, University of Tu¨bingen 51. Siehl, H.-U. and Brixner, S. (2004). J. Phys. Org. Chem. 17, 1039 52. Siehl, H.-U., Mu¨ller, T., Gauss, J., Buzek, P. and Schleyer, P.v.R. (1994). J. Am. Chem. Soc. 116, 6384 53. Mu¨ller, T., Lennartz, R. and Siehl, H.-U. (2000). Angew. Chem. 112, 3203–3206; Angew. Chem. Int. Ed. Engl. 39, 3074–3077 54. Krossing, I. and Raabe, I. (2004). Angew. Chem. 116, 2116; Angew. Chem. Int. Ed. 43, 2066 55. Siehl, H.-U., Kaufmann, F.P., Apeloig, Y., Braude, V., Danovich, D., Berndt, A. and Stamatis, N. (1991). Angew. Chem. 103, 1546; Angew. Chem. Int. Ed. 30, 1479 56. Uhl, W., Grunenberg, J., Hepp, A., Matar, M. and Vinogradov, A. (2006). Angew. Chem. Int. Ed. 45, 4358 57. Myhre, P.C., Kruger, J.D., Hammond, B.L., Lok, S.M., Yannoni, C.S., Macho, V., Limbach, H.H. and Vieth, H.M. (1984). J. Am. Chem. Soc. 106, 6079 58. (a) Schleyer, P.v.R., Carneiro, J.W. de M., Koch, W. and Raghavachari, K. (1989). J. Am. Chem. Soc. 111, 5475; (b) Botkin, J.H., Forsyth, D.A. and Sardella, D.J. (1986). J. Am. Chem. Soc. 108, 2797 59. Mayr, H. and Koschinsky, R. (1989). J. Am. Chem. Soc. 111, 2305 60. Kronja, O., Koehli, T.-P., Mayr, H. and Saunders, M. (2000). J. Am. Chem. Soc. 122, 8067

COMPUTATIONAL NMR SPECTROSCOPY OF CARBOCATIONS

163

61. (a) Kirchen, R.P., Ranganayakulu, K. and Sorensen, T.S. (1987). J. Am. Chem. Soc. 109, 7811; (b) Kirchen, R.P. and Sorensen, T.S. (1978). J. Am. Chem. Soc. 100, 1487 62. (a) Rauk, A., Sorensen, T.S., Maerker, C., Carneiro, J.W. de M., Sieber, S. and Schleyer, P.v.R. (1996). J. Am. Chem. Soc. 118, 3761; (b) Rauk, A., Sorensen, T.S. and Schleyer, P.v.R. (2001). J. Chem. Soc. Perkin Trans. 2, 869 63. Vrcek, V., Kronja, O. and Siehl, H.-U. (2002). J. Chem. Soc. Perkin Trans. 2, 106 64. Kirchen, R.P., Sorensen, T.S. and Wagstaff, K. (1978). J. Am. Chem. Soc. 100, 6761 65. Buzek, P., Schleyer, P.v.R., Vancik, H. and Sunko, D.S. (1991). J. Chem. Soc. Chem. Commun. 1538 66. Galasso, V. (1998). Int. J. Quant. Chem. 70, 313 67. (a) Vinkovic-Vrcek, I., Vrcek, V. and Siehl, H.-U. (2002). J. Phys. Chem. A 106, 1604; (b) Buzek, P., Schleyer, P.v.R., Sieber, S., Koch, W., Carneiro, J.W. de M., Vancik, H. and Sunko, D.S. (1991). J. Chem. Soc. Chem. Commun. 671 68. Siehl, H.-U. and Vrcek, V. Unpublished results 69. Cioslowski, J. (1993). J. Am. Chem. Soc. 115, 5177 70. Taeschler, C. and Sorensen, T.S. (2000). J. Mol. Model. 6, 217 71. Gallaso, V. (1999). Chem. Phys. 241, 247 72. Sun, F. and Sorensen, T.S. (1993). J. Am. Chem. Soc. 115, 77 73. Tantillo, D.J. and Hoffmann, R. (2003). J. Am. Chem. Soc. 125, 4042 74. Staral, J.S., Yavari, I., Roberts, J.D., Prakash, G.K.S., Donovan, D.J. and Olah, G.A. (1978). J. Am. Chem. Soc. 100, 8016 75. (a) Saunders, M., Laidig, K.E., Wiberg, K.B. and Schleyer, P.v.R. (1988). J. Am. Chem. Soc. 110, 7625; (b) Koch, W., Liu, B. and DeFrees, D.J. (1988). J. Am. Chem. Soc. 110, 7325; (c) McKee, M.L. (1986). J. Phys. Chem. 90, 4908 76. Myhre, P.C., Webb, G. and Yannoni, C.S. (1990). J. Am. Chem. Soc. 112, 8992 77. (a) Saunders, M. and Siehl, H.-U. (1980). J. Am. Chem. Soc. 102, 6868; (b) Brittain, W.J., Squillacote, M.E. and Roberts, J.D. (1984). J. Am. Chem. Soc. 106, 7280 78. FuX, M. (1997). Ph.D. Thesis, University of Tu¨bingen 79. (a) Saunders, M. and Kates, M.H. (1980). J. Am. Chem. Soc. 102, 6867; (b) Koch, C.W., Liu, B., de Frees, D.J., Vancik, H. and Sunko, D.E. (1990). Angew. Chem. Int. Ed. 19, 183; (c) Myhre, P.C., Webb, G.G. and Yannoni, C.S. (1990). J. Am. Chem. Soc. 112, 8991 80. Schleyer, P.v.R. and Sieber, S. (1993). Angew. Chem. Int. Ed. 32, 1606 81. Perera, S.A. and Bartlett, R.J. (1996). J. Am. Chem. Soc. 118, 7849 82. Werstiuk, N.H. and Muchall, H.M. (2000). J. Phys. Chem. A 104, 2054 83. Brunelle, P., Sorensen, T.S. and Taeschler, C. (2001). J. Org. Chem. 66, 7294 84. Mu¨ller, T., Bauch, C., Ostermeier, M., Bolte, M. and Auner, N. (2003). J. Am. Chem. Soc. 125, 2158 85. Schleyer, P.v.R., Laidig, K., Wiberg, K.B., Saunders, M. and Schindler, M. (1988). J. Am. Chem. Soc. 110, 300 86. Prakash, G.K.S., Rasul, G. and Olah, G.A. (1998). J. Phys. Chem. A 102, 2579 87. Bremer, M., Schleyer, P.v.R., Scho¨tz, K., Kausch, M. and Schindler, M. (1987). Angew. Chem. 99, 795–796; Angew. Chem. Int. Ed. 26, 761 88. Bremer, M., Schleyer, P.v.R. and Fleischer, U. (1989). J. Am. Chem. Soc. 111, 1147 89. Cheeseman, J.R., Trucks, G.W., Keith, T.A. and Frisch, M.J. (1996). J. Chem. Phys. 104, 5497 90. Schleyer, P.v.R., Lenoir, D., Mison, P., Liang, G., Prakash, G.K.S. and Olah, G.A. (1980). J. Am. Chem. Soc. 102, 683 91. Olah, G.A., Head, N.J., Rasul, G. and Prakash, G.K.S. (1995). J. Am. Chem. Soc. 117, 875 92. Levy, J.B. (1999). Struct. Chem. 10, 121 93. Schleyer, P.v.R., Buzek, P., Mu¨ller, T., Apeloig, Y. and Siehl, H.-U. (1993). Angew. Chem. Int. Ed. 32, 1471 94. Olah, G.A., Rasul, G., Buchholz, H.A., Li, X.Y. and Prakash, G.K.S. (1995). Bull. Soc. Chim. Fr. 132, 569

164

H-U. SIEHL

95. (a) Lambert, J.B., Zhang, S., Stern, C.L. and Huffmann, J.C. (1993). Science 260, 1917; (b) Lambert, J.B. and Zhang, S. (1993). J. Chem. Soc. Chem. Commun. 383 96. Lambert, J.B. and Zhao, Y. (1997). Angew. Chem. Int. Ed. 36, 400 97. Mu¨ller, T., Zhao, Y. and Lambert, J.B. (1998). Organometallics 17, 278; For similar results see: (a) Kraka, D., Sosa, C.P. and Cremer, D. (1997). Chem. Phys. Lett. 9, 279; (b) Kim, K.-C., Reed, C.A., Elliott, D.W., Mueller, L.J., Tham, F., Lin, L. and Lambert, J.B. (2002). Science 297, 825 98. Mu¨ller, T. (2001). Angew. Chem. Int. Ed. 40, 3033 99. Meyer, R., Werner, K. and Mu¨ller, T. (2002). Chem. Eur. J. 8, 1163 100. Siehl, H.-U., Mu¨ller, B., Malkina, O. (1997). In Organosilicon Chemistry III. , Auner, N. and Weiss, J. (eds), p. 25. Wiley-VCH, Weinheim 101. Mu¨ller, B. (1995). Ph.D. thesis, University of Tu¨bingen 102. Olah, G.A., Shamma, T., Burrichter, A., Rasul, G. and Prakash, G.K.S. (1997). J. Am. Chem. Soc. 119, 12923 103. (a) Laali, K.K., Hollenstein, S., Galembeck, S.E., Nakamura, Y. and Nishimura, J. (1999). J. Chem. Soc. Perkin Trans. 2, 2129; (b) Laali, K.K., Okazaki, T. and Galembeck, S.E. (2002). J. Chem. Soc. Perkin Trans. 2, 621; (c) Laali, K.K., Hollenstein, S., Galembeck, S.E. and Coombs, M.M. (2000). J. Chem. Soc. Perkin Trans 2, 211; (d) Laali, K.K., Okazaki, T., Galembeck, S.E. and Siegel, J.S. (2001). J. Org. Chem. 66, 8701 104. (a) Mills, N.S., Benish, M.A. and Ybarra, C. (2002). J. Org. Chem. 67, 2003; (b) Mills, N.S. (2002). J. Org. Chem. 67, 7029; (c) Levy, A., Rakowitz, A. and Mills, N.S. (2003). J. Org. Chem. 68, 3990 105. Cremer, D., Svensson, P., Kraka, E. and Ahlberg, P. (1993). J. Am. Chem. Soc. 115, 7445 106. Cremer, D., Svensson, P., Kraka, E., Konkoli, Z. and Ahlberg, P. (1993). J. Am. Chem. Soc. 115, 7457 107. Bremer, M., Scho¨tz, K., Schleyer, P.v.R., Fleischer, U., Schindler, M., Kutzelnigg, W., Koch, W. and Pulay, P. (1989). Angew. Chem. Int. Ed. 28, 1042 108. Cremer, D., Reichel, F. and Kraka, E. (1991). J. Am. Chem. Soc. 113, 9459 109. (a) Lambert, J.B., Lin, L. and Rassolov, V. (2002). Angew. Chem. Int. Ed. 41, 1429; (b) Mu¨ller, T. (2002). Angew. Chem. Int. Ed. 41, 2277; (c) Otto, M., Scheschkewitz, D., Kato, T., Midland, M.M., Lambert, J. and Bertrand, G. (2002). Angew. Chem. Int. Ed. 41, 2275; (d) Lambert, J.B. (2002). Angew. Chem. Int. Ed. 41, 2278 110. Haw, J.F., Nicholas, J.B., Xu, T., Beck, L.W. and Ferguson, D.B. (1996). Acc. Chem. Res. 29, 259 111. (a) Haw, J.F., Nicholas, J.B., Song, W., Deng, F., Wang, Z., Xu, T. and Heneghan, C.S. (2000). J. Am. Chem. Soc. 122, 4763; (b) Song, W., Nicholas, J.B. and Haw, J.F. (2001). J. Am. Chem. Soc. 123, 121 112. Xu, T. and Haw, J.F. (1994). J. Am. Chem. Soc. 116, 7753 113. Song, W., Nicholas, J.B. and Haw, J.F. (2001). J. Chem. Phys. B 105, 4317 114. (a) Xu, T., Barich, D.H., Torres, P.D. and Haw, J.F. (1997). J. Am. Chem. Soc. 119, 406; (b) Xu, T., Barich, D.H., Goguen, P.W., Song, W., Wang, Z., Nicholas, J.B. and Haw, J.F. (1998). J. Am. Chem. Soc. 120, 4025; (c) Song, W., Nicholas, J.B., Sassi, A. and Haw, J.F. (2002). Cat. Lett. 81, 49 115. Fernandez, L., Marti, V. and Garcia, H. (1999). Phys. Chem. Chem. Phys. 1, 3689 116. (a) Sorensen, T.S. (1998). Angew. Chem. Int. Ed. 37, 603; (b) de Rege, P.J.F., Gladysz, J.A. and Horvath, I.T. (1997). Science 276, 776 117. Olah, G.A. Unpublished results 118. Olah, G.A., Burrichter, A., Mathew, T., Vankar, Y.D., Rasul, G. and Prakash, G.K.S. (1997). Angew. Chem. Int. Ed. 36, 1875 119. Christe, K.O., Hoge, B., Boatz, J.A., Prakash, G.K.S., Olah, G.A. and Sheehy, J.A. (1999). Inorg. Chem. 38, 3132 120. Olah, G.A., Burrichter, A., Rasul, G., Gnann, R., Christe, K.O. and Prakash, G.K.S. (1997). J. Am. Chem. Soc. 119, 8035

COMPUTATIONAL NMR SPECTROSCOPY OF CARBOCATIONS

165

121. Olah, G.A., Burrichter, A., Rasul, G., Yudin, A.K. and S Prakash, G.K. (1996). J. Org. Chem. 61, 1934 122. (a) Prakash, G.K.S., Wang, Q., Rasul, G. and Olah, G.A. (1998). J. Organomet. Chem. 550, 119; (b) Prakash, G.K.S., Bae, C., Rasul, G. and Olah, G.A. (2002). J. Org. Chem. 67, 1297 123. Prakash, G.K.S., Rasul, G., Liang, G. and Olah, G.A. (1996). J. Phys. Chem. 100, 15805 124. Head, N.J., Rasul, G., Mitra, A., Bashir-Heshemi, A., Prakash, G.K.S. and Olah, G.A. (1995). J. Am. Chem. Soc. 117, 12107 125. Xu, T., Barich, D.H., Torres, P.D., Nicholas, J.B. and F Haw, J. (1997). J. Am. Chem. Soc. 119, 396 126. Prakash, G.K.S., Rasul, G., Burrichter, A., Laali, K.K. and Olah, G.A. (1996). J. Org. Chem. 61, 9253 127. (a) Olah, G.A., Rasul, G., Heiliger, L. and Prakash, G.K.S. (1996). J. Am. Chem. Soc. 118, 3580; (b) Olah, G.A., Rasul, G., Yudin, A.K., Burrichter, A., Prakash, G.K.S., Chistyakov, A.L., Stankevich, I.V., Akhrem, I.S., Gambaryan, N.P. and Vol’pin, M.E. (1996). J. Am. Chem. Soc. 118, 1446 128. Kaupp, M., Malkina, O.L. and Malkin, V.G. (1997). Chem. Phys. Lett. 265, 55 129. Rasul, G., Prakash, G.K.S. and Olah, G.A. (1999). J. Mol. Struct. 466, 245 130. Reddy, V.P., Rasul, G., Prakash, G.K.S. and Olah, G.A. (2003). J. Org. Chem. 68, 3507 131. Olah, G.A., Rasul, G., York, C. and Prakash, G.K.S. (1995). J. Am. Chem. Soc. 117, 11211 132. Olah, G.A., Calin, M. and O’Brien, D.H. (1967). J. Am. Chem. Soc. 89, 3582 133. Pa´linko´, I., Burrichter, A., Rasul, G., To¨ro¨k, B., Prakash, G.K.S. and Olah, G.A. (1998). J. Chem. Soc. Perkin Trans. 2, 379 134. Prakash, G.K.S., Rasul, G., Olah, G.A., Liu, R.H. and Tidwell, T.T. (1999). Can. J. Chem. 77, 525 135. Olah, G.A., Bausch, J., Rasul, G., George, H. and Prakash, G.K.S. (1993). J. Am. Chem. Soc. 115, 8060 136. Olah, G.A. and White, A.M. (1968). J. Am. Chem. Soc. 90, 6067 137. Olah, G.A., White, A.M. and O’Brien, D.H. (1970). Chem. Rev. 70, 561 138. (a) Rasul, G., Prakash, G.K.S. and Olah, G.A. (1994). J. Org. Chem. 59, 2552; (b) Olah, G.A., Burrichter, A., Rasul, G., Christe, K.O. and Prakash, G.K.S. (1997). J. Am. Chem. Soc. 119, 4345 139. Rasul, G., Prakash, G.K.S. and Olah, G.A. (2000). J. Org. Chem. 65, 8786 140. (a) Olah, G.A., Heiner, T., Rasul, G. and Prakash, G.K.S. (1998). J. Org. Chem. 63, 7993; (b) Rasul, G., Reddy, V.P., Zdunek, L.Z., Prakash, G.K.S. and Olah, G.A. (1993). J. Am. Chem. Soc. 115, 2236 141. Olah, G.A., Burrichter, A., Rasul, G., Hachoumy, M. and Prakash, G.K.S. (1997). J. Am. Chem. Soc. 119, 12929 142. Rasul, G., Prakash, G.K.S. and Olah, G.A. (1994). J. Am. Chem. Soc. 116, 8985 143. Jemmis, E.D., Subramanian, G., Kos, A.J. and Schleyer, P.v.R. (1997). J. Am. Chem. Soc. 119, 9504 144. Apeloig, Y., Schleyer, P.v.R. and Pople, J.A. (1977). J. Am. Chem. Soc. 99, 5901 145. Helgaker, T. and Pecul, M. (2004). Spin–spin coupling constants with HF and DFT methods. In Calculation of NMR and EPR Parameters, Theory and Applications, Kaupp, M., Bu¨hl, M. and Malkin, V.G. (eds), Chapter 7, pp. 101–121. Wiley-VCH, Weinheim, Germany 146. Gauss, J. (1994). Chem. Phys. Lett. 229, 198 147. Auer, A.A., Gauss, J. and Stanton, J.F. (2003). J. Chem. Phys. 118, 10407 148. For historical overviews on the 2-norbornyl cation, the bicyclobutonium ion and related hypercoordinated carbocations see: (a) Nonclassical Ions, Reprints and Commentary, P. D. Bartlett, W.A. Benjamin, Inc., New York and Amsterdam 1965; (b) The Nonclassical Ion Problem, Brown, H.C. with comments by Schleyer, P.v.R. Plenum Press, New York, 1977

This page intentionally left blank

Dynamics of guest binding to supramolecular systems: techniques and selected examples TAMARA C.S. PACE and CORNELIA BOHNE Department of Chemistry, University of Victoria, PO Box 3065, Victoria, BC, V8W 3V6, Canada 1 Introduction 167 2 Techniques 169 Theoretical background for relaxation kinetics 169 Stopped-ﬂow experiments 171 Temperature jump experiments 172 Ultrasonic relaxation measurements 174 Laser ﬂash photolysis and time-resolved ﬂuorescence Fluorescence correlation spectroscopy 178 Nuclear magnetic resonance 181 Surface plasmon resonance 183 3 Examples of supramolecular dynamics studies 185 DNA 186 Cyclodextrin 204 4 Conclusions 216 Acknowledgements 217 References 217

1

175

Introduction

Supramolecular systems have the capability of achieving functions not accessible by using molecular components.1,2 Supramolecular chemistry provides the framework to move from the molecular world to the nano- and meso-scale world. In this respect, the development of supramolecular systems provides a bottom-up approach for the construction of nanoscale objects where complexity is paramount to achieve multi-step function. Detailed knowledge of the system’s thermodynamics and dynamics is required to rationally design new structures and to modify known ones. The building blocks of supramolecular systems are held together by intramolecular interactions and these systems are reversible. This intrinsic property is not only a consequence of the more labile interactions within supermolecules, compared to covalent bonds in molecules, but reversibility is essential for the function expressed by supramolecular systems. Kinetics can never be inferred from thermodynamic studies. For example, the knowledge of a host–guest equilibrium constant does not Corresponding author. E-mail address: [email protected] (C. Bohne).

167 ADVANCES IN PHYSICAL ORGANIC CHEMISTRY VOLUME 42 ISSN 0065-3160 DOI: 10.1016/S0065-3160(07)42004-4

r 2008 Elsevier Inc. All rights reserved

168

T.C.S. PACE AND C. BOHNE

provide any information on how fast this complex is formed nor on how long the complex exists, since a combination of association rate constants between the guest and host and dissociation rate constants for the exit of the guest from the complex can lead to the same value for the equilibrium constant. The ﬁeld of supramolecular chemistry has been driven by the synthesis of increasingly complex systems. Structural variation and complexity is required to achieve and control parameters such as selectivity, molecular recognition, pre-organization, cooperativity, and multivalency,1,3–16 which are required for functions such as catalysis, sensing or transport. Structural knowledge is required to understand how 3-dimensional organization affects function, thermodynamic studies are used to establish the efﬁciency of complex formation and to determine the types of species present at equilibrium, and dynamic investigations provide a description of how the system achieves equilibrium. The structural sophistication currently achieved is greater than the details known on the dynamics of supramolecular systems. The investigation of the dynamics of these systems requires kinetic measurements in realtime,17,18 which provided early challenges because the dynamics are in general fast. The objective of this monograph is to provide a conceptual overview of the techniques available to measure the dynamics of supramolecular systems, and in particular host–guest complexes. The advantages and disadvantages of each technique will be discussed and their application to the binding of small molecules to cyclodextrins (CDs) or DNA will be described for selected examples. Our focus is on kinetic events that are faster than can be measured by manually mixing solutions (o1 s). Most techniques for which fast real-time kinetics can be measured rely on a perturbation of an equilibrium followed by the measurement of the relaxation kinetics that take the system to a new equilibrium. Stopped-ﬂow, time-resolved ﬂuorescence, laser ﬂash photolysis (LFP), temperature jump, and ultrasonic relaxation spectroscopy fall into this category. Fluorescence correlation spectroscopy (FCS) is based on a different concept, where temporal ﬂuctuations of the measured intensity are observed because of changes that occur to one molecule. In the case of NMR, information about the relaxation kinetics is obtained from the line width of the peak when the relaxation kinetics for the chemical system occur on a time scale similar to the intrinsic relaxation of the NMR experiment. For the last two techniques, i.e. FCS and NMR, the measurements are performed without the perturbation of the equilibrium. Finally, surface plasmon resonance (SPR) is a technique where kinetics are obtained when one reactant is ﬂowed over a surface containing a second immobilized reactant, and therefore corresponds to a system with phase transfer (Fig. 1). We chose to show the application of techniques employed to study supramolecular dynamics to host systems that have deﬁned binding sites and, therefore, form host–guest complexes with deﬁned stoichiometries. CDs were chosen because they represent a host with only one binding site, and therefore can be viewed as a model system for hosts with the lowest degree of complexity. DNA was chosen because it provides multiple binding sites for small molecules, i.e. intercalative and groove binding, while the DNA is structurally fairly well deﬁned.

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

169

Fig. 1 Time ranges accessible to techniques commonly used to study supramolecular dynamics.

2

Techniques

THEORETICAL BACKGROUND FOR RELAXATION KINETICS

The dynamics of a supramolecular system are deﬁned by the association and dissociation rate constants of the various components of the system. The time-scale for the dynamic events is inﬂuenced by the size (length-scale) and by the complexity of the system. The fastest time for an event to occur in solution is limited by the diffusion of the various components to form encounter complexes. This diffusion limit provides an estimate for the shortest time scale required for kinetic measurements. The diffusion of a small molecule in water over a distance of 1 nm, which is the length-scale for the size of small host systems such as CDs or calixarenes, is 3 ns at room temperature. In general terms, one can deﬁne that mobility within host systems can occur on time scales shorter than nanoseconds, while the association/dissociation processes are expected to occur in nanoseconds or on longer time scales. The complexity of a system also inﬂuences its dynamics, since various kinetic events can occur over different time scales. An increase in complexity can be related to an increase in the number of building blocks within the system, or complexity can be related to the presence of more than one binding site. The most common methodology for measuring fast kinetics in real time is to perturb a system at equilibrium for a time duration that is much shorter than the relaxation kinetics that follow perturbation. This perturbation can be achieved by changing the concentration of chemicals through fast mixing (stopped-ﬂow), changing the temperature of the solution (temperature jump), simultaneously changing the

170

T.C.S. PACE AND C. BOHNE

pressure and temperature (ultrasonic relaxation), or by creating a new chemical (excited-state molecule, which can be detected by ﬂuorescence or laser ﬂash photolysis studies). The simplest kinetic scheme is that for the formation of a 1:1 complex between the guest (G) and the host (H):

G+H

k+ k–

HG

K11=

k [HG] = + k– [H][G]

(1)

The relaxation kinetics after perturbation are deﬁned by the following rate law: d½HG ¼ kþ ½H½G k ½HG dt

(2)

If one of the components for the bimolecular reaction is in excess (e.g. [H]b[G]) and the changes in concentrations due to the perturbation are small, the kinetics follow an exponential function where the observed rate constant (kobs) is given by kobs ¼ kþ ½H þ k

(3)

The concentration of free host ([H]) is related to the total concentration of host ([H]T) by the mass balance equation: ½HT ¼ ½H þ ½HG

(4)

In general, the relationship [H]T[H] holds and the analytical concentration of the host is employed. Otherwise, a quadratic expression for the concentration is obtained from the equations for the equilibrium constant and the mass balance. The consequence of Equation (3) is that the relaxation process is related to the sum of the rate constant for the pseudo-ﬁrst-order association process and the rate constant for the dissociation process. The association process can be inﬂuenced by changes in the concentration of H, but the value of k– is intrinsic to the system and cannot be manipulated by external parameters, such as concentrations of reactants. The relaxation process can be dominated by the association or dissociation process depending on the relative value of k+[H] compared to k– . The lifetime for the relaxation process is the inverse of the observed rate constant (tobs ¼ 1/kobs). Kinetic schemes involving sequential and coupled reactions, where the reactions are either ﬁrst-order or pseudo-ﬁrst order, lead to expressions for concentration changes with time that can be modeled as a sum of exponential functions where each of the exponential functions has a speciﬁc relaxation time. More complex equations have to be derived for bimolecular reactions where the concentrations of reactants are similar.19,20 However, the rate law is always related to the association and dissociation processes, and these processes cannot be uncoupled when measuring a relaxation process.

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

171

STOPPED-FLOW EXPERIMENTS

Stopped-ﬂow is a chemical relaxation technique in which a concentration jump is used to study the kinetics of fast processes in solution on the millisecond to minute time scale. In this method two reactants, placed in separate syringes, are driven through a mixing chamber where they are thoroughly mixed. The resulting mixture then ﬂows into an observation cell, where ﬂow is rapidly stopped and further changes in the system are followed.21,22 The mixing of the two solutions results in a perturbation of the initially equilibrated chemical system in each syringe, and relaxation to the new equilibrium state is monitored as a function of time. Since the solutions are mixed prior to observation, the mixing or dilution of the initial systems is not the observable detected. Two different modes are employed to measure the relaxation kinetics. In one mode, the components of the system (H and G) are kept in separate syringes and are mixed, whereas in the second mode one syringe contains HG and is diluted by mixing with the homogeneous solution contained in the second syringe (Fig. 2). Most commonly absorption or ﬂuorescence spectroscopy is used for detection of the changes in the concentration of G or HG. The monitoring wavelength is chosen so that the difference between the molar absorptivities, in case of absorption, or emission quantum yields, in the case of ﬂuorescence detection, between G and HG is maximized. The amplitude of the relaxation process depends on the difference in the molar absorptivities or ﬂuorescence quantum yields, but the observed rate constants are the same at all observation wavelengths when the kinetics are ﬁrst- or pseudo-ﬁrst order (Fig. 3). The time resolution of stopped ﬂow experiments is typically 1–2 ms,21 and is determined by the time required to mix the solutions, ﬂow the mixed solution to the detection chamber, and stop the ﬂow. Smaller detection cells can be used to decrease the time resolution at the expense of the signal-to-noise ratio of the detected signals. Various kinetic traces have to be averaged to achieve good kinetic proﬁles and sample volumes of milliliters with concentrations of micromolar to millimolar are required.

Fig. 2 Schematic for a stopped ﬂow experiment.

172

T.C.S. PACE AND C. BOHNE

Fig. 3 Relaxation kinetics after a perturbation at t ¼ 0. The observable commonly corresponds to changes in absorption, ﬂuorescence intensity, or scattering intensity.

The detection of concentration changes by absorption or ﬂuorescence requires that the host or guest contain chromophores and that changes in the absorption or emission spectra occur for the species free in solution or contained in the complex. Fluorescence detection often allows very low concentration changes to be monitored.22 This is especially important since low concentrations often need to be used to make association kinetics slow enough to be accessible by stopped-ﬂow.23 TEMPERATURE JUMP EXPERIMENTS

The concept for temperature jump experiments is that an equilibrium is perturbed in a fast manner by changing the temperature of the solvent. Three methods are available to achieve the pulsed heating of a solution: (i) Joule heating, (ii) microwave heating, and (iii) heating by absorption of laser light.24 The basic premise for temperature jump experiments is that the equilibrium being studied has to change when the temperature of the system is raised. The relaxation kinetics are followed by changes over time of the absorption, ﬂuorescence, or scattering intensities of the solutions. In addition, conductimetry can be employed when the heating of the solution is achieved through microwave or laser pulses. The high electrical conductance of samples used for Joule heating experiments (see below) precludes, in this case, the use of conductance measurements.24 Joule heating is the most commonly used method for temperature jump studies. The solution is heated by the discharge of a capacitor through a solution containing electrolytes, which are responsible for the heat transfer within the solution. The rise time for the heating process depends on the electrolyte concentration. For example, a typical rise time of 2 ms was obtained for a discharge through 100 mL of aqueous solution containing 0.1 M salt leading to a temperature increase of 5 K.24 With a carefully designed cell the time resolution for Joule heating was decreased to 0.2 ms,25 but in general most systems have a microsecond time resolution. The suitability of Joule heating for temperature jump studies involving charged macromolecules, such as DNA, has been disputed on the basis that the high electric ﬁelds applied lead to

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

173

artifactual signals that overlay the relaxation signal of interest.26–28 However, it was shown that these artifacts can be minimized by using magic angle detection to monitor the changes in absorption or ﬂuorescence signals.29 Microwave heating is achieved with a pulse of microwave energy generated within a magnetron. The only requirement for the use of this heating method is that the solvent has a permanent dipole moment.24 This technique can be employed to study reactions in non-aqueous solvents since the presence of electrolytes is not necessary.30 However, the temperature changes attained were much lower than in the case of Joule or laser heating.24 The use of lasers makes it possible to achieve heating times of pico- to nanoseconds,31–34 and this is the heating method of choice when fast kinetics are studied. The challenge in using laser irradiation to achieve a temperature jump is to efﬁciently transfer the optical energy into thermal energy. An early design used dyes that when excited in the visible region of the spectrum by a laser led to heat release from the non-radiative decay of the dye.35 This method is inefﬁcient and care has to be taken that the dye does not interact with the solutes for the reaction being studied. Direct absorption of the laser light by water is preferred since the heating rise time is fast and does not depend on the photophysical decay of the excited state dye. The absorption coefﬁcient of water at the laser irradiation wavelength deﬁnes the design of the sample cell and the volume that can be irradiated homogeneously. The absorption coefﬁcient of water at the typical 1064 nm laser band of a YAG laser is too low to achieve efﬁcient heating. Two different types of lasers have been used with irradiation wavelengths in the 1.3–2.0 mm range. Raman shifted YAG lasers using for example H2,36 D2,37 CH4,38 or N239 led to laser irradiation wavelengths of 1.89, 1.56, 1.54, and 1.41 mm, respectively. The absorption coefﬁcient of water in this wavelength region is high, which means that the irradiation pathlength has to be short to avoid non-homogeneous heating. In general, laser excitation and the monitoring beam are nearly collinear and counter propagating. The volumes excited are on the order of 20 mL38 and the cooling time, i.e. the time for the heated volume to equilibrate with the surrounding solution is ca. 10 ms.37 The iodine laser emitting at 1315 nm led itself to the irradiation of a larger volume (up to 500 mL) because the molar absorption at this wavelength is closer to the optimum value of 0.5 cm1.24,31 At this wavelength, an irradiation path of 2 mm can be used and with a dual path of the laser beam 50% of the laser light is absorbed with a heat absorption proﬁle that has a homogeneity better than 90%. The heating of a larger volume led to cooling times in excess of 1 s. Alternatively, in the case of the YAG/Raman shifted lasers the molar absorptivity can be changed by working in H2O/D2O mixtures39 since the D2O absorption coefﬁcient at the irradiation wavelengths of these lasers is lower than the coefﬁcient for water. In summary, the necessary condition for temperature jump experiments is that the equilibrium for the chemical system of interest changes with a change in temperature. The advantages of temperature jump experiments are that the perturbation is achieved by a change in a property of the solvent, a fast time resolution can be achieved, as short as picoseconds when using lasers, and a time domain over more than 6 orders of magnitude can be probed with the same technique. The disadvantage of the technique

174

T.C.S. PACE AND C. BOHNE

is that for most systems a change in absorption or ﬂuorescence is probed and for this reason the chemical system has to contain a chromophore for which the absorption or emission intensities vary for the component bound to the host or free in the homogeneous solvent. In the case of Joule heating, an electrolyte always has to be present leading in cases to artifactual signals, while when using lasers the irradiation wavelength is chosen for one type of solvent, in general water, and the system cannot be employed for other solvents.

ULTRASONIC RELAXATION MEASUREMENTS

In ultrasonic relaxation measurements perturbation of an equilibrium is achieved by passing a sound wave through a solution, resulting in periodic variations in pressure and temperature.40,41 If a system in chemical equilibrium has a non-zero value of DHo or DVo then it can be cyclically perturbed by the sound wave. The system cannot react to a sound wave with a frequency that is faster than the rates of equilibration of the system, and in this case only classical sound absorption due to frictional effects occurs. When the rate for the host–guest equilibration is faster than the frequency of the sound wave the system re-equilibrates during the cyclic variation of the sound wave with the net result of an absorption of energy from the sound wave to supply heat to the reaction (Fig. 4). The sound absorption coefﬁcient, a, is increased when the dynamics of the chemical system are of the same order of magnitude as the frequency of the sound wave,41 and experimentally this quantity is measured as a function of frequency of the ultrasonic sound wave (Fig. 4). When the frequency of the sound wave is of the same order as the frequency for the relaxation process, effects due to relaxation of the equilibrium give rise to characteristic changes in the quantity a/f2, where a is the sound absorption coefﬁcient measured at frequency f.40 The variation of a with frequency, f, has an inﬂection point at the relaxation frequency of the system, fr, which is related to 1/t, where t is the relaxation time (1/t ¼ 2pfr).40,41 The expression relating the quantity

Fig. 4 Dependence for a single relaxation process of the parameters a/f2 and alU

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

175

a/f2 to the relaxation frequency is a A ¼ þB 2 f 1 þ ðf =f r Þ2

(5)

where A is the amplitude of the relaxation, which depends on the thermodynamic parameters DHo and DVo and B measures the contribution from other processes including sound absorption processes that have relaxation frequencies higher than those being measured.41,42 If there are n equilibria in the system then the ﬁrst term on the right side of Equation (5) becomes a sum of terms, and the spectrum will exhibit n relaxation times. These rate constants can only be separated if each relaxation process has a signiﬁcant amplitude for ultra-sound absorption and the rate constants for the relaxation differ by more than a factor of 4–5.40 A broad range of wavelengths needs to be swept for ultrasonic relaxation measurements. Two techniques using an acoustic resonator cavity and a pulsed system need to be employed.40–43 The resonator technique is employed to cover the low-frequency domain (0.1 to 10–20 MHz) where the pulsed acoustic wave is continuously applied to the cell and the resulting sound waves are detected by a receiving detector crystal. The pulsed technique is employed to cover the frequency range from 10 to 220 MHz and the attenuation of the ultrasound pulse between the emitting and the detecting crystals is measured. For both techniques, reference measurements have to be employed because the acoustic properties of the sample cell are not ideal. The common lifetime range accessible by the ultrasonic relaxation method is from 1 ms to 1 ns, though the technique has been expanded to relaxation times from 10 ms to 0.1 ns.42 The biggest advantage of ultrasonic relaxation measurements is that the presence of a chromophore is not required as for the relaxation techniques based on measuring changes in absorption or ﬂuorescence intensities. Therefore, the technique is in principle applicable to a wide range of systems provided that the reaction enthalpy and volume changes are non-zero. One of the disadvantages is that to use this relaxation method the molecular process that is being perturbed must be identiﬁed.40 There are a number of processes, such as conformational changes, which can also be observed by this method. All other observable processes must be identiﬁed and accounted for in the analysis of the relaxation data, especially when the relaxation processes have close or overlapping relaxation times. LASER FLASH PHOTOLYSIS AND TIME-RESOLVED FLUORESCENCE

Photophysical techniques are employed to measure the decay kinetics of excited states and reactive intermediates with short lifetimes. For studies on the guest–host binding dynamics the formation of excited states or reactive intermediates provides the fast perturbation of the system, since a new chemical is formed after the absorption of light. Either the guest or host is excited by a short pulse of light typically from a laser (in the discussion below we assume that the guest is excited). The relaxation kinetics can be directly measured if the equilibrium constants for the ground and excited state species in the host–guest complex are sufﬁciently different. Alternatively, the

176

T.C.S. PACE AND C. BOHNE

dynamics can be obtained by introducing a competitive deactivation pathway for the excited states, i.e. quenching experiments. The relationship between the excited state lifetime of the guest and the relaxation times for the host–guest system need to be considered.17,18 If the binding dynamics are slower than the guest’s excited state lifetime, the excited molecule will not move from its environment and no information on the entry and exit rate constants can be obtained. Dynamic information on the guest binding to the host is only obtained when the excited state lifetime is of the same order of magnitude as the host–guest relaxation time. The lifetime for singlet excited states of organic molecules are in general shorter than 100 ns; the exception being pyrene44 and 2,3-diazabicyclo[2.2.2] oct-2-ene (DBO).45 For this reason, time-resolved ﬂuorescence measurements are limited to measurements where the entry/exit rate constants are very fast. Triplet excited states are much longer lived with lifetimes which, in aqueous solution, can readily be hundreds of nanoseconds. The kinetics for triplet states are measured using laser ﬂash photolysis. Several different methods exist for the measurement of time-resolved ﬂuorescence,46,47 but the one generally used to study the dynamics of supramolecular systems is single-photon counting.48 Excitation sources with a high repetition rate excite the ﬂuorophore in the sample and the time for the detection of single photons is measured. The probability distribution in time of the detected photons follows the statistical distribution of all emitted photons, and a histogram of number of photons vs. time detected is built. The advantage of using single photon counting detection is that a large dynamic range for the intensity can be accumulated which is helpful when differentiating ﬂuorophores that decay with different lifetimes. In most studies using ﬂuorescence, the lifetime of the excited state was too short for it to move from the supramolecular system to the homogeneous solvent. However, the entry and exit rate constants for quencher molecules, i.e. molecules that deactivate the excited state, have been investigated.18 This methodology was extensively employed to study the dynamics of guests with micelles,18,49 where the ﬂuorophore and quencher can be bound to the same host. In addition, ﬂuorescence is an excellent technique to study the dynamics at short times (femtoseconds to picoseconds) within host systems, as was recently described for the binding of guests to cyclodextrins.50 Laser ﬂash photolysis experiments48,51 are based on the formation of an excited state by a laser pulse. Time resolutions as short as picoseconds have been achieved, but with respect to studies on the dynamics of supramolecular systems most studies used systems with nanosecond resolution. Laser irradiation is orthogonal to the monitoring beam used to measure the absorption of the sample before and after the laser pulse, leading to measurements of absorbance differences (DA) vs. time. Most laser ﬂash photolysis systems are suitable to measure lifetimes up to hundreds of microseconds. Longer lifetimes are in general not accessible because of instabilities in the lamp of the monitoring beam and the fact that the detection system has been optimized for nanosecond experiments. Methods using excited states to measure the dynamics of supramolecular systems have the additional dimension that the excited state has a lifetime and this decay has to be considered in the kinetic treatment of the data. The decay processes are parallel

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

177

Fig. 5 Schematic representation of the binding dynamics of an excited state guest. deactivation events (ko and kH o ) with the association ðkþ Þ and dissociation ðk Þ processes (Fig. 5). The concentration of guest is in general much lower than the concentrations for the host or quencher and for this reason the bimolecular reactions shown in Fig. 5 are pseudo-ﬁrst-order reactions. The relaxation kinetics can be followed without the need of addition of quenchers when the absorption spectra for the excited guest in the host and in the homogeneous phase are different and the equilibrium constants of the ground- and excited-state guest with the host are also different. The excited guest is located in two different compartments, i.e. host and homogeneous phase, and it can move between these compartments during its excited state lifetime. The general solution for this mechanism corresponds to a decay of the sum of two exponentials where the two exponential factors include the rate constants for all four processes (k þ , k , ko and kH o ). The solution is equivalent to the one described for excimer and exciplex emission processes.52,53 When the host–guest binding dynamics are much faster than the decay of the excited state the expression for the observed rate constant for the fast component can be approximated to the sum of the association process for the excited state guest with the host and the excited state guest dissociation process (Equation (3) described in the theoretical section above). This approximation is only valid when the difference between the two exponential factors measured is at least two orders of magnitude.17,18 In most cases, the absorption spectra for the free and bound guest are not sufﬁciently different and a quenching methodology is employed to determine the values of k þ and k 2 . This quenching methodology, initially developed to investigate the binding dynamics of guests with micelles,54 is based on the much more efﬁcient quenching of the excited guest in the homogeneous phase than when bound to the host.17,18,55,56 The decay follows a mono-exponential function when the concentration of excited guest in the homogeneous phase is much smaller than bound to the host. In this case, the dependence of the observed rate constant (kobs) is ﬁt to Equation (6), where the parameters are those shown in Fig. 5: H kobs ¼ kH o þ k þ k q ½Q

k k þ ½H ko þ k þ ½H þ kq ½Q

(6)

178

T.C.S. PACE AND C. BOHNE

The values for ko and kq are determined from quenching experiments in the homogeneous solvent, while the value of kH o is determined from the triplet state decay in the presence of host but absence of quencher. These values are ﬁxed when ﬁtting the quenching plot in the presence of host to Equation (6). In summary, ﬂuorescence measurements are useful to study supramolecular dynamics because of the high sensitivity of emission measurements and the fact that single photon counting experiments provide a large dynamic range for intensity measurements leading to an easier differentiation between similar relaxation constants (i.e. lifetimes). The drawback of ﬂuorescence experiments is that the lifetime of singlet excited states of most organic molecules are short and only fast binding events can be measured directly. Kinetic measurements using triplet excited states circumvent the drawback of a short lifetime and this technique is suitable to measure dynamics that occur from tenths of nanoseconds to hundreds of microseconds. However, laser ﬂash photolysis experiments require the use of excited state guests that are inert toward the host system and guest molecules with a chromophore. In addition, it is important to note that with photophysical methods the association and dissociation rate constants of the excited guest are measured and these dynamics can be signiﬁcantly different from those observed for the ground state guest.57

FLUORESCENCE CORRELATION SPECTROSCOPY

This technique does not require the equilibrium of the chemical system to be perturbed to measure the host–guest binding dynamics. Fluorescence correlation spectroscopy is based on the measurement of changes in the ﬂuorescence intensity of individual molecules, which lead to intensity ﬂuctuations.58–63 For this reason, the measurements are made by detecting the emission from a small sample volume (femtoliters to microliters) containing a small number of ﬂuorophores. The sample is continuously irradiated and the ﬂuctuations in the ﬂuorescence intensity arise due to any event which makes the ﬂuorophore unavailable to be excited to the emissive singlet excited state, such as diffusion of the ﬂuorophore out of the detection volume, formation of a dark state, such as a triplet excited state, or photoreaction. The concentration of ﬂuorophore in the detection volume has to be low (10–13–10–8 M) so that the ﬂuctuation in the intensity for one molecule is observable over any background emission. The high concentration limit is a consequence of the fact that the correlated photons from single molecules scale with the number of molecules in the detection volume, while the contribution from uncorrelated photons, arising from the emission from different molecules, scales with the square of the number of molecules. The lowest concentration is determined by the probability of ﬁnding a molecule in the detection volume.58 The ﬁrst ﬂuorescence correlation spectroscopy experiments were carried out several decades ago,62,64 but the general use of the technique was made possible with the introduction of lasers with high beam quality and long-term temporal stability, low noise detectors, and high-quality microscope objectives with high numeric apertures.58,63 The most common set-up is using a confocal inverted epi-ﬂuorescence

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

179

microscope.58,60,61,63 Aqueous solutions of ﬂuorescent molecules are placed on top of a cover-slide and excitation is carried out by a laser focused through the objective into a small region of the sample solution. The emitted light is collected by the same objective and imaged onto a confocal aperture that restricts observation to a small region in the sample, where light is refocused onto the detector. The detector output is recorded and ﬂuctuations from the average signal are autocorrelated. The autocorrelation function G(t) corresponds to the correlation of a time-shifted replica of itself at various time-shifts (t) (Equation (7)).58,65 This autocorrelation deﬁnes the probability of the detection of a photon from the same molecule at time zero and at time t. Loss of this correlation indicates that this one molecule is not available for excitation, either because it diffused out of the detection volume or it is in a dark state different from its ground state. Two photons originating from uncorrelated background emission, such as Raman scattering, or emission from two different molecules do not have a time correlation and for this reason appear as a time-independent constant offset for G(t).58 GðtÞ ¼ IðtÞIðt þ tÞ (7) The biggest advantage in using ﬂuorescence correlation spectroscopy is that a wide time range is recorded, which recently was shown to cover more than 12 orders of magnitude from picoseconds to seconds.66 On the nanosecond time scale the correlation curve is related to the antibunching phenomenon, which has a time constant deﬁned by the ﬂuorescence lifetime and the excitation rate. Antibunching reﬂects the fact that directly after emission of a photon from the excited state the molecule needs to be re-excited before being able to emit the next photon.58,60,66 Antibunching appears as a growth in the correlation curves (Fig. 6). Any event that decreases the probability of forming the singlet excited state of the molecule appears as a decrease

Fig. 6 Correlation curve showing the various processes that occur in different time regimes. Supramolecular dynamics can be measured in any of the time regimes between the antibunching and diffusion phenomena.

180

T.C.S. PACE AND C. BOHNE

in the correlation curve (Fig. 6). The correlation eventually disappears when the molecule diffuses out of the detection volume. Indeed, one of the most common applications of ﬂuorescence correlation spectroscopy is to characterize the diffusion parameters of molecules in solution.58,61 Fluorescence correlation spectroscopy can be used to measure the binding dynamics of host–guest complexes when the ﬂuorescence quantum yields for the free and bound hosts are different. Analysis of ﬂuorescence correlation spectra depends on the proﬁle for the excitation pulse, which impacts the shape of the emission proﬁle and mechanistic assumptions are made with respect to the diffusion of the various species in solution.58 For each chemical system different assumptions are made. The formation of a 1:1 complex between a guest and cyclodextrin was assumed for one recent study using ﬂuorescence correlation spectroscopy.65 The illumination proﬁle was assumed to be Gaussian, and in principle information on the host–guest entry and exit rate constants for the ground-, singlet- and triplet-excited states is contained in the correlation curve. The model can be simpliﬁed to include only the binding dynamics for the ground state. In the absence of a binding event when the correlation times for diffusion (tD) are much slower than for the triplet state (tT), which is turn is slower than the ﬂuorescence dynamics (tF) the correlation function is the product of the individual functions: GðtÞ ¼ G F GT G D

(8)

The association and dissociation processes of the triplet state guest were not incorporated into the derivation of the correlation function for the host–guest complex because at the low excitation energies and the low intersystem crossing rate constants for the guest the contribution from the triplet state dynamics to the correlation curve was negligible. The model derived assumes that the guest interconverts between species having different diffusion coefﬁcients and that the relaxation time for the complex formation (tR Equation (9), note that this equation is equivalent to Equation (3) stated in the theoretical background) is much faster than the average diffusion time for the guest in the homogeneous solvent or complexed to the host ð¯tD Þ. tR ¼

1 kþ ½H þ k

(9)

The derived function for the correlation function including the host–guest dynamics (GR(t)) is given by 2 !1=2 wxy 1 t 1 t G R ðtÞ ¼ 1þ 1þ ð1 þ AR et=tR Þ N G þ N HG t¯ D t¯ D wz

(10)

where NG and NHG are the mean numbers of free guest and complexed guest in the sample volume, AR is the amplitude for the reaction term of the correlation function and wxy/wz is the aspect ratio of the sampling volume. Fitting of the correlation curve to Equation (10) yields the value for tR. A similar model including the presence of the

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

181

triplet state was derived for the interaction of Rhodamine 6G with dGTP, a nucleotide.67 In summary, the advantage of using ﬂuorescence correlation spectroscopy is that very small sample volumes can be employed and the system is studied at equilibrium without the need to perturb the chemical equilibrium. The disadvantages of the technique are difﬁculties in differentiating between artifactual signals and relaxations related to chemical phenomena and the development of suitable models for ﬁtting the correlation curves.

NUCLEAR MAGNETIC RESONANCE

Nuclear magnetic resonance (NMR) experiments are used to study the exchange kinetics of chemical systems in equilibrium.28,68,69 As is the case for ﬂuorescence correlation spectroscopy no perturbation of the chemical system in equilibrium is required to obtain kinetic information from NMR experiments. However, NMR is not very sensitive to concentration changes. The kinetic information for NMR experiments is contained in the line broadening observed for a nucleus that resides in two different magnetic environments, and values for rate constants can be obtained using line-shape analysis.28,68,69 Line broadening experiments obtained using 1D NMR is the method of choice when analyzing the kinetics of a molecule in two sites. 2D NMR techniques, such as 2D EXYS are employed when the kinetics are sequential, i.e. more than one step, or multiple sites are analyzed.69 For example in the case of supramolecular systems this technique was employed to measure the exchange kinetics in capsules.70 For a host–guest equilibrium as deﬁned in Equation (1) the NMR signal for an active nuclei either in the guest (or host) has to have a different chemical shift when free in solution or bound to the host–guest complex. The line shape for the NMR peak(s) for the guest depends on the difference between the frequency for the NMR peak of the guest free in solution (nG) and the guest bound to the host–guest complex (nHG) and the residence time for the guest in each environment,28 deﬁned in Equation (1) by the entry (k+[H], assuming [H]4[G]) and exit (k– ) rate constants. The two extreme situations are when the host–guest binding dynamics are slow compared to spin–spin relaxation time (T2) and separate peaks are observed for the guest in the homogeneous phase and in the host–guest complex, and when the dynamics are fast compared to T2 leading to the observation of one peak (Fig. 7). In the slow exchange limit the residence time of the guest in the homogeneous phase and the host complex is long and no exchange between the two sites occurs during the relaxation process of the NMR experiment. In this scenario, the residence time of the guest is longer than T2 and the peak intensity corresponds to the fractional population of the guest in each site and the width of the two signals is determined by T2 (1/pT2) (Fig. 7a). As the residence time of the guest in each environment is shortened the two lines broaden (Fig. 7). When the value of the residence times for G in the homogeneous phase (k+[H]) and within the complex (k– ) are intermediate between T2 and (nG – nHG)1 the NMR peaks for G and HG are

182

T.C.S. PACE AND C. BOHNE

Fig. 7 NMR spectra for a guest in water (G) and in the host–guest complex (HG) when the residence time in the host–guest complex is long (a), progressively shorter and comparable residence times to T2 (b and c) and faster than T2 (d).

separate but broadened. The observed additional line width is related to the association (k+[H]/p) and dissociation processes (k– /p) (Fig. 7b).28 When the relaxation time for the binding dynamics (k+[H]+k– )1 is on the order of the reciprocal frequency difference the two lines coalesce and a very broad single line is observed (Fig. 7c). Further increase in the relaxation time for the binding dynamics leads to a sharpening of the single line. The term due to the broadening of the line because of exchange of the guest between two environments is deﬁned by 28 kþ ½H þ k 2 2 2 Dn1=2exchange ¼ 4pðnG nHG Þ f G f HG (11) kþ ½Hk where fG is the fraction of guest in the homogeneous solution ([G]/([G]+[H])) and fHG is the fraction of guest bound to H (1fG). Finally, when the exchange rate is much faster than the reciprocal frequency difference the line is sharpened and no kinetic information can be obtained. The position of the line in the spectrum is determined by the values of nG and nHG weighted by the fractions of the free and bound guest. At the intermediate broadening cases the kinetic information can be obtained from ﬁtting the shape of the spectrum to the appropriate equations.28 Changes in temperature are frequently used to change the broadening of the spectra. The range of rate constants accessible for NMR techniques is narrow and covers relaxation process from 0.1 ms to seconds for line shape analysis and the range from 10 ms to 100 s for 2D EXYS.69 The biggest advantage of NMR experiments is that the assignment of the signal to a particular moiety of the guest and host can be achieved with great accuracy from structural NMR assignments.

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

183

SURFACE PLASMON RESONANCE

Surface plasmons are oscillating surface charge density waves that propagate parallel to the surface at the interface between a metal and a dielectric.71 Surface plasmon resonance (SPR) is the basis for sensors measuring equilibrium constants of biomolecular interactions72 and has also been applied to the measurement of binding dynamics. This technique is based on the immobilization of one of the reactants on a metal surface, and the ﬂow of an excess of the other reactant over this surface. We will arbitrarily assume that the guest is bound to the surface and the host is the excess reagent in the solution that ﬂows over the surface. When the two reactants interact at the surface there are changes in the refractive index at the metal surface and these changes can be measured. An SPR instrument therefore can monitor, in real time, changes at the metal surface that are consequences of complex formation between the host and guest. The conﬁguration most often used in SPR instruments relies on the phenomenon of total internal reﬂectance and was developed by Kretchmann (Fig. 8).71,73 Total internal reﬂectance occurs when light traveling from a medium of higher refractive index toward a medium of lower refractive index reaches the interface and is reﬂected back completely into the higher refractive index medium. An important side effect of total internal reﬂection is the propagation of an evanescent wave across the interface into the medium of lower refractive index. In SPR experiments p-polarized light of a certain wavelength strikes the interface between the two media,71,72 which is coated with a thin metal ﬁlm. The wave vector of the evanescent wave is given by the following equation:71 wo Z sin y K ev ¼ (12) c g where wo is the frequency of the incident light, Zg the refractive index of the glass, y the angle of incidence of light and c the speed of light in a vacuum. The surface plasmon wave vector can be approximated to:71 sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ wo m Z2s (13) K sp ¼ c m þ Z2s where em is the dielectric constant of the metal ﬁlm and Zs is the refractive index of the dielectric medium. This latter parameter is the one that will vary when a host binds to a guest immobilized on the metal surface.

Fig. 8 Schematic representation of a surface plasmon resonance experiment.

184

T.C.S. PACE AND C. BOHNE

The evanescent wave produced by the total internal reﬂection of the incoming light is able to couple with the plasmons in the metal ﬁlm and resonantly excite the surface plasmons when Ksp ¼ Kev,71,73 which corresponds to a speciﬁc angle of incidence. This process causes the incident light to lose energy to the metal ﬁlm, and results in the reduction of intensity of the reﬂected light at this speciﬁc angle. The energy loss of the reﬂected light is observed as a sharp minimum in the angle-dependent reﬂectance.73 Ksp is dependent on the refractive index of the medium above the metal ﬁlm. Changes in the refractive index immediately above the surface therefore cause the angle of incidence which will result in the excitation of a surface plasmon to change.71,72 Often the incident light used has a wide range of incident angles on the interface and a 2D detector is used to simultaneously record all the reﬂected light intensity. The angle of incidence that corresponded to the excitation of surface plasmons can be determined from the position where the detected light intensity is less than expected. The change in the angle of incidence is reported as resonance units (RU), and a response of 103 RU represents an angle change of 0.1 degree.71,72 In a typical SPR experiment real-time kinetic study, solution ﬂows over the surface, so desorption of the guest immobilized on the surface due to this ﬂow must be avoided.72 In the ﬁrst stage of a typical experiment the mobile reactant is introduced at a constant concentration ([H]o) into the buffer ﬂowing above the surface-bound reactant. This favors complex association, and the progress of complex formation at the surface is monitored. The initial phase is then followed by a dissociation phase where the reactant is removed from the solution ﬂowing above the surface, and only buffer is passed over the surface to favor dissociation of the complex.72–74 The obtained binding curves (sensograms) contain information on the equilibrium constant of the interaction and the association and dissociation rate constants for complex formation (Fig. 9). In the simplest case a 1:1 complex is formed between the host in solution and the guest immobilized on the surface. The response of the SPR sensor, R, is proportional to the concentration of the complex formed, and thus pseudo-ﬁrst-order rate equations can be used to analyze the data.73 If no host is initially bound the function for R

Fig. 9 Sensogram for a surface plasmon resonance experiment where the growth corresponds to the association process, while after to the dissociation process is measured.

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

185

is given by RðtÞ ¼ Req ½Ho ð1 ekobs t Þ

(14)

where Req is the response value reached at equilibrium, [H]o is the concentration of host in the solution that is ﬂowing over the metal surface containing the guest, and kobs is deﬁned as the sum of the association process (k+[H]) and the dissociation process (k– ) (Equation (3)). If the host is removed from the solution ﬂowing over the metal surface at time to the dissociation of the host from the immobilized guest follows an exponential function: RðtÞ ¼ Req ek ðtto Þ

(15)

The advantages of SPR experiments are that only small amounts of sample are required,72 often hundreds of microliters of solutions with nanomolar to micromolar concentration of reactants and the substrate attached to the surface can oftentimes be reused after washing in buffer. The fact that changes in the refractive index values are measured avoids the need to use absorption or ﬂuorescence markers to follow the binding kinetics. The largest SPR signals are obtained for large changes in mass, and for this reason the best approach is to immobilize the lowest molecular weight component of the host–guest complex on the surface. However, available attachment chemistries may make it necessary to attach the higher molecular weight molecule, thus reducing the detected signal.74 It is also important to ensure that surface immobilization does not alter the structure of the metal bound reactants, neither does it restrict the access for binding of the reactant in the solution phase.73,75 From the kinetic point of view SPR experiments have the advantage that both the association and dissociation processes can be measured from the two phases in one sensogram. However, it is possible for artifacts to arise from refractive index mismatch during the buffer change and, for this reason, in general the initial parts of the association and dissociation phases are excluded from the kinetic analysis.73 When multiexponential decays are observed it is important to distinguish between kinetics related to the chemistry and potential artifacts, such as conformational changes of the bound reactant or effects due to mass transport limitations.73,75 The upper limit of detectable association rate constants has been estimated to be on the order of 105–106 M1 s1.73,75,76

3

Examples of supramolecular dynamics studies

The application of the techniques discussed above to the binding of guests to DNA or cyclodextrins (CDs) is described below. The intent of this section is not to provide an exhaustive review and analysis of the available data, but the objective is to use examples to show how the different techniques were employed in studies of supramolecular dynamics. The values for rate and equilibrium constants stated below

186

T.C.S. PACE AND C. BOHNE

were reported in the original literature with different precisions (i.e. signiﬁcant ﬁgures) and the error analysis, when provided, was variable. We have chosen to report data below with two signiﬁcant ﬁgures and we have omitted all reported errors. This procedure insured homogeneity in the presentation of data and we did not have to try to evaluate the error analysis in each report. The latter was not possible because of the reporting differences from the various studies. For this reason, readers interested in the accuracy and/or precision of individual values stated in the tables below should consult the original literature cited. DNA

There are several binding modes for small molecules with DNA including intercalation between the base pairs, groove binding in the major and minor grooves, as well as non-speciﬁc electrostatic interactions with the phosphate backbone.77 Intercalators generally feature planar tricyclic aromatic backbones containing heteroatoms and up to two ﬂexible cationic side chains. Minor groove binders tend to have a shape similar to that of the DNA groove and consist of repeating aromatic units with cationic end groups.78 Minor groove binders tend to show speciﬁcity for AT sequences due to speciﬁc hydrogen bonding interactions.79 The dynamics of intercalation of small molecules with DNA, groove binding and binding to speciﬁc sites, such as base pair mismatches have been studied by stopped-ﬂow,23,80–108 temperature jump experiments,26,27,94,109–120 surface plasmon resonance,121–129 NMR,86,130–135 ﬂash photolysis,136–138 and ﬂuorescence correlation spectroscopy.64 The application of the various techniques to study the binding dynamics of small molecules will be analyzed for speciﬁc examples of each type of binding. Intercalative guests Ethidium bromide (1) is a widely used guest molecule (Scheme 1). The property that makes 1 a good probe for DNA binding is that its ﬂuorescence quantum yield is very low in water and increases signiﬁcantly when 1 intercalates between the base pairs of DNA.139 The binding dynamics of 1 with DNA have been mostly analyzed assuming two different mechanisms. In one mechanism, a simple 1:1 stoichiometry was assumed and the observed rate constant is equal to the sum of the association and dissociation processes (Equation (3) where [H] ¼ [DNA base pairs]4[1]). An alternate mechanism assumes that 1 binds to two distinct sites in DNA and an interconversion between the two sites, A and B, can occur by the reaction with a second DNA molecule (Scheme 2). In general, the concentration of DNA is expressed as concentration of base pairs. At low [guest]/[DNA] ratios the binding of the guest can be assumed to be independent of the binding of other guests to DNA, whereas at high concentrations of DNA the data have to take into account that binding of one guest molecule precludes the binding of a second guest at adjacent base pairs and a sizeexclusion model is used.95,112,140

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS H2N

NH2

187

H2N

N

N C2H5

C2H5

(1)

H2N

H2N

(2)

NH2 H2N

N

NH2

CH3

O

N CH3

(3)

(4)

O

Scheme 1 Structure of the ethidium cation (1) and the derivatives ethyl phenidium cation (2), carboxy dimidium cation (3), and desphenyl dimidium cation (4). k+A

(G-DNA)A k–A

G + DNA

kBA

kAB

k+B k–B

(G-DNA)B

Scheme 2 Mechanism for the binding of ethidium cation (G) to two binding sites of DNA (A and B). For the intramolecular conversions, kAB and kBA are bimolecular rate constants where (G-DNA)A or (G-DNA)B react with a second DNA molecule leading to transfer of the guest.

In the original development of ﬂuorescence correlation spectroscopy the binding of 1 to DNA was investigated.64 Although the authors acknowledged that binding of 1 to Calf-thymus DNA (ct-DNA) was probably complex, their data were analyzed assuming a 1:1 stoichiometry. The association rate constant was determined to be more than two orders of magnitude smaller than the rate constant for a diffusional process141 and the lifetime of 1 intercalated within DNA was 37 ms (Table 1). The equilibrium constant calculated from the ratio of the rate constants was determined to be 5.6 105 M1, a value in agreement with the equilibrium constant determined by conventional ﬂuorescence titration (6 105 M1).64 This agreement suggests that the assumption made as to the binding stoichiometry was valid. A stopped-ﬂow methodology using deuterium exchange was employed to study the binding dynamics of 1 with ct-DNA.93 This method relies on the fact that the absorption spectrum of 1 changed upon deuteration, and the rate for deuteration was slowed down when 1 was bound to DNA when compared to the rate in aqueous solution. This method was validated by comparing the rate constant values with

188

T.C.S. PACE AND C. BOHNE

Table 1 Association (k+) and dissociation rate constants (k– ) for the binding of ethidium bromide to DNA assuming a 1:1 complexation stoichiometry Technique Fluorescence correlation spectroscopya Stopped ﬂowb Temperature jumpc Stopped ﬂowc Pressure jumpd

k+ (106 M1 s1)

k (102 s1)

Reference

15 1.6 6.4 5.4 7.4

0.27 0.30 0.16 0.39 0.21

64 93 94 94 142

a

ct-DNA. Sonicated ct-DNA, deuterium exchange method, pH ¼ 6.5. c Sonicated ct-DNA (average chain lengths of 500 bp), 24 1C, k– values calculated from reported values of K. d Poly [d(G-C)], 0.1 M NaCl, pH ¼ 7.2, 20 1C. b

Table 2 Rate constants for the binding of ethidium bromide to DNA assuming the mechanism shown in Scheme 2 determined in temperature jump (TJ) or stopped ﬂow (SF) experiments Guest Technique

1 1 2 3 4 1 1

TJa TJb TJc TJc TJc SFd SFe

kA þ (106 M1 s1)

1.4 2.6 3.4 4.1 7.3 40

2 1 kA (10 s )

1.6 1.7 76 4.3 20 0.39 0.70

kAB kB (102 s1) kBþ (105 M1 s1) (106 M1 s1) 0.48 0.26 3.2 0.20 1.6 1.1 3.0

0.35 0.59 3.4 0.47 1.7 0.14 0.25

21 1.3 34 3.5 12 6.0 18

kBA Reference (105 M1 s1)

2.5 1.2 6.5 2.7 15 86

112 120 120 120 120 94 94

a

1 M NaCl, sonicated ct-DNA, 19 1C. 23 1C personal communication by Ryan and Crothers in ref. 120. c Sonicated ct-DNA, 1 M NaCl, 25 1C. d Sonicated ct-DNA (average chain lengths of 500 bp), 24 1C. e Poly[d(A-T)], 24 1C. b

those obtained from stopped-ﬂow dilution experiments. The value for the dissociation rate constant was the same as measured in ﬂuorescence correlation spectroscopy experiments (Table 1) and for one of the binding sites when more than one relaxation process was observed (see below, Table 2), while a large difference was observed for the association rate constant. The H-D exchange method has not been widely employed probably due to the restriction that the guest needs to have exchangeable hydrogens, the absorption or ﬂuorescence spectra have to change with deuteration and the kinetic analysis is more complex because the deuteration reactions are included. Temperature jump studies were performed in combination with stopped-ﬂow experiments94 in response to criticism as to the reliability of temperature jump experiments26,142 and divergent interpretation of the kinetics. In particular, the suggestion from pressure jump experiments that the association step is diffusive in nature

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

189

led to the reinvestigation of this system.142 Addition of co-solvents was used to change the viscosity of the solvent and pressure jump experiments indicated that the association rate constants were dependent on viscosity, suggesting that the association process was diffusive in nature but only a small fraction of DNA sites had the ability to provide an intercalation site.142 The subsequent temperature jump and stopped-ﬂow experiments re-investigated this point by using DNA of different lengths and no dependence of the association and dissociation rate constants on size of the DNA was observed.94 The values for k+ and k– were compared for temperature jump and stopped-ﬂow conditions for DNA concentrations where the decay followed a mono-exponential function and no migration between DNA molecules occurred (see below).94 This report shows the importance of detecting ﬂuorescence signals at the magic angle, which eliminated the fast components in the kinetics due to artifacts. The values for the association and dissociation rate constants obtained by both techniques are similar. The complexity of the binding dynamics of 1 with DNA became apparent in subsequent temperature jump experiments, where three relaxation processes were observed.112 The fastest relaxation process had a small amplitude (o14%) and its kinetics were uncoupled from the second and third relaxation processes. This fast process was assigned to the binding of 1 to a minor site with a k+ value of 1.5 107 M1 s1 and a k– value of 6.9 103 s1.112 This assignment was problematic because of the possible interference of artifacts for temperature jump experiments when the ﬂuorescence detection is not performed at the magic angle,29 and this kinetic component was not observed in later studies.94,120 The second and third relaxation processes were coupled, where the observed rate constants differed by a factor of 3 to 7 and the rate constant for each relaxation process varied linearly with the DNA concentration.112 This dependence is consistent with the mechanism shown in Scheme 2, where 1 binds to 2 different sites in DNA and an interconversion between the sites is mediated in a bimolecular reaction with a second DNA molecule. For such coupled kinetics, the sum and the product of the two relaxation rate constants are related to the individual rate constants shown in Scheme 2. Such an analysis led to the values for the dissociation rate constants from each binding site, one of the interconversion rate constants and the association rate constant for the site with slowest binding dynamics (Table 2).112 The dissociation rate constant from one of the sites was similar to the values that were determined assuming a 1:1 binding stoichiometry (Table 1). A more detailed temperature jump experiment was performed in which the binding of 1 and its derivatives (2–4)120 was analyzed using the mechanism shown in Scheme 2 with a kinetic treatment along the same lines as described above.112 This study provides a good example of the detail of information that can be gained from kinetic studies. The results were interpreted as the binding of guests to intercalation sites accessible from the major (site A) and minor grooves (site B) of DNA.120 The larger changes in the association rate constant for site B were suggested to be due to the more restricted environment in the minor groove leading to a larger steric component for the association process. The association to site B is slower than for site A for

190

T.C.S. PACE AND C. BOHNE

guests containing amino substituents in the 3- and 8-position, while the reverse is observed when one amino group is removed as is the case for 2, suggesting that for this guest the steric restrictions and/or desolvation requirements are less restrictive. A large effect of the position of the amino substituents and the presence of the phenyl ring was observed on the residence time of the ethidium derivatives within site A, suggesting that these molecular motifs enhance the stability of complex formation. In contrast, the residence time in site B shows less dependence on speciﬁc molecular features. The transfer rate constant from site A to site B shows a larger dependence on the guest’s structure than the reverse rate constant. This result was interpreted to mean that for transfer into site B similar steric constraints were involved as for the binding of the free guest, while for the transfer from site B to site A rearrangements of the DNA structure may be more important.120 The effect of high DNA concentrations on the binding dynamics of 1 with DNA was also compared for temperature jump and stopped-ﬂow experiments, where at low concentrations both techniques yielded the same results (see above, Table 1).94 At high DNA concentrations, three relaxation processes were observed in the temperature jump experiments where the one occurring on the longest time scale was assigned to temperature-dependent conformational changes and not to the binding dynamics of guests with DNA. The other two processes were apparent in both types of measurement and were interpreted with the interconversion model (Scheme 2). The values for the rate constants were obtained from the analysis of the stopped-ﬂow experiments, because of the better signal-to-noise ratio for this technique when compared to the temperature jump experiments (Table 2). Finally, no dependence on the temperature, DNA concentration and salt concentration was observed for a temperature jump study using ct-DNA that was not sonicated.27 Based on these results the authors concluded that only large-scale dynamics of the DNA were responsible for the binding kinetics of 1 to DNA, and they suggested that studies with short length DNA may not be relevant for in vivo situations. In summary, the description for the binding of 1 to DNA above shows that kinetics studies with different techniques led to a mechanistic picture for the binding dynamics of this guest with DNA where the association process is several orders of magnitude slower than a diffusional process and the residence time within the intercalated site is of several tenths of milliseconds. Experiments with good signal to noise ratios showed that two different binding sites exist for 1 and at high concentrations of DNA direct transfer from one intercalating site to the second type of site in a different DNA molecule is possible. A variation of the values for k+ and k– between the different studies is apparent, which in part is due to different experimental conditions such as temperature and pH. In addition, the uncertainty as to the homogeneity of the DNA samples, i.e. composition and length, probably also plays a role in the differences reported. Acridine derivatives, such a proﬂavine (5) and acridine orange (6) (Scheme 3), are a second class of intercalative DNA guest molecules for which binding dynamics have been extensively studied. Temperature jump studies on the binding dynamics of 5 with ct-DNA and T2 Bacteriophage DNA showed two lifetimes in the relaxation kinetics.117 The observed

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

H 2N

N

NH2

N

N

H

191

N

H (6)

(5)

Scheme 3 Acridine cation derivatives: Proﬂavine (5), Acridine Orange (6). K1 G + DNA

k2 (G-DNA)A

k–2

(G-DNA)B

Scheme 4 Mechanism for the binding of proﬂavine (5) with DNA.

rate constant for the slowest relaxation process increased with the concentration of DNA but at high concentrations of the host the value for the relaxation rate constant leveled off. Two possible mechanisms can explain this result. In one mechanism guest 5 forms a weak complex in a fast manner followed by the intercalation of the guest between the DNA base pairs (Scheme 4). In an alternate mechanism, 5 is bound to two independent sites in a parallel fashion (Scheme 2 without the possibility for interconversion of the guest between sites). The latter mechanisms could not be rigorously ruled out, but it was deemed unlikely since it would require a very high association rate constant for the intercalation process. The linear dependence of the rate constant for the ﬁrst relaxation process on the concentration of DNA was analyzed using Equation (3) (Table 3). The dependence of the observed rate constant for the second relaxation process on the DNA concentration is given by117 kobs ¼

K 1 k2 ½H þ k2 1 þ K 1 ½H

(16)

where the overall equilibrium constant between 5 and DNA is related to the equilibrium constants for each step by K ¼ K 1 ð1 þ K 2 Þ

(17)

The non-linear dependence of the relaxation process on the DNA concentration was also observed in stopped-ﬂow experiments and the same mechanism, i.e. fast pre-equilibrium followed by a slow intercalation step, was proposed.99 This latter study did not report values for the individual rate constants. The mechanism proposed in Scheme 4 was employed in subsequent studies despite the criticism on the accuracy for the data related to the fast kinetic component (see below). The original temperature jump study also showed that the relaxation kinetics depend on the structure of the DNA.117 The slower intercalation rate for 5 with T2 Bacteriophage DNA when compared to ct-DNA was ascribed to the glucosylation of the former DNA (Table 3).

192

T.C.S. PACE AND C. BOHNE

Table 3 Rate constants for the binding of proﬂavine (5) to DNA assuming the mechanism shown in Scheme 4. Guest 5a 5a 5b 5c 5d 5d 5d 5e 5*f

DNA ([NaCl])

k1 (106 M1 s1)

k– 1 (104 s1)

K1 (103 M1)

k2 (103 s1)

k– 2 (102 s1)

K2 (M1)

Reference

ct (0.2 M) T2 Bacteriophage ct (0.24 M) ct (0.11 M) ct (0.5 M) ct (0.2 M) ct (0.05 M) poly[d(A-T)] poly[d(A-T)]

14

1.3

28

1.1

1.1 4.6 2.5 4.2 3.5 6.7 7.5 4.6 4.2

6.9 0.70 1.8 6.8 3.3 3.3 4.3 3.6 0.81

4.2 1.6 1.1 4.3 3.0 2.6 2.5 0.8 2.5

16 4.4 16 16 11 13 17 45 3.2

117 117 116 111 26 26 26 119 113

a

Sonicated DNA, pH 6.9, 25 1C. Sonicated DNA, 10 1C. c Sonicated DNA, pH 7.0, and 25 1C. d pH 6.8, 25 1C. e pH 7.0, 17 1C. f Excited state of 5. b

A second temperature jump study on the dynamics of 5 with DNA showed that the fast relaxation process was overlaid with the artifactual signal from the electrical discharge used to raise the temperature.116 This effect was more prominent at the lower ionic strength values employed. The relaxation kinetics were measured at the higher salt concentration to minimize the effects from the artifacts. The same qualitative dependence for the observed rate constants for the fast and slow components on the DNA concentration was observed as in the earlier study.117 The rate constants were of the same order of magnitude in both studies, but comparisons are not warranted because the experiments were performed at different temperatures (Table 3). The same chemical system was reinvestigated with the same experimental technique but detection at the magic angle was employed to eliminate the artifact due to electrical discharge.111 The kinetics observed were mono-exponential and the observed rate constant leveled off to a constant value at high DNA concentrations. The kinetic data were also analyzed with Equation (16) assuming the mechanism in Scheme 4 (Table 3). A further study used laser temperature jump measurements where the solution is heated up by absorption of light, which circumvents the possibility of artifacts due to electrical discharge.26 The DNA concentration was varied by two orders of magnitude at high [DNA]/[5] ratios to avoid the formation of aggregates. The kinetics followed a mono-exponential function and the values of kobs leveled off at high concentrations of DNA. The kinetics were analyzed assuming the sequential mechanism and using Equation (16) (Table 3). A parallel mechanism was ruled out in this study because the rate constant for the intercalation step would be too high to be consistent with a process where the guest is bound to a constrained site. The laser temperature jump experiment had time resolution to observe the relaxation process for the formation of the loose complex, which was not observed. The lack of detection of this relaxation process was ascribed to the similarity of the absorption spectra of 5 free in solution and when bound to the outer surface of DNA.

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

193

The studies described above with one chemical system showed that early temperature jump experiments suffered from artifacts, which were eliminated in subsequent experiments, by the use of the detection at the magic angle. From the qualitative point of view all experimental results are in agreement with the leveling off of the rate constant for the slow relaxation process at high DNA concentrations. Some variability was observed for the values of the rate constants and equilibrium constants, but considering the variations in experimental conditions the agreement is good. The effect of the base sequence on the binding dynamics of 5 with DNA was investigated using temperature jump experiments by measuring the relaxation kinetics of 5 with polydeoxynucleotides (poly[d(G-C)], poly[d(A-T)]) and natural DNAs, Micrococcus lysodeicticus (ml-DNA, 72% GC content) and Bacillus megaterium (bm-DNA, 70% AT content), with different contents of base pairs.119 For poly[d(A-T)] one relaxation time was observed that leveled off at high DNA concentrations and the data were analyzed using the sequential mechanism (Table 3). Noticeable is the much longer residence time (lower k2 value) for 5 in the intercalated site when compared to measurements with natural DNAs. For poly[d(G-C)] only one relaxation process was observed and the assignment was made that 5 did not intercalate into the base pairs of this polydeoxynucleotide. This assignment was later revised in another study which suggested that intercalation occurs into the GC base pairs and the linear dependence observed was due to the low concentrations of DNA employed.118 The important point of this study119 is that the binding dynamics of 5 are different for the two types of polydeoxynucleotides. When samples of poly[d(G-C)] and poly[d(A-T)] were mixed a further longer lived relaxation time was observed, which was assigned to the exchange of 5 between GC and AT sites.119 Kinetics with positive and negative amplitudes were observed for the natural ml-DNA and bm-DNA samples binding to 5.119 The kinetics were monitored by the ﬂuorescence of 5 and the emission quantum yield for 5 bound to GC sites was much lower than for 5 bound to AT sites. The positive amplitudes were related to relocation of 5 from GC sites to AT sites. The different kinetics observed for ml-DNA compared with bm-DNA are in line with the observation that the binding dynamics of 5 with poly[d(G-C)] and poly[d(A-T)] were very different. The effect of the DNA sequence dependence on the binding dynamics of 5 and 6 with ct-DNA (42% GC content) and ml-DNA (72% GC content) was investigated using laser temperature jump experiments.118 Only one relaxation process was observed for both guests, but the presence of the leveling off effect at high DNA concentration was dependent on the guest and the type of DNA. No values for the rate constants were reported in this study. Flash photolysis studies with absorption or delayed ﬂuorescence detection were performed to compare the binding of ground and excited state guests with DNA.113,136 The triplet lifetimes for 5 and 6 were shown to be lengthened in the presence of DNA.136 The decays were mono-exponential with the exception of the high excitation ﬂux conditions where the triplet–triplet annihilation process, a bimolecular reaction, contributed to the decay. The residence time for the excited guest was estimated to be shorter than for the ground state, but no precise values for the rate constants were reported. However, the estimated equilibrium constants for the

194

T.C.S. PACE AND C. BOHNE

excited state136 were similar to the values reported by others for the ground state complexation of the guests with DNA. The binding dynamics of 5 in the excited state were studied with laser ﬂash photolysis experiments where the kinetics for the triplet state were followed by the decay of the delayed ﬂuorescence.113 The analysis of the kinetics included the coupling of the kinetics for the excited state binding dynamics with poly[d(A-T)] and the decay for the excited state back to the ground state. The data were analyzed assuming a sequential or a parallel reaction model. The two mechanisms cannot be differentiated in the analysis of the kinetics for the ground state. However, the data for the excited state guest kinetics were only consistent with the sequential mechanism (Scheme 4). The value for the equilibrium constant in the pre-equilibrium step was the same for the ground and excited state of 5 with poly[d(A-T)], while the intercalation step was less favored for the excited state (Table 3).113 Analysis of the decay of the triplet excited state of 6 bound to ct-DNA was employed to determine the mobility of cations along the DNA backbone.138 The kinetics were measured either by absorption or delayed ﬂuorescence detection in a ﬂash photolysis system. The metal ions employed, Ag+ and Mn2+, are known quenchers of triplet states. The decay for the triplet state followed a mono-exponential function in the absence and presence of metal cations indicating that the binding of 6 with the DNA was homogenous. A non-exponential decay would be expected if more than one site with different quenching efﬁciencies were present. The lifetime for the intercalated guest is of tens of milliseconds. The quenching events by the metal ions were faster than the residence time of the guest in the DNA suggesting that 6 bound to DNA can be viewed as static during the lifetime of the excited state and the unimolecular quenching observed was due to the mobility of the cations along the DNA backbone. The mobility of Mn2+ was shown to be faster (X105 s1) than the mobility for Ag+. This result agreed well with the fact that Ag+ binds more tightly to DNA than Mn2+.138 9-Aminoacridine carboxamide derivatives have been developed as potential antitumor agents.80,104,105 The binding kinetics of this class of molecule were studied by stopped-ﬂow, and for the experimental conditions used the guests were believed to be in their dicationic form. The SDS micelle sequestration technique was employed which was initially developed to study the binding dynamics of actinomycin with DNA.95 In this method a solution containing the 9-aminoacridine carboxamide derivative/DNA sample is mixed with a solution containing SDS micelles.80,104,105 The negatively charged micelles sequester the positively charged guest as soon as it dissociates from the DNA. For this reason, the lifetimes measured are related to the dissociation of the guest from the DNA and correspond to residence lifetimes of the guest bound to the DNA. Control experiments were performed where the relaxation rate constants measured were independent of the SDS concentration, suggesting that the addition of the micelles did not affect the kinetics. In addition, the total change in amplitude in the stopped-ﬂow experiments was compared to the amplitude change of the system in equilibrium. A value of 100% indicates that all the dissociation processes are accounted for in the stopped ﬂow experiments whereas smaller percentages indicate that fast processes are present that occur during the mixing time of

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

195

the stopped ﬂow experiment. A recent study showed that caution should be exercised when using the SDS sequestration technique, especially for large guest molecules, since a dependence of the values for the relaxation rate constants on the concentration of SDS was observed.106 The dissociation of 7 (Scheme 5) from poly[(G-C)] showed three relaxation times and the amplitude corresponded to the total signal, while the dissociation of 7 from poly[(A-T)] was faster and only two relaxation times, corresponding to 70% of the total signal were observed in the stopped-ﬂow experiment. The biological activity of this class of molecules was correlated to the presence of four relaxation times when the dissociation process is measured with DNA, in particular the presence of the longlived component of hundreds of milliseconds.86,104,105,132,143 The difference in the dissociation kinetics observed with the two polydeoxynucleotides indicates that intercalation into G-C sites is responsible for the biological activity. The dissociation of 7 from ct-DNA led to four relaxation times, a result that is in line with the relaxation times observed with poly[(G-C)] and poly[(A-T)]. A very large number of compounds were studied by this technique,104,105 and recently a correlation of the kinetics with structural aspects determined from X-ray crystallography studies was presented.80 The intercalation of 7 occurs through the minor groove and the function of the side chain is to form speciﬁc hydrogen-bonding interactions with the G residues.80,104,105 A small number of examples will be highlighted here to illustrate the change in kinetics with changes in the structure of the guest (Table 4). The methylene chain length for the carboxamide side chain is critical for the observation of the long residence time (7 vs. 8). Guest 9 does not have the amino group on the acridine ring and this moiety is not charged. Despite the lower values for the equilibrium constants the dissociation processes were slower. These slow kinetics were suggested to be due to the bulkiness of the methyl group and the requirement that it thread through the DNA for intercalation and dissociation.104 Most of the binding dynamics of compounds 10–12 were faster than the R X N O

N H

(CH2)n N

O

N H

7

R = NH2

n=2

10

X=N

8

R = NH2

n=4

11

X = CH

9

R = CH3

n=2

O

O 12

O

N H

Scheme 5 9-Aminoacridine carboxamide derivatives.

N

N

196

T.C.S. PACE AND C. BOHNE

Table 4 Relaxation lifetimes for the dissociation of 7–12 from DNA or polydeoxynucleotides104,105 Guest 7 7 7 8 9 10 11 12 a

DNA

t1 (ms)

t2 (ms)

t3 (ms)

poly[(G-C)] poly[(A-T)] ct-DNA ct-DNA ct-DNA ct-DNA ct-DNA ct-DNA

3 4 6 4 7 7 3 3

76 8 28 16 30 38 8 16

230 86 52 90

t4 (ms)

428 530

53

F/%a 106 70 90 85 97 32 33 38

Fraction of the total signal change that is slower than the time resolution of the experiment.

millisecond time resolution of the stopped-ﬂow experiment, showing that the presence of the carboxamide side chain is not sufﬁcient to ensure a long residence time in the DNA. These results indicated that the mechanism for the dissociation is different for 10–12 than that for guests 7–9. The fast dynamics for the xanthone derivative (12) is supported by laser ﬂash photolysis experiments with 2-aminoxanthone where a residence time between 1 and 140 ms was estimated for the intercalated guest.137 The dynamics of bisacridines, i.e. two acridine moieties linked by several different spacers, with DNA and polydeoxynucleotides were studied using stopped ﬂow and NMR.86,103,130 The binding dynamics were shown to depend on the type of spacer. Mobility along the DNA was observed for some guests, where this mobility was faster than complete dissociation into the aqueous phase.86 The studies using NMR compared the binding dynamics of 9-aminoacridine with that of bis-9-aminoacridine compounds containing linkers with different chain lengths and charge with d(AT)5d(AT)5.130 Measurements were performed for the hydrogen bonded imino protons of thymine, which permitted the monitoring of the intercalated guest in combination with the detection of base pair opening rates. The effect of temperature on the broadening of the spectra and on the spin–lattice and spin–spin relaxation rates was investigated. The former provides information on the guest’s residence time inside DNA from the line broadening of the spectra, while the changes in the relaxation rates provide information on the exchange of the imino protons with protons of water molecules in the absence and presence of guest. The mono-intercalation of 9-aminoacridine was interpreted using a model where a pre-equilibrium is followed by intercalation (Scheme 4), while for the bis-intercalators a second intercalation step was proposed for the second intercalating moiety on the guest. The data were interpreted in terms of jump rate constants which correspond to mobility of the guest along the d(AT)5-d(AT)5 backbone. The authors comment that the jump rate constants may not correspond to the dissociation rate constants measured by other techniques if the dissociation of the guest from the initial pre-equilibrium step is slow when compared to the de-intercalation step.130 Although this argument is correct it is in contradiction with previous experiments with other intercalating guests where the de-intercalation step was shown to be much slower than dissociation from the initial

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

197

complex (see above). The jump rate constant for 9-aminoacridine was determined to be 300 s1 (ca. 3 ms, 17 1C),130 a value similar to that determined by NMR for ethidium bromide.132 Although the rate constant determined by NMR is of the same order of magnitude as the de-intercalation rates observed for other guests such as ethidium bromide and proﬂavine (Tables 1 and 3), NMR does not provide the detailed mechanistic information when kinetics are multiexponential, and the determined rate constants by NMR are probably better interpreted as average rate constants. The biggest advantage for the NMR experiments is that the effect of guest interaction into d(AT)5-d(AT)5 enabled the investigation of the effect binding has on the base pair dynamics in the host.130 Intercalation of 9-aminoacridine and the bisacridines with neutral linking chains did not change the exchange rate for the protons on the bases of d(AT)5-d(AT)5. A relatively fast jump rate constant between 100 and 500 s1 was observed for these guests. However for the bis-intercalator with a positively charged linking chain an opposite effect was observed on the dynamics of the imino protons. The protons at the ends of the d(AT)5-d(AT)5 backbone exchanged faster, while the one close to the binding site exchanged slower. Daunomycin (13, Scheme 6) is an anthracycline derivative, which corresponds to a class of anticancer drugs that form intercalative complexes with DNA. Temperature jump, stopped ﬂow and stopped ﬂow combined with SDS sequestration experiments were employed to study the binding dynamics of 13 and derivatives with DNA.84,91,96–98 The experimental challenges were similar to those described above for structurally simpler guests. However, the kinetics were more complex and mechanisms involving the formation of a pre-equilibrium complex followed by sequential unimolecular steps, or parallel steps were proposed. Studies with polydeoxynucleotides96 and with 16 base pair deﬁned oligonucleotides containing the central sequence of CGTACG, TAGCTG, TCATCC and TATATA98 suggested that the complex kinetics observed are due to binding of 13 to sites with different afﬁnities, i.e. strong and weak binding sites. These results underscore the importance of studying the binding kinetics of guests with DNA using complementary techniques and deﬁned base pair sequences. The latter is important when guests have several different binding motifs. O

OH

O

OH

OH O

O O

13

NH2 HO

Scheme 6 Daunomycin (daunorubicin), an anthracycline derivative.

198

T.C.S. PACE AND C. BOHNE

R8

O

R5

O

H

N

N

H

R4

14 R5 = R8 = H R4 = NH(CH2)2NH+(CH2CH3)2 15 R4 = R8 = H R5 = NH(CH2)2NH+(CH2CH3)2 16 R4 = R5 = H R8 = NH(CH2)2NH+(CH2CH3)2 O R2 R6 O 17 R2 = R6 = NHCO(CH2)2NH+(CH2CH3)2

Scheme 7 Anthraquinone derivatives.

The effect of the position of side chains on the intercalation kinetics of anthraquinones,101 which are related to the aromatic moiety of daunomycin (13), was studied with the stopped ﬂow SDS sequestration technique. Guest molecules 14–17 can have two different intercalation modes, a classic mode where both side chains are located in the same groove of the DNA helix (14 and 16) or a threading mode where the side chains are located in opposite grooves of the DNA (15 and 17) (Scheme 7). The relative position of intercalated 14 with respect to the DNA bases was suggested to be the same as for 13. The kinetics for all guests were ﬁt to the sum of two exponentials. The recovered observed rate constants differed by a factor of 4 for the kinetics of these guests with ct-DNA and by a factor of 10 for the kinetics of the guests with the polydeoxynucleotides. For this reason, the kinetics were analyzed by determining an apparent observed rate constant deﬁned by the fractional amplitudes (Ai) and the individual rate constants: kapp obs ¼ A1 k1 þ A2 k2

(18)

The apparent rate constant increased with the concentration of DNA and values for the association (ka) and dissociation (kd) rate constants were determined from the linear relationship of kapp obs with the sum of the association process (ka[DNA]) and dissociation process (kd), equivalent to Equation (3) (Table 5). Note that the parameters ka and kd will be employed when these values were determined from overall apparent rate constants to differentiate them from rate constants (k+ and k– ) determined for an elementary process. The most important conclusion from this work is that despite similar equilibrium constants the association and dissociation rate

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

199

Table 5 Association and dissociation rate constants for anthraquinone derivatives with ct-DNA and polydeoxynucleotides from ref. 101 Guest 14 15 16 17

a

DNA

ka (105 M1 s1)

kd (s1)

K ¼ ka/kd (105 M1)

ct-DNA poly[(A-T)] poly[(G-C)] ct-DNA poly[(A-T)] poly[(G-C)] ct-DNA poly[(A-T)] poly[(G-C)] ct-DNA poly[(A-T)] poly[(G-C)]

24

11

19 2.1 22 1.5 7.0 71 3.9 1.5 12 0.92

9.3 0.93 3.3 1.3 18 58 7.6 3.2 3.5 4.7

2.2 – 2.0 2.3 6.7 1.2 0.39 1.2 0.51 0.47 3.4 0.20

a

a

Too fast to be measured.

constants can differ signiﬁcantly, in some cases by an order of magnitude. Both the association and dissociation processes were faster for guests (14 and 16) where the side chain with a positive charge does not have to be threaded through the DNA helix. The dynamics were slower for the threading guests because the distortion of the DNA to let the side chain pass through to the other side of the helix requires considerable structural rearrangement. The dynamics for all guests were faster with poly[(A-T)] than with poly[(G-C)], and with the exception of the dissociation of 17 the values for ka and kd were higher for poly[(A-T)] than for poly[(G-C)]. Compounds 15–17 showed a slight preference for the AT sequence over the GC sequence. The dynamics of a naphthalene diimide with alkylamino substituents (18) were studied using stopped ﬂow measurements.102 The kinetics were ﬁt to the sum of two exponentials and the data were analyzed as described above for the anthraquinone derivatives (Table 6).101 The binding dynamics of 18 to polydeoxynucleotides showed that the dissociation was much slower from the GC sequence when compared to the AT sequence. This effect is primarily responsible for the higher equilibrium constant of 18 with poly[d(G-C)]2, leading to a selectivity of this guest for the GC sequence. The binding dynamics for 18 with the polydeoxynucleotides were much slower than observed for monocationic intercalating guests with similar equilibrium constants. The slow dynamics for 18 were ascribed to threading of one of the alkylamino chains through the polydeoxynucleotide helix.102 The threading binding motif was supported by the effect of changes in the salt concentration on the binding kinetics (Scheme 8). Surface plasmon resonance studies were employed to measure the equilibrium constants and association and dissociation rate constants of bisnaphthalimide derivatives (20, 21) with hairpin DNA immobilized on the metal surface.123 The equilibrium constants were higher and the dynamics slower for compounds 20 and 21 when compared to the equilibrium constants and dynamics of the model monomer (19). The values for ka and kd were determined from the change in the surface plasmon resonance signal when, respectively, the ligand solution was ﬂowed over the

200

T.C.S. PACE AND C. BOHNE

Table 6 Association and dissociation rate constants for compounds 18–21 with DNA Guest 18 19 20 21

DNA

ka (105 M1 s1)

poly[d(G-C)]2 poly[d(A-T)]2 [AT]4b [CG]4b [AT]4b [CG]4b [AT]4b [CG]4b

1.4 3.4

kd (s1)

K ¼ ka/kd (105 M1)

Reference

6.7 (5.0)a 0.46 (0.25)a (0.47)a (0.98)a 97 (31)a 880 (460)a (14)a 160 (46)

102 102 123 123 123 123 123 123

0.21 7.4

c

c

c

c

3.3 2.2

0.034 0.0025

4.5

0.029

c

c

a

K values in parentheses were determined from equilibrium measurements. 5 -Biotin-labeled DNA hairpins, central sequence indicated. c Too fast to be measured. b 0

O

O

O H N

N H N N H O

N O

O

N

18 O N

R NH

R NH

19

O N

N

N O N

20 R = H 21 R = CH3

O N

Scheme 8 Naphthalene diimide (18), pyrazinonaphthalimide (19), and bis pyrazinonaphthalimide derivatives (20, 21).

metal surface containing DNA and a buffer solution was used for the dissociation phase. It is worth noting that this technique is suitable for kinetic measurements that are slower than measurable with stopped-ﬂow. In addition, a larger discrepancy is found for the surface plasmon resonance data when compared to the stopped ﬂow experiments between the equilibrium constant determined from the ratio of rate constants or from thermodynamic (equilibrium) experiments. A better correlation between the equilibrium constants determined from kinetic and thermodynamic measurements obtained from surface plasmon resonance investigations were observed for the binding of sandramycin derivatives to DNA, where the values for the association rate constants varied from 95 to 1.2 104 M1 s1 and the dissociation rate constants were smaller than 2 103 s1.121 Addition of the methyl group between the two intercalating moieties in 21 led to a signiﬁcant decrease in the binding afﬁnity and slower dynamics.123 This effect is more pronounced for the GC sequence than the AT sequence and is primarily due to a faster dissociation rate constant for 21 compared to the dynamics for 20. The

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

201

methyl group is believed to restrict the molecular contacts between the linker chain and the sugar-phosphate backbone of the polydeoxynucleotide. This example indicates that molecular interactions exterior to the intercalating site can be responsible for determining the binding dynamics. The selectivity for the GC sequence was conﬁrmed with DNase I footprinting experiments which conﬁrmed the preferential binding of 20 and 21 to the GC sequence. Groove binding guests Distamycin (22) is known to bind to the minor groove of DNA with a preference for AT-rich regions. The binding occurs with two different stoichiometries, 1:1 (guest:DNA) and 2:1 (guest:DNA). Despite the large number of studies on the binding thermodynamics and structural characterization for the interaction of 22 with DNA, kinetic studies were limited by the small changes observed for the absorption spectra of free and bound guest. Fluorescence was used as the monitoring technique in stopped ﬂow SDS sequestration studies because it was shown that the weak ﬂuorescence of 22 in water is enhanced when this guest is bound to DNA.81 The complementary double helix with the sequence 50 -GCGATTAGCG-30 was used to study the kinetics for the 1:1 binding mode, while the sequence 50 -GCGAAGTTGCG-30 was used to study the binding dynamics for the 2:1 binding site. The kinetics for the 1:1 binding mode followed a monoexponential function and the observed rate constant varied linearly with the concentration of DNA for low [DNA]/[22] ratios. The association rate constant (k1) was obtained from the plot of kobs vs. [DNA], while the dissociation rate constant (k– 1) was obtained from dissociation experiments using the SDS sequestration technique (see above) (Table 7). The latter approach is necessary because of the small values for the dissociation rate constants. In the presence of excess DNA with respect to the concentration of 22, two relaxation times were observed and the slower process became slower as the concentration of DNA was increased. This unusual result is consistent with the presence of reactive and unreactive conformations of 22, where the slow component corresponds to the conformational change, or the binding of 22 to incorrect binding sites followed by relocation. In the case of the 2:1 complex studied with the longer oligonucleotide the ﬂuorescence intensity ﬁrst increased followed by slower kinetics with an emission intensity decrease. The fast relaxation process varied linearly with Table 7 Association and dissociation rate constants for the binding of 22 to d(GCG-YGCG) complementary base paired oligonucleotides at 25 1C Y -ATTA-AAGTT-AAGTT-AAACT-ATATA-AAAAT-

Stoichiometry

k1 (107 M1 s1)

k– 1 (s1)

k2 (s1)

Reference

1:1 2:1 2:1 2:1 2:1 2:1

6.6 7.5

3.1 0.08 0.028 2.0 0.14 2.5

15 5.1

81 81 85 85 85 85

202

T.C.S. PACE AND C. BOHNE

N O

NH H

O

N

O

HN

N H

N O

22

HN

NH2 H2N

Scheme 9 Distamycin structure.

the DNA concentration and a value for the association rate constant similar to the value determined for the 1:1 complex was measured (Table 7). This result suggests that the binding of the ﬁrst molecule of 22 was rate determining and the binding of the second molecule of 22 was faster. The slow relaxation process only depends slightly on the concentration of DNA. These results were interpreted with a sequential mechanism where the initial binding of two molecules of 22 was fast followed by a rearrangement to the more stable complex (Scheme 9). The dissociation of the 2:1 complex was much slower than for the 1:1 complex. The effect of the binding site sequence on the preference for 1:1 or 2:1 binding was explored by using 4- and 5-base pair sites in oligonucleotides. The residence times for 22 in the DNA complexes were determined in stopped-ﬂow experiments using absorption detection and SDS sequestration.85 The 4-base pair site can only bind one molecule of 22, while two guest molecules can be bound to the 5-base pair site. Binding to the 5-base pair site can occur sequentially with the initial formation of the 1:1 complex (e.g. –AAAAT- site), while for other sites binding as a 2:1 complex occurs at low 22/DNA ratios (e.g. -ATATA-, -AAACT- and -AAGTT-) suggesting that cooperative binding occurs. The rates for association and dissociation of 22 for the 1:1 complex and the rate of association for the 2:1 complex were too fast to be measured by the technique employed. However, the dissociation rate constants from the 2:1 complex were determined and showed a change of two orders of magnitude depending on the sequence of the binding site (Table 7). With the exception of the -AAACTbinding site the residence time for 22 was longer for those sites where the cooperativity for the 2:1 binding is largest, i.e. lack of initial formation of the 1:1 complex. NMR studies have provided great detail on the binding mode of 22 and derivatives to oligonucleotides and also provided estimates for the residence time of the guest in the complex.131,133–135 The upper limit for k– 1 from the -AATT- binding site was estimated to be 4 s1,133 a value that is similar to the value observed for the -ATTAbinding site determined by stopped ﬂow (Table 7).81 The dissociation rate constant of 22 from the 2:1 complex with the complementary d-(CGCGAATTCGCG) base pair oligonucleotide was estimated to be 0.2 s1 (30 1C),134 a value which is in the range determined by stopped-ﬂow experiments (Table 7).85 The effect of structural modiﬁcation to 22 on the binding efﬁciency and binding dynamics of hairpin DNA immobilized on metal surfaces were studied using surface plasmon resonance.124,125 The association and dissociation kinetics were analyzed using a sequential model for the binding of the imidazole containing polyamides

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

203

with DNA where the formation of the 1:1 complex was followed by the complexation of a second guest molecule. The association rate constants and dissociation rate constants observed for 22 with the hairpin DNA containing the sequence AAATTT were much lower124 than observed for DNA experiments in solution (Table 7).81,85 This difference could be due to the different types of DNA or the immobilization of the DNA on the metal surface. Studies with a series of polyamides124,125 showed than the terminal formamido group is essential to achieve long residence times for the guest in the minor groove of DNA. In the absence of the formamido group the association and dissociation kinetics were too fast to be measured with the surface plasmon resonance technique.125 In addition, thermodynamic and kinetic results showed that these polyamides have a selectivity for mismatched DNA sequences. This selectivity is due to slow binding kinetics in particular a slower dissociation rate constant.124,125 Cyanine dyes are a second class of molecules that bind to the minor groove of DNA. However, depending on their structure they can also intercalate into DNA.144–146 In the case of minor groove binding there seems to be a selectivity for AT sites, while no selectivity was observed for cyanine dyes that bind by intercalation. The methodologies employed to study the dynamics of these dyes with DNA have been described above and only a brief overview of the studies with this class of molecules will be provided. Guests 23 and 24 intercalate with poly(dA-dT)2 and poly(dG-dC)2.109,110 The binding dynamics were analyzed assuming the formation of an equilibrium complex followed by sequential rearrangements of the complex. The kinetics were measured using temperature jump experiments. Two relaxation processes were observed for both guests. In the case of 24 the binding dynamics for both processes were faster for poly(dG-dC)2 than for poly(dA-dT)2, while for 23 the same rate constants were measured for both polydeoxynucleotides (Scheme 10).110 Longer cyanine dyes, such as 25 bind to the minor groove of DNA and these types of molecules have been developed as ﬂuorescent DNA staining dyes, and kinetic information has been obtained to optimize the binding ability of these molecules to DNA.87,88 The binding dynamics of derivatives of 25 with ct-DNA were determined using the stopped-ﬂow SDS sequestration methodology. The decays were ﬁt to the sum of two exponentials and the apparent observed rate constants were calculated using Equation (18) (see above). The dissociation rate of a derivative of 25 with a charged amino alkyl chain was slower than for 25, whereas addition of a neutral methoxyethoxyethyl chain increased the dissociation rate. The dynamics of a series of Hoechst dyes (26) with hydroxy and methoxy substituents on the phenyl ring with poly[d(A-T)], d(CGCGAATTCGCG)2, and its corresponding T4-looped 28-mer hairpin were studied with stopped ﬂow methods.23,83 The kinetics followed an exponential decay and the association rate constant was obtained from the slope of the plot of kobs as a function of DNA concentration. The intercept for this plot, which corresponds to k– was in most cases too small to be determined with any precision. For this reason, the dissociation rate constants were obtained by using poly[d(A-5BrU)] as a sequestering agent since it was shown that the derivatives of 26 had high equilibrium and association rate constants with this polynucleotide. The afﬁnities of all derivatives of 26 were higher

204

T.C.S. PACE AND C. BOHNE

S

N

N 23 S

S N

N 24 N N

S O

N

25 N

N N

N H

N H

R3

N H

26

R2 R1

Scheme 10 Structures of Cyan40 (23), CCyan2 (24), Boxto (25) and Hoechst dyes (26, Hoechst 33258 for R1 ¼ R 2 ¼ H, R3 ¼ OH).

with the -AATT- binding sites than with poly[d(A-T)] as a result of the slow down of the dissociation process. In the case of the 28-mer hairpin DNA, the association rate constants varied from 2.0 108 to 2.9 108 M1 s1, whereas the dissociation rate constants varied from 0.012 to 0.42 s1. The longest residence time was observed for the bis-m-hydroxy-substituted derivative of 26. The association rate constants observed were close to the diffusion controlled limit in water and were signiﬁcantly higher than for distamycin (22) which also binds to the minor groove of DNA. This result suggests that very little rearrangement of the DNA backbone is required for the binding of 26.

CYCLODEXTRIN

Cyclodextrins are cyclic hosts made of D-glucose units (6, 7, or 8 for a-, b-, and g-CD) that form host–guest complexes with a number of organic and inorganic molecules. The size of the cavity depends on the number of glucose units and increases from 5 A˚ for a-CD, containing 6 glucose units to 8 A˚ for g-CD with 8 glucose units.147,148 Complex formation between CDs and small molecules is determined by steric constraints for the guest to ﬁt into the CD cavity and by the hydrophobicity of the guest. The characterization of these complexes from the thermodynamic point of view is

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

205

extensive,149,150 and both enthalpic and entropic factors can contribute to complex formation. Studies on the dynamics of complexation for guests with cyclodextrins have been carried out using ultrasonic relaxation,40,151–168 temperature jump experiments,57,169–183 stopped-ﬂow,170,178,184–197 ﬂash photolysis,57,198–202 NMR,203–205 ﬂuorescence correlation spectroscopy,65 phosphorescence measurements,56,206 and ﬂuorescence methods.45,207 In contrast to the studies with DNA described above, there are only a few examples in which different techniques were employed to study the binding dynamics of the same guest with CDs. This probably reﬂects that the choice of technique was based on the properties of the guests. The examples below are grouped either by a type of guest or under the description of a technique. Azo dyes The importance of understanding the guest binding dynamics with CD was recognized early on when CDs were being developed as host systems, and the ﬁrst comprehensive study involved the use of azo dyes (27).169 Temperature jump experiments were used in these studies. The binding dynamics were determined to be 1:1 for the phenol and phenolate derivatives and one relaxation process was observed for all guests investigated. A linear relationship was observed between the observed relaxation rate constant and the concentration of a-CD and the association and dissociation rate constants were obtained by using Equation (3), i.e. kobs corresponds to the sum of the association (k+[H], where the host in excess concentration is the CD) and the dissociation (k– ) process (Scheme 11). In the case where the arylsulfonate group is a benzene instead of a naphthalene the relaxation kinetics for guest complexation with a-CD measured by stopped-ﬂow showed either one or two relaxation processes.185,190 When one relaxation process was observed the dependence of the observed rate constant on the concentration of CD was linear and the values for the association and dissociation rate constants were determined using Equation (3). When two relaxation processes were observed the observed rate constant for the fast process showed a linear dependence on the

Scheme 11 Hydroxyphenylazo arylsulfonate derivatives 27 and 28.

206

T.C.S. PACE AND C. BOHNE

CD concentration, while the dependence for the slow process reached a limiting value at high concentrations of host. These results were interpreted as the fast formation of a pre-equilibrium complex, which rearranged to the stable complex over longer periods of time. It is important to note that this interpretation is based on the fact that a complex with a 1:1 stoichiometry is formed. The two-step mechanism is comparable to the mechanism proposed for the formation of the complex between proﬂavine and DNA (Scheme 4). The expression for kobs is given by Equation (16) where the limiting value for kobs at high host concentration is the sum of k2 and k– 2. The overall equilibrium constant is given by Equation (17). The kinetic studies with the derivatives of 27 showed that despite similar equilibrium constants the values for the rate constants varied by more than three orders of magnitude (Table 8).169 For this guest only the phenol moiety is small enough to ﬁt into the a-CD cavity, the naphthyl moiety being too large. Small changes to the bulk of the guest have dramatic consequences on the complexation kinetics. The equilibrium constants obtained from binding isotherms agree well with the calculated equilibrium constants from the kinetic data. This result suggests that the kinetics correspond to a measurement leading to the equilibrium state. The addition of alkyl groups to the position ortho to the phenol decreased the binding dynamics, with a correlation with the size of the substituent (Table 8). This result suggests that as the size of the guest increases the entry of the guest into the CD cavity becomes more difﬁcult and probably a distortion of the cavity has to occur to accommodate the guest. However, once inside the CD the exit rate constant of the guest also decreased. In the case of di-ortho substitution no complexation with a-CD was observed suggesting that the guest is too large to ﬁt inside the cavity. An interesting result was observed for the complexation of the phenolate derivative of 27 to a-CD (Table 8). The equilibrium constant was higher than for the phenol derivative, which at ﬁrst glance is anti-intuitive since the phenolate guest is expected to be better solvated than the phenol derivative in the homogeneous solution. However, the kinetics were consistent with the stronger solvation of the guest Table 8 Equilibrium constants and association and dissociation rate constants for derivatives 27 and 28 with a-CD determined by temperature jump (27, T ¼ 14 1C, I ¼ 0.5 M), stopped-ﬂow (28, T ¼ 25 1C, I ¼ 0.1 M) and with pressure jump experiments (29, T ¼ 25 1C, I ¼ 0.15 M) Compound 27 27 27 27 28 28 28 28 28 28 28 29 a

R¼H R ¼ H, phenolate R ¼ CH3 R ¼ CH2CH3 R¼H R ¼ CH3 R ¼ CH2CH3 R ¼ CH2CH2CH3 R ¼ CH(CH3)2 R ¼ CH(CH3)CH2CH3 R ¼ C(CH3)3

K (103 M1)

k1þ (104 M1 s1)

k1 (102 s1)

k2þ (s1)

k2 (s1)

K (103 M1)a

0.27 0.65 0.42 0.46 6.3 8.3 10 8.3 4.5 5.6 1.9 5.8

1300 17 12 0.6 4 103 70 2.1 2.0 1.2 1.1 0.046 1.4

550 2.6 3.5 0.19 4 10 0.77 0.065 0.060 0.094 0.14 0.0055 0.13

– – – – – –

– – – – – –

b

b

0.87 0.58 0.8

0.55 0.26 0.16

0.24 0.65 0.34 0.32 4 10 9.1 3.2 8.6 4.1 4.7 0.84 4.3

b

b

0.70

0.22

Calculated from kinetic data. The changes in absorbance were too small to determine rate constants.

b

Reference 169 169 169 169 190, 190, 190, 185, 185, 185, 190, 192

208 208 208 190, 208 190, 208 190, 208 208

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

207

in the bulk phase leading to a much lower association rate constant for the phenolate guest. The higher equilibrium constant is then a consequence of the lower dissociation rate constant for the phenolate derivative, probably because the negative charge has to pass through a relatively hydrophobic environment in the CD cavity before the guest exits into the aqueous phase. The derivatives of 28 had higher equilibrium constants185,190 than compounds 27, suggesting that the benzylsulfonic moiety for 28 also interacts with the CD. In contrast to compounds 27, the derivatives of 28 can incorporate with different directionalities through the wide and narrow entrances of the a-CD. The complexation was assumed to occur from the sulfonate side of the guest.190 Increase in the size of the substituent slowed the binding dynamics (Table 8). For the parent 28 and its methyl derivative the guest is small enough to slip into the cavity of the CD leading to the observation of one relaxation process. For larger substituents the presence of two relaxation times was interpreted as the fast formation of a complex followed by a slow rearrangement to the thermodynamically stable complex.185,190 Once the consecutive two-step mechanism was in operation the rate constants showed a much smaller dependence on the size of the substitutents (Table 8). More detail on the consecutive mechanism for the complex formation between azo derivatives with a-CD was obtained from the combination of high pressure stoppedﬂow studies with structural characterization by NMR.192,193 The advantage of using pressure as a variable is that information on volume changes for each relaxation process can be obtained. The one compound in this study that is similar to the compounds in the previously described studies is the derivative of 28 with a di-ortho methyl substitution with respect to the phenol group (29).192 For the mechanistic interpretation of the data it was important to ensure that a unidirectional inclusion of the guest occurred, which is achieved by the steric crowding around the phenol group. NMR data support that the benzylsulfonic moiety of 29 was included from the wider entrance of a-CD. In the stopped-ﬂow experiments two relaxation processes were observed, where the observed rate constant for the fast process showed a linear dependence on the concentration of CD and saturation at high host concentrations was observed for the second process. The effect of changes in temperature and pressure on both rate constants was investigated. In the case of the pressure studies, the changes in volumes of activation ðDa V Þ were obtained by using Equation (19) and by ﬁtting the data for both relaxation processes simultaneously.192 k ¼ ko e

Da V P RT

(19)

The addition of the second methyl group on the phenol ring led to the observation of the consecutive inclusion process with a decrease in the dynamics for complex formation (Table 8, cf. 29 with 28 (RQCH3)). This result supports the previous suggestion190 that small guests can slip into the CD cavity and in one process form the stable host–guest complex. Negative volumes of activation for the forward (23.6 cm3 mol1) and reverse (12.6 cm3 mol1) reactions were observed for the formation of the initial complex between 29 and a-CD. The overall volume change for this complex formation was

208

T.C.S. PACE AND C. BOHNE

also negative (11.0 cm3 mol1).192 For the second process the volumes of activation for the forward (12.4 cm3 mol1) and reverse (16.1 cm3 mol1) reactions were negative with the overall reaction having a slightly positive volume change (3.7 cm3 mol1). The combination of the kinetic results with NMR structural characterization led to the following mechanistic interpretation: The initial complexation is driven by hydrophobic interactions with the inclusion of the sulfonate moiety into the CD cavity. In this process, partial desolvation of the sulfonate head group and CD cavity occurs. In the initial complex, the sulfonate group displaces internal water molecules and this charged group is located inside the cavity. The second process is characterized by the protrusion of the sulfonate group through the narrow rim of the CD cavity and re-solvation, leading to an overall positive change in the reaction volume for this rearrangement process. A recent study,209 in which previous results on the complexation of a series of non-centrosymmetrical guests with CDs were re-evaluated, suggested that the two observed relaxation processes could possibly be interpreted as a mechanism involving two parallel reactions: inclusion of the guest through either the wide or narrow rim of the cyclodextrin. This mechanism was shown to lead to the same dependence of observed rate constants on concentration of cyclodextrin as the consecutive mechanism. This study showed that even for seemingly simple host systems the mechanistic details for complexation can be quite complex and still controversial. Stopped-flow Stopped-ﬂow experiments have been used in several instances to study the binding dynamics of guests with CDs.170,178,184–197 The types of experiments and mechanisms have been exempliﬁed above in the case of azo dyes. In addition to kinetic studies with organic guest molecules the complexation kinetics of Cu2+ with a-CD194 and a competition of the binding of a guest with several CDs and Ni2+ and/or Zn2+ were slow enough to be studied using stopped-ﬂow and temperature jump methods.171,184 CDs can form complexes with stoichiometries different from 1:1. Stopped-ﬂow experiments were employed to study the binding dynamics of a 2:2 complex between pyrene and g-CD.196 Both, 1:1 and 2:2 complexes are formed and the 2:2 complex exhibits excimer-like emission. The association rate constant for the 2:2 complex was found to be 6 107 M1 s1, while the dissociation rate constant was 73 s1. These values correspond to a decrease of up to 5 orders of magnitude when compared to the dynamics for the 1:1 complex. Temperature jump The ﬁrst comprehensive kinetic study on the binding dynamics of guests with CDs involving azo dyes169 was described above in which one relaxation process was observed. A study with different azo dyes showed two relaxation processes and the data were consistent with a sequential mechanism.170 The kinetics of a series of guests that form 2:1 (guest:CD) complexes were studied using temperature jump methods.179–183,210 The formation of the 1:1 complex was determined to be fast and to occur within the time resolution of the equipment, while

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS K11

G + H

209

HG

21

k+ G + HG

k–21 K22

H + HG2

HG2

H2G2

Scheme 12 Formation of guest–host complexes with multiple stoichiometries.

N

N N

SO3

30

R

R N

O

R N R

31 R = CH3 32 R = CH2CH3

Scheme 13 Methyl orange (30), pyronine Y (31), and pyronine B (32).

the kinetics for the formation of the 2:1 (guest:CD) complex were measured as a relaxation process. In some cases, the existence of a 2:2 complex formed in a fast process was also inferred (Scheme 12). Compounds 30–32 formed 2:1 complexes with CDs (Scheme 13). The formation of the 1:1 complex was fast and for this reason only one relaxation process was observed. In the cases where the 2:2 complex was present its formation was also fast and only one relaxation process for the 2:1 complex was observed in the temperature jump experiments. Since the equilibria are coupled the expression for the observed 21 rate constant includes K11 (and K22 when the 2:2 complex is present), k21 þ ; k , and the 180–182 concentrations of guest, 1:1 complex and CD. The values for the association and dissociation rate constants and equilibrium constants were obtained from the non-linear ﬁt of the dependence of kobs on the total concentration of CD (Table 9). The formation of the 2:1 (guest:CD) complexes with g-CD for all three guests was more efﬁcient than the formation of the 1:1 complex. The 2:1 complex was also more stable than the dimerization of the guests in aqueous solution at high concentrations of guest. These results suggest that the cavity of g-CD is too large when one guest is incorporated inside the cavity leading to a loose ﬁt. The rate constant for the association of the second guest to the 1:1 complexes was close to the diffusion controlled limit, while the residence time for the guest dimer varied from 70 to 200 ms. No great variation was observed for the rate constants as was reported for the binding dynamics of azo dyes to a-CD (see above), suggesting that the substituents

210

T.C.S. PACE AND C. BOHNE

Table 9 Equilibrium constants and association and dissociation rate constants for guests that form complexes with CDs with multiple stoichiometries Guest (CD) 30 (g-CD)a 31 (g-CD)b 32 (g-CD)c

K11 (102 M1)

K21 (105 M1)

k21 þ (10 M1s1)

k21 (103 s1)

0.45 10 4.3

20 1.2 1.3

9.4 1.7 0.82

4.8 14 6.4

9

Reference K22 (102 M1) 61 0.52 –

180 181 182

a

0.1 M Na2PO4, pH ¼ 9.0, 25 1C. 1 M NaCl, pH ¼ 6.1, 25 1C. c 1 M NaCl pH ¼ 5.7, 25 1C. b

at the perimeter of the aromatic rings did not have a major inﬂuence on the complexation process. The complexation of 31 with b-CD with a 1:1 stoichiometry had a much higher efﬁciency (4.2 103 M1) than the complex where 2 CDs bind to one guest (2.7 102 M1). The binding dynamics for the 1:1 complex were fast with an association rate constant of 1.1 108 M1 s1 and a dissociation rate constant of 2.6 104 s1.181 The association of the second b-CD molecule (5.4 106 M1 s1) was much slower than a diffusion controlled process, while the dissociation rate constant (2.0 104 s1) was similar to the one observed for the 1:1 complex. In the case of 32, only the 1:1 complex was detected with b-CD, with association and dissociation rate constants similar to those reported for 31 (1.1 108 M1 s1, 1.5 104 s1).182 Ultrasonic relaxation Ultrasonic relaxation is an excellent technique to study the binding dynamics of guests that do not have chromophores, and this technique was employed to measure the binding dynamics of inorganic anions,165,166 alcohols,151,154,156,157,160,163 alkylammonium ions,155,158 carboxylic acids and derivatives,152,153,159 and amino acids.161,162,164 These data have been interpreted assuming a CD complex stoichiometry of 1:1, mainly because one relaxation process was observed. It is important to note that relaxation processes were observed for solutions containing only cyclodextrins,211–213 which were attributed to various processes related to changes in solvation of the CD cavity and bond rotations in the glucose structure. The amplitudes of these relaxations were seen to increase linearly with CD concentration, and decreased when a guest was included into the CD cavity. These results underline the importance of performing control experiments for the CD solutions in the absence of guests. The ultrasonic absorption spectrum for a series of inorganic salts with b-CD showed one relaxation process.166 No absorption was observed for solutions only containing b-CD. The equilibrium constants determined from competitive binding isotherms were relatively low (2–30 M1). The relaxation frequency (fr) was related to the observed relaxation rate constant, which is equal to the sum of the association and dissociation processes. The association rate constants for all salts with the exception of perchlorate were similar and this result was interpreted to mean that

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

211

Table 10 Association and dissociation rate constants of inorganic anions with b-CD at 25 1C using a frequency range from 15 to 205 MHz166 Anion

k+ (107 M1 s1)

k– (106 s1)

ClO 4 I SCN Br NO 3 Cl

200 6.5 4.4 4.5 4.5 5.4

74 3.6 4.4 6.9 8.2 21

conformational changes to the CD were the rate limiting step for the complexation process when anions ﬁt inside the CD cavity. The changes in the dissociation rate constant paralleled the changes in the equilibrium constants (Table 10). When similar concentrations of guest and host are employed the equation for kobs includes the sum of the concentrations of free host and free guest in the aqueous solution: kobs ¼ 2pf r ¼ kþ ð½H þ ½GÞ þ k

(20)

The concentration for free CD ([H]) and free guest ([G]) can be substituted by the analytical concentration of CD and guest ([H]O and [G]O) and the association rate constant can be related to the equilibrium constant between the guest and host ðkþ ¼ K k Þ leading to Equation (21). This form of the equation is necessary when neither the host or guest concentrations are in excess. qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ kobs ¼ k ð1 þ K½Ho þ ½Go Þ2 4K 2 ½Ho ½Go (21) The equilibrium constant and dissociation rate constant were determined simultaneously by non-linear least-squares ﬁtting, unless the absorption signal was too low157 or no dependence of relaxation frequency on concentration was observed.159,161,162 The association rate constant was then calculated from the deﬁnition of the equilibrium constant. The equilibrium constants determined from the dynamics in this manner agree fairly well with equilibrium constants determined independently. Another approach that has been used to extract rate constants from ultrasonic relaxation data involves using independently determined equilibrium constants to determine concentrations of each species at equilibrium163 or, for special cases where molecule-speciﬁc electrodes exist, direct determination of equilibrium concentrations.167,168 A selection of data for a series of alcohols with a- and b-CD is shown in Table 11. The association rate constants remained constant, showing no dependence on chain length of the alcohols for binding in both a- and b-CD. However, the dissociation rate constant did decrease with the alcohol’s chain length in both sizes of cyclodextrin. This decrease in dissociation rate constant could be related to the fact that the larger guests are more hydrophobic. The dissociation rate constants were smaller

212

T.C.S. PACE AND C. BOHNE

Table 11 Equilibrium constants and association and dissociation rate constants for alcohol/ CD complexes determined by ultrasonic relaxation at 251C Guest 1-Propanol 1-Butanol 2-Butanol tert-Butanol Ethanol 1-Propanol 1-Butanol 2-Butanol 1-Butanol 1-Pentanol a

CD

Freq (MHz)

K (M1)

k+ (108 M1 s1)

b b b b a a a a a a

3–220 1–220 0.8–95 0.8–95 1–95 1–95 0.8–95 0.8–95 2–95 2–95

3.7–4.5 16–17 16–22

5.1 2.8 3.2 3.6 2.8 2.7 5.5 3.5 4.2 3.6

9.7 13–29 73 300

k K (M1)a Reference (107 s1) 12 3.8 2.8 0.85 2.9 1.7 0.46 1.0 0.55 0.12

4.2 7.2 12 43 – 16 120 34 – –

156 151 154 160 157 157 160 160 163 163

Calculated from kinetic data.

for complexes with a-CD than with b-CD, indicating that when the cavity is smaller the hydrophobic interaction between the chain and the CD is greater. Ethanol was too small to form an inclusion complex with b-CD and no relaxation process was observed,156 while in the presence of the smaller a-CD a fast relaxation process was detected for this alcohol.157 The comparison of isomers of butanol suggests that as the bulkiness of the guest increases the dissociation rate constant decreases leading to larger equilibrium constants (Table 11), probably because the hydrophobic interactions of the bulkier guest with the CD cavity are improved. In the case of the binding of 1-butanol with a-CD, similar values were obtained for the association and dissociation rate constants using a method in which the equilibrium constant was recovered from the kinetic measurements or determined independently (Table 11). NMR NMR has been used for measurements of dynamics in few examples of CD complexes, and most applications are related to slow processes, such as those observed for rotaxanes. An example of such a slow process is the threading of a-CD onto a monomeric model of ionene, which could be followed directly by measuring the intensity changes of the signals due to complexed and uncomplexed material to give an association rate constant of 0.036 M1 s1.204 Line shape analysis was performed for the binding of some dihydroxycholate ions to b-cyclodextrin.205 The dihydroxycholates show different 18-CH3 signals for the complexed and free dihydroxycholate ions. To extract exchange rate constants from the NMR spectra, a complete line-shape simulation was performed. The simulation requires input of the chemical shift difference between the two sites, the line width in the absence of exchange, the residence time in each site (tHG and tG), and the relative population (fHG and fG) of each site (Equation (11)). The values were varied until the simulated and experimental spectra could be superimposed. The dissociation rate

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

213

constant for the different dihydroxycholate ions from b-CD at 18 1C ranged from 8 to 22 s1, while the association rate constants were between 4.0 104 and 6.3 104 M1 s1. The 2:2 complex formed between b-CD and reduced tetracyanoquinodimethane shows separate signals for the free and bound CD.203 2D 1H exchange spectroscopy gave an exchange rate of 0.9 s1 at 30 1C for the exchange between the free guest and the 2:2 complex. The exchange may occur via numerous steps, but no resolution of the intermediate steps could be achieved from the data treatment. Fluorescence correlation spectroscopy This technique was employed to study the binding dynamics of Pyronine Y (31) and B (32) with b-CD.65 The theoretical background for this particular system has been discussed with the description of the technique above. Separate analysis of the individual correlation curves obtained was difﬁcult since the diffusion time for the complex could not be determined directly because, even at the highest concentration of CD employed, about 20% of the guest molecules were still free in solution. The curves were therefore analyzed using global analysis to obtain the dissociation rate constant for the 1:1 complex (Table 12). The association rate constant was then calculated from the deﬁnition of the equilibrium constant. The association rate constants were the same within experimental error. The dissociation rate constant for 31 was however an order of magnitude larger than that for 32. The association rate constants determined with ﬂuorescence correlation spectroscopy were similar to the rate constants determined using temperature jump experiments (see above). However, a signiﬁcant difference was observed for the dissociation rate constants where, for the 1:1 complex, values of 2.6 104 and 1.5 104 s1 were determined in the temperature jump experiments for 31 and 32, respectively.181,182 The reasons for this difference were not discussed by the authors of the study with ﬂuorescence correlation spectroscopy. One possibility is that the technique is not sensitive enough to detect the presence of higher-order complexes, such as the 1:2 (31:CD) complex observed in the temperature jump experiments. One other possibility is the fact that the temperature jump experiments were performed in the presence of 1.0 M NaCl. Fluorescence In most cases, the lifetimes for the singlet excited states are too short for the excited state to move between the binding site in a supramolecular system and the Table 12 Guest

Association and dissociation rate constants for pyronine/CD complex at 21 1C65 k+ (108 M1s1)a

k– (104 s1)

K (103 M1)a

2 1.5

50 7.6

0.40 2.1

31 32 a

The K values were ﬁxed.

214

T.C.S. PACE AND C. BOHNE

OH N 33

N

34

Scheme 14 2,3-diazabicyclo[2.2.2]oct-2-ene (33) and 2-naphthol (34).

Table 13 CD a b g

Association rate constants and equilibrium constants for DBO bound to CDs45 k þ (108 M1 s1)

K (M1)

1.9 4.0 0.8

50 1100 6

homogeneous solvent.18 In very special cases time-resolved ﬂuorescence experiments can be used to study dynamics of guest binding with CDs. One such case is the use of 2,3-diazabicyclo[2.2.2]oct-2-ene (DBO, 33, Scheme 14).45 This compound has a mono-exponential decay in water with a very long lifetime (730 ns in dearated D2O). The lifetime of the singlet excited state of 33 is shortened by a quenching mechanism involving an ‘aborted’ hydrogen transfer leading to the ground state of 33 with no net reaction occurring. In the presence of a- and b-CD the decay for the ﬂuorescence intensity was ﬁt to the sum of two exponentials. The fast exponential term, which had lifetimes of 33 and 95 ns for a- and b-CD, respectively, did not depend on the concentration of the CDs and was assigned to the lifetime of 33 within the CD cavity. These lifetimes were much shorter than in water because the CDs provide appropriate H-moieties to quench the singlet excited state of 33. The second decay process became shorter as the concentration of CD was increased, and this component was assigned to the association of free 33 to the CD cavities where quenching occurred. The second-order rate constants constitute the association rate constants for the excited state ðk þ Þ of 33 with the various CDs (Table 13). Though the association rate constants show the same trends as the equilibrium constants the changes in k þ are smaller, indicating that the values for the dissociation rate constants probably also show changes with the CD structure. Global compartmental analysis can be used to recover association and dissociation rate constants in some speciﬁc cases when the lifetimes are much shorter than the lifetimes for the association and dissociation processes. An example is the study for the binding dynamics of 2-naphthol (34, Scheme 14) with b-CD.207 Such an analysis is possible only if the observed lifetimes change with CD concentration and at least one of the decay parameters is known independently, in this case the lifetime of the singlet excited state of 33 (5.3 ns). From the analysis the association and dissociation rate constants, as well as intrinsic decay rate constants and iodide quenching rate constants, were recovered. The association and dissociation rate constants were found to be 2.5 109 M1 s1 and 520 s1, respectively.207

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

215

Laser flash photolysis Laser ﬂash photolysis can be used to follow the kinetics of triplet excited state guests when bound to a host system. The mobility of this type of excited state between the host and the homogeneous phase occurs because triplet excited states are much longer lived than singlet excited states.18 The relocation of the excited state guest can be measured directly if it has a different absorption spectrum than observed for the ground state, or a quenching methodology can be employed to determine the association ðk þ Þ and dissociation ðk Þ rate constants of the excited guest by ﬁtting the curved quenching plot to Equation (6). Xanthone (35) is a guest for which the binding dynamics with CDs can be measured by both the direct and quenching methodology. In addition, laser temperature jump experiments were performed to measure the binding dynamics of ground-state xanthone with b-CD (Table 14).57,200–202 The equilibrium constant for the triplet state of xanthone was signiﬁcantly lower than measured for the ground state.57,202 The comparison between the dynamics for the ground and the excited state of xanthone showed that the association rate constants were the same for both electronic states, while the dissociation rate constant for the excited state increased signiﬁcantly, probably because of the higher dipole moment and basicity of the excited state. In the use of the triplet methodologies it is inherently easier to obtain values for k and sometimes the only way to estimate k þ is by calculating this parameter from the values of the dissociation rate constant for the excited state and the equilibrium constant for the ground state. The example for ground and excited state xanthone shows that such a calculation would lead to an error in the association rate constant of at least one order of magnitude. Therefore, calculations involving parameters for different electronic states should be done with caution. To date xanthone is the only guest molecule for which relocation of the excited state has been directly monitored. Studies of other guest molecules rely on the quenching methodology and the dynamics of xanthone complexation to b-CD have also been studied in this manner.200 Using this methodology both rate constants were higher than determined directly (Table 14). Though the reasons for the overestimation of the rate constants are unknown comparison of binding dynamics for different guests can be carried out as long as similar quenchers are employed.17

Table 14 Equilibrium constants and association and dissociation rate constants for the ground-state and triplet excited state of xanthone (35) with CDs CD b- (ground state)a b- b,c b- b,d a

K (M1) KT (M1) k+ (108 M1 s1) k– (106 s1) k þ (108 M1 s1) k (106 s1) Reference 1100

Laser temperature jump. Laser ﬂash photolysis. c Direct method. d Quenching method. b

4 70 90

0.3 6.0 11

8.1 12

57 202 200

216

T.C.S. PACE AND C. BOHNE O

O 35

O

O O

36

O 37

Scheme 15 Xanthone (35), chromone (36), and ﬂavone (37).

Table 15 Equilibrium constants in the ground (K) and excited state (KT) and association and dissociation rate constants for guest molecules with bCD200 Guest

K (M1)

KT (M1)

k þ (108 M1 s1)

k (106 s1)

1100 240 1090

90 140 550

11 30 24

12 21 4.4

Xanthone Chromone Flavone

Chromone (36), ﬂavone (37), and xanthone (35) (Scheme 15) were used as guests to study how the guest size affects the binding dynamics with b-CD (Table 15).200 Chromone exhibited fast dynamics that can be related to the fact this guest does not ﬁt tightly into the CD cavity. Flavone on the other hand had an exit rate constant that was signiﬁcantly smaller than that of xanthone, which may be due to the pendant phenyl ring being included deep into the CD cavity hindering the exit of the guest. This example shows that although the equilibrium constants for ground state 35 and 37 are the same the binding dynamics and equilibrium constants for the triplet state are quite different. The binding efﬁciency of excited states is very different from the ground states and no inferences can be made from the values of equilibrium constants as to possible trends for the association and dissociation rate constants.

4

Conclusions

This monograph described the techniques that can be used for the study of fast processes in supramolecular dynamics. Some of the techniques have been widely used while others have only been applied recently. Many of the techniques rely on the presence of a chromophore in the guest or host that can be used to monitor changes in absorption or ﬂuorescence as the reaction proceeds. This feature is a reﬂection of the ease in building detection systems for absorption and ﬂuorescence measurements, and in the case of ﬂuorescence it is also related to the high sensitivity of this method. For systems that do not contain a chromophore the dynamics can be

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

217

studied by ultrasonic relaxation, NMR or surface plasmon resonance. However, the later two techniques are limited to fairly slow kinetic measurements, while ultrasonic relaxation can be applied to a narrow time domain. In order to broaden the types of systems to be studied it would be desirable to develop detection systems that do not rely on chromophores. In this respect, conductimetry is a good alternate technique that has been used in some cases. The examples provided for the binding of guests to DNA and CDs showed that even for simple binding stoichiometries kinetic measurements can uncover mechanistic details that are not apparent from equilibrium measurements.

Acknowledgements CB thanks the Natural Sciences and Engineering Research Council of Canada (NSERC) for support of her research program in supramolecular dynamics in the form of operating and equipment grants, and TCSP thanks NSERC for a Canada Graduate Scholarship.

References 1. Lehn, J.M. (1995). Supramolecular Chemistry: Concepts and Perspectives. VCH, Weinheim 2. Atwood, J.L., Davies, J.E.D., MacNicol, D.D., Vo¨gtle, F. and Lehn, J.M. (1996). Comprehensive Supramolecular Chemistry, Vols. 1–11, Pergamon, New York 3. Badjic, J.D., Nelson, A., Cantrill, S.J., Turnbull, W.B. and Stoddart, J.F. (2005). Acc. Chem. Res. 38, 723–732 4. De Fetyer, S. and De Schryver, F.C. (2003). Chem. Soc. Rev. 32, 139–150 5. Ercolani, G. (1998). J. Phys. Chem. B 102, 5699–5703 6. Ercolani, G. (2003). J. Am Chem. Soc. 125, 16097–16103 7. Ercolani, G., Mandolini, L., Mencarelli, P. and Roelens, S. (1993). J. Am. Chem. Soc. 115, 3901–3908 8. Lavigne, J.J. and Anslyn, E.V. (2001). Angew. Chem. Int. Ed. 40, 3118–3130 9. Linton, B. and Hamilton, A.D. (1997). Chem. Rev. 97, 1669–1680 10. Mulder, A., Huskens, J. and Reinhoudt, D.N. (2004). Org. Biomol. Chem. 2, 3409–3424 11. Prins, L.J., Reinhoudt, D.N. and Timmerman, P. (2001). Angew. Chem. Int. Ed. 40, 2382–2426 12. Rebek Jr., J. (1999). Acc. Chem. Res. 32, 278–286 13. Rebek Jr., J. (2005). Angew. Chem. 44, 2068–2078 14. Rowan, S.J., Cantrill, S.J., Cousins, G.R.L., Sanders, J.K.M. and Stoddart, J.F. (2002). Angew. Chem. 41, 898–952 15. Sauvage, J.-P. and Hosseini, M.W. (1996). Templating, Self-Assembly, and SelfOrganization, Vol. 9, Pergamon, New York 16. Vo¨gtle, F. (1996). Molecular Recognition: Receptors for Molecular Guests, Vol. 2, Pergamon, New York 17. Bohne, C. (2006). Langmuir 22, 9100–9111 18. Kleinman, M.H. and Bohne, C. (1997). In Molecular and Supramolecular Photochemistry, Ramamurthy, V. and Schanze, K.S. (eds.), Vol. 1, pp. 391–466. Marcel Dekker Inc., New York 19. Bernasconi, C.F. (1976). Relaxation Kinetics. Academic Press, Inc., New York

218

T.C.S. PACE AND C. BOHNE

20. Czerlinski, G.H. (1966). Chemical Relaxation: An Introduction to Theory and Application of Stepwise Perturbation. Marcel Dekker Inc., New York 21. Berger, R.L., Balko, B., Borcherdt, W. and Friauf, W. (1968). Rev. Sci. Instr. 39, 486–493 22. Robinson, B.H. (1974). In Chemical and Biological Applications of Relaxation Spectrometry, Wyn-Jones, E. (ed.), D. Reidel Publishing Co., Boston 23. Breusegem, S.Y., Sadat-Ebrahimi, S.E., Douglas, K.T., Clegg, R.M. and Loontiens, F.G. (2001). J. Mol. Biol. 308, 649–663 24. Crooks, J.E. (1983). J. Phys. E: Sci. Instrum. 16, 1142–1147 25. Reich, R.M. and Sutter, J.R. (1977). Anal. Chem. 49, 1081–1085 26. Marcandalli, B., Winzek, C. and Holzwarth, J.F. (1984). Ber. Bunsenges. Phys. Chem. 88, 368–374 27. Monaco, R.R. and Gardiner, W.C. (1993). Biochem. Biophys. Res. Commun. 196, 975–983 28. Strehlow, H. (1992). Rapid Reactions in Solution. VCH, Weinheim 29. Porschke, D. (1996). Ber. Bunsenges. Phys. Chem. 100, 715–720 30. Caldin, E.F. and Field, J.P. (1982). J. Chem. Soc. Faraday Trans. 1 78, 1923–1935 31. Holzwarth, J.F., Schmidt, A., Wolff, H. and Volk, R. (1977). J. Phys. Chem. 81, 2300–2301 32. Callender, R. and Dyer, R.B. (2002). Curr. Opin. Struc. Biol. 12, 628–633 33. Gruebele, M., Sabelko, J., Ballew, M. and Ervin, J. (1998). Acc. Chem. Res. 31, 699–707 34. Ma, H., Wan, C. and Zewail, A.H. (2006). J. Am. Chem. Soc. 128, 6338–6340 35. Caldin, E.F., Crooks, J.E. and Robinson, B.H. (1971). J. Phys. E: Sci. Instrum. 4, 165–189 36. Ameen, S. (1975). Rev. Sci. Instr. 46, 1209–1215 37. Yamamoto, K., Mizutani, Y. and Kitagawa, T. (2000). Biophys. J. 79, 485–495 38. Williams, A.P., Longfellow, C.E., Freier, S.M., Kierzek, R. and Turner, D.H. (1989). Biochemistry 28, 4283–4291 39. Turner, D.H., Flynn, G.H., Sutin, N. and Beitz, J.V. (1972). J. Am Chem. Soc. 94, 1554–1559 40. Thurn, T., Bloor, D.M. and Wyn-Jones, E. (2005). In Mixed Surfactant Systems (2nd edn), Abe, M. and Scamehorn, J.F. (eds), pp. 709–768. CRC Press 41. Verrall, R.E. (1995). Chem. Soc. Rev. 24, 135–142 42. Eggers, F. and Funck, T. (1976). Naturwissenschaften 63, 280–285 43. Rassing, J. (1975). In Chemical and Biological Applications of Relaxation Spectrometry, Wyn-Jones, E. (ed.), pp. 1–16. D. Reidel Publishing Company, Dordrecht 44. Kalyanasundaram, K. and Thomas, J.K. (1977). J. Am. Chem. Soc. 99, 2039–2044 45. Nau, W.M. and Zhang, X. (1999). J. Am. Chem. Soc. 121, 8022–8032 46. O’Connor, D.V. and Phillips, D. (1984). Time-Correlated Single Photon Counting. Academic Press, Orlando 47. Birch, D.J.S. and Imhof, R.E. (1991). In Topics in Fluorescence Spectroscopy, Lakowisz, J. (ed.), Vol. 1, pp. 1–95. Plenum Press 48. Bohne, C., Redmond, R.W. and Scaiano, J.C. (1991). In Photochemistry in Organized and Constrained Media, Ramamurthy, V. (ed.), pp. 79–132. VCH Publishers, New York 49. Gehlen, M.H. and De Schryver, F.C. (1993). Chem. Rev. 93, 199–221 50. Douhal, A. (2004). Acc. Chem Res. 104, 1955–1976 51. Scaiano, J.C. (2004). In Reactive Intermediate Chemistry, Moss, R.A., Platz, M.S. and Jones, M. Jr. (eds), Wiley, New York 52. Birks, J.B. (1970). Photophysics of Aromatic Molecules. Wiley-Interscience, London 53. Cheung, S.T. and Ware, W.R. (1983). J. Phys. Chem. 87, 466–473 54. Almgren, M., Grieser, F. and Thomas, J.K. (1979). J. Am. Chem. Soc. 101, 279–291 55. Rinco, O., Nolet, M.-C., Ovans, R. and Bohne, C. (2003). Photochem. Photobiol. Sci. 2, 1140–1151 56. Turro, N.J., Okubo, T. and Chung, C.J. (1982). J. Am. Chem. Soc. 104, 1789–1794

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

219

57. Liao, Y., Frank, J., Holzwarth, J.F. and Bohne, C. (1995). J. Chem. Soc. Chem. Commun. 199–200 58. Enderlein, J., Gregor, I., Patra, D. and Fitter, J. (2004). Curr. Pharm. Biotech. 5, 155–161 59. Go¨sch, M. and Rigler, R. (2005). Adv. Drug Deliv. Rev. 57, 169–190 60. Krichevsky, O. and Bonnet, G. (2002). Rep. Prog. Phys. 65, 251–297 61. Thompson, N.L. (1991). In Topics in Fluorescence Spectroscopy, Vol. 1, Techniques, Lakowicz, J.R. (ed.), pp. 337–378. Plenum Press, New York 62. Webb, W.W. (2001). In Fluorescence Correlation Spectroscopy, Rigler, R. and Elson, E.S. (eds), pp. 305–330. Springer, Berlin 63. Widengren, J. (2001). In Fluorescence Correlation Spectroscopy, Rigler, R. and Elson, E.S. (eds), pp. 277–301. Springer, Berlin 64. Magde, D., Elson, E.S. and Webb, W.W. (1972). Phys. Rev. Lett. 29, 705–708 65. Al-Souﬁ, W., Reija, B., Novo, M., Felekyan, S., Ku¨hnemuth, R. and Seidel, C.A.M. (2005). J. Am. Chem. Soc. 127, 8775–8784 66. Felekyan, S., Ku¨hnemuth, R., Kudryavtsev, V., Sandhagen, C., Becker, W. and Seidel, C.A.M. (2005). Rev. Sci. Instr. 76, 083104 67. Widengren, J., Dapprich, J. and Rigler, R. (1997). Chem. Phys. 216, 417–426 68. Freeman, R. (1988). A Handbook of Nuclear Magnetic Resonance. Wiley, New York 69. Perrin, C.L. and Dwyer, T.J. (1990). Chem. Rev. 90, 935–967 70. Palmer, L.C. and Rebek Jr., J. (2004). Org. Biomol. Chem. 2, 3051–3059 71. Green, R.J., Frazier, R.A., Shakesheff, K.M., Davies, M.C., Roberts, C.J. and Tendler, S.J.B. (2000). Biomaterials 21, 1823–1835 72. Englebienne, P., Van Hoonacker, A. and Verhas, M. (2003). Spectroscopy 17, 255–273 73. Schuck, P. (1997). Annu. Rev. Biophys. Biomol. Struct. 26, 541–566 74. Wilson, W.D. (2002). Science 295, 2103–2104 75. O’Shannessy, D.J. and Winzor, D.J. (1996). Anal. Biochem. 236, 275–283 76. Hall, D.R., Cann, J.R. and Winzor, D.J. (1996). Anal. Biochem. 235, 175–184 77. Kumar, C.V. (1991). In Photochemistry in Organized and Constrained Media, Ramamurthy, V. (ed.), pp. 783–816. VCH Publishers, Inc., New York 78. Denny, W.A. (2003). In DNA and RNA Binders: From Small Molecules to Drugs, Demeunynck, M., Bailly, C. and Wilson, W.D. (eds), pp. 482–502. VHC, Weinheim 79. Lacy, E.R., Madsen, E.M., Lee, M. and Wilson, W.D. (2003). In DNA and RNA Binders: From Small Molecules to Drugs, Demeunynck, M., Bailly, C. and Wilson, W.D. (eds), pp. 384–413. VHC, Weinheim 80. Adams, A., Guss, J.M., Collyer, C.A., Denny, W.A., Prakash, A.S. and Wakelin, L.P.G. (2000). Mol. Pharmacol. 58, 649–658 81. Baliga, R. and Crothers, D.M. (2000). Proc. Natl. Acad. Sci. USA 97, 7814–7818 82. Bittman, R. and Blau, L. (1975). Biochemistry 14, 2138–2145 83. Breusegem, S.Y., Loontiens, F.G., Regenfuss, P. and Clegg, R.M. (2001). In Methods in Enzymology: Drug– Nucleic Acid Interactions, Chaires, J.B. and Waring, M.J. (eds), Vol. 340, pp. 212–233. Academic Press, San Diego 84. Chaires, J.B., Dattagupta, N. and Crothers, D.M. (1985). Biochemistry 24, 260–267 85. Chen, F.M. and Sha, F. (1998). Biochemistry 37, 11143–11151 86. Denny, W.A., Atwell, G.J., Baguley, B.C. and Wakelin, L.P.G. (1985). J. Med. Chem. 28, 1568–1574 87. Eriksson, M., Karlsson, J., Westman, G. and A˚kerman, B. (2003). Nucl. Acids Res. 31, 6235–6242 88. Eriksson, M., Westerlund, F., Mehmedovic, M., Lincoln, P., Westman, G., Larsson, A. and A˚kerman, B. (2006). Biophys. Chem. 122, 195–205 89. Fox, K.R., Brassett, C. and Waring, M.J. (1985). Biochim. Biophys. Acta 840, 383–392 90. Fox, K.R. and Waring, M.J. (1984). Biochemistry 23, 2627–2633 91. Gabbay, E.J., Grier, D., Fingerle, R.E., Reimer, R., Levy, R., Pearce, S.W. and Wilson, W.D. (1976). Biochemistry 15, 4209–4215

220

T.C.S. PACE AND C. BOHNE

92. Lombardy, R.L., Tanious, F.A., Ramachandran, K., Tidwell, R.R. and Wilson, W.D. (1996). J. Med. Chem. 39, 1452–1462 93. Mandal, C., Englander, S.W. and Kallenbach, N.R. (1980). Biochemistry 19, 5819–5825 94. Meyer-Almes, F.J. and Porschke, D. (1993). Biochemistry 32, 4246–4253 95. Mu¨ller, W. and Crothers, D.M. (1968). J. Mol. Biol. 35, 251–290 96. Phillips, D.R., Greif, P.C. and Boston, R.C. (1988). Mol. Pharmacol. 33, 225–230 97. Rizzo, V., Sacchi, N. and Menozzi, M. (1989). Biochemistry 28, 274–282 98. Roche, C.J., Thomson, J.A. and Crothers, D.M. (1994). Biochemistry 33, 926–935 99. Sakoda, M., Hiromi, K. and Akasaka, K. (1971). Biopolymers 10, 1003–1012 100. Tanious, F.A., Ding, D., Patrick, D.A., Tidwell, R.R. and Wilson, W.D. (1997). Biochemistry 36, 15315–15325 101. Tanious, F.A., Jenkins, T.C., Neidle, S. and Wilson, W.D. (1992). Biochemistry 31, 11632–11640 102. Tanious, F.A., Yen, S.-F. and Wilson, W.D. (1991). Biochemistry 30, 1813–1819 103. Veal, J.M., Li, Y., Zimmerman, S.C., Lamberson, C.R., Cory, M., Zon, G. and Wilson, W.D. (1990). Biochemistry 29, 10918–10927 104. Wakelin, L.P.G., Adams, A. and Denny, W.A. (2002). J. Med. Chem. 45, 894–901 105. Wakelin, L.P.G., Atwell, G.J., Rewcastle, G.W. and Denny, W.A. (1987). J. Med. Chem. 30, 855–861 106. Westerlund, F., Wilhelmsson, L.M., Norde´n, B. and Lincoln, P. (2003). J. Am. Chem. Soc. 125, 3773–3779 107. Wilson, W.D., Grier, D., Reimer, R., Bauman, J.D., Preston, J.F. and Gabbay, E.J. (1976). J. Med. Chem. 19, 381–384 108. Wilson, W.D., Tanious, F.A., Ding, D., Kumar, A., Boykin, D.W., Colson, P., Houssier, C. and Bailly, C. (1998). J. Am. Chem. Soc. 120, 10310–10321 109. Biver, T., De Biasi, A., Secco, F., Venturini, M. and Yarmoluk, S. (2005). Biophys. J. 89, 374–383 110. Biver, T., Pulzonetti, M., Secco, F., Venturini, M. and Yarmoluk, S. (2006). Arch. Biochem. Biophys. 451, 103–111 111. Biver, T., Secco, F., Tine`, M.R. and Venturini, M. (2003). Arch. Biochem. Biophys. 418, 63–70 112. Bresloff, J.L. and Crothers, D.M. (1975). J. Mol. Biol. 95, 103–123 113. Corin, A.F. and Jovin, T.M. (1986). Biochemistry 25, 3995–4007 114. Dodin, G. (1989). J. Med. Chem. 32, 1127–1134 115. Dodin, G., Schwaller, M.-A., Aubard, J. and Paoletti, C. (1988). Eur. J. Biochem. 176, 371–376 116. Dourlent, M. and Hogrel, J.F. (1976). Biochemistry 15, 430–436 117. Li, H.J. and Crothers, D.M. (1969). J. Mol. Biol. 39, 461–477 118. Marcandalli, B., Stange, G. and Holzwarth, J.F. (1988). J. Chem. Soc. Faraday Trans. 1 84, 2807–2819 119. Ramstein, J., Ehrenberg, M. and Rigler, R. (1980). Biochemistry 19, 3938–3948 120. Wakelin, L.P.G. and Waring, M.J. (1980). J. Mol. Biol. 144, 183–214 121. Boger, D.L. and Saionz, K.W. (1999). Bioorg. Med. Chem. 7, 315–321 122. Carrasco, C., Facompre, M., Chisholm, J.D., Van Vranken, D.L., Wilson, W.D. and Bailly, C. (2002). Nucl. Acids Res. 30, 1774–1781 123. Carrasco, C., Joubert, A., Tardy, C., Maestre, N., Cacho, M., Brana, M.F. and Bailly, C. (2003). Biochemistry 42, 11751–11761 124. Lacy, E.R., Cox, K.K., Wilson, W.D. and Lee, M. (2002). Nucl. Acids Res. 30, 1834–1841 125. Lacy, E.R., Le, N.M., Price, C.A., Lee, M. and Wilson, W.D. (2002). J. Am. Chem. Soc. 124, 2153–2163 126. Nakatani, K., Kobori, A., Kumasawa, H., Goto, Y. and Saito, I. (2004). Bioorg. Med. Chem. 12, 3117–3123

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

221

127. Nguyen, B., Lee, M.P.H., Hamelberg, D., Joubert, A., Bailly, C., Brun, R., Neidle, S. and Wilson, W.D. (2002). J. Am. Chem. Soc. 124, 13680–13681 128. Tanious, F.A., Hamelberg, D., Bailly, C., Czarny, A., Boykin, D.W. and Wilson, W.D. (2004). J. Am. Chem. Soc. 126, 143–153 129. Wolf, L.K., Fullenkamp, D.E. and Georgiadia, R.M. (2005). J. Am. Chem. Soc. 127, 17453–17459 130. Assa-Munt, N., Leupin, W., Denny, W.A. and Kearns, D.R. (1985). Biochemistry 24, 1449–1460 131. Blasko`, A., Browne, K.A. and Bruice, T.C. (1994). J. Am. Chem. Soc. 116, 3726–3737 132. Feigon, J., Denny, W.A., Leupin, W. and Kearns, D.R. (1984). J. Med. Chem. 27, 450–465 133. Klevit, R.E., Wemmer, D.E. and Reid, B.R. (1986). Biochemistry 25, 3296–3303 134. Pelton, J.G. and Wemmer, D.E. (1990). J. Am. Chem. Soc. 112, 1393–1399 135. Ragg, E., Mazzini, S., Bortolini, R., Mongelli, N. and D0 Alessio, R. (1998). J. Chem. Soc. Perkin Trans. 2, 149–159 136. Geacintov, N.E., Waldmeyer, J., Kuzmin, V.A. and Kolubayev, T. (1981). J. Phys. Chem. 85, 3608–3613 137. Pace, T.C.S., Monahan, S.L., MacRae, A.I., Kaila, M. and Bohne, C. (2006). Photochem. Photobiol. 82, 78–87 138. Prusik, T. and Geacintov, N.E. (1976). FEBS Lett. 71, 236–240 139. Waring, M.J. (1965). J. Mol. Biol. 13, 269–282 140. McGhee, J.D. and von Hippel, P.H. (1974). J. Mol. Biol. 86, 469–489 141. Murov, S.L., Carmichael, I. and Hug, G.L. (1993). Handbook of Photochemistry; 2nd revised and expanded ed. Marcel Dekker, Inc., New York. 142. Macgregor Jr., R.B., Clegg, R.M. and Jovin, T.M. (1987). Biochemistry 26, 4008–4016 143. Atwell, G.J., Cain, B.F., Baguley, B.C., Finlay, G.J. and Denny, W.A. (1984). J. Med. Chem. 27, 1481–1485 144. Larsson, A., Carlsson, C. and Jonsson, M. (1995). Biopolymers 36, 153–167 145. Netzel, T.L., Naﬁsi, K., Zhao, M., Lenhard, J.R. and Johnson, I. (1995). J. Phys. Chem. 99, 17936–17947 146. Yarmoluk, S.M., Lukashov, S.S., Ogul’Chansky, T.Y., Losytskyy, M.Y. and Kornyushyna, O.S. (2001). Biopolymers 62, 219–227 147. Szejtli, J. (1996). In Cyclodextrins, Szejtli, J. and Osa, T. (eds), Vol. 3, pp. 5–40. Elsevier Science Ltd., New York 148. Szejtli, J. (1998). Chem. Rev. 98, 1743–1753 149. Connors, K.A. (1997). Chem. Rev. 97, 1325–1357 150. Rekharsky, M.V. and Inoue, Y. (1998). Chem. Rev. 98, 1875–1917 151. Nishikawa, S. (1997). Bull. Chem. Soc. Jpn. 70, 1003–1007 152. Nishikawa, S., Fukahori, T. and Ishikawa, K. (2002). J. Phys. Chem. A 106, 3029–3033 153. Nishikawa, S. and Ugawa, T. (2000). J. Phys. Chem. A 104, 2914–2918 154. Nishikawa, S., Ugawa, T. and Fukahori, T. (2001). J. Phys. Chem. B 105, 7594–7597 155. Nishikawa, S., Yamaguchi, K. and Fukahori, T. (2003). J. Phys. Chem. A 107, 6415–6418 156. Nishikawa, S. and Yamaguchi, S. (1996). Bull. Chem. Soc. Jpn. 69, 2465–2468 157. Nishikawa, S., Yokoo, N. and Kuramoto, N. (1998). J. Phys. Chem. B 102, 4830–4834 158. Yamaguchi, K., Fukahori, T. and Nishikawa, S. (2005). J. Phys. Chem. A 109, 40–43 159. Fukahori, T., Kondo, M. and Nishikawa, S. (2006). J. Phys. Chem. B 110, 4487–4491 160. Fukahori, T., Nishikawa, S. and Yamaguchi, K. (2004). Bull. Chem. Soc. Jpn. 77, 2193–2198 161. Fukahori, T., Nishikawa, S. and Yamaguchi, K. (2004). J. Acoust. Soc. Am. 115, 2325–2330 162. Fukahori, T., Ugawa, T. and Nishikawa, S. (2002). J. Phys. Chem. A 106, 9442–9445 163. Hall, D., Bloor, D., Tawarah, K. and Wyn-Jones, E. (1986). J. Chem. Soc. Faraday Trans. 1 82, 2111–2121

222

T.C.S. PACE AND C. BOHNE

164. Ugawa, T. and Nishikawa, S. (2001). J. Phys. Chem. A 105, 4248–4251 165. Haller, J., Miecznik, P. and Kaatze, U. (2006). Chem. Phys. Lett. 429, 97–102 166. Rohrbach, R.P., Rodriguez, L.J., Eyring, E.M. and Wojcik, J.F. (1977). J. Phys. Chem. 81, 944–948 167. Takisawa, N., Hall, D.G., Wyn-Jones, E. and Brown, P. (1988). J. Chem. Soc. Faraday Trans. 1 84, 3059–3070 168. Thomason, M.A., Mwakı´ bete, H. and Wyn-Jones, E. (1990). J. Chem. Soc. Faraday Trans. 86, 1511–1515 169. Cramer, F., Saenger, W. and Spatz, H.-C. (1967). J. Am. Chem. Soc. 89, 14–20 170. Hersey, A. and Robinson, B.H. (1984). J. Chem. Soc. Faraday Trans. 1 80, 2039–2052 171. Hersey, A., Robinson, B.H. and Kelly, H.C. (1986). J. Chem. Soc. Faraday Trans. 1 82, 1271–1287 172. Okubo, T. and Kuroda, M. (1989). Macromolecules 22, 3936–3940 173. O¨rstan, A. and Wojcik, J.F. (1988). Carbohydr. Res. 176, 149–154 174. O¨rstan, A. and Wojcik, J.F. (1985). Carbohydr. Res. 143, 43–50 175. Taguchi, K. (1986). J. Am. Chem. Soc. 108, 2705–2709 176. Yoshida, N. and Fujimoto, M. (1980). Chem. Lett. 231–234 177. Yoshida, N. and Fujimoto, M. (1980). Chem. Lett. 1377–1380 178. Yoshida, N., Shirai, T. and Fujimoto, M. (1989). Carbohydr. Res. 192, 291–304 179. Clarke, R.J., Coates, J.H. and Lincoln, S.F. (1984). J. Chem. Soc. Faraday Trans. 1 80, 3119–3133 180. Clarke, R.J., Coates, J.H. and Lincoln, S.F. (1984). Carbohydr. Res. 127, 181–191 181. Schiller, R.L., Lincoln, S.F. and Coates, J.H. (1987). J. Chem. Soc. Faraday Trans. 1 83, 3237–3248 182. Schiller, R.L., Lincoln, S.F. and Coates, J.H. (1986). J. Chem. Soc. Faraday Trans. 1 82, 2123–2132 183. Villani, R.P., Lincoln, S.F. and Coates, J.H. (1987). J. Chem. Soc. Faraday Trans. 1 83, 2751–2756 184. Demont, P.M. and Reinsborough, V.C. (1991). Aus. J. Chem. 44, 759–763 185. Seiyama, A., Yoshida, N. and Fujimoto, M. (1985). Chem. Lett. 1013–1016 186. Yoshida, N. (1995). J. Chem. Soc. Perkin Trans. 2, 2249–2256 187. Yoshida, N. and Fujimoto, M. (1987). J. Phys. Chem. 91, 6691–6695 188. Yoshida, N. and Fujita, Y. (1995). J. Phys. Chem. 99, 3671–3677 189. Yoshida, N. and Hayashi, K. (1994). J. Chem. Soc. Perkin Trans. 2, 1285–1290. 190. Yoshida, N., Seiyama, A. and Fujimoto, M. (1990). J. Phys. Chem. 94, 4246–4253 191. Yoshida, N., Seiyama, A. and Fujimoto, M. (1984). Chem. Lett. 703–706 192. Abou-Hamdan, A., Bugnon, P., Saudan, C., Lye, P.G. and Merbach, A.E. (2000). J. Am. Chem. Soc. 122, 592–602 193. Bugnon, P., Lye, P.G., Abou-Hamdan, A. and Merbach, A.E. (1996). Chem. Commun. 2787–2788 194. Mochida, K. and Matsui, Y. (1976). Chem. Lett. 963–966 195. Saudan, C., Dunard, F.A., Abou-Hamdan, A., Bugnon, P., Lye, P.G., Lincoln, S.F. and Merbach, A.E. (2001). J. Am. Chem. Soc. 123, 10290–10298 196. Dyck, A.S.M., Kisiel, U. and Bohne, C. (2003). J. Phys. Chem. B 107, 11652–11659 197. Okubo, T., Maeda, Y. and Kitano, H. (1989). J. Phys. Chem. 93, 3721–3723 198. Barra, M., Bohne, C. and Scaiano, J.C. (1990). J. Am. Chem. Soc. 112, 8075–8079 199. Barros, T.C., Stefaniak, K., Holzwarth, J.F. and Bohne, C. (1998). J. Phys. Chem. A 102, 5639–5651 200. Christoff, M., Okano, L.T. and Bohne, C. (2000). J. Photochem. Photobiol. A 134, 169–176 201. Liao, Y. and Bohne, C. (1996). J. Phys. Chem. 100, 734–743 202. Okano, L.T., Barros, T.C., Chou, D.T.H., Bennet, A.J. and Bohne, C. (2001). J. Phys. Chem. B 105, 2122–2128

DYNAMICS OF GUEST BINDING TO SUPRAMOLECULAR SYSTEMS

223

203. Hartzell, C.J., Mente, S.R., Eastman, N.L. and Beckett, J.L. (1993). J. Phys. Chem. 97, 4887–4890 204. Herrmann, W., Keller, B. and Wenz, G. (1997). Macromolecules 30, 4966–4972 205. Yim, C.T., Zhu, X.X. and Brown, G.R. (1999). J. Phys. Chem. B 103, 597–602 206. Turro, N.J., Bolt, J.D., Kuroda, Y. and Tabushi, I. (1982). Photochem. Photobiol. 35, 69–72 207. van Stam, J., De Feyter, S., De Schryver, F.C. and Evans, C.H. (1996). J. Phys. Chem. 100, 19959–19966 208. Yoshida, N., Seiyama, A. and Fujimoto, M. (1990). J. Phys. Chem. 94, 4254–4259 209. Park, J.W. (2006). J. Phys. Chem. B 110, 24915–24922 210. Schiller, R.L., Coates, J.H. and Lincoln, S.F. (1984). J. Chem. Soc. Faraday Trans. 1 80, 1257–1266 211. Jobe, D.J., Verrall, R.E. and Reinsborough, V.C. (1990). Can. J. Chem. 68, 2131–2136 212. Kato, S., Nomura, H. and Miyahara, Y. (1985). J. Phys. Chem. 89, 5417–5421 213. Rauh, S. and Knoche, W. (1985). J. Chem. Soc. Faraday Trans. 1 81, 2551–2559

This page intentionally left blank

Mechanisms of oxygenations in zeolites EDWARD L. CLENNAN Department of Chemistry, University of Wyoming, 1000 East University Avenue, Laramie, WY 82071, USA 1 Introduction 225 2 Zeolites 226 Structures 226 Synthesis 229 Properties 229 3 Experimental considerations 230 Reagent/substrate loading 230 Reaction monitoring 231 Product recovery 232 4 Intrazeolite photooxygenations 232 Wagnerova classiﬁcation I oxygenations 233 Wagnerova classiﬁcation II oxygenations 253 Acknowledgement 261 References 262

1

Introduction

Conducting reactions in nanospace where the dimensions of the reaction vessel are comparable to those of the reactants provides a new tool that can be used to control the selectivity of chemical transformations.1 This dimensional aspect of nano-vessels has been referred to as shape selectivity.2 The effect of spatial conﬁnement can potentially be exerted at all points on the reaction surface but its inﬂuence on three stationary points along the reaction coordinate (reactants, transition states, and products) deserve special attention.3,4 (1) Molecular sieving of the reactants, excluding substrates of the incorrect dimension from the reaction site can occur (reactant selectivity). (2) Enzyme-like size selection or shape stabilization of transition states can dramatically inﬂuence reaction pathways (transition state selectivity). (3) Finally, products can be selectively retained that are too large to be removed via the nano-vessel openings/pores (product selectivity). In addition to shape selectivity, which is primarily a steric directing effect, orbital conﬁnement,5 a quantum6 electronic7 effect, can also dramatically inﬂuence reactivity in nanospace.8–10 This concept, ﬁrst introduced by Corma and coworkers,8 points out that when molecular orbitals are conﬁned and not allowed to extend over all space that their energies increase. The HOMO is more sensitive than the LUMO to size restrictions resulting in a decrease of the frontier molecular orbital band gap. This effect can be experimentally demonstrated in systems where the size of the 225 ADVANCES IN PHYSICAL ORGANIC CHEMISTRY VOLUME 42 ISSN 0065-3160 DOI: 10.1016/S0065-3160(07)42005-6

r 2008 Elsevier Inc. All rights reserved

226

E.L. CLENNAN

molecule and the dimensions of the available space are comparable. For example, anthracene11 and naphthalene12 upon conﬁnement both exhibit bathochromic shifts of their 0–0 transitions and decreases in their ﬂuorescence lifetimes. In addition, phosphorescence, which cannot be observed in solution, can be observed at low temperature or even at room temperature in tightly conﬁned situations.13 Shape selectivity and orbital conﬁnement effects are direct results of the physical dimensions of the available space in microscopic vessels and are independent of the chemical composition of nano-vessels. However, the chemical composition in many cases cannot be ignored because in contrast to traditional solution chemistry where reactions occur primarily in a dynamic solvent cage, the majority of reactions in nano-vessels occur in close proximity to a rigid surface of the container (vessel) and can be inﬂuenced by the chemical and physical properties of the vessel walls. Consequently, we begin this review with a brief examination of both the shape (structure) and chemical compositions of a unique set of nano-vessels, the zeolites, and then we will move on to examine how the outcome of photochemical reactions can be inﬂuenced and controlled in these nanospace environments.

2

Zeolites

Zeolites are crystalline inorganic solids that are better known as molecular sieves.14 The word zeolite has Greek origin and means ‘‘boiling stone’’. The ﬁrst identiﬁcation in 1756 of a natural occurring zeolite has been attributed to a Swedish mineralogist, Baron Axel F. Cronstedt, who noted its ability to swell in the presence of water.14 The ﬁrst laboratory synthesis of a zeolite has been attributed to St. Claire Deville in 1862.15 However, modern zeolite chemistry began in the 1940s with the seminal studies of Richard Barrer and Robert Milton.15 By the mid-1950s these groups had successfully made several synthetic and natural occurring zeolites including zeolites A and X (vide infra). These novel porous materials have found widespread industrial use especially as heterogeneous acid catalysts in petroleum processing. STRUCTURES

Zeolites are aluminosilicates characterized by a network of silicon and aluminum tetrahedra with the general formula MX(AlO2)X(SiO2)Y. The M are cations that are necessary to balance the formal negative charge on the aluminum atoms. The tetrahedra are linked to form repeating cavities or channels of well-deﬁned size and shape. Materials with porous structures similar to zeolites but with other atoms in the framework (P, V, Ti, etc.), as a class are referred to as zeotypes. The structure committee of the International Zeolite Association (IZA; http://www.iza-online.org/) has assigned, as of July 1st 2007, 176 framework codes (three capital letters) to these materials. These mnemonic codes do not depend on the composition (i.e. the distribution of different atom types) but only describe the three-dimensional labyrinth of framework atoms. One of the most popular zeolites for photochemical studies faujasite (FAU) is depicted in Fig. 1. Fig. 1a is a ball and stick drawing which not only shows the

MECHANISMS OF OXYGENATIONS IN ZEOLITES

227

II II' C III I' I I' C

a

b

Fig. 1 (a) Ball and stick structure of faujasite (FAU). (b) Conventional stick drawing of faujasite (FAU).

tetrahedral aluminum and silicon atoms but also the oxygen atoms which bridge (link) the tetrahedral units. Fig. 1b is conventional stick drawing that differs from standard organic structural drawings by virtue of the fact that the lines represent the oxygen bridges rather than bonds. The connectivity of the aluminum and silicon tetrahedra in FAU generates a honeycomb like structure with local Td symmetry consisting of large 13 A˚ diameter supercages that can be accessed via tetrahedrally arranged 7.4 A˚ diameter windows. The two commercially available synthetic forms of FAU are called zeolite X and Y. They differ in their silicon/aluminum ratio. Zeolite X is more aluminum rich and has a typical unit cell composition of M86(AlO2)86(SiO2)106 264(H2O). Zeolite Y is more silicon rich with a typical unit cell composition of M56(AlO2)56(SiO2)136 253(H2O). The unit cell contains 8 of the large supercages. The silicon/aluminum ratio is actually a function of the synthesis method with zeolite X having a Si/Al ratio from 1.0 to 1.4 and zeolite Y from 1.4 to 3.0. This corresponds to between 77 and 96 aluminum atoms in a unit cell of zeolite X and between 48 and 76 aluminum atoms in a unit cell of zeolite Y. The lower limit of the Si/Al ratio of 1 is in accord with Lo¨wenstein’s rule that prohibits Al–O–Al linkages, presumably as a result of electrostatic repulsion between proximal negatively charged tetrahedral aluminum atoms. However, nonLo¨wenstein aluminum distribution has occasionally been suggested.16 The cations needed to provide electrical neutrality and balance the charge on the framework aluminum atoms are known to exist at discrete locations (sites) in FAU as depicted in Fig. 1b.17 Site I and I0 (facing into larger cage) are located at the center of the face of a hexagonal prism. Typically there are 16 of these sites in both zeolite X and Y. Sites II (facing into supercage) and II0 are located on non-hexagonal prism 6-ring faces. Typically there are 32 of these sites in both zeolite X and Y. Sites III are on the face of the 4-ring on the supercage wall. Typically there are 38 of these sites in zeolite X but only 8 of these sites in zeolite Y. Site C is at the center of a smaller cage known as the sodalite cage (vide infra). The identity of the cation determines the

228

E.L. CLENNAN

7.1Å

A

C

B

D

Fig. 2 (A) Zeolite L (LTL). (B) ZSM-5 (MFI). (C) Zeolite A (LTA). (D) Mordenite (MOR).

supercage void volume.18 For example, in zeolite X the supercage void volume for Li, Na, K, Rb, Cs, and Tl cations are 873, 852, 800, 770, 732, and 775 A˚3, respectively. On the other hand, the corresponding void volumes in zeolite Y are 834, 827, 807, 796, 781, and 798 A˚3, respectively. Several two- and three-dimensional projections of other zeolites of interest to photochemists are shown in Fig. 2. Zeolite L (Fig. 2A) is a unidirectional channel system with a 12-tetrahedral atom ring opening of 7.1 A˚ in diameter. Zeolite ZSM-5 (Fig. 2B) is characterized by nearly perpendicular 10-tetrahedral atom intersecting straight (5.4 A˚ 5.6 A˚) and sinusoidal (5.1 A˚ 5.5 A˚) channels. Zeolite A (Fig. 2C) is characterized by a 11.4 A˚ supercage accessible via 4.1 A˚ octahedrally arranged 8-tetrahedral atom windows. Mordenite (Fig. 2D) is a unidirectional 12-tetrahedral atom elliptical channel system similar to zeolite L but with a framework consisting of an array of 5-membered rings. It is convenient to visualize zeolites in terms of secondary building units (SBUs). Several of the SBUs currently recognized by the IZA are depicted in Fig. 3 along with other useful structural units, the sodalite and pentasil cages. The sodalite unit is

MECHANISMS OF OXYGENATIONS IN ZEOLITES

3

6-2

spiro-5

4-4

4-4=1

sodalite cage

229

6-6

pentasil cage

Fig. 3 SBUs and the sodalite and pentasil structural units.

found in both faujasites and in zeolite A. On the other hand, mordenite and ZSM-5 both contain the pentasil structural unit. The SBUs contain up to 16 tetrahedral atoms and are always found in an integral number within the unit cell.

SYNTHESIS

Zeolites are formed by crystallization at temperatures between 80 and 200 1C from aqueous alkaline solutions of silica and alumina gels in a process referred to as hydrothermal synthesis.15,19 A considerable amount is known about the mechanism of the crystallization process, however, no rational procedure, similar to organic synthetic procedures, to make a speciﬁcally designed zeolite topology is available. The products obtained are sensitive functions of the reaction conditions (composition of gel, reaction time, order of mixing, gel aging, etc.) and are kinetically controlled. Nevertheless, reproducible procedures have been devised to make bulk quantities of zeolites. Procedures for post-synthetic modiﬁcations have also been described.20–22 Marcus and Cormier23 have classiﬁed synthetic zeolites as 1st, 2nd, or 3rd generation. The 1st generation zeolites (e.g. zeolites X, Y, A, and mordenite) were synthesized prior to 1960 using only inorganic reagents and the classical hydrothermal process and are characterized by low silica to aluminum ratios. The 2nd generation zeolites (e.g. ZSM-5 and silicalite) were synthesized using structure directing agents (templates) and are characterized by higher silicon aluminum ratios. The 3rd generation zeolites are zeotypes (e.g. aluminophosphates, AlPO4, SAPO, and MeAlPO) that contain framework atoms other than silicon or aluminum. PROPERTIES

Two important zeolite properties are; (1) the intra-pore electrostatic ﬁeld, and (2) its acid–base character. As discussed below post-synthetic modiﬁcations of many zeolites to ﬁne-tune these properties are possible and provide a unique opportunity to inﬂuence reaction outcome.

230

E.L. CLENNAN Weaker Lewis Base

Brønsted Acid Stronger Lewis Base O

Si

H O

Si

O

Al

O

Si

M

O

Lewis Acid

Si

Al

O

Si

O

Si

O

Lewis Acid

Fig. 4 Structures of Lewis and Brønsted acid sites in zeolites.

The intrazeolite cations necessary to balance the negative charge on the framework aluminum atoms are poorly shielded and as a result high electric (electrostatic) ﬁelds on the order of 1–10 V/nm are found in their vicinity. The magnitudes of the electric ﬁelds can be calculated from measured effects on the vibrational frequencies or intensities of IR bands of small diatomics such as CO or N2.24 They can also be determined from difference electron density maps determined by X-ray diffraction methods.25 These high electric ﬁelds can dramatically inﬂuence the stabilities of transition states with signiﬁcant charge separations. Zeolites are amphoteric hosts with well-established acid and base properties. The nature and extent of these sites have been extensively probed with both NMR and IR techniques coupled with, and without, probe molecules (e.g. amines and phosphines).26 Both Brønsted and Lewis acid sites have been identiﬁed. The Brønsted acid sites have been associated with bridging OH groups and the Lewis acid sites with three coordinate aluminum defects, and with the counterions (Fig. 4).27 The bridging oxygen atoms function as Lewis base sites with the more potent sites residing adjacent to a negatively charged tetra-coordinate aluminum. In general the acid character of a given zeolite framework increases with an increasing Si/Al ratio and with the decreasing size of the charge-balancing alkali metal cation. Conversely, the basic character increases with a decreasing Si/Al ratio and with increasing cation size. A very convenient method to quantitatively determined the number of Brønsted acid sites in the often used photochemical nano-vessels, zeolites X and Y, is available.28 This method take advantage of indicator/probe molecules which undergo an intense color change upon protonation within the zeolite pore network. The amount of a base necessary to quench the color change gives a direct measure of the ‘‘concentration’’ of acidic sites. The base used to titrate the Brønsted sites must be more basic than the probe molecule and sufﬁciently basic to be completely protonated.

3

Experimental considerations

REAGENT/SUBSTRATE LOADING

Standard reactant and reagent loading protocols are necessary to produce samples that behave reproducibly. The level of hydration for example, if not precisely controlled, can lead to different experimental observations on otherwise identical

MECHANISMS OF OXYGENATIONS IN ZEOLITES

231

samples.27 The popular X and Y zeolites can be conveniently dehydrated over a period of 24 h at temperatures between 100 and 200 1C on a vacuum line at 104 to 105 torr without damage to the zeolite framework. Re-hydration is rapid and subsequent manipulations must be conducted under absolutely dry conditions. Reactants and reagents can be conveniently loaded into the dry zeolite by adsorption. This can be accomplished by intimately mixing the solid or liquid reactant and the powdered zeolite, by absorption from the gas phase, or by diffusion in a solvent slurry containing the zeolite and dissolved reactant. The choice of solvent for the slurry method is critical. It must be volatile enough to be removable at a pressure and temperature that does not result in evacuation of the reactant or its decomposition. In addition, the reactant must have a greater afﬁnity for the interior of the zeolite than for the slurry solvent itself. The lack of afﬁnity for the interior of the zeolite is an acute problem for non-polar hydrocarbons that lack binding sites for the intrazeolitic cations. The use of ﬂuorocarbons such as perﬂuorohexane as slurry solvents takes advantage of the ﬂuorophobicity of many hydrocarbons and has alleviated this problem to some extent.29 Reagents or reactants that are too large to diffuse through the zeolite pore opening, but are of the correct size to be housed in intrazeolite cavities, have also been incorporated by very novel ‘‘ship-in-a-bottle’’ syntheses. These procedures involve synthesis of the reactant or reagent, using smaller precursor molecules, directly in the interior of the zeolite. Successful ‘‘ship-in-a-bottle’’ syntheses require the use of efﬁcient reactions that either produce no by-products or small by-products that can be removed by solid–liquid extraction.30 For example, the ‘‘ship-in-a-bottle’’ synthesis of the electron-transfer sensitizer, 2,4,6-triphenylpyrylium cation, 1,31 was reported in 1994 by Garcı´ a et al.32 by heating the acid form of faujasite Y, HY, at 383 K for 7 days in isooctane containing chalcone and acetophenone (Fig. 5). Solid–liquid extraction was employed in this case to remove the dihydrochalcone and 1,3,5-triphenyl-1,5-pentanedione by-products. Subsequently, a ‘‘camel through the eye of a needle’’ procedure was reported for the incorporation of 1.33 This procedure, however, is also a ‘‘ship-in-a-bottle’’ synthesis since it involved incorporation of the in situ generated smaller precursor, 1,3-5-triphenylpent-2-en-1,5dione, followed by intrazeolite condensation dehydration (Fig. 5). Molecular modeling of this caged pyrylium salt reveals a very tight ﬁt with the phenyl rings partially penetrating the windows into the adjacent supercage.32

REACTION MONITORING

Zeolites are devoid of chromophores that prevent interrogation of intercalated substrates in the IR, between approximately 3000 and 1200 cm1 and in the UV–Vis regions of the electromagnetic spectrum.14 As a consequence, both time-resolved34,35 or static monitoring36 of intrazeolite reactions are possible by using either diffuse reﬂectance techniques27,35 or absorption techniques employing thin wafers.37,38 In addition, other optical methods such as emission spectroscopy are available

232

E.L. CLENNAN

O Ph +

O

Ph chalcone

O

HY

extract

Ph

Ph

OO

Ph

H2O

O Ph "ship-in-a-bottle"

NaY O 1

H2O 40 °C

O

Ph O

Ph

O

+

H2O

Ph

"camel-through-the-eye-of-a-needle"

Fig. 5 Intrazeolite synthesis of incarcerated guests.

to monitor intrazeolite reactions.39 Solid-state magic angle spinning (MAS) NMR experiments have also been extensively utilized to examine both reaction products and intermediates in intrazeolite reactions.40–43 PRODUCT RECOVERY

Reliable mechanistic conclusions require high intrazeolite yields that account for the majority of the substrate mass balance. This can be a challenge because of the smallscale reactions often conducted for mechanistic studies. In addition, rapid removal of the products from the zeolite, and/or low conversions to decrease residence time, is occasionally necessary because of the sensitivity of the reaction products to the zeolite environment.44,45 Intrazeolite products are generally recovered by extractive techniques from either the intact zeolite, or from a mixture formed after mild digestion of the zeolite. Polar solvents such as tetrahydrofuran or acetonitrile coupled with a continuous extraction technique is in particular an effective means to remove polar products with an afﬁnity for the interior of the zeolite.44 Zeolite digestion with mineral acids, in order to liberate the products, must be conducted with care in order to prevent acid catalyzed product decomposition or reaction.46,47

4

Intrazeolite photooxygenations

Oxidations are indisputably one of the most important chemical transformations known and improvement in oxidation technology would have a signiﬁcant impact

MECHANISMS OF OXYGENATIONS IN ZEOLITES

233

Spin Inversion Energy Transfer

I

RH 3

III

O2

Coordination 2-Electron Transfer

II Radical Formation 1-Electron Transfer

Fig. 6 Wagnerova oxygenation classiﬁcation scheme.

on the chemical and biochemical industries. In particular, reﬁned methods to use molecular oxygen as a terminal oxidant would provide an economically viable and environmentally friendly alternative to metal-based oxidants that are widely utilized in these endeavors. However, a signiﬁcant impediment to the use of oxygen is its inherent lack of selectivity and the difﬁculty associated with development of strategies designed to control the reactivity of this small molecule. Nevertheless, a myriad of oxygenation methods with different characteristics are known, which can be conveniently classiﬁed using the Wagnerova classiﬁcation scheme (Fig. 6),48 are available. The oxygenations are grouped into classes I, II, and III based upon the method used to overcome the spin barrier for reaction of the organic substrate with triplet oxygen. This classiﬁcation is further reﬁned using the method of activation, light or heat, and whether oxygen or the substrate is activated to induce reaction.49 In this review we will discuss examples of the most common Wagnerova Class I, and II intrazeolite oxygenations and compare them to their homogeneous analogs. WAGNEROVA CLASSIFICATION I OXYGENATIONS

The ﬁrst Wagnerova Class I intrazeolite photooxygenation was reported in 1996 by Li and Ramamurthy.50 In these heterogeneous hexane-slurry reactions they produced singlet oxygen by energy transfer51 from a zeolite-embedded sensitizer which then proceeded to react with the substrate via an ene reaction to give allylic hydroperoxides (Fig. 7). This reaction, and in particular the reaction selectivity, is critically dependent upon the experimental conditions. Although a variety of different singlet oxygen sensitizers have been incarcerated in zeolites (e.g. C60@NaY52 and Co(II)phthalocyanine@NaY53) the majority of intrazeolite photooxygenations have been conducted using either thionin, 2, or methylene blue, 3, as sensitizers.49 These thiazine dyes are most often loaded into the zeolite by diffusion from aqueous solution to form highly colored zeolite powders. Drying of the zeolite to remove residual water is essential. Water forces the sensitizers to aggregate and as a consequence self-quenching, rather than singlet oxygen formation, occurs and no ene reactions are observed.54 Typically, the colored zeolite powders are carefully dried at moderate temperatures (100 1C) and

234

E.L. CLENNAN

at 104 torr for 24 h in order to prevent sensitizer decomposition. Singlet oxygen formation was veriﬁed upon irradiation of these ‘‘dry’’ samples by detection of its characteristic emission at 1268 nm. Product-based estimates29 and direct measurement55 were used to determine a singlet oxygen lifetime of approximately 7.5 ms in NaY (Si/Al ¼ 2.4) under the reaction conditions (hexane slurry). This lifetime is sufﬁcient to allow the metastable singlet oxygen to migrate more than 370 A˚ within its lifetime. As a consequence, it can sample more than 5000 supercages surrounding the locus of its generation in order to ﬁnd a substrate.

H2N

N

Cl

N

S

NH2

N

Cl

S

2

N

3

The choice of slurry solvent for these reactions is also critical. The hydrocarbon substrates shown in Fig. 7 exhibit a limited afﬁnity for the interior of the zeolite. Consequently, at the beginning of the reaction a signiﬁcant fraction of the substrate resides in the hexane. Some of the reaction must occur in the interior of the zeolite since the unique product selectivity (Fig. 7) is not observed on the surface of silica gel or when the sensitizer is size excluded from migrating into the zeolite.50 In fact, the observation that the product selectivity for the reaction of 4 is identical in perﬂuorohexane, where its ﬂuorophobicity drives it completely into the zeolite, and in hexane, where a substantial portion resides outside the zeolite,29 requires that in hexane it must migrate into the zeolite where the hydroperoxide product accumulates and 100% of the reaction occurs (Fig. 8). The higher percentage of conversion of 4 in perﬂuorohexane (Fig. 8) is due to its average higher substrate intrazeolite concentration, in comparison to the hexane experiments, during the course of the irradiations. At the other extreme, the use of a solvent that has a much higher intrazeolite afﬁnity than the substrate can force 100% of the reaction to occur in the bulk solvent. 1

O2

hexane

+ OOH

OOH

Thionin@NaY

100

--

Thionin/CH3CN

40

60 OOH

1

O2

OOH

OOH

hexane Thionin@NaY

100

--

--

Thionin/CH3CN

10

47

43

Fig. 7 Wagnerova Type I singlet oxygen ene reactions.

MECHANISMS OF OXYGENATIONS IN ZEOLITES

235

1

O2

+

hexane

4 C6H14/MB@NaY C6F14/MB@NaY

OOH

OOH

% conversion 35 ± 3 81 ± 6

92 95

8 5

*MB - methylene blue OOH

OOH

1

O2

+

OOH +

5 Ru(bpy)32+/MeOH Ru(bpy)32+@NaY/MeOH

50

17

31

51

16

33

Thionin@NaY/hexane

--

--

100

Thionin/CH3CN

48

45

6

Fig. 8 A comparison of singlet oxygen ene product ratios in different solvent slurries.

That was the situation encounter in 1986 when Pettit and Fox56 reported nearly identical selectivity in the singlet oxygen ene reaction of 1-methylcylclohexene, 5, using either a soluble sensitizer (Ru(bpy)2+ 3 /MeOH) or a zeolite-embedded sensitizer (Ru(bpy)2+ @NaY) in a methanol slurry (Fig. 8). 3 Irradiation times must also be selected carefully in order to maximize the product yield. Allylic hydroperoxides are susceptible to decomposition upon extended irradiation.45 Control reactions indicate that both light and oxygen are necessary for the decomposition and that reactivity appears to track increasing nucleophilicity of the double bond in the allylic hydroperoxide. These observations led to the suggestion that over-oxidation with singlet oxygen is responsible for the decreasing yields with increasing irradiation times. Fortunately, these intrazeolite singlet oxygen reactions with hydrocarbon substrates (Fig. 7) are very rapid in comparison to their homogeneous analogs and short irradiation times can lead to large conversions.57 In addition, use of ﬂuorocarbon slurries (vide supra) can also increase the rate of reaction and prevent decomposition. Short irradiation times are also advantageous in order to prevent sensitizer bleaching which is known to be a problem with methylene blue.58 Finally, decomposition can be promoted by residual Brønsted acid sites in the zeolite.45 Addition of a small amount of pyridine to remove these acid sites is known to prevent rearrangements of allylic alcohols and allow their intrazeolite photooxygenations.59 Conventional singlet oxygen sensitizers such as Thiazine dyes, 2 and 3, have been reported to promote Wagnerova Class II electron-transfer photooxygenations.60 For example, the participation of singlet oxygen in the reaction depicted in Fig. 9

236

E.L. CLENNAN R CHO 3

O2

hexane C6H14/MB@NaY hν R

R

R = H, Me

Fig. 9 A Wagnerova Class II electron-transfer photooxygenation.

O

1

O2

CH3

H H H H

O

Ha Ha

H Na

Na

H A

B

Fig. 10 Complexation, A, and polarization, B, models for intrazeolite photooxygenations.

has been ruled out since neither 2,3-dimethyl-2-butene nor b-carotene quench product formation.60 Consequently, sensitizer loading levels, /SSSen (molecules per supercage), for Wagnerova Class I photooxygenations are kept low in order to prevent aggregation and potential competition with sensitized Wagnerova Class II reactions. Methylene blue loading levels between /SSMB ¼ 0.01 and 0.0001 are sufﬁcient for intrazeolite photooxygenation of simple alkenes as shown in Fig. 7.44 From both a synthetic and physical organic viewpoint the most intriguing feature of the intrazeolite reactions depicted in Fig. 7 are the regioselectivities. In both cases hydrogen abstraction occurs exclusively from the disubstituted end of the alkene, which is a reversal of the preference observed in solution. In order to rationalize this selectivity two models as shown in Fig. 10 were suggested. Both models invoke complexation of the alkene to the sodium cation; a phenomenon that is well established in both the condensed and gas phases.61,62 Model A provides an explanation for the regiochemistry in the acyclic alkene (Fig. 7).63 In this model complexation forces the allylic methyl group on the face of the alkene approached by singlet oxygen and thereby prevents abstraction of Ha. Model B provides an explanation for the regiochemistries observed during intrazeolite photooxygenations of cyclic alkenes (Fig. 7).64 In this model the cation polarizes the double bond placing a greater positive charge on the di-substituted rather than mono-substituted sp2 hybridized carbon. As a consequence the electrophilic singlet oxygen adds to the mono- rather than di-substituted site.

MECHANISMS OF OXYGENATIONS IN ZEOLITES R

HOO

hν O2

R

237 OOH R

+

R = Me

CH3CN NaMBY

I (32.9±3.3%) [39.5±3.7%]

II (67.1±3.3%) [60.5±3.7%]

I/II 0.49 0.65

R = Et

CH3CN NaMBY

(42.0±1.7%) [52.7±1.7%]

(58.0±1.7%) [47.3±1.7%]

0.72 1.11

R = iPr

CH3CN NaMBY

(52.6±1.1%) [51.8±2.8%]

(47.4±1.1%) [48.2±2.8%]

1.1 1.07

R = tBu

CH3CN NaMBY

(70.6±1.1%) [61.2±1.1%]

(29.4±1.1%) [38.8±1.1%]

2.4 1.58

Fig. 11 Steric effect in the intrazeolite singlet oxygen ene reaction.

R

O

O

Na

+

Na

+

O O R

b

D3 C

CD3

H3 C

CH3

kH/kD= 1.04 ± 0.02

a III

IV

O O

CD3 CH3

D3C H3C V

C

Fig. 12 Sodium complexed perepoxides III and IV and isotope effect negating zwitterion V as an intrazeolite intermediate.

In 1999, Clennan and Sram65 pointed out that neither of these models successfully rationalized the intrazeolite singlet oxygen ene regiochemistry observed in a series of tetrasubstituted alkenes (Fig. 11). In solution the I/II ratio increases as the size of R increases. This is consistent with increasing steric interaction with the approaching oxygen that enhances formation of the C–OO bond at the remote, less congested, sp2 carbon. Model A or B (Fig. 10) would predict that this steric interaction would be greater in the zeolite and the I/II ratio should increase. In fact it does increase for R ¼ Me (0.49 ) 0.65) and for R ¼ Et (0.72 ) 1.11), however, it decreases for both R ¼ iPr (1.1 ) 1.07) and R ¼ tBu (2.4 ) 1.58). To rationalize the results in Fig. 11 Model C was suggested that invoked complexation of the cation to the perepoxide-like intermediate or transition state (Fig. 12). The Stephenson isotope effect of 1.0470.02 for the intrazeolite photooxygenation of Z-2,3-dimethyl-1,1,1,4,4,4-hexadeutero-2-butene (Fig. 12) is consistent with an intermediate or transition state with perepoxide symmetry.65 A zwitterionic intermediate like V (Fig. 12) would be expected to collapse with preferential abstraction of hydrogen rather than deuterium. In this new model as the R group gets larger the population of complexed perepoxide III increases at the expense of complexed perepoxide IV. Since hydrogen abstraction from the two cis-methyl groups in III (a and b) is

238

E.L. CLENNAN

equally likely, the I/II ratio (Fig. 11) should approach 1.0 as experimentally observed. In addition, while sodium complexation to the double bond in models A or B (Fig. 10) would decrease reactivity toward the potent electrophile, singlet oxygen, complexation to the incipient perepoxide in III or IV (Fig. 12) would be expected to increase reactivity consistent with the experimental observation. The Markovnikov-like preference for abstraction of hydrogen from the most substituted end of the alkene (Fig. 7) is diagnostic of a substantial amount of positive charge on the carbons in the perepoxide ring. The sodium cation in III and IV (Fig. 12) acts as an electron sink placing more positive charge on the carbon framework than in the non-complexed analog. As a consequence the C–O bond in III and IV is longer to the carbon best able to accommodate the positive charge and it preferentially breaks to give hydrogen abstraction from the most substituted end of the alkene. A comparison of the intrazeolite and solution photooxygenations of a series of trimethylstyrenes (Fig. 13) that showed enhanced sensitivity to substitutents in intrazeolite in comparison to solution reactions provided independent conﬁrmation of increased positive charge on the carbon framework of the intrazeolite embedded perepoxide.65 Despite the fact that Model C65 does not explicitly invoke complexation of the interstitial cation with the alkene linkage, extensive experimental61,62 and computational66 precedent suggests that complexation is thermodynamically favorable. The interaction between the alkene and the cation serves to anchor the substrate near the supercage wall where the cation is located (Fig. 1). The proximity to the framework wall has experimental consequences as revealed by examination of the 0.15 0.1

solution intrazeolite

0.05 0 -0.05 -0.1

OOH

HOO

1

O2

+

-0.15 X

-0.2 -0.4

Y

X

Y

X

A

-0.2

0

Y

B

0.2

0.4

0.6

0.8

Hammett sigma (s)

Fig. 13 Substituent effects on the regiochemistry of the intrazeolite singlet oxygen ene reactions of trimethylstyrenes.

MECHANISMS OF OXYGENATIONS IN ZEOLITES twin CD3 6

1O

1O

CD3

HOO

10

+

HOO

49% 57%

50% 32% +

19% 56% HOO

2

CD3CN NaMBY

18% 9% CD2

HOO

8% 33%

OOH

CD3 + 48% 61%

1.09 10.2

+

+ 0% 0%

CD3 44% 6% OOH

HOO

HOO

0.29 0.26

63% 35%

+

CD3CN NaMBY

0.98 1.78

OOH

HOO +

1O 2

0.69 1.71

+

1% 11% HOO

CD3 55% 21% OOH

+

2

twix/lone

OOH

38% 36%

2

1O

lone CD3

7% 43%

CD3CN NaMBY

8

HOO +

CD3CN NaMBY

7

twix CD2

2

CD3CN NaMBY 1O

9

HOO

239

57% 75%

43% 25%

1.33 3.00

Fig. 14 Intrazeolite product distributions in a series of trialkyl substituted alkenes.

data in Fig. 14.44 The trisubstituted alkenes 6–10 are either labeled or have a substituent pattern that allows comparison of the extent of hydrogen abstraction from the same side of the alkene (i.e. twix and lone). When two methyls compete as the site of hydrogen abstraction, the twix/lone ratio increases when the reaction is moved from solution into the zeolite by approximately a factor of 2–3 (e.g. 6 0.69 ) 1.71 (factor 2.5); 7 0.98 ) 1.78 (factor 1.8); 10 1.33 ) 3.00 (factor 2.25)). This increase is a measure of the Markovnikov directing effect, which is more important in the zeolite than in solution. These three alkenes differ in the identity of the twin substituent (i.e. 6-CD3; 7-ethyl; 10-isopropyl), however, these substitutents have very similar electronic properties and as a result stabilize the positive charge in the perepoxide, which is responsible for the magnitude of the Markovnikov effect, to the same extent. This increase of a factor of 2–3, however, is not observed when hydrogen abstraction from a methyl and ethyl group compete (e.g. 8 and 9). The decrease in the twix/lone ratio for 8 and the large increase by a factor of 9.4 for reaction of 9 both reﬂect a substantial decrease in abstraction from the ethyl group in the zeolite in comparison to solution. This is consistent with the Ramamurthy suggestion that complexation of sodium to the alkene forces the alkyl group on the same face as the approaching singlet oxygen thereby making the hydrogens on the methylene carbon inaccessible to the pendant oxygen in the perepoxide. In Model C (Fig. 12), however, as the singlet oxygen approaches the face of the alkene the cation

240

E.L. CLENNAN [2 + 2]

[4 + 2]

O O Ph

Ph

1O

CHO

2

O O

Ph +

+

O O +

OOH

11

+ H O O

O O H O O

ene [4 + 2] 20% 92%

solution thionin@NaY

8%

HOO 1O 2

12

O O

+

CH2Cl2/MB*

15%

thionin@NaY

100%

85% *Methylene Blue

Fig. 15 Comparison of product mode (ene versus [2+2] and [4+2]) in Wagnerova Type I photooxygenations in solution and in the interior of a zeolite.

moves to stabilize the incipient perepoxide and it is the close proximity of the supercage wall that continues to enforce the conformation that is unfavorable for hydrogen abstraction. A striking feature of the intrazeolite singlet oxygen ene reaction is the rate enhancement often observed in these reactions.57,67 This rate enhancement is nicely accounted for by stabilization of the incipient perepoxide as depicted by Model C in Fig. 12. This rate enhancement can be used to promote the ene reaction at the expense of other reaction modes.68 An interesting example of this was reported by Stratakis and Rabalakos.69 Photooxygenations of alkenylarenes, 11 and 12 (Fig. 15) are dominated by [2+2] and [4+2] reactions in solution but react predominately by the ene mode in intrazeolite reactions. A second striking feature of the singlet oxygen ene reaction in solution is the ‘‘cis-effect’’ which refers to the propensity for abstraction of hydrogen from the most substituted side of the alkene70 (Fig. 14). This unique aspect of the reaction has been attributed to; (1) a gearing effect which lowers the rotational barrier to place the allylic hydrogen in the preferred, perpendicular to the p-plane, geometry for abstraction,71 and (2) a stabilizing interaction between the HOMO localized on the trailing oxygen in the perepoxide and the LUMO on the allylic hydrogens.72,73 The importance of the ‘‘cis-effect’’ drops precipitously in intrazeolite photooxygenations (e.g. 6 93% ) 57%; 7 99% ) 89%; 8 81% ) 44%; 9 92% ) 67%). This decrease is a natural consequence of the cation complexation of the perepoxide, which for steric reasons is preferred on the least substituted side of the alkene. Model C depicted in Fig. 12 works well for simple alkyl-substituted alkenes (e.g. Figs. 7 and 14). However, when alternative binding sites are provided different

MECHANISMS OF OXYGENATIONS IN ZEOLITES

241

O

O O

1O

2

HOO

O O

O

+ HOO

13 NaMBY (CD3)2CO

94% 97%

O

6% 3%

O O

1O 2

HOO

O O

O

+ HOO

14 NaMBY CD3CN

94±1% 98±2%

6±1% 2±2%

O

O O

O

O O

1O 2

NaMBY

O

+ O HOO

15

71±4% 85±1%

NaMBY CD3CN O

29±4% 15±1% O

O 1O2

HOO

16 NaMBY CD3CN

O O

OOH 85% 53% O

1O 2

O

+

15% 47% O

17 NaMBY CD3CN

OOH

O

HOO

O +

13±7% 47±1%

OOH 87±7% 53±1%

Fig. 16 Comparison of intrazeolite and solution singlet oxygen ene reactions of electron poor alkenes.

binding motifs can be observed and different regiochemistries and rate enhancements/ reductions can be realized. An example of this phenomenon is illustrated in Fig. 16.74 In compounds 13–17 the carbonyl group is expected, based upon limited binding data,75,76 to provide the preferential binding site for sodium. Consistent with this suggestion of increased binding afﬁnities is the observation that these compounds, in contrast to simple alkyl substituted alkenes, migrate rapidly and completely from hexane into NaMBY. However, despite the higher intrazeolite concentrations, photooxidations of these alkenes were sluggish and reasonable conversions were only

242

E.L. CLENNAN

O R1

O

(CH2)n

1

O2 k1

R2 VI

O

Na+

O (CH2)n

1

O2

R1

kobs = ∑niki i

k2

R2 VII

O O

Na+

Na+ O

VII-VIII O

R1

(CH2)n

R2

1

O2 k3

VIII

Fig. 17 Curtin-Hammett dynamic binding in the interior of a zeolite.

realized at higher doping levels of methylene blue (/SSMB ¼ 0.01). In addition, compounds 13 and 14 reacted with identical regiochemistry in solution and in the zeolite, while compounds 16 and 17 reacted with regiochemistries reminiscent of those observed for simple alkenes (Figs. 7 and 14), and compound 15 exhibited intermediate behavior. The key to understanding the reactivities of these carbonyl containing substrates, and all substrates with multiple binding sites, lies in the recognition that complexation is a rapid dynamic process.76 In these particular compounds three different species are in rapid equilibrium within the zeolite supercage and their reactivities are those expected based upon the Curtin-Hammett Principle and the Winstein-Holness Equation77 (Fig. 17). The three species are the uncomplexed substrate, VI, and the two possible complexes, VII and VIII, formed by binding to the two functional groups. The observed rate constant for reaction of the substrate will be a weighted

MECHANISMS OF OXYGENATIONS IN ZEOLITES

243

average as given by the equation in Fig. 17 where ni is the mole fraction of the ith complex and ki is the rate constant for reaction of singlet oxygen with that complex. The observed rates constant will therefore depend on both the thermodynamically determined population of the complex and its rate of reaction. In the case of the carbonyl compounds, 13 and 14, depicted in Fig. 16 the very slow rate of reaction can therefore be attributed to preferential population of complex VIII. In this complex, the double bond is even less nucleophilic than the double bond in the uncomplexed substrate, VI, and the sodium is not aligned as in complex VII for stabilization of the incipient perepoxide (Fig. 12). In the case of substrate 15 the additional alkyl group on the double bond increases the population of complex VII to a sufﬁcient extent to allow observation of the unique regiochemistry induced in this complexation motif. The difference in the regiochemistry in the solution and intrazeolite photooxygenation of 16 and 17 suggests that the tether connecting the two binding sites allows simultaneous binding to give complex VII–VIII (Fig. 17). A second example of an intrazeolite photooxygenation in which the Curtin-Hammett model (Fig. 17) provides a satisfying rationale was provided by Ramamurthy and coworkers.78 Intrazeolite photooxygenations of 18 and 19 (Fig. 18) exhibit the same preference for formation of the secondary hydroperoxide as observed with simple alkyl substituted alkenes despite the fact that cation binding to the phenyl ring is energetically more favorable than to a double bond. Within the Curtin-Hammett formalism despite the lower population of the phenyl analog of VII (Fig. 17) relative to a phenyl/cation complex, k2 is sufﬁciently large to allow observation of intrazeolite ene regiochemistry reminiscent of simple alkyl substituted alkenes. These workers also recognized that a folded conformation in which the cation is simultaneously bound to the aryl ring and the alkene (or perepoxide) could be involved. Experimental veriﬁcation of simultaneous cation binding to a phenyl ring and double bond was provide by Stratakis and coworkers79 who noted that the intrazeolite ene regiochemistries of a series of phenyl substituted alkenes, 20, were dependent upon the distance separating the two functional groups (Fig. 18). A similar effect on regiochemistry was not observed when the phenyl rings were replaced by cyclohexyls, 21. Stratakis80 and coworkers have also elegantly demonstrated that addition of a second binding site for the cation can have dramatic effects on the regiochemistry of the intrazeolite ene reaction. The intrazeolite photooxygenation of a-terpinyl acetate, 22, generated the endocyclic allylic hydroperoxide D as the overwhelming product (Fig. 19). This is in stark contrast to the results with limonene, 23, and p-menth-1-ene, 24, which give exclusively the exocyclic allylic hydroperoxide. The results with 23 and 24 conform to expectations based upon model C designed for simple alkenes (Fig. 12 and/or 17). However, 22 reacts predominantly by hydrogen abstraction from the most hindered side and from the least substituted end of the alkene. The ‘‘cis-effect’’, ([C]+[D]+[E])/[A]+[B]), is enhanced from 56% in solution (CH2Cl2/methylene blue) to 91% in the zeolite! (Fig. 19) To account for this unusual result Stratakis and coworkers80 suggested the intramolecular dual binding site model shown in the box in Fig. 19.

244

E.L. CLENNAN 1O

HOO

2

+ OOH

18 NaThioninY CH3CN

88% 53%

12% 47%

1O 2

+ OOH 90% 51%

19 NaThioninY CH3CN

(CH2)n D3C

1O 2

21 n=0 n=1 n=2

(CH2)n

(CH2)n + D3C

n=0 n=1 n=2 n=3

D3C

10% 49%

NaThioninY

20

(CH2)n

HOO

OOH

D2C

33% 65% 38% 39%

1O

57% 28% 18% 39%

(CH2)n

NaThioninY D3C

OOH 41% 40% 42%

D2C

OOH 37% 35% 37%

(CH2)n

D3C

OOH

(CH2)n +

2

HOO +

10% 7% 44% 22%

HOO +

(CH2)n

D3C 22% 25% 21%

Fig. 18 Inﬂuence of a phenyl ring as an alternative binding site in intrazeolite singlet oxygen ene reactions.

In addition to the change in regiochemistry, the dual binding complexation model also forces a change in diastereoselectivity, ([D]/[C]), from 5.7 in solution to 43.5 in the zeolite81 (Fig. 19). Unfortunately, the lack of availability of chiral zeolites has severely hampered efforts to perform enantioselective intrazeolite singlet oxygen ene reactions. Nevertheless, in 1997 Ramamurthy and coworkers reported a modest enantiomeric excess of approximately 15% for the formation of the secondary allylic hydroperoxide from photooxygenation of 18 in (+)-Ephedrine (a-1-(methylaminoethyl)benzylalcohol hydrochloride doped [email protected] The effect of cations on ene regiochemistry of several intrazeolite photooxygenations have been examined83 (Fig. 20). The effect is modest but the unique

MECHANISMS OF OXYGENATIONS IN ZEOLITES

245

HOO 1O 2

HOO

OOH

HOO

HOO +

+

+

HOO

+

+

NaThioninY OAc

OAc

22

5%(27) A solution values in parenthesis

OAc

OAc

OAc

4%(17) B

2%(7) C

87%(40) D

OAc

OAc 2%(9) E

OOH

1O 2

NaThioninY Na Ac 23

O

O O

OOH

1O 2

NaThioninY

24

Fig. 19 Inﬂuence of oxygen as an alternative binding site in intrazeolite singlet oxygen ene reactions.

hν

R

O2 R = Et

R = Ph

+

R

OOH

R

OOH

CH3CN/RB*

49

51

LiY/Thionin/Hexane

74

26

NaY/Thionin/Hexane

74

26

RbY/Thionin/Hexane

59

41

CsY/Thionin/Hexane

50

50

CH3CN/RB*

53

47

LiY/Thionin/Hexane

95

5

NaY/Thionin/Hexane

88

12

RbY/Thionin/Hexane

70

30

CsY/Thionin/Hexane

52

48 * Rose Bengal

Fig. 20 Cation effect on an intrazeolite singlet oxygen ene reaction in Y-zeolite.

246

E.L. CLENNAN hν 1 TPP + 3O2 O2

OOH

+ OOH Isooctane NaY ZSM-5

HOO

45% 100% 25%

55% 0% 75% +

Isooctane NaY ZSM-5

+ HOO

HOO

OOH 38% 32% 19%

61% 57% 48%

1% 11% 33% +

OOH Isooctane NaY ZSM-5

+ HOO

HOO

55% 35% 29%

12% 56% 36%

33% 9% 35% OOH

+

+ OOH

OOH Isooctane NaY ZSM-5

9% 1% 50%

35% 2% 30%

56% 88% 20%

Fig. 21 Comparison of singlet oxygen regiochemistry in solution to that observed in two different zeolites.

intrazeolite regiochemistry appears to decrease with decreasing alkene-cation binding afﬁnity.66 In the weakly binding CsY (Fig. 20) near solution-like behavior is observed. These results are also consistent with the Curtin-Hammett like model in which reaction via the unbound substrate (e.g. VI in Fig. 17) becomes the dominant reaction pathway. This could be a consequence of the higher population of the unbound alkene or alternatively a result of the diminished stabilization of the incipient perepoxide and the concomitant smaller k2 for reaction of the bound substrate (Fig. 17). The majority of intrazeolite photooxygenations have been conducted in NaY,84,85 however, one study in the pentasil zeolite ZSM-5 demonstrates that steric conﬁnement effects can play important roles.84 A comparison of the reactions of a series of trisubstituted alkenes in isooctane, NaY, and in ZSM-5 is given in Fig. 21. The reactions

MECHANISMS OF OXYGENATIONS IN ZEOLITES

247

Ph PhCHO

Ph +

+

+ Ph

OHC

Ph

Ph

O Ph

Ph

O O O

Ph +

+

Ph PTE**/DCA*

73%

73%

16%

4%

4%

ZSM-5/DCA*

0

0

0

0

0

O O Ph 3% 100%

* 9, 10-dicyanoanthracene ** pentaerythritol trimethyl ether

Fig. 22 Effect of zeolite enforced physical separation of substrate and sensitizer on product distribution in photooxygenation of 1,4-diphenyl-1,3-butadiene.

in ZSM-5 were conducted by adding the zeolite containing the substrates to an isooctane solution of tetraphenylporphyrin (TPP). Isooctane and TPP are both too large to migrate into either the straight (5.4 A˚ 5.6 A˚) or sinusoidal (5.1 A˚ 5.5 A˚) channels of ZSM-5 and as a consequence the reaction occurs by migration of singlet oxygen from isooctane where it is formed by energy transfer from excited TPP into the zeolite pores containing the substrates (Fig. 21). The formation of the allylic hydroperoxides encased in the boxes in Fig. 21 experience a dramatic increase as the reactions are moved from NaY into ZSM-5.84 In each case these compounds represent hydrogen abstraction from the largest substituent group present in the alkene substrate. The authors suggest, and verify by molecular modeling, that the peroxide diastereomer with the pendant oxygen on the side of the alkene bearing the largest group are more compact.84 Comparison of the dimensions of the two peroxide diastereomers reveals that only the more compact isomer is of the correct size to be accommodated in the straight channels (5.4 A˚ 5.6 A˚) of ZSM-5. Tung and coworkers86 have also taken advantage of the ability to physically isolate the sensitizer from the substrate to discriminate against competing Wagnerova Type II processes (Fig. 6). Experimental measures to promote singlet oxygen reactions such as the use of low concentrations of sensitizer, low temperatures, and use of deuterated solvents in order to enhance the singlet oxygen lifetime, are not always successful.51 This is in particular the case with electron deﬁcient sensitizers such as 9,10-dicyanoanthracene (DCA).87 For example, photooxygenation of 1,4-diphenyl1,3-butadiene with DCA in pentaerythritol trimethyl ether (PTE) leads to a very complicated reaction mixture86 (Fig. 22). The formation of the ozonide is especially diagnostic of an electron-transfer pathway initiated by formation of the epoxide radical cation and DCA radical anion.88 Subsequent reaction of either the closed, or distonic ring opened, epoxide radical cation with superoxide, formed by reduction with DCA radical anion, is a well established reaction.88 In contrast to the very complicated product mixture obtained in solution, photooxygenation of 1,4-diphenyl1,3-butadiene (Fig. 22) in ZSM-5 with physically isolated DCA resulted in exclusive formation of the endoperoxide from [4+2] cycloaddition of singlet oxygen that migrated into the zeolite.

248

E.L. CLENNAN Ph +

PhCHO PhCHO

hν DCA*, O2

Ph

OHC

hν DCA*, O2 Ph

PTE**, ZSM-5

O

hν HA***, O2 PTE** Ph O

O O

O

Ph

+ Ph

O Ph

O Ph

* 9, 10-dicyanoanthracene ** pentaerythritol trimethyl ether *** hypocrellin A

Fig. 23 Compartmentalization of substrates and sensitizers as a tool to determine mechanistic origin of products.

The ability of zeolites to compartmentalize substrates and sensitizers provides a novel tool to determine the mechanistic origin of photooxygenation products. For example, photooxygenation of trans-stilbene (Fig. 23) with DCA in solution leads to a mixture of products. However, conducting the same reaction in ZSM-5 produces only benzaldehyde (Fig. 23) establishing it as a singlet oxygen product in the zeolite and as a potential singlet product in the solution reaction. Interestingly, the conﬁned intrazeolite space sterically prevents formation of the preferred singlet oxygen product in solution, the bis-endoperoxide (Fig. 23). Intrazeolite singlet oxygen Wagnerova Type I photooxygenations of organosulfur compounds have also been examined.89 A comparison of several of these reactions to their homogeneous analogs are presented in Fig. 24.46,90 Singlet oxygen as the reactive intermediate in these reactions was implicated by the fact that 2,3-dimethyl2-butene, which can migrate into the zeolite, quenched formation of the products while b-carotene, which is too large to migrate into the zeolite had no inﬂuence on the reaction. In these examples a dramatic increase in the sulfone at the expense of the sulfoxide yield occurs upon moving the reaction from solution into the zeolite. In addition, formation of the S–C cleavage (Pummerer) product is either completely suppressed or drastically reduced in the zeolite.91,92 The solution reactions have been extensively investigated and have been postulated to occur via the mechanism shown in Fig. 25.93,94 A notable feature of this mechanism is the formation of two intermediates, the persulfoxide (PS), and the hydroperoxysulfonium ylide (HPSY). Extensive computational studies have demonstrated that PS forms on a very ﬂat potential energy surface and experimentally PS has been shown to decompose primarily along the physical quenching channel, kq95 (Fig. 25). It is this decomposition that is responsible for the very low quantum yields (Fo0.05) observed in many of these reactions. Intramolecular hydrogen abstraction can also occur to form HPSY that subsequently reacts with sulﬁde to give the sulfoxide product. Experimental evidence which supports the mechanism shown in Fig. 25 is extensive and includes: (1) trapping studies with Ph2S and Ph2SO which are themselves inert to singlet oxygen (Fig. 25); (2) substituent effects on these trapping reactions which demonstrate that Ar2SO is trapping a nucleophilic intermediate and Ar2S an

MECHANISMS OF OXYGENATIONS IN ZEOLITES

249

O hν O2

S

CH3CN/MB NaMBY

10% 44% O

S 25

S 26

87% 0%

O O S

S

acetone/RB NaMBY

98% 46% O

+ 2% 54% O O S

S

O

+

+

S

H

2

CH3CN/MB NaMBY

59% 67%

S

O

19% 33%

21% 0% O O S

S

hν O2

+

S

+

2

CH3CN/MB NaMBY

28

S

H

2

hν O2

27

O +

3% 56%

hν O2

S

O O S

+

29% 69%

15% 25%

43% 3%

Fig. 24 Comparison of Wagnerova Type I singlet oxygen reactions of sulﬁdes in solution and in the interior of zeolite-Y.

R

S

CH3

3O

+

2

kq HO 1O

R

S

CH3

O

2

kT

R H3C

O S

PS kSO Ph2SO

kX H

O R S

R

CH3

kS

H HPSY

O 2 R

S

CH3

kPhS Ph2S

O + S CH3 R

S

O Ph2SO2

Ph2SO + R

S

CH3

Fig. 25 Mechanism of the reaction of singlet oxygen with sulﬁdes in solution.

250

E.L. CLENNAN

O O S

R R

Ph2S % conversion 98.7

hν, O2 0.02M zeolite/MB NaMBY

M

M+--PS Ph2SO

+

Ph2SO2

89.1%

10.9%

LiMBY

91.6%

8.4%

38.3

KMBY

95.2%

4.8%

4.8

RbMBY

100%

0%

0.5

CsMBY

100%

0%

50.8

BaMBY

98.6%

1.4%

95.2

Fig. 26 Cation complexed persulfoxide and a measure of its stability as manifested by % conversions as a function of interstitial cation.

electrophilic intermediate; (3) kinetic evidence that Ph2S but not Ph2SO competes with the sulﬁde substrate for an intermediate; (4) kinetic evidence that Ph2SO but not Ph2S increases the efﬁciency of the reaction (i.e. competes with kq); and (5) product isotope effect studies that require abstraction of the a-hydrogen. In order to rationalize the dramatic inﬂuence of the zeolite environment (Fig. 24) the solution mechanism has been modiﬁed by replacing the persulfoxide with a cation complexed persulfoxide M+–PS46 (Fig. 26). The dramatic effect of the cation identity on the percentage of conversion of Ph2S to its oxidized products (Fig. 26) provides direct evidence for the important role of this complexation. Diphenylsulﬁde, which is inert to singlet oxygen under typical irradiation conditions, is converted quantitatively to the sulfoxide and sulfone after 1 h of irradiation in NaMBY. In addition, the percent conversions decrease precipitously as the sodium is replaced with cations that generate decreasing electrostatic ﬁelds (Fig. 26). The modest percentage of conversion in BaMBY, despite the ability of Ba2+ to generate high electrostatic ﬁelds, was attributed to its large size that hindered approach of the second sulﬁde necessary to produce the two sulfoxide products. The effect of the complexation is to stabilize the persulfoxide, PS. This stabilization can potentially manifest itself in a more rapid rate of formation, or alternatively, in suppression of physical quenching (kq in Fig. 25). Evidence that the latter phenomenon is of paramount importance was provided by examination of a series of sulﬁdes with tethered oleﬁnic linkages.96 In these cases the enhancement of reaction at the sulfur at the expense of the ene reaction at the oleﬁnic linkages could be quantitatively predicted based upon the assumption that kq ¼ 0. The cation stabilization of the persulfoxide extends its lifetime sufﬁciently to allow trapping with adventitious water and with sulﬁde substrate46 (Fig. 27). These reactions provide a satisfying explanation for the enhanced intrazeolite sulfone yields

MECHANISMS OF OXYGENATIONS IN ZEOLITES

-H2O S O O

O O S

OOH H2O

S

251

OH

M S

M+--PS

2

S O

7 y = 0.010663 + 2.4021x R= 0.99459

[C5H10SO] / [C5H10SO2]

6 5 4 3 2 1O

1

S

+

2

NaY

S O

S O O

0 0

0.5

1

1.5

2

2.5

3

<S>5 (molecules per supercage)

Fig. 27 Competition by water and sulﬁde for cation complexed persulfoxide and its effect on the sulfoxide/sulfone ratio as a function of zeolite loading.

(Fig. 24) and the unprecedented effect of loading level on sulfoxide/sulfone ratio (Fig. 27). Suppression of the Pummerer reaction (Fig. 24) could also be a manifestation of the stabilization of the persulfoxide which prevents its interconversion to the hydroperoxysulfonium ylide, HPSY (Fig. 25), which is the intermediate that has been suggested to undergo a 1,2-shift of the hydroperoxy group and ultimately produces the S–C bond cleavage products.92 However, the situation is probably more complex since the intrazeolite reaction of b-chlorosulﬁde, 29 (Fig. 28A), requires a-hydrogen abstraction. The complexation motif (Fig. 28B) which favors the extended rather than folded M+–PS may also play an important role. A complete understanding of these reactions will require additional studies. The complexities associated with these reactions were also evident during an intrazeolite product isotope effect study of 2,2,6,6-tetradeutero-1,4-dithiane97 (Fig. 29). The absence of an isotope effect during oxidation with m-chloroperbenzoic is consistent with the lack of C–H(D) abstraction on the reaction surface. The substantial isotope effect of 1.1570.02 for the reaction of singlet oxygen is consistent with the

252

E.L. CLENNAN O

A

hν

S

Cl

29

O

S

O2 NaMBY

+

Cl

S

58%

42%

B S

X

O H HO Na

destabilizing steric interaction

S O H HO

S H H

O O

Na

extended

folded

Fig. 28 A. Intrazeolite photooxygenation of a b-chlorosulﬁde that requires a hydroperoxy sulfonium ylide; and B. The formation of an extended cation complexed persulfoxide that can potentially inhibit hydroperoxy sulfonium ylide formation.

O D D

S

D D

O

S + D D

S

D D

S kH

O Cl

OOH

D D

kH/kD

hν, O2, NaMBY

1.15 ± 0.01

1O 2

D D H S D H Na+ S D H H 1O

kD

kD

0.996 ± 0.004 1.15 ± 0.02

Keq

D D

S

hν, O2 Rose Bengal acetone-d6

D D H S Na+ D H S D H H

S

2

kH

Fig. 29 A comparison of product isotope effects in MCPBA and singlet oxygen reactions in solution and zeolite media.

preferential abstraction of hydrogen (kX in Fig. 25) and its enhance ability in comparison to deuterium to compete with physical quenching (kq in Fig. 25). However, the observation of a substantial isotope effect (1.1570.01) in NaMBY is surprising given the demonstrated ability of the intrazeolite medium to suppress physical

MECHANISMS OF OXYGENATIONS IN ZEOLITES

253

quenching.96 To account for this observation the authors96 suggested the equili brium shown in the box in Fig. 29. Consequently, two limiting cases can explain the observed isotope effect; (1) physical quenching is suppressed and as a result kH/kD ¼ Keq, or (2) physical quenching is not suppressed because the counterion (sodium) is not available to stabilize the persulfoxide because it is complexed to the D remote sulfur and kH =kD ¼ K eq ðkH X =kX Þ. WAGNEROVA CLASSIFICATION II OXYGENATIONS

Blatter and Frei reported the ﬁrst intrazeolite Wagnerova Class II Oxygenation in 199398 (Fig. 30). The activation to overcome the spin barrier was provided by light and it acted on a reaction intermediate rather than on either of the starting materials. The key intermediate is a charge-transfer (CT) species involving transfer of electron density from the alkene to oxygen. Absorption of light by this intermediate induces electron transfer to form an alkene radical cation and superoxide (A in Fig. 30).99 Collapse of this ion pair by hydrogen abstraction generates a resonance stabilized allyl radical and the hydroperoxy radical which subsequently undergo bond formation to form the allyl hydroperoxide product100,101 (Fig. 30). Oxygen-organic molecule contact CT complexes are well-established species in solution,102–104 in solid oxygen matrices105–106, and in the gas phase107 under high oxygen pressure. These species have been most often identiﬁed by a reversibly formed low-intensity featureless absorbance on the bathochromic edge of organic absorption spectra. For example, a new absorbance at approximately 380 nm in an oxygen matrix of 2,3-dimethyl-2-butene is not assignable to either the matrix or alkene absorbance but has been attributed to CT absorbance.108,109 The most notable feature of these intrazeolite photooxygenations (Fig. 30) is that the oxygen CT band experiences a dramatic bathochromic shift in comparison to solution. This was detected initially by recording the product growth as a function of irradiation wavelength (laser reaction excitation spectrum)98,110 and was later veriﬁed by direct observation using diffuse reﬂectance UV–Vis spectroscopy.111 For example, 2,3-dimethyl-2-butene CT-absorbance is shifted to lower energy by more than 300 nm O2 @ NaY

O2

@ NaY CT Complex

hν 633 nm -50°C

(*)

O2

O O @ NaY

A 2

OOH OOH @ NaY B

Fig. 30 A Wagnerova Type II photooxygenation of an alkene.

O

254

E.L. CLENNAN O2

O2 @ NaY

Absorption Onset

8.30

O2 Absorption Onset IP(eV)

@ NaY

~750 nm

IP(eV)

@ BaY

~500 nm 8.80

O2

O2 @ NaY

@ BaY

~500 nm

~450 nm

~500 nm

8.67

9.13

9.73

O2

@ CaY

H

O2

O2

@ BaY

@ NaY

~600 nm

~500 nm

~500 nm

8.80

10.57

9.80

Fig. 31 Substrate-oxygen charge transfer band onsets as a function of the ionization potentials of the substrates.

by enclosure in NaY.98 These signiﬁcant intrazeolite shifts of oxygen-substrate CT bands were subsequently shown to be a general phenomenon and have now been reported for a wide spectrum of alkenes,111 aromatics,112 and alkanes113,114 (Fig. 31). These intrazeolite bathochromic shifts for oxygen-organic compound CT complexes are much larger than those observed for viologen–aromatic CT complexes.115,116 Blatter and Frei98 attributed these remarkable shifts to stabilization of the excited CT state (A and the transition state leading to it in Fig. 30) by the large electrostatic ﬁelds generated by the very poorly shielded cations in the zeolites. This suggestion is supported by several experimental observations including; (1) The longest wavelength light capable of inducing photooxygenation of 2,3,-dimethyl-2-butene in NaY is 760 nm in comparison to only 600 nm in a high-silica faujasite with fewer charge-balancing counterions.110 (2) Dramatic solvent effects on intramolecular CT absorbance’s serve as the basis of several solvochromatic scales (e.g. ET and Z).117 (3) The percent conversions for the intrazeolite photooxygenation of toluene increase monotonically with increasing intrazeolite electrostatic ﬁeld.118 (4) An alternative explanation invoking formation of singlet oxygen was ruled out by noting that the acetone/allylic hydroperoxide branching ratio (Fig. 30) dramatically increased upon replacement of NaY by a high-silica faujasite.98 Such a change would be unprecedented for singlet oxygen; decreases, not increases, in dioxetane formation (the precursor to acetone) have been observed by going to less polar solvents.119 On the other hand, deprotonation to give the radical pair, B in Fig. 30, is expected to be disfavored in the less polar high-silica faujasite and as a consequence acetone formation is expected to increase. (5) An alternative explanation which attributed the unusual reactivity to turnover at a small number of defect sites was eliminated by a time resolved FT-IR study.120 This study demonstrated that product growth was complete on a time scale too short to allow the cage-to-cage diffusion of substrates necessary to locate the defect sites. The intrazeolite electrostatic ﬁeld, however, is not universally accepted as the major factor contributing to the success of this reaction. Pidko and van Santen121

MECHANISMS OF OXYGENATIONS IN ZEOLITES

255

have recently suggested, based upon a series of density functional theory (DFT) calculations, that the ability of the cations to provide a suitable orientation for the oxygen in the organic substrate CT complex is the overriding factor enhancing intrazeolite photooxygenations. The minimum energy geometries of oxygen-2,3dimethyl-2-butene CT complexes were located and their energies determined in a series of zeolite-cluster models. The CT energies were then calculated with a clusterfree oxygen/2,3-dimethyl-2-butene CT complex in the same geometry as found in the cluster. These calculations demonstrated that the energies of the CT complex were nearly identical in the presence or absence of a zeolite-cluster model (i.e. in the presence or absence of cations). However, the red shifts (energies) of the CT bands were a sensitive function of the distance between the alkene and oxygen in the zeolite-cluster model. Consequently, this model suggests that it is destabilization of the CT complex by forcing it into a favorable pre-transition state orientation rather than stabilization of the excited state CT complex (A in Fig. 30) which is responsible for the observed intrazeolite bathochromic shifts. This new CT intrazeolite photooxygenation holds great promise as a general, exceedingly mild, environmentally green, protocol for the formation of value-added oxidized hydrocarbons (Fig. 32). The ability to irradiate at low temperatures (o50 1C) prevents thermal decomposition of the sensitive peroxidic products. In addition, competitive photolytic decomposition of the products, via difﬁcult to control chain reactions, is prevented by the long wavelengths necessary to activate the CT intermediate. Despite these attractive features, the afﬁnity of the oxidized hydrocarbons for the interior of the zeolite, the substantial scattering of visible light by the zeolite,122 and the design of an appropriate photochemical reactor provide major impediments/challenges for scale up and transfer of this new technology to an industrial setting.123,124 Although we have only illustrated the Frei intrazeolite photooxygenation with 2,3-dimethyl-2-butene (Fig. 30), it is a general phenomenon that allows controlled introduction of oxygen into alkenes, alkanes, and aryl substrates. Examples 1–4 in Fig. 32 demonstrate that even di-substituted (Examples 1 and 2) and mono-substituted (Example 4)125,126 alkenes react to give primarily allylic hydroperoxides with high selectivity. In contrast to the work reported in Ref.125, Tang and coworkers127 report that 1-butene reacts primarily via the dioxetane rather than via the allylic hydroperoxide. This suggestion was made based upon isolation of propanal as the major product rather than direct observation of the dioxetane. We note that allylic hydroperoxides can also give the same cleavage products as dioxetanes via Hock Cleavage.128 As the temperatures are increased the allylic hydroperoxides decompose by loss of water to give a,b-unsaturated carbonyl compounds. In addition, if any residual alkene remains during the warm-up the allylic hydroperoxide can transfer an oxygen to produce an epoxide. Photochemical cis–trans isomerizations of alkenes do not appear to compete with the photooxygenation reactions. For example, no intrazeolite photo-interconversions of cis- and trans-2-butene were detected and their epoxides were formed stereospecﬁcally.110 The lack of reactivity of 3,3-dimethyl-1-butene (Example 3, Fig. 32) reﬂects the hypsochromic shift of the CT onset with decreasing ionization potential of the

256

E.L. CLENNAN O hν > 600nm HOO

@ NaY

1.

@ NaY

0°C

O2

+

OH

0°C

H2O

O + O

2.

@ NaY

HOO

hν 514nm

@ NaY

0°C

O2

+

OH

0°C

H2O

O +

O2

@ NaY

3.

No Reaction at any Wavelength

hν 488nm @ BaY

4.

O2 -100°C

O

HOO

H +

0°C O

O

OOH

@ BaY

@ NaY

OOH

488 or 514nm O2 hν 514nm HOO

H @ BaY HO

O RT, dark

CH3CH2CH3 @ BaY

CH3CH3 @ CaY

O @ BaY

O2

+

9.

O

hν, 458nm

6.

@ BaY

@ BaY

O2

8.

CHO

hν < 515nm

5.

7.

O

H 21°C t1/2 = 6h

H H

H2O

hν < 500nm

RT, dark O

50°C 6h O

RT

O2 hν < 500nm O2

+

CH3OH O

+ H2O

OOH H

+

H2O

Fig. 32 Examples of Wagnerova Type II photooxygenations of alkene, arene, and alkane substrates.

alkene.110 This shift is that anticipated based upon the theory of CT transitions.129 However, increasing the intrazeolite electrostatic ﬁeld by ion exchange to form BaY generates a sufﬁcient bathochromatic shift that even 1-propene (Example 4, Fig. 32) can be induced to react.125 Other zeolites (e.g. BaZSM-5) as hosts have also been

MECHANISMS OF OXYGENATIONS IN ZEOLITES O hν > 400nm

O2 @ BaY

:

0.5

+

H

hν > 400nm @ BaY

O2 RT

:

1.0

Mechanism

O2

O

O +

O2 RT

257

+

Saturated Aldehydes

0.5

HOO

HOO

@ BaY

O2

@ BaY

@ BaY HOO HOO @ BaY

@ BaY

Fig. 33 Product ratios in the Wagnerova Type II photooxygenation of 1-pentene and a mechanistic rationale for their formation.

examined but appear to produce more complicated reaction mixtures presumably due to acid catalyzed polymerization of the alkenes.126 A closer examination by ex situ analysis using NMR or gas chromatography illustrates that intrazeolite reaction mixtures can get complex. For example photooxygenation of 1-pentene leads to three major carbonyl products plus a mixture of saturated aldehydes (valeraldehyde, propionaldehyde, butyraldehyde, acetaldehyde)38 (Fig. 33). Ethyl vinyl ketone and 2-pentenal arise from addition of the hydroperoxy radical to the two different ends of the allylic radical (Fig. 33). The ketone, E-3-penten-2-one, is formed by intrazeolite isomerization of 1-pentene followed by CT mediated photooxygenation of the 2-pentene isomer. Dioxetane cleavage, epoxide rearrangement, or presumably even Hoch cleavage130,131 of the allylic hydroperoxides can lead to the mixture of saturated aldehydes. The industrial process (Co3+ autooxidation)124 for oxidation of toluene to benzaldehyde is plagued by over-oxidation to form benzoic acid. However, a remarkably selective oxidation to the aldehyde is observed in both BaY and CaY (Example 5, Fig. 32). The higher ionization potential of benzaldehyde (9.5 eV) relative to toluene (8.8 eV) and the resulting hypsochromically shifted CT absorbance protects it from over-oxidation at the long wavelengths used in the intrazeolite experiments. Benzylhydroperoxide was not directly observed by FT-IR during this reaction, however, the fact that only 16O-benzaldehyde is produced when 16O2 is replaced with 18O2 after irradiation and before the zeolite is allowed to warm to room temperature, provides compelling evidence that it is a precursor to the benzaldehyde. Product extraction after irradiation demonstrated that over 90% selectivity for formation of benzaldehyde can be realized in NaY with formation of only traces of benzylalcohol, phenol, and 2-hydroxytoluene.118 Signiﬁcantly reduced selectivity for benzaldehyde formation was observed in the more acidic zeolites BaHY, BaZSM-5, and BaBeta, presumably because of acid catalyzed rearrangement of the benzylhydroperoxide intermediate.118

258

E.L. CLENNAN

Examples 6–9 (Fig. 32) illustrate intrazeolite Wagnerova Type II photooxygenation of alkanes. Given the exceedingly high ionization potentials of saturated alkanes (9.8–11.5 eV) the fact that any reaction at all is observed is remarkable. Nevertheless, isobutane (I.P. 10.57 eV) is converted to tert-butyl hydroperoxide in 98% selectivity in BaY even at greater than 50% conversion113 (Example 7, Fig. 32). In addition, intrazeolitic tert-butyl hydroperoxide can be used to stereospeciﬁcally epoxidize cis- and trans-2-butene (Example 7, Fig. 32). In the case of ethane (Example 9, Fig. 32) the higher electrostatic ﬁelds in CaY were necessary to induce this very difﬁcult to oxidize (I.P. 11.5 eV) hydrocarbon to react.124 Especially noteworthy in this reaction is the total absence of CO2 formation which is the major product in catalytic thermal processes.132 The reactions depicted in Fig. 32 are most often carried out at low temperatures. The incursion of a thermal process at elevated temperatures has occasionally been observed. In some cases the thermal oxygenation products are identical to the photochemical products and in other cases are different. For example, when 2,3dimethyl-2-butene/O2@NaY is warmed above 20 1C a reaction was observed which led to pinacolone (3,3-dimethyl-2-butanone) as the major product.98,110 Pinacolone is not formed in the photochemical reaction at the same temperature. On the other hand, identical products were observed in the thermal and photochemical intrazeolite oxygenations of cyclohexane.114,133–135 These intrazeolite thermal processes occur at temperatures well below that necessary to induce a classical autooxidation process in solution. Consequently, the strong electrostatic stabilization of oxygen CT complexes may also play a role in the thermal oxygenations. Indeed, the increase in reactivity of the thermal oxygenation of cyclohexane with increasing intrazeolite electrostatic ﬁeld led to the conclusion that initiation of both the thermal and photochemically activated processes occur by the same CT mechanism.134 Identical kinetic isotope effects (kH/kD ¼ 5.570.2) for the thermal and photochemical processes appears to support this conclusion.133 The exact mechanistic details of these thermal processes can be very difﬁcult to determine since a large number of factors including substrate and product diffusion rates can play a role.136 For example, after the electrostatic ﬁeld initiation, different loading levels of substrates can dictate different mechanisms and products. At low loading levels, as often used with gas phase reactions, the Frei mechanism (Fig. 30) can dominate. At higher loading levels the concentration of substrate can enhance the rates of the propagation steps of a classical chain autooxidation process. Initiation of this chain can occur by zeolite enhanced homolytic decomposition of a small amount of hydroperoxide that could form in an initial Frei oxygenation. The Frei oxygenation of cyclohexane can be changed to an autooxidation by simply conducting the thermal reaction at 42072 K in a high-pressure reactor using a cyclohexane slurry of zeolite Y rather than by using zeolite powders which had been doped with cyclohexane from the gas phase.136 The Frei photooxygenations in Fig. 32 were all conducted by irradiation of the zeolite powder in the absence of solvent. Solvents shield the substrate from the electrostatic ﬁeld of the cation and reduces the magnitude of the bathochromic shift.122 Kojima and coworkers reported that irradiation of trans-stilbene@NaY in

MECHANISMS OF OXYGENATIONS IN ZEOLITES

259

an oxygen saturated cyclohexane slurry resulted predominantly in cis– trans isomerization, a trace amount of phenanthrene, and no oxygenation.137–139 On the other hand the oxygenation product benzaldehyde was produced upon irradiation of the dry zeolite powder, trans-stilbene@NaY. These workers suggested that competitive oxygenation of the cyclohexane rather than shielding of the cations was responsible for the absence of substrate oxygenation in the slurry environment. Nevertheless, intrazeolite oxygenations of 1,1-diarylethylenes (e.g. 30, and 31) have been reported in hexane slurries.36,140 Ph hν

Ph @NaY O2

O Ph2CH2 +

Ph2CO + Ph2CH

O +

H

Ph Ph

30 <S>

λ(nm)

0.27

>420

66

34

0.27

366 254

42

58

36 61

37 14

0.27 0.27*

27

>420

Ph hν

Ph @NaY O2

O Ph2CO

+

Ph2CH

31 <S>

λ(nm)

0.27

>420

49

51

0.27

366 254

37±2 27

62±3

0.27 0.27*

>420

42

22

63

O +

CH3

25

CH3

Ph Ph Ph Ph

O +

Ph Ph

10 35

*pyridine included

In order to rationalize the complex reaction mixtures in these slurry reactions the authors suggested that irradiations of the oxygen CT complexes resulted in simultaneous formation of an epoxide and dioxetane36 (Fig. 34). The epoxide products were isolated only when pyridine was co-included in the zeolite during the reaction. Collapse of the 1,1-diarylethylene radical cation superoxide ion pair provides a reasonable explanation for the formation of the dioxetane, however, epoxide formation is more difﬁcult to rationalize. However, we do point out that photochemical formation of oxygen atoms has previously been observed in other systems.141 All the other products were formed either thermally or photochemically from these two primary photoproducts (Fig. 34). The thermal (acid catalyzed) formation of 1,1diphenylacetaldehyde from the epoxide during photooxygenation of 30 (Fig. 34) was independently veriﬁed by addition of an authentic sample of the epoxide to NaY. The formation of diphenylmethane in the reaction of 30 but not 31 is also consistent with the well-established facile (at 254 nm but not 366 or 420 nm) Norrish Type I

260

E.L. CLENNAN Ph

O2

O

hν

Ph Ph

Ph @NaY hν

O hν

O O

Ph2CH

Ph Ph

Ph2CH

CHO

H or O

Ph2CO

Ph2CH

H

A

Ph2CHOOH

Ph2CH

O2

or hν O+

heat

-H2O

Ph2CH

Ph2CH CO H

Ph2CH2

_

O2 H

O

CO OOH

Ph2CH

OOH

Fig. 34 Mechanism of product formation in the Wagnerova Type II photooxygenations of 1,1-diarylethylenes.

OH R

R

H

O2

OH

OH

hν R

R

H

O2

R

O R

OH R

+ H2O2

OH @ Ti-MCM-41

OOH R

R

R

OOH

O hν , O2, CH3CN

hν , O2, CH3CN S O S

+

O O S

Fig. 35 Mechanism of formation of hydrogen peroxide in the Wagnerova Type II photooxygenation of alcohols and its in-situ use to epoxidize alkenes and oxidized sulﬁdes.

MECHANISMS OF OXYGENATIONS IN ZEOLITES

O

hν

TP.@Y

H2O

+

H2O

261

H

+

HO

TP+@Y

HO

+

O2

OO

OOH +

Fig. 36 A Wagnerova Type II hydroperoxy radical chain initiated autooxidation.

decarbonylation of aldehydes but reluctant decarbonylation of methyl ketones.142 Finally, the formation of the oxygen CT complex A (Fig. 34), was directly observed by diffuse reﬂectance UV–Vis spectroscopy and the competency of 1,1-diphenylacetaldehyde as a precursor to diphenylmethane and benzophenone was veriﬁed by use of an authentic sample. The functional groups in the Frei substrates shown in Fig. 32 are limited to alkenes and aromatic rings. However, Frei photooxygenation of an alcohol has been reported143 (Fig. 35). The reaction generates a hydroperoxy ketal which decomposes to form hydrogen peroxide and the ketone. This reaction can be coupled with the wellestablished ability of ‘‘redox molecular sieves (zeolites)’’ to utilize hydrogen peroxide or alkyl hydroperoxides as terminal oxidants.144 For example, irradiation of a benzhydrol encapsulated titanium substituted zeolite in the presence of alkenes or sulﬁdes resulted in oxygen transfer to generate epoxides and sulfoxides, respectively (Fig. 35). This reaction was unsuccessful with alcohols such as isopropanol or 1-phenylethanol consistent with a Frei-type electron transfer as the key step in the process. Non-Frei Wagnerova Type II photooxygenations have also been reported. An especially elegant example is illustrated in Fig. 36.145 This reaction is initiated by electron transfer formation of hydroxyl radical. Rehm–Weller analysis146 indicates that electron transfer from water to excited 2,4,6-triphenylpyrylium, TP+, is exergonic. In addition, laser ﬂash photolysis of wet benzene slurries TP+@Y allowed direct detection of TP.147 The formation of the hydroxyl radical has also been independently veriﬁed by irradiation of TP+@Y in aqueous acetonitrile and observation of an EPR signal indicative of the hydroxy radical adduct of 5,5-dimethyl-1-pyrroline N-oxide.147 The hydroxyl radical then initiates a chain reaction incorporating oxygen into the product (Fig. 36).

Acknowledgement We thank the National Science Foundation for support of our zeolite photochemistry.

262

E.L. CLENNAN

References 1. Ramamurthy, V. (1991). Photochemistry in Organized and Constrained Media. VCH Publishers, Inc., New York, NY 2. Sen, S.E., Smith, S.M. and Sullivan, K.A. (1999). Organic transformations using zeolites and zeotype materials. Tetrahedron 55(44), 12657–12698 3. Csicsery, S.M. (1986). Catalysis by shape selective zeolites – science and technology. Pure Appl. Chem. 58, 841–856 4. Csicsery, S.M. (1984). Shape-selective catalysis in zeolites. Zeolites 4, 202–213 5. Sastre, G., Viruela, P.M. and Corma, A. (1997). Quantum chemistry calculations on the effect of electron conﬁnement upon the frontier molecular jorbitals of ethylene and benzene in sodalite. Implications on reactivity. Chem. Phys. Lett. 264, 565–572 6. Zicovich-Wilson, C.M., Planelles, J.H. and Jasko´lski, W. (1994). Spatially conﬁned simple quantum mechanical systems. Int. J. Quantum Chem. 50, 429–444 7. Zhang, L.Z., Sun, W. and Cheng, P. (2003). Spectroscopic and theoretical studies of quantum and electronic conﬁnement effects in nanostructured materials. Molecules 8, 207–222 8. Corma, A., Garcia, H., Sastre, G. and Viruela, P.M. (1997). Activation of molecules in conﬁned spaces: an approach to zeolite-guest supramolecular systems. J. Phys. Chem. B 101, 4575–4582 9. Rachdi, F., Hajji, L., Goze, C., Jones, D.J., Maireles-Torres, P. and Rozie´re, J. (1996). Quantum size effects induced by conﬁnement of C60 in MCM41. Solid State Commun. 100, 237–240 10. Zicovich-Wilson, C.M., Corma, A. and Viruela, P. (1994). Electronic conﬁnement of molecules in microscopic pores. A new concept which contributes to explain the catalytic activity of zeolites. J. Phys. Chem. 98, 10863–10870 11. Ma´rquez, F., Garcı´ a, H., Palomares, E., Ferna´ndez, L. and Corma, A. (2000). Spectroscopic evidence in support of the molecular orbital conﬁnement concept: case of anthracene incorporated in zeolites. J. Am. Chem. Soc. 122, 6520–6521 12. Ma´rquez, F., Zicovich-Wilson, C.M., Corma, A., Palomares, E. and Garcia, H. (2001). Naphthalene included within all-silica zeolites: inﬂuence of the host on the naphthalene photophysics. J. Phys. Chem. B 105, 9973–9979 13. Ma´rquez, F., Martı´ , V., Palomares, E., Garcia, H. and Adam, W. (2002). Observation of azo chromophore ﬂuorescence and phosphorescence emissions from DBH by applying exclusively the orbital conﬁnement effect in siliceous zeolites devoid of charge-balancing cations. J. Am. Chem. Soc. 124, 7264–7265 14. Dyer, A. (1988). An Introduction to Zeolite Molecular Sieves, p. 149, John Wiley & Sons, New York 15. Cundy, C.S. and Cox, P.A. (2003). The hydrothermal synthesis of zeolites: history and development from the earliest days to the present time. Chem. Rev. 103, 663–701 16. Klinowski, J., Carr, S.W., Tarling, S.E. and Barnes, P. (1987). Magic-angle-spinning NMR shows the aluminosilicate framework of ultramarine to be disordered. Nature (London) 330, 56–58 17. Dutta, P.K. and Ledney, M. (1997). Charge-transfer processes in zeolites: toward better artiﬁcial photosynthetic models. In Molecular Level Artificial Photosynthetic Materials, Meyer, G.J. (ed.), Vol. 44, pp. 209– 271. John Wiley & Sons, New York 18. (a) Ramamurthy, V., Corbin, D.R. and Eaton, D.F. (1990). Norrish Type I and Type II reactions of ketones as photochemical probes of the interior of zeolites. J. Org. Chem. 55, 5269–5278; (b) Ramamurthy, V. (1991). Photochemistry in Organized and Constrained Media. VCH Publishers, Inc., New York, NY

MECHANISMS OF OXYGENATIONS IN ZEOLITES

263

19. Catlow, C.R.A. (1992). Zeolites: structure, synthesis and properties – an introduction. In Modelling of Structure and Reactivity in Zeolites, Catlow, C.R.A. (ed.), pp. 1– 17. Academic Press, London 20. Kerr, G.T. (1973). Hydrogen zeolite Y, ultrastable zeolite Y, and aluminum-deﬁcient zeolites. Adv. Chem. Ser. 121, 219–229 21. Klinowski, J., Anderson, M.W., Fyfe, C.A. and Gobbi, G.C. (1983). Dealumination of mordenite using silicon tetrachloride vapor. Zeolites 3, 5–7 22. Skeels, G.W. and Breck, D.W. (1983). In Zeolite chemistry. V-substitution of silicon for aluminum in zeolites via reaction with aqueous fluorosilicate, Proceedings of the 6th International Conference on Zeolites, 1984, 1983, Olson, D. and Bisio, A. (eds), pp. 87–96. Butterworth, Guiford. 23. Marcus, B.K. and Cormier, W.E. (1999). Going green with zeolites. Chem. Eng. Prog. 95, 47–53 24. Li, P., Xiang, Y., Grassian, V.H. and Larsen, S.C. (1999). CO adsorption as a probe of acid sites and the electric ﬁeld in alkaline earth exchanged zeolite beta using FT-IR and ab initio quantum calculations. J. Phys. Chem. B 103(24), 5058–5062 25. Spackman, M.A. and Weber, H.-P. (1988). Electrostatic potential in dehydrated sodium zeolite A from low resolution X-ray diffraction data. J. Phys. Chem. 92, 794–796 26. Kao, H.-M. and Grey, C.P. (1997). Determination of the 31P–27Al J-coupling constant for trimethylphosphine bound to the Lewis acid site of zeolite HY. J. Am. Chem. Soc. 119, 627–628 27. Hashimoto, S. (2003). Zeolite photochemistry: impact of zeolites on photochemistry and feedback from photochemistry to zeolite science. J. Photochem. Photobiol. C 4, 19–49 28. Rao, V.J., Perlstein, D.L., Robbins, R.J., Lakshminarasimhan, P.H., Kao, H.-M., Grey, C.P. and Ramamurthy, V. (1998). Detection of low levels of Brønsted acidity in Na+Y and Na+X zeolites. J. Chem. Soc. Chem. Commun. 269–270 29. Pace, A. and Clennan, E.L. (2002). A new experimental protocol for intrazeolite photooxidations. The ﬁrst product-based estimate of an upper limit for the intrazeolite singlet oxygen lifetime. J. Am. Chem. Soc. 124, 11236–11237 30. Alvaro, M., Ferrer, B., Forne´s, V., Garcı´ a, H. and Scaiano, J.C. (2002). Bipyridinium macroring encapsulated within zeolite Y supercages. Preparation and intrazeolitic photochemistry of a common electron acceptor component of rotaxanes and catenanes. J. Phys. Chem. B 106, 6815–6820 31. Miranda, M.A. (1994). 2,4,6-Triphenylpyrylium tetraﬂuoroborate as an electrontransfer photosensitizer. Chem. Rev. 94, 1063–1069 32. Corma, A., Forne´s, V., Garcı´ a, H., Miranda, M.A., Primo, J. and Sabater, M.-J. (1994). Photoinduced electron transfer within zeolite cavities: cis-Stilbene isomerization photosensitized by 2,4,6-triphenylpyrylium cation imprisoned inside zeolite Y. J. Am. Chem. Soc. 116, 2276–2280 33. Amat, A.M., Arques, A., Bossmann, S.H., Braun, A.M., Gob, S. and Miranda, M.A. (2003). A ‘‘Camel through the eye of a needle’’: direct introduction of the TPP+ ion inside Y-zeolites by formal ion exchange in aqueous medium. Angew Chem. Int. Ed. Engl. 42, 1653–1655 34. Vasenkov, S. and Frei, H. (2000). Time-resolved study of acetyl radical in zeolite NaY by step-scan FT-IR spectroscopy. J. Phys. Chem. A 104, 4327–4332 35. Wilkinson, F. and Kelly, G. (1989). Diffuse reﬂectance ﬂash photolysis. In CRC Handbook of Organic Photochemistry, Scaiano, J.C. (ed.), Vol. 1, pp. 293– 314. CRC Press, Boca Raton, FL 36. Clennan, E.L. and Pan, G.L. (2003). Zeolite-promoted oxidations of 1,1-diarylethylenes. Org. Lett. 5, 4979–4982

264

E.L. CLENNAN

37. Basu, P., Ballinger, T.H. and Yates Jr., J.T. (1988). Wide temperature range IR spectroscopy cell for studies of adsorption and desorption on high area solids. Rev. Sci. Instrum. 59, 1321–1327 38. Xiang, Y., Larsen, S.C. and Grassian, V.H. (1999). Photooxidation of 1-alkenes in zeolites: a study of the factors that inﬂuence product selectivity and formation. J. Am. Chem. Soc. 121, 5063–5072 39. Cano, M.L., Cozens, F.L., Garcı´ a, H., Martı´ , V. and Scaiano, J.C. (1996). Intrazeolite photochemistry. 13. Photophysical properties of bulky 2,4,6-triphenylpyrylium and tritylium cations within large- and extra-large-pore zeolites. J. Phys. Chem. 100, 18152–18157 40. Haw, J.F., Nicholas, J.B., Song, W., Deng, F., Wang, Z., Xu, T. and Heneghan, C.S. (2000). Roles for cyclopentenyl cations in the synthesis of hydrocarbons from methanol on zeolite catalyst HZSM-5. J. Am. Chem. Soc. 122, 4763–4775 41. Haw, J.F., Song, W., Marcus, D.M. and Nicholas, J.B. (2003). The mechanism of methanol to hydrocarbon catalysis. Acc. Chem. Res. 36, 317–326 42. Song, W., Nicholas, J.B. and Haw, J.F. (2001). Acid–base chemistry of a carbenium ion in a zeolite under equilibrium conditions: veriﬁcation of a theoretical explanation of carbenium ion stability. J. Am. Chem. Soc. 123, 121–129 43. Zhang, J., Krawietz, T.R., Skloss, T.W. and Haw, J.F. (1997). 13C MAS NMR observation of photoproducts in zeolites and their further reaction on acid sites. J. Chem. Soc. Chem. Commun. 685–686 44. Clennan, E.L. and Sram, J.P. (2000). Photochemical reactions in the interior of a zeolite. Part 5. The origin of the zeolite induced regioselectivity in the singlet oxygen ene reaction. Tetrahedron 56, 6945–6950 45. Shailaja, J., Sivaguru, J., Robbins, R.J., Ramamurthy, V., Sunoj, R.B. and Chandrasekhar, J. (2000). Singlet oxygen mediated oxidation of oleﬁns within zeolites: selectivity and complexities. Tetrahedron 56, 6927–6943 46. Clennan, E.L., Zhou, W. and Chan, J. (2002). Mechanistic organic chemistry in a microreactor. Zeolite-controlled photooxidations of organic sulﬁdes. J. Org. Chem. 67, 9368–9378 47. Kupfer, R., Poliks, M.D. and Brinker, U.H. (1994). Carbenes in constrained systems. 2. First carbene reactions within zeolites-solid state photolysis of adamantane-2-spiro-30 diazirine. J. Am. Chem. Soc. 116, 7393–7398 48. Wagnerova, D.M. (2001). Classiﬁcation of dioxygen reactions: a unifying view. Z. Phys. Chem. Int. 215 (Pt 1), 133–138 49. Clennan, E.L. (2003). Molecular oxygenations in zeolites. In Photochemistry of Organic Molecules in Isotropic and Anisotropic Media, Ramamurthy, V. and Schanze, K.S. (eds), Vol. 9, pp. 275– 308. Marcel Dekker, New York 50. Li, X. and Ramamurthy, V. (1996). Selective oxidation of oleﬁns within organic dye cation-exchanged zeolites. J. Am. Chem. Soc. 118(43), 10666–10667 51. Foote, C.S. and Clennan, E.L. (1995). Properties and reactions of singlet dioxygen. In Active Oxygen in Chemistry, Foote, C.S., Valentine, J.S., Greenberg, A. and Liebman, J.F. (eds), Vol. 2, pp. 105– 140. Blackie Academic & Professional, New York, NY 52. Galletero, M.S., Garcia, H. and Bourdelande, J.L. (2003). Dramatic persistence (minutes) of the triplet excited state and efﬁcient singlet oxygen generation for C-60 incorporated in Y zeolite and MCM-41 silicate. Chem. Phys. Lett. 370, 829–833 53. Ray, S. and Vasudevan, S. (2003). Encapsulation of cobalt phthalocyanine in zeolite-Y: evidence for nonplanar geometry. Inorg. Chem. 42, 1711–1719 54. Ramamurthy, V., Sanderson, D.R. and Eaton, D.F. (1993). Control of dye assembly within zeolites: role of water. J. Am. Chem. Soc. 115, 10438–10439 55. Jockusch, S., Sivaguru, J., Turro, N.J. and Ramamurthy, V. (2005). Direct measurement of the singlet oxygen lifetime in zeolites by near-IR phosphorescence. Photochem. Photobiol. Sci. 4, 403–405

MECHANISMS OF OXYGENATIONS IN ZEOLITES

265

56. Pettit, T.L. and Fox, M.A. (1986). Photoassisted oxygenation of oleﬁns: an exchanged zeolite as a heterogeneous photosensitizer. J. Phys. Chem. 90, 1353–1354 57. Stratakis, M., Rabalakos, C., Mpourmpakis, G. and Froudakis, G.E. (2003). Ene hydroperoxidation of isobutenylarenes within dye-exchanged zeolite Na-Y: control of site selectivity by cation–arene interactions. J. Org. Chem. 68, 2839–2843 58. Hoppe, R., Schulz-Ekloff, G., Wo¨hrle, D., Kirschhock, C., Fuess, H. (1994). Location and photostability of faujasite-incorporated methylene blue. In Zeolites and Related Microporous Materials: State of the Art 1994, Weitkamp, J., Karge, H.G., Pfeifer, H. and Ho¨lderich, W. (eds), Vol. 84, pp. 821– 827. Elsevier Science BV, Amsterdam 59. Clennan, E.L., Zhang, D. and Singleton, J. (2006). A comparison of intrazeolite and solution singlet oxygen ene reactions of allylic alcohols. Photochem. Photobiol. 82, 1226–1232 60. Li, X. and Ramamurthy, V. (1996). Electron transfer reactions within zeolites: photooxidations of stilbenes. Tetrahedron Lett. 37, 5235–5238 61. Barich, D.H., Xu, T., Zhang, J. and Haw, J.F. (1998). Modeling of benzene adsorption in metal exchanged zeolites by calculation of 7Li chemical shifts. Angew Chem. Int. Ed. Engl. 37, 2530–2531 62. Ma, J.C. and Dougherty, D.A. (1997). The cation–p interaction. Chem. Rev. 97, 1303–1324 63. Robbins, R.J. and Ramamurthy, V. (1997). Generation and reactivity of singlet oxygen within zeolites: remarkable control of hydroperoxidation of alkenes. J. Chem. Soc. Chem. Commun. 1071–1072 64. Ramamurthy, V., Lakshminarasimhan, P., Grey, C.P. and Johnston, L.J. (1998). Energy transfer, proton transfer and electron transfer reactions within zeolites. J. Chem. Soc. Chem. Commun. 2411–2424 65. Clennan, E.L. and Sram, J.P. (1999). Photooxidation in zeolites. Part 2: a new mechanistic model for reaction selectivity in singlet oxygen ene reactions in zeolitic medsia. Tetrahedron Lett. 40, 5275–5278 66. Froudakis, G.E. and Stratakis, M. (2003). A DFT study on the interaction of Li+ and Na+ with alkyl-substituted ethenes. Eur. J. Org. Chem. 359–364 67. Stratakis, M. (2003). Selective photooxygenation of alkenes by zeolite conﬁnement. In 13th European Symposium on Organic Chemistry, Medimond, V. (ed.), pp. 15– 20. Medimond Publishing Co., Bologna 68. Stratakis, M. and Soﬁkiti, N. (2002). Intrazeolite photo-oxygenation of (R)-(–)-alphaphellandrene. J. Chem. Res. S 374–375 69. Stratakis, M. and Rabalakos, C. (2001). Chemoselective hydroperoxidation of alkenylarenes within thionin-supported zeolite Na-Y. Tetrahedron Lett. 42, 4545–4547 70. Clennan, E.L. (2005). Photo-oxygenation of the ene-type. In Synthetic Organic Photochemistry, pp. 365–390. Marcel Dekker, New York 71. Houk, K.N., Williams Jr., J.C., Mitchell, P.A. and Yamaguchi, K. (1981). Conformational control of reactivity and regioselectivity in singlet oxygen ene reactions: Relationship to the rotational barriers of acyclic alkylethylenes. J. Am. Chem. Soc. 103, 949–951 72. Inagaki, S., Fujimoto, H. and Fukui, K. (1976). The selection of O2(1Dg)–oleﬁn reaction courses. Intermolecular nonbonded attraction and p bond polarity of oleﬁns. Chem. Lett. 749–752 73. Stephenson, L.M. (1980). The mechanism of the singlet oxygen ene reaction. Tetra hedron Lett. 21, 1005–1008 74. Clennan, E.L., Sram, J.P., Pace, A., Vincer, K. and White, S. (2002). Intrazeolite photooxidations of electron-poor alkenes. J. Org. Chem. 67, 3975–3978 75. Caldwell, J.W. and Kollman, P.A. (1995). Cation–p interactions: nonadditive effects are critical in their accurate representation. J. Am. Chem. Soc. 117, 4177–4178

266

E.L. CLENNAN

76. Staley, R.H. and Beauchamp, J.L. (1975). Intrinsic acid–base properties of molecules. Binding energies of Li+ to p and n-donor bases. J. Am. Chem. Soc. 97, 5920–5921 77. Eliel, E.L. and Wilen, S.H. (1994). Stereochemistry of Organic Compounds, p. 1267, John Wiley & Sons, New York, NY 78. Kaanumalle, L.S., Shailaja, J., Robbins, R.J. and Ramamurthy, V. (2002). Cation controlled singlet oxygen mediated oxidation of oleﬁns within zeolites. J. Photochem. Photobiol. A: Chem. 153, 55–65 79. Stratakis, M., Nencka, R., Rabalakos, C., Adam, W. and Krebs, O. (2002). Thioninsensitized intrazeolite photooxygenation of trisubstituted alkenes: substituent effects on the regioselectivity as probed through isotopic labeling. J. Org. Chem. 67, 8758–8763 80. Stratakis, M., Soﬁkiti, N., Baskakis, C. and Raptis, C. (2004). Dye-sensitized intrazeolite photooxygenation of 4-substituted cyclohexenes. Remote substituent effects in regioselectivity and diastereoselectivity. Tetrahedron Lett. 45, 5433–5436 81. (b) Stratakis, M. and Kosmas, G. (2001). Enhanced diastereoselectivity of an ene hydroperoxidation reaction by conﬁnement within zeolite Na-Y; a stereoisotopic study. Tetrahedron Lett. 42, 6007–6009 82. Joy, A., Robbins, R.J., Pitchumani, K. and Ramamurthy, V. (1997). Asymmetrically modiﬁed zeolite as a medium for enantioselective photoreactions: reactions from sping forbidden excited states. Tetrahedron Lett. 38, 8825–8828 83. Ramamurthy, V., Shailaja, J., Kaanumalle, L.S., Sunoj, R.B. and Chandrasekhar, J. (2003). Controlling chemistry with cations: photochemistry within zeolites. Chem. Commun. 1987–1999 84. For an exception seeChen, Y.Z., Wu, L.Z., Zhang, L.P. and Tung, C.H. (2005). Conﬁned space-controlled hydroperoxidation of trisubstituted alkenes adsorbed on pentasil zeolites. J. Org. Chem. 70, 4676–4681 85. For an exception seeFeng, K., Zhang, R.Y., Wu, L.Z., Tu, B., Peng, M.L., Zhang, L.P., Zhao, D. and Tung, C.H. (2006). Photooxidation of oleﬁns under oxygen in platinum(II) complex-loaded mesoporous molecular sieves. J. Am. Chem. Soc. 128, 14685–14690 86. Tung, C.-H., Wang, H. and Ying, Y.-M. (1998). Photosensitized oxidation of alkenes adsorbed on pentasil zeolites. J. Am. Chem. Soc. 120, 5179–5186 87. Araki, Y., Dobrowolski, D.C., Goyne, T.E., Hanson, D.C., Jiang, Z.Q., Lee, K.J. and Foote, C.S. (1984). Chemistry of singlet oxygen. 47. 9,10-Dicyanoanthracene-sensitized photooxygenation of alkyl-substituted oleﬁns. J. Am. Chem. Soc. 106, 4570–4575 88. Schaap, A.P., Siddiqui, S., Prasad, G., Palomino, E. and Sandison, M. (1985). The photochemical preparation of ozonides by electron-transfer photo-oxygenation of epoxides. Tetrahedron 41, 2229–2235 89. Iliev, V., Ileva, A. and Bilyarska, L. (1997). Oxidation and photooxidation of sulfurcontaining compounds in the resence of immobilized phthalocyanine complexes. J. Mol. Catal. A: Chem. 126, 99–108 90. Zhou, W. and Clennan, E.L. (1999). Organic reactions in zeolites. 1. Photooxidations of sulﬁdes in methylene blue doped zeolite Y. J. Am. Chem. Soc. 121, 2915–2916 91. Toutchkine, A., Aebisher, D. and Clennan, E.L. (2001). Substituent dictated partitioning of intermediates on the sulﬁde singlet oxygen reaction surface. A new mechanism for oxidative C–S bond cleavage in a-hydroperoxysulﬁdes. J. Am. Chem. Soc. 123, 4966–4973 92. Toutchkine, A. and Clennan, E.L. (2000). A new mechanism for oxidative C–S bond cleavage during reactions of singlet oxygen with organic sulﬁdes: electronically dictated reaction selectivity in the persulfoxide intermediate. J. Am. Chem. Soc. 122, 1834–1835 93. Clennan, E.L. (2001). Persulfoxide: key intermediate in reactions of singlet oxygen with sulﬁdes. Acc. Chem. Res. 34(11), 875–884 94. Clennan, E.L. and Pace, A. (2005). Advances in singlet oxygen chemistry. Tetrahedron 61, 6665–6691

MECHANISMS OF OXYGENATIONS IN ZEOLITES

267

95. Jensen, F., Greer, A. and Clennan, E.L. (1998). Reaction of organic sulﬁdes with singlet oxygen. A revised mechanism. J. Am. Chem. Soc. 120, 4439–4449 96. Zhou, W. and Clennan, E.L. (2000). Relative reactivities of tethered functional groups in the interior of a zeolite. Org. Lett. 2, 437–440 97. Clennan, E.L. and Pan, G.-L. (2003). Isotope effects as mechanistic probes in solution and in intrazeolite photooxygenations. The formation of a hydroperoxysulfonium ylide. J. Org. Chem. 68, 5174–5179 98. Blatter, F. and Frei, H. (1993). Very strong stabilization of alkene-O2 charge transfer states in zeolite NaY: red-light-induced photooxidation of 2,3-dimethyl-2-butene. J. Am. Chem. Soc. 115, 7501–7502 99. It is not known if A is a stationary point on the ground or excited state potential energy surface. Consequently, we enclose the symbol for its excited state conﬁguration in parenthesis (*). Whether an excited state species directly partitions to products (i.e. the dioxetane and the radical pair) or ﬁrst undergoes radiationless conversion to a superoxide/alkene radical cation pair on the ground state potential energy surface is irrelevant to the ensuing discussion 100. Blatter, F., Sun, H., Vasenkov, S. and Frei, H. (1998). Photocatalyzed oxidation in zeolite cages. Catal. Today 41, 297–309 101. Frei, H. (1997). Highly selective photochemical and dark oxidation of hydrocarbons by O2 in zeolites. In Studies in Surface Science and Catalysis. 3rd World Congress on Oxidation Catalysis, Grasselli, R.K., Oyama, S.T., Gaffney, A.M. and Lyons, J.E. (eds), Vol. 110, pp. 1041– 1050. Elsevier Science, New York 102. Evans, D.F. (1953). Molecular association of oxygen and aromatic substances. J. Chem. Soc. 345–347 103. Onodera, K., Furusawa, C.-I., Kojima, M., Tsuchiya, M., Aihara, S., Akaba, R., Sakuragi, H. and Tokumaru, K. (1985). Mechanistic considerations on photoreaction of organic compounds via excitation of contact charge transfer complexes with oxygen. Tetrahedron 41, 2215–2220 104. Tsubomura, H. and Mulliken, R.S. (1960). Molecular complexes and their spectra. XII. Ultraviolet absorption spectra caused by the interaction of oxygen with organic molecules. J. Am. Chem. Soc. 82, 5966–5974 105. Hashimoto, S. and Akimoto, H. (1989). UV absorption spectra and photochemical reactions of simple aromatic hydrocarbons in the cryogenic oxygen matrix. J. Phys. Chem. 93, 571–577 106. Rest, A.J., Salisbury, K. and Sodeau, J.R. (1977). Contact charge transfer interactions between aromatic molecules and oxygen in gas matrixes at 10 K. J. Chem. Soc. Faraday Trans. 2: Mol. Chem. Phys. 73, 265–273 107. Itoh, M. and Mulliken, R.S. (1969). Singlet-triplet absorption bands of methyl-substituted ethylenes. J. Phys. Chem. 73, 4332–4334 108. Hashimoto, S. and Akimoto, H. (1986). Visible light photooxygenation of 2,3-dimethyl2-butene at the contact charge-transfer band in the cryogenic oxygen matrix. J. Phys. Chem. 90, 529–532 109. Hashimoto, S. and Akimoto, H. (1987). Absorption spectra of contact-charge-transfer bands and photochemical reactions of simple alkenes in the cryogenic oxygen matrix. J. Phys. Chem. 91, 1347–1354 110. Blatter, F. and Frei, H. (1994). Selective photooxidation of small alkenes by O2 with red light in zeolite Y. J. Am. Chem. Soc. 116, 1812–1820 111. Blatter, F., Moreau, F. and Frei, H. (1994). Diffuse reﬂectance spectroscopy of visible alkene-O2 charge transfer absorptions in zeolite Y and determination of photooxygenation quantum efﬁciencies. J. Phys. Chem. 98(50), 13403–13407 112. Sun, H., Blatter, F. and Frei, H. (1994). Selective oxidation of toluene to benzaldehyde by O2 with visible light in barium(2+)- and calcium(2+)-exchange zeolite Y. J. Am. Chem. Soc. 116, 7951–7952

268

E.L. CLENNAN

113. Blatter, F., Sun, H. and Frei, H. (1996). Highly selective formation of tert-butyl hydroperoxide from the reaction of isobutane and O2 in a zeolite under visible light. Chem. Eur. J. 2, 385–389 114. Sun, H., Blatter, F. and Frei, H. (1996). Cyclohexanone from cyclohexane and O2 in a zeolite under visible light with complete selectivity. J. Am. Chem. Soc. 118, 6873–6879 115. Clennan, E.L. (2004). Viologen embedded zeolites. Coord. Chem. Rev. 248, 477–492 116. Yoon, K.B. (1993). Electron- and charge-transfer reactions within zeolites. Chem. Rev. 93, 321–339 117. Reichardt, C. (1992). Solvatochromism, thermochromism, piezochromism, halochromism, and chiro-solvatochromism of pyridinium N-phenoxide betaine dyes. Chem. Soc. Rev. 21, 147–153 118. Panov, A.G., Larsen, R.G., Totah, N.I., Larsen, S.C. and Grassian, V.H. (2000). Photooxidation of toluene and p-xylene in cation-exchanged zeolites X, Y, ZSM-5, and beta: the role of zeolite physicochemical properties in product yield and selectivity. J. Phys. Chem. B 104, 5706–5714 119. Clennan, E.L. (2000). New mechanistic and synthetic aspects of singlet oxygen chemistry. Tetrahedron 56, 9151–9179 120. Vasenkov, S. and Frei, H. (1998). Time-resolved FT-infrared spectroscopy of visible light-induced alkene oxidation by O2 in a zeolite. J. Phys. Chem. B 102, 8177–8182 121. Pidko, E.A. and van Santen, R.A. (2006). Conﬁned space-controlled oleﬁn-oxygen charge transfer in zeolites. J. Phys. Chem. B 110, 2963–2967 122. Vasenkov, S. and Frei, H. (1997). UV–visible absorption spectroscopy and photochemistry of an alkene-O2 contact charge-transfer system in large NaY crystals. J. Phys. Chem. B 101, 4539–4543 123. Frei, H. (2006). Selective hydrocarbon oxidation in zeolites. Science 313, 309–310 124. Frei, H., Blatter, F. and Sun, H. (1996). Photocatalyzed oxidation of hydrocarbons in zeolite cages. Chemtech 26, 24–30 125. Blatter, F., Sun, H. and Frei, H. (1995). Selective oxidation of propylene by O2 with visible light in a zeolite. Catal. Lett. 35, 1–12 126. Myli, K.B., Larsen, S.C. and Grassian, V.H. (1997). Selective photooxidation reactions in zeolites X, Y and ZSM-5. Catal. Lett. 48, 199–202 127. Tang, S.L.Y., McGarvey, D.J. and Zholobenko, V.L. (2003). Photooxidation and dark thermal oxidation of 1-butene on cationic forms of zeolite Y: a spectroscopic study. Phys. Chem. Chem. Phys. 5, 2699–2705 128. (i) For examples of Hock Cleavages see the following referencesChan, Y.-Y., Li, X., Zhu, C., Liu, X., Zhang, Y. and Leung, H.-K. (1990). Sensitized photooxygenation. 3. Mechanistic studies on the singlet oxygenation of 5,6-disubstituted 3,4-dihydro-2Hpyrans. J. Org. Chem. 55, 5497–5504; (ii) Frimer, A.A., Weiss, J., Gottlieb, H.E. and Wolk, J.L. (1994). Preparation and photosensitized oxidation of isopropylidenecyclobutanes and cyclobutenes. J. Org. Chem. 59, 780–792; (iii) Sharon, O. and Frimer, A.A. (2003). Synthesis and photosensitized oxygenation of cyclopropylidenecyclobutenes. Tetrahedron 59, 8153–8162 129. Foster, R. (1969). Organic Charge-Transfer Complexes, Vol. 15, p. 470, Academic Press, New York 130. Frimer, A.A. (1979). The reactions of singlet oxygen with oleﬁns: the question of mechanism. Chem. Rev. 79, 359–387 131. Frimer, A.A. and Stephenson, L.M. (1985). The singlet oxygen ene reaction. In Singlet O2. Reaction Modes and Products, Frimer, A.A. (ed.), Vol. 2, pp. 67– 91. CRC Press, Boca Raton, FL 132. Sun, H., Blatter, F. and Frei, H. (1997). Oxidation of propane to acetone and of ethane to acetaldehyde by O2 in zeolites with complete selectivity. Catal. Lett. 44, 247–253

MECHANISMS OF OXYGENATIONS IN ZEOLITES

269

133. Larsen, R.G., Saladino, A.C., Hunt, T.A., Mann, J.E., Xu, M., Grassian, V.H. and Larsen, S.C. (2001). A kinetic study of the thermal and photochemical partial oxidation of cyclohexane with molecular oxygen in zeolite Y. J. Catal. 204, 440–449 134. Vanoppen, D.L., De Vos, D.E., Jacobs, P.A. (1997). Cation effects in the oxidation of adsorbed cyclohexane in Y zeolite: an in situ IR study. In Progress in Zeolite and Microporous Materials. Studies in Surface Science and Catalysis, Chon, H., Ihm, S.-K. and Uh, Y.S. (eds), Vol. 105, pp. 1045– 1051. Elsevier Science BV, Amsterdam 135. Li, G., Xu, M., Larsen, S.C. and Grassian, V.H. (2003). Photooxidation of cyclohexane and cyclohexene in BaY. J. Mol. Catal. A: Chem. 194, 169–180 136. Vanoppen, D.L., De Vos, D.E. and Jacobs, P.A. (1998). Ion-exchanged Y zeolites as catalysts for the liquid phase autooxidation of cyclohexane. J. Catal. 177, 22–28 137. Kojima, M., Nakajoh, M., Matsubara, C. and Hashimoto, S. (2002). Photooxygenation of aromatic alkenes in zeolite nanocavities. J. Chem. Soc. Perkin Trans. 2, 1894–1901 138. Matsubara, C. and Kojima, M. (2001). Photo-oxygenation and photodimerization of 4-methoxystyrene in NaY. The role of co-adsorbed water and ankene-oxygen chargetransfer complex. Res. Chem. Int. 27, 975–989 139. Takeya, H., Kuriyama, Y. and Kojima, M. (1998). Photooxygenation of stilbenes in zeolite by excitation of their contact charge transfer complexes with oxygen. Tetrahedron Lett. 39, 5967–5970 140. Lakshminarasimhan, P., Thomas, K.J., Johnston, L.J. and Ramamurthy, V. (2000). Wavelength dependent oxygen mediated electron-transfer reactions within M+Y zeolites: photooxidation and reduction of 1,1-diarylethylenes. Langmuir 16, 9360–9367 141. Wan, Z. and Jenks, W.S. (1995). Oxenoid reactivity observed on the photolysis of certain aromatic sulfoxides. J. Am. Chem. Soc. 117, 2667–2668 142. Gilbert, A. and Baggott, J.E. (1991). Essentials of Molecular Photochemistry, p. 538, CRC Press, Boca Raton 143. Khenkin, A.M. and Neumann, R. (2000). Aerobic photochemical oxidation in mesoporous Ti-MCM-41: epoxidation of alkenes and oxidation of sulﬁdes. Catal. Lett. 68(1–2), 109–111 144. Arends, I.W.C.E., Sheldon, R.A., Wallau, M. and Schuchardt, U. (1997). Oxidative transformation of organic compounds mediated by redox molecular sieves. Angew Chem. Int. Ed. Engl. 36, 1145–1163 145. Sanjuan, A., Alvaro, M., Corma, A. and Garcia, H. (1999). An organic sensitizer within Ti-zeolites as photocatalyst for the selective oxidation of oleﬁns using oxygen and water as reagents. Chem. Commun. 1641–1642 146. Rehm, D. and Weller, A. (1969). Ber. Bunsenges. Phys. Chem. 73, 834–839 147. Sanjua´n, A., Alvaro, M., Aguirre, G., Garcı´ a, H. and Scaiano, J.C. (1998). Intrazeolite photochemistry. 21. 2,4,6-Triphenylpyrylium encapsulated inside zeolite Y supercages as heterogeneous photocatalyst for the generation of hydroxyl radical. J. Am. Chem. Soc. 120, 7351–7352

This page intentionally left blank

Metal-catalyzed alcoholysis reactions of carboxylate and organophosphorus esters R. STAN BROWN and ALEXEI A. NEVEROV Department of Chemistry, Queen’s University, Kingston, Ontario K7L 3N6, Canada 1 2 3 4 5

Introduction 271 Background theory 274 Titrations in alcohol 276 Metal ion alcoholysis and titration in alcohol 278 Transition metal ion and Ln3+ catalysts of transesteriﬁcations of neutral carboxylate and organophosphate esters 284 Mechanism of alcoholysis of carboxylate esters 288 Mechanism of alcoholysis of neutral phosphate esters 294 6 Transition metal ion and La3+-catalysis of the alcoholysis of phosphate diesters 308 Metal-catalyzed alcoholysis of an RNA model 310 Zn2+ ligand models for dinuclear enzymes promoting the cleavage of RNA 316 Exhalted catalysis of methanolysis of HPNPP promoted by a dinuclear complex in methanol 318 7 Conclusions 324 Acknowledgements 325 References 325

1

Introduction

Metal ion-catalyzed hydrolytic reactions of esters, amides1 and organophosphorus esters2 have been studied for many years. In general these are relatively well-understood processes where the reactions involve catalytically active Mx+(OH) species with the metal ion serving several possible purposes. These include: (1) decreasing the pKa of the metal associated HOH so that an Mx+-coordinated hydroxide can be formed at near neutral pH; (2) a bifunctional role where the Mx+(OH) acts as a Lewis acid coordinating transiently to the CQO or PQO group, and then delivering the coordinated hydroxide and (3) possibly promoting accelerated breakdown of any intermediates through coordination of the leaving group, particularly a poor one, to the metal ion. While water is required as a reagent for the hydrolytic processes, in bulk media it solvates the ionic species very heavily so as to impede the interaction of the catalytic species with hydrophobic substrates such as neutral esters. In addition, the high dielectric constant dampens the attractive interaction between species of opposite charge. Not surprisingly, Mx+(OH) species in water are relatively poor catalysts for acyl or phosphoryl transfer from neutral substrates. It seems reasonable that moving to less polar solvents with lower dielectric constants might ameliorate these problems and allow a more full realization of metal 271 ADVANCES IN PHYSICAL ORGANIC CHEMISTRY VOLUME 42 ISSN 0065-3160 DOI: 10.1016/S0065-3160(07)42006-8

r 2008 Elsevier Inc. All rights reserved

272

R.S. BROWN AND A.A. NEVEROV

ion’s ability to catalyze acyl and phosphoryl transfer reactions, although there seems to be little known about the mechanisms of these processes. Among organic solvents the lower alcohols such as methanol and ethanol, while being close to water in terms of chemical properties, solvation effects and hydrogen-bonding characteristics, should offer opportunities to investigate this thesis. Transition metal or lanthanide metal ion-catalyzed alcoholysis reactions, while being important industrial processes for transesteriﬁcation3 of carboxylate esters, are less well understood than the analogous hydrolysis reactions and have not been applied to phosphoryl transfer processes except in our labs. Buncel and co-workers4 and Mandolini and co-workers3f–i have extensively investigated the ethanolysis of phosphinate and phosphate esters promoted by alkali metal ethoxides and alkaline earth ethoxides (Ba2+) respectively, but these reactions occur stoichiometrically under highly basic conditions whereas the transition and lanthanide M2+- and M3+-catalyzed alcoholysis reactions we discuss below occur under essentially neutral pH conditions. One might expect that Mx+-promoted alcoholysis reactions will be subject to the same mechanistic generalities as Mx+-promoted hydrolysis reactions. The bulk of our studies show that this is the case although there are important mechanistic and practical differences which in certain ways simplify mechanistic study and circumvent at least two of the problems associated with metal hydroxides. First, in water at pH values above the pKa of the Mx+-coordinated H2O, the active Mx+-hydroxides form oligomeric and polymeric gels or precipitates that complicate the mechanistic analyses. In some cases this can be overcome by complexing the metal ion to ligands that block the oligomerization and prevent the precipitation of the Mx+-hydroxides. However, the latter approach is known to introduce other biases because the ligands employed are generally those which are readily available or easily synthesized and not necessarily the ideal ones for controlling geometry and reactivity of the complexes. Second, Mx+-hydroxides in water are not particularly effective intermolecular catalysts and so the most effective systems are ones where the metal ion is held, through transient or ﬁxed association with some proximal binding group, close to the scissile CQO or PQO linkage. As will be seen later, one of the major stumbling blocks in creating highly active Mx+(OH) catalysts is the fact that a metal-coordinated hydroxo group is less basic and nucleophilic than free hydroxide, a problem that is difﬁcult to overcome in water but one that seems to be readily overcome in alcohols where the active species involve one or more Mx+(OR) units. Moving to alcohol solvents substantially solves some of the above problems but requires due consideration of the control and measurement of pH in neat alcohol.5 Bosch and co-workers6 have described simpliﬁed methods for determining the ss pH in methanol solution which have greatly facilitated our work. Recently, we have For the designation of pH in non-aqueous solvents, we use the forms described by Bosch and coworkers6 based on the recommendations of the IUPAC, In Compendium of Analytical Nomenclature. Deﬁnitive Rules 1997, 3rd edn, Blackwell, Oxford, UK, 1998. If one calibrates the measuring electrode with aqueous buffers and then measures the pH of an aqueous buffer solution, the term w w pH is used; if the electrode is calibrated in water and the ‘pH’ of the neat buffered methanol solution then measured, the term sw pH is used, and if the electrode is calibrated in the same solvent and the ‘pH’ reading is made, then the term ss pH is used.

METAL-CATALYZED ALCOHOLYSIS

273

shown that one can titrate M2+- and M3+-ions in methanol7 and in ethanol8 with the end result being that now one can perform detailed mechanistic studies of reactions as a function of ss pH in the presence of metal ions in these alcohols under buffered conditions as simply as one can in water. Importantly most of the M2+ and M3+ systems are completely soluble in alcohols throughout the ss pH region where formation of the metal ion alkoxides occurs, obviating the solubility problems encountered in water and in some cases making it unnecessary to employ ligands for successful study. The fact that the autoprotolysis constants for methanol and ethanol are 1016.77 and 1019.1 means that neutral ss pH in the two solvents is 8.46 and 9.558 while that in water is, of course, 7.00. These alcohols have additional beneﬁts related to their low dielectric constants and polarity relative to that of water (31.5 (CH3OH), 24.3 (EtOH) vs. 78 (H2O))9 which favor enhanced organic substrate solubility, greater ion pairing between the catalytic metal ion and any anionic or dipolar substrates and greater Lewis acid interaction between neutral substrates and metal ions. This not only enhances binding, but also the rates of the subsequent acyl and phosphoryl transfer reactions. Interestingly, in the case of La3+ catalytically active dimers, such as La3+ 2 ( OR)1,2,3 y, are formed spontaneously in methanol or ethanol without the need for any binding ligands. As will be shown, many acyl and phosphoryl transfer reactions to alcohol can be greatly accelerated by Mx+(OR) species in contrast to the hydrolytic reactions in water where the Mx+(OH) species often have little or no activity. Much of our earlier work on the Mx+-catalysis of acyl or phosphoryl transfer to methanol has been summarized10 and so will only be mentioned here when necessary for a brief background or when relevant to the more recent work. In what follows we will ﬁrst consider some general background theoretical aspects of the inﬂuence of reduced polarity and dielectric constant media on metal ion-catalyzed reactions. Subsequently, a brief introduction is given for titration in alcohol to determine the x+ s (HOR) systems as well as the speciation of s pK a values for ionization of various M the metal ions, and when in the presence of various ligands, the formation constants for the ligand:Mx+ and ligand:Mx+(OR) species in solution. Subsequently we deal with the effectiveness of some Mx+-ions, notably La3+, Eu3+, and some Zn2+ and Cu2+:ligand systems in promoting the alcoholysis of carboxylate esters and various neutral phosphate, phosphonate, phosphonothioate and phosphorothioate esters. These are cases where the substrates, being neutral but dipolar, are considered to be weakly binding to the metal ion. Thus, the kinetics do not exhibit saturation binding although chemical intuition states that complexes must be formed along the reaction pathway. Almost all these reactions exhibit spectacular rate accelerations when promoted by the Mx+(OR) catalysts in the neutral ss pH regions relative to the background reactions which are promoted by lyoxide. Where we have studied the effect of varying the leaving group on the rates of transesteriﬁcation, we now have some fairly detailed mechanistic information. Subsequently, we will deal with some anionic phosphate diester substrates that might be taken as models for RNA and DNA. These bind very tightly to the metal ion and then are subject to rapid metal-catalyzed decomposition at rates that exceed anything so far reported for the analogous reactions in water.

274

2

R.S. BROWN AND A.A. NEVEROV

Background theory

Aside from being fundamentally interesting and industrially important, phosphoryl and acyl transfer reactions are key biological processes. Numerous enzymes containing one or more metal ions (notably Zn2+, but in some cases Mg2+, Ca2+ and Fe2+) catalyze ester and peptide hydrolysis as well as phosphoryl transfer processes.11 It has been stated in the context of short, strong hydrogen bonds, that the ‘active sites of enzymes are non-aqueous, and the effective dielectric constants resemble those in organic solvents rather than that in water’,12a,b and it is perhaps surprising how sparse is the experimental literature concerning the effect of reduced dielectric constant of a solvent on simple reactions of biological relevance. An early example demonstrated that the decarboxylation of pyruvate promoted by 3,4-dimethylthiazolium ion (a model for thiamine pyrophosphate-catalyzed reactions in an enzyme’s interior) occurs 104–105 faster in ethanol than in water.13 From this Lienhard concluded that a large part of the thiamin-dependent enzymes’ catalysis may occur from the hydrophobicity of the active site promoting reactions where charge neutralization of the zwitterionic intermediate is neutralized in the transition state. It is an intriguing possibility that metal ion-catalyzed reactions of the sort we discuss herein, where charge is dissipated in the transition states, might be subject to large rate accelerations in solvents of reduced polarity and dielectric constant. Solvent effects can have important accelerating or decelerating effects on organic transformations depending on the overall solvation of the ground state starting materials and transition states for the rate-limiting steps of the reaction,14 but it is interesting to us how little application of reduced polarity medium effects there is for reactions of biological interest including metal ion-catalyzed processes. The generally accepted process for metal ion-catalyzed reactions of the sort we consider here involves pre-equilibrium binding with the substrate, followed by a reaction of the complex as schematized in Equation (1). Whether the metal ion is free or complexed by ligands, or bears an associated lyate, or whether the substrate is neutral or anionic, these appear to be just the sort of processes one might expect to experience large rate accelerations in passing from water to a medium of reduced dielectric constant such as alcohols or other lower polarity solvents. Lig: MX+ + Subst.

Kb

Lig: MX+:Subst.

kcat

Lig:MX+ + P

(1)

The Debye-Hu¨ckel theory for association of spherical ions in a medium of dielectric constant Dr posits that the electrostatic potential energy of interaction between oppositely charged ions is PE ¼ ðzþ eÞðz eÞ=ð4pD0 Dr rÞ

(2)

where r is the distance between the centers of the ions, z+e and ze are their charges in coulombs (e is the proton charge), D0 the permittivity of a vacuum, and Dr is the relative permittivity of dielectric constant of the medium.15 When the electrostatic attraction energy is greater than the thermal kinetic energy of the species in solution (given as its average translational kinetic energy of 3/2kT, where k is the Boltzmann

METAL-CATALYZED ALCOHOLYSIS

275

Table 1 Changes in hypothetical potential energies of attraction of oppositely charged spherical ions based on reductions in dielectric constant in passing from water to methanol and ethanol, and the computed changes in binding constant assuming the changes in potential energy are translated into free energies of bindinga Electrostatic PE (H2O, kJ mol1)

Electrostatic PE (CH3OH, kJ mol1)

Increase in Kb H2O methanol

Electrostatic PE (C2H5OH, kJ mol1)

Increase in Kb H2O - ethanol

10.45 20.9 31.35 41.8 52.25

12.5 158 2 103 2.5 104 3.2 105

13.38 26.75 41.13 53.5 66.9

41 1.7 103 6.9 104 2.8 106 1.24 108

4.18 8.36 12.54 16.72 20.9 a

Kb computations made under assumption that all the P.E. of attraction is reﬂected in the DG of binding.

constant and T is the Kelvin constant) then the encounter complex lasts long enough for multiple collisions and achieves the status of an ‘ion pair’. A change from water (Dr ¼ 78) to methanol or ethanol (Dr ¼ 31.5, 24.39) increases the potential energy of the attraction for oppositely charged ions by a factor of 2.5 and 3.2, respectively. Ignoring speciﬁc changes in solvent effects, the data in Table 1 show the dramatic effect on the calculated binding constant according to log Kb ¼ (DGH2O+DGROH)/ 2.303RT for a hypothetical process of Mx++Ay Mx+:Ay where the elec1 trostatic potential energy of binding is 4.18–20.9 kJ mol in water, all of which is expressed in the free energy in passing to methanol and ethanol ignoring speciﬁc changes in solvent effects. While increasing the extent of binding of the reactants is one way to enhance the rate for a process such as that given in Equation (1), it is absolutely required that the association of the metal ion to the substrate must lower the activation energy for the acyl or phosphoryl transfer reaction, otherwise no increase in rate will be observed. Our earlier10 and more recent studies indicated that metal ion catalysis of the methanolyses of neutral carboxylate esters and activated amides16 and neutral phosphate, phosphorothioate, phosphonate and phosphonothioate esters17 were profoundly accelerated in methanol relative to water, the main accelerating effects of the solvent change being proposed as increased pre-association Mx+/CQO or Mx+/PQO binding, and a changed activity of the metal bound methoxide which is manifested in the actual acyl/phosphoryl transfer. Amis proposed Equation (3)18 for the reaction rate constant involving a limiting case of head-on approach of an ion to a neutral dipolar molecule from electrostatic considerations, as would be the case for metal ion catalysts and neutral CQO or PQO substrates. This expression relates the natural log of the rate constant in a medium of dielectric constant (D) to the charge on the ion (ZA), the dipole moment ln k0D¼D ¼ ln k0D¼1 þ

Z2 m DkTr2

(3)

of the molecule (m), the separation of the ion and head of the dipole (r), Boltzmann’s constant (k) and absolute temperature (T). It indicates that the rate constant of the

276

R.S. BROWN AND A.A. NEVEROV

reaction for an oppositely charged metal ion and negative dipole head will increase with a reduction in the dielectric constant of the medium. A more general, but complex, treatment was developed by Landskroener and Laidler19 describing the effect of dielectric constant on the velocities for ion–dipole, dipole–dipole and ion–ion reactions. The general approaches used for these treatments are highly simpliﬁed since they are really pertinent only to cases where electrostatic interactions on reactions are more important than speciﬁc solvation and other effects.18,19 Admittedly the catalytic processes generated by the association of the metal ions and neutral or anionic CQO or PQO ester substrates is probably more complicated than can be analyzed quantitatively through application of the simple form of Equation (3), but the general idea that a reduced dielectric constant does lead to a greater interaction energy for bringing the reactants together is well rooted in theory. Up to a certain point, larger Kb binding constants could be a prelude to enhanced rates of the subsequent chemical reactions since the ion-paired intermediate complexes now exist long enough for multiple intra-complex collisions from which a productive chemical reaction could occur. This will be particularly so for a metal-catalyzed reaction where there is charge neutralization of the substrate (as in the reaction of a negatively charged phosphate anion) in the rate-limiting transition state for the reaction. Of course, the attraction energy cannot be so great as to prevent separation of the catalyst and reaction product after the reaction has occurred, else wise true turnover cannot occur.

3

Titrations in alcohol

To conduct meaningful mechanistic and kinetic studies in alcohol media reliable and simple measurement and control of the solution ss pH is essential. Potentiometric titration is the method of choice for obtaining acid dissociation constants or metal ion complex stability constants and in favorable cases the speciation of mixtures of metal-ion-containing complexes in solution can be proposed.20 Titrations in nonaqueous solvents are not nearly as widely reported as those in aqueous media, particularly in cases with metal ions21 and determination of ‘pH’ in a non-aqueous solvent referenced to that solvent is complicated due to the lack of a way to relate the electrode EMF readings to absolute ss pH (see footnote * and ref. 6) so nonaqueous solvents are generally inconvenient to use22 for detailed studies of reaction mechanisms where pH control is required. The measurement of and determination of the ss pK a values for ionization of mono and dibasic acids in methanol6,7 and ethanol8 is now relatively straightforward. deLigny and Rehbach23 empirically determined a method for measuring the ss pH in methanol where one subtracts a correction constant of –2.34 (on the molality scale) from a measured electrode reading. Bosch and coworkers6 subsequently reported a method for determining ss pH on the molarity scale which, for our purposes, is relatively simple: if a glass electrode is calibrated using standard aqueous buffers but the potentiometric measurements are made in a non-aqueous solvent, the values are termed sw pH. Subsequently one computes ss pH ¼ sw pH – d, where d is a correction

METAL-CATALYZED ALCOHOLYSIS

277

factor of –2.24 on the molarity scale for measurements made as above in methanol.6 The measurement in ethanol is done analogously, although the correction factor and autoprotolysis constant are different from, and more uncertain than, those of methanol. Bates et al.24 and Grunwald et al.25 independently determined the d-correction constant which is related to the electrode junction potential between the two solvents (E¯ j ) and the primary medium effect for the solvent ðm gH Þ by the equation d ¼ E¯ j m gH .26a The practical measurement of ss pH in ethanol is accomplished by subtracting this constant (d ¼ 2.91 or 2.3626b on the molality scale for Bates24 and Grunwald25b) from the measured electrode reading (ss pH ¼ sw pH d) provided that d is determined under ‘ideal conditions’24 of low ionic strength and an intermediate value of ss pH. The Bates and Grunwald d-correction factors differ substantially but there appear to be no criteria to determine which is better so we have chosen for our work to use the mean of these (d ¼ 2.54).8 Details about standardization of electrodes and how to conduct the titrations in methanol and ethanol are published6,7,8 and need not be repeated here. Most of our analyses of the titration data (d[titrant]/dss pH) were conducted using the commercially available ﬁtting program Hyperquad 2000 NT27 setting the respective autoprotolysis constants for methanol and ethanol at 1016.77 and 1019.1 at 25 1C as described.7,8 Data given in Table 2 are some ss pK a values of some monobasic acids and aminium ions that are useful as buffering agents or ligands which show general trends in the ionization constants in passing from water to methanol and then ethanol. The ss pK a s for carboxylic and phenolic (neutral) acids are signiﬁcantly higher in ethanol than in water or even methanol, consistent with what is expected for the decreased dielectric constant. The ss pK a values for the cationic aminium ions do not vary greatly in passing from water to ethanol and, depending on the structure of the base, the acidity in ethanol can be higher or lower than in water. Since neutral s s pH in EtOH is 9.55 while in methanol and in water it is 8.39 and 7.0, a species having the same numerical ss pK a is considerably more acidic in ethanol. The trends in ss pK a in passing from water to the less polar alcohols can be explained in part by considering the equilibria for carboxylic acid and aminium ion dissociation shown in Equations (4) and (5). Carboxylic acid dissociation creates two opposite charges, while aminium ion dissociation simply relocates (H+) from the amine to the solvent as ROH+ 2 , which may be R1C(=O)OH + HOR

R1C(O)O− + +H2OR

(4)

R2R3R4NH+ + HOR

R2R3R4N: + +H2OR

(5)

differently solvated in a given medium than the aminium ion depending on the various alkyl groups. Speciﬁc solvation that stabilizes charge separated ions is expected to be poorer in less polar solvents than in water, thus accounting for the increased ss pK a s of carboxylic acids and the relative indifference in the ss pK a s of aminium ions. The solvation effects in play comprise more than the dielectric constant and include speciﬁc interactions like hydrogen bonding, lone-pair ion

278

R.S. BROWN AND A.A. NEVEROV

Table 2 ss pK a values of various monobasic acids and aminium ions as determined by Hyperquad analysis of titration data at 1 103 M in anhydrous ethanol, T ¼ 25.0 1C Acid Acetic acid Benzoic acid 2,4-Dinitrobenzoic acid 3,5-Dinitrobenzoic acid 4-Nitrophenol 2,6-Lutidine 1-Methylimidazole 4-Ethylmorpholine Triethylamine 1,10-Phenanthroline 2,9-Dimethyl-1,10-phenanthroline 1,5,9-Triazacyclododecane

pKa (water) 4.76b 4.21b 1.42c 2.82c 7.15d 6.72f 6.95d 7.76f 11.01d 4.86h,i 6.17h,i 7.49k,i 12.60l

s s pK a

(methanol)a

9.7770.10 9.1270.02 6.4870.03 7.1470.02 11.30e 6.86e 7.60g 8.28 10.78e 5.63j,i 6.43j,i 14.92j,l

s s pK a

(ethanol)a

10.5870.08 10.4070.08 7.2970.05 8.0170.05 11.7970.17 6.3770.06 7.5070.14 7.9870.04 10.2270.05 5.3670.23i 6.1570.01i 6.3270.06i 12.5470.05l

Source: Data from ref. 8. a Errors given are standard deviations of the mean of 3-5 determinations. b Isaacs, N. (1995). Physical Organic Chemistry (2nd edn), pp. 238–239. Pearson Education Ltd., Harlow, Great Britain. c Gluck, S.J., Steele, K.P. and Benko, M.H. (1996). J. Chromatog. A 745, 117. d Handbook of Chemistry and Physics, 48th edn. (1967–68). CRC Press, Cleveland, Ohio, pp. D88–89. e Rived, F., Rose´s, M. and Bosch, E. (1998). Anal. Chim. Acta 374, 309. f Andon, R.J.L., Cox, J.D. and Herington, E.F.G. (1954). Trans. Faraday Soc. 50, 918. g Neverov, A.A. and Brown, R.S. (2001). Inorg. Chem. 40, 3588. h Schilt, A.A. and Smith, G.F. (1956). J. Phys. Chem. 60, 1546. is s pK a 2. j Desloges, W., Neverov, A.A. and Brown, R.S. (2004). Inorg. Chem. 43, 6752. k Bell, T.W., Choi, H.-J., Harte, W. and Drew, M.G.B. (2003). J. Am. Chem. Soc. 126, 12196. ls s s pK a 3. May be a lower limit since s pK a is close to solvent ionization.

interactions, steric effects and enforced ion pairing that can have dramatic effects on the acidity and reactivity.

4

Metal ion alcoholysis and titration in alcohol

The titration of metal ions in alcohol solvents28 follows the same sort of rules as titrations of metal ions in water29 but poses additional problems due to the lower polarity that increases ion pairing and oligomerization of the metal ions. We have performed several such titrations with the analysis of the potentiometric data depending on the level of information one requires. More complete and timeconsuming analyses are reserved for the most effective catalytic metals, namely La3+, and for the transition metal ion Zn2+ and Cu2+ along with some simple complexes of the latter two which we describe a little later. For the other metal ions described in our titration papers,7,8 we only present the data in terms of the

METAL-CATALYZED ALCOHOLYSIS

279

apparent, or conditional, ss pKa values for formation of the metal-bound alkoxides at an [Mx+]total of 1 mM. The data give titration constants30 ss G01 ; ss G02 ; ss G03 . . . ; etc., that are deﬁned as the conditional half-neutralization points determined for a solution containing the metal ion, and are simply interpreted as the solution ss pH at which the [OR]/[Mx+]total ratio is 0.5, 1.5 and 2.5. The metal ions are generally introduced as triﬂate or perchlorate salts as these counterions are generally found to have the least propensity for ion pairing. In most cases, knowledge of the halfneutralization values simpliﬁes setting the solution ss pH at a desired value due to the buffering capabilities of the Mx+q(OR)y species present in solution. However, the conditional half-neutralization values reveal nothing of the intricate details of the acid/base behavior involved and the latter constants are complicated composites of various equilibria that include ion pairing between the Mx+ and solution counter ions (which we generally ignore) and metal dimerizations/oligomerizations that can affect the ionizations of (Mx+)n(HOR)m forms. These details can be worked out with considerable effort which requires computer ﬁttings of the data obtained through titrations at different [Mx+] to various models comprising species for which there is corroborative evidence from other studies such as kinetics or spectroscopic ones. For other cases, such as La3+ where more detail is required about the nature of the species present in solution, titration data can be computer ﬁt to more complicated multi-equilibrium models containing Mx+q(OR)y forms whose stoichiometry is suggested by information gained from independent spectroscopic or kinetic techniques. One must be mindful of the pitfalls of simply ﬁtting the potentiometric data to complex multi-component models for which there is no independent evidence for the various species. Without some evidence for the species put into the ﬁt, the procedure simply becomes an uncritical mathematical exercise of adding and removing various real and proposed components until the goodness of ﬁt is satisfactory. A case in point concerns the speciation of La3+ in methanol and ethanol which is of interest because of the very large rate enhancements for La3+-catalyzed acyl and phosphoryl transfer to alcohol solvents which will be described later. Earlier kinetic studies of the catalysis of methanolysis of carboxylate esters16a and the activated amide acetyl imidazole31 indicated that the process was largely bimolecular in each of [La3+] and [OCH3], suggesting that a dimeric species (La3+ 2 ( OCH3)2) was largely responsible for the great rate accelerations observed. Electrospray MS on solutions of La(OTf)3 in the presence of 1 equivalent of NaOCH3 also indicated the presence of dimeric La3+-containing species containing two and three methoxides as well as three and two, respectively, triﬂate counterions and varying numbers of solvent methanols.16a Finally the potentiometric titration data7 depended on the concentrations of the La(OTf)3, with higher concentrations driving the apparent s s pK a s to lower values, a phenomenon suggestive of the equilibrium formation of dimers or higher-order aggregates.29a Analysis of these potentiometric titration data according to the model suggested in Equations (6) and (7) using the program Hyperquad27 allows one to compute the various stability constants ðss K n Þ which

280

R.S. BROWN AND A.A. NEVEROV

Table 3 Stability constants for the formation of La2(OR)n species and microscopic ss pK a values in methanol and ethanol (25.0 1C) Equilibrium (Kneq)

log ss K neq a

[La2(OR)1]/[La]2[OR] [La2(OR)2]/[La]2[OR]2 [La2(OR)3]/[La]2[OR]3 [La2(OR)4]/[La]2[OR]4 [La2(OR)5]/[La]2[OR]5

11.6670.04c 14.8570.06d 20.8670.07c 27.3670.08d 27.5270.09c 38.7770.22d 34.5670.20c 49.8470.16d 39.3270.26c

Microscopic ss pK a b

2 s s pK a

¼ 7.5670.10c ¼ 6.6070.03d 3 c s s pK a ¼ 10.1170.02 d ¼ 7.6970.15 4 c s s pK a ¼ 9.7370.11 d ¼ 8.0370.22 5 c s s pK a ¼ 12.0070.07

a

Derived from ﬁts of the potentiometric titration data in methanol and in ethanol using the program Hyperquad.27 b Deﬁned as log ss pK a for La2(OR)n(HOR)x La2(OR)n+1(HOR)y1+H+, calculated from appropriate data in column 2 as ss pK auto (autoprotolysis constant of solvent) – ss pK neq ss pK n1 eq where 16.77 is the ss pK auto for pure methanol and 19.1 is the ss pK auto for ethanol. c Computed constants for [La(OTf)3] ¼ 2 103 M in methanol, from ref. 7 d Computed constants for [La(OTf)3] ¼ 1 103 M in ethanol, from ref. 8.

are given in Table 3 along with the analogous stability constants subsequently determined8 for the − La3+ 2 ( OCH3)n s n s K eq

2La3+ + nOCH3−

n 3þ 2 ¼ ½La3þ 2 ð OCH3 Þn =½La ½OCH3 La3+ 2 ( OEt)1,2,3,4,5

(6) (7)

in ethanol. From these can be derived the individformation of ual microscopic ss pK a values for the ionizations La3+ 2 ( OR)n(HOR) 3+ + La2 ( OR)n+1+H : these too are listed in Table 3. From the formation constants derived by ﬁtting the potentiometric titration data as above can be computed the speciation for the Ln3+ 2 ( OR)14 forms by HySS, 27 another program in the Hyperquad suite. The speciation diagram for La3+ at 2 103 M in ethanol is shown in Fig. 1. According to this treatment the two dominant dimeric La3+ species have even numbers of attached ethoxides, namely s La3+ 2 ( OEt)2 between s pH 5.9 and 8.4 (maximum concentration of 48% relative to 3+ s [La2 ]total at ss pH 7.3), and La3+ 2 ( OEt)4 above s pH 7. Species with odd numbers of 3+ 3+ ethoxides, La2 ( OEt)1 and La2 ( OEt)3, are also present to a lesser extent (maximum concentrations of 17% and 37% respectively reached at ss pH values of 6.7 and 7.8). The computed speciation for La3+ in ethanol is similar to the calculated speciation diagram shown in Fig. 2 for 2.0 103 M La3+ in methanol as a function of a more extended ss pH range.7 The two dominant species again have even numbers 3+ s of attached methoxides, e.g. La3+ 2 ( OMe)2 (80% at s pH 8.9), and La2 ( OMe)4 3+ 3+ s (80% at s pH 11). La2 ( OCH3)1 and La2 ( OCH3)3, are also present to a lesser extent (25% in each at respective ss pH values of 7.5 and 10).

METAL-CATALYZED ALCOHOLYSIS

281

0.0006 La2(OEt)4 0.75

La2(OEt)2

0.0004

La2(OEt)3 0.50

0.0003

0.0002

La2(OEt)1

k2obs (dm3mol-1s-1)

[(La3+)2(-OEt)n] (M)

0.0005

0.25

0.0001

0.0000

0.00 5

6

7

8

9

s s pH – Fig. 1 A calculated speciation diagram for La3+ n ¼ 1–4, total 2 ( OEt)n, [La3+] ¼ 2 103 M, Overlaid are the observed second-order rate constants (kobs 2 ) for the La3+-catalyzed ethanolysis of paraoxon (1) as a function of ss pH. Reprinted with permission from ref. 8.

0.0010 La2(OCH3)2

La2(OCH3)4

40

30 0.0006

20 0.0004 La2(OCH3)1

La2(OCH3)3

k2obs (dm3mol-1s-1)

[La3+2(-OCH3)n] (M)

0.0008

10

0.0002

0.0000

0 5

6

7

8

9

10

11

12

s s pH

Fig. 2 Speciation diagram for 2.0 103 M La3+ in methanol as a function of ss pH. Data superimposed on the ﬁgure as () are second-order rate constants for La3+-catalyzed methanolysis of p-nitrophenyl acetate (2) as a function of ss pH.

282

R.S. BROWN AND A.A. NEVEROV

Having established the speciation, we now have a very powerful tool for analyzing the kinetic data for the ss pH dependence of the La3+catalysis of the alcoholysis of various substrates. Included in the Figs 1 and 2 plots are the second-order rate constants for La3+-catalysis of the ethanolysis of paraoxon (1) and the methanolysis of p-nitrophenyl acetate (PNPA, 2) as a function of ss pH in ethanol and methanol, respectively. The kinetic data mainly follow the rise/fall behavior of the La3+ 2 ( OR)2 3+ 3+ species with some involvement of the other species, La2 ( OR)1, La2 ( OR)3 and La3+ 2 ( OR)4. O EtO P O OEt 1

O NO2

H3C

O EtO P S NO2 OEt

O 2

3

O NO2

NO2

N 4

obs To determine the activities for the various La3+ data 2 ( OR)n we analyze the k2 as a linear combination of individual rate constants (Equation 8), where k2:1 2 , 2:n 3+ k2:2 y k are the second-order rate constants for each La ( OR) promoting 2 2 n 2 ethanolysis and methanolysis of 1 and 2 respectively.

2:1 2:2 3þ 3þ kobs 2 ¼ ðk2 ½La2 ð ORÞ1 þ k 2 ½La2 ð ORÞ2 þ . . . 3þ k2:n 2 ½La2 ð ORÞn Þ=½LaðOTfÞ3 t

ð8Þ

When ﬁt to this expression, the kinetic data for the ethanolysis8 of 1 yields the contributions of the relevant La3+-dimers shown in Fig. 3 where the solid line represents the sum of the contributions and the three dotted and dashed lines give the contributions of the individual dimers at each ss pH. The individual species rate constants are determined as 0.5970.05, 2.2470.19 and 1.1170.09 dm3 mol1 s1 2:2 2:3 for k2:1 2 , k2 and k2 . 3+ Fits for the La -catalyzed methanolysis of p-nitrophenyl acetate (2),7 paraoxon (1), its thio derivative, O,O-Diethyl S-p-nitrophenyl phosphorothioate (3)10b and N-p-nitrophenyl b-lactam (4)32 have also been computed, with the various k2:n 2 the values being given in Table 4. The catalysis afforded by the La3+ system for the transesteriﬁcations of paraoxon in ethanol and methanol is quite spectacular relative to the background reactions that are assumed to be promoted by the lyoxide. The reaction rate constant of ethoxide with paraoxon in ethanol at 5.1 103 dm3 mol1 s133 is roughly a factor of two lower than the rate constant of methoxide with paraoxon in methanol (1.1 102 dm3 mol1 s1).17a However a solution 2 mmol dm3 in total [La3+], 4 1 which contains 1 mmol dm3 of La3+ s 2 , has a maximum rate constant of 7 10 for decomposition of 1 in ethanol at ss pH of 7.3, and accelerates the rate of ethanolysis of paraoxon by a factor of 4.4 1011-fold relative to the ethoxide reaction at the same ss pH.34 By way of comparison, the acceleration afforded by a 1 mmol dm3 solution of the La3+ dimer catalyzing the methanolysis of 1 at the 2 maximal ss pH of 8.3 (kobs ¼ 0.0175 s1) is 109-fold greater than its background methoxide reaction. On this simple basis La3+ in ethanol appears to be catalytically 2 superior to La3+ in methanol, but this stems almost exclusively from the ss pH values 2

METAL-CATALYZED ALCOHOLYSIS

283

0.8 0.7

k2obs (dm3mol-1s-1)

0.6 0.5 0.4 0.3 0.2 0.1 0.0 5

6

7

8

9

s s pH 3+ obs Fig. 3 Plot of the contributions of various La3+ -catalyzed 2 ( OEt)n forms to the k2 for La ethanolysis of paraoxon (1) as a function of ss pH. Solid line gives the combined effects of – various species; ( — —), contribution of La3+ 2 ( OEt)1; (- - - - - - -), contribution of obs – 3+ – La3+ ( OEt) ; (— — — ), contribution of La ( OEt) 2 3; included as (’) are the actual k2 2 2 kinetic data. Reprinted with permission from ref. 8.

3+ Table 4 Computed second-order rate constants (k2:n 2 ) for the methoxide and La2 ( OCH3)na species catalyzing methanolysis of 1–4, T ¼ 25 1C

k2/dm3 mol1 s1 paraoxon (1)a

k2/dm3 mol1 s1 p-nitrophenyl acetate (2)b

k2/dm3 mol1 s1 phosphorothioate (3)

k2/dm3 mol1 s1 N-p-nitrophenyl b-lactam (4)

kOCH3 1 ¼ 0.011

kOCH3 2 ¼ 2207100c

kOCH3 3 ¼ 0.12

kOCH3 4 ¼ 0.43

k2:1 2 ¼ 15.872.9

d k2:1 2 NA

k2:1 2 ¼ 17.978.7

k2:1 ¼ 0.0670.01 2

k2:2 2 k2:3 2 k2:4 2

k2:2 2 k2:3 2 k2:4 2

k2:2 2 k2:3 2 k2:4 2

¼ 27.971.4

k2:2 2 ¼ 0.9870.11

NAb

k2:3 2 ¼ 1.2370.13

¼ 16.271.8

d k2:4 2 NA

¼ 51.171.1 ¼ 35.676.5 ¼ 49.771.4

¼ 91.878.8 NAd ¼ 38.075.6

a

Data from Ref. 10b. kOCH3 is observed rate constant for the attack of CH3O on indicated substrate. c At 27.4 1C; Schowen, R.L. and Behn, C.G. (1968). J. Am. Chem. Soc. 90, 5839. d Indicated rate constant not required to obtain satisfactory ﬁts to data. b

where the comparison background reactions are made and the far lower [OEt] relative to [OMe]. However, it is clear that in terms of the second-order rate constants the methanolysis reaction is more effective by about 25-fold at the maximum ss pH values for 2:2 the La3+ 2 -catalyzed reactions having the same amount of catalyst. For example k2

284

R.S. BROWN AND A.A. NEVEROV

for (La3+(OCH3))2 promoted methanolysis of paraoxon is 51.1 dm3 mol1s1, while that for (La3+(OCH2CH3))2 promoted ethanolysis is 2.24 dm3 mol1s1. While more study is required to determine the reasons why (La3+(OCH3))2 is more active than (La3+(OCH2CH3))2 this probably stems from the fact that coordination to the La2+ cations stabilizes the alkoxide better in ethanol than in methanol, as judged by ðss pK EtOH ss pK MeOH Þ and a steric effect that may retard the catalytic a a reaction with the latter larger complex. It is important to note that the stabilization of the alkoxides through coordination to two La3+-ions does not render them less reactive than free alkoxide as might be expected from a simple Brønsted relationship. This supports the oft-stated premise that there is a dual role for the Mx+(OR) species involving Lewis acid activation and delivery of the coordinated alkoxide. While this is rarely observed with Mx+(OH) species in water, our work shows that this is the norm in alcohols.

5 Transition metal ion and Ln3+ catalysts of transesterifications of neutral carboxylate and organophosphate esters As summarized previously,10 virtually all the transition metal and lanthanide ions we have investigated have some catalytic activity in promoting the transesteriﬁcation of carboxylate esters and neutral phosphate esters such as paraoxon (1). The active forms of these are routinely generated in situ by adding 1 equivalent of NaOR to an alcohol solution containing 1 equivalent of the metal ion, which gives predominantly species having a net stoichiometry of (Mx+(OR))1,2.35 For La3+ there is evidence that the most active form for methanolysis and ethanolysis is 3+ the dimer La3+ 2 ( OR)2. The other forms La2 ( OR)n where n ¼ 1, 3, 4 may also 3+ have activity (Table 4), but the La2 ( OR)2 form has the highest activity in the neutral ss pH domains due to its appreciable second-order rate constant and its high prevalence. For other metal ions there are some interesting variations where the Mx+(OR) monomeric form is the most active, while the dimeric forms are inactive. For example, the Eu3+-catalyzed methanolysis of the esters p-nitrophenyl acetate (2), phenyl acetate (5) and ethyl acetate has been studied as a function of ss pH along with its titration in methanol.16e All the kinetic evidence indicates that the solution behavior is consistent with the species shown in Equations (9–11), with Eu3+(OCH3)(CH3OH)x1 being the reactive one (ss pK 1a ¼ 6.33) having a kmax rate 2 constant of 4272 dm3 mol1 s1 for methanolysis of 2 and 11.771.5 dm3 mol1 s1 for methanolysis of 5. The ss pH rate proﬁle for the Eu3+-catalyzed methanolysis of 2 and 5 exhibit bell-shaped proﬁles with a second apparent ss pK 2a of 8.02 for conversion of Eu3+(OCH3) into inactive Eu3+(OCH3)2 which we believe exists as a heavily methoxy-bridged oligomer of repeating Eu3+(OCH3)2 subunits as in 6. O H3C

O 5

METAL-CATALYZED ALCOHOLYSIS *

O R

Eu

285

R

R

R

R

O

O

O

O

O

Eu

R

O R

Eu

O

Eu

R

S H3CO P O H3CO 7

x*

NO2 CH3

6 (charges ommitted for clarity)

Ka1

3+

Eu (CH3OH)x

3+

−

Eu3+(CH3O− )(CH3OH)x−1 + H+

Ka2

Eu (CH3O )(CH3OH)x−1 nEu3+(CH3O− )2(CH3OH)x−2

Kolig

Eu3+(CH3O−)2(CH3OH)x−2 + H+

(9)

(10)

[Eu3+(CH3O− )2(CH3OH)y]n + n(CH3OH)(x−2−y)

(11) 2+36

2+37

and Cu there is evidence that the monomeric On the other hand, for Zn forms (M2+(OR)) are catalytically active toward transesteriﬁcation of carboxylate esters like 2 and paraoxon (1). The situation is more complicated because the maximum rate constants for the reactions are achieved when the [OR]/[M2+]total ratio is 0.1–0.3, and then starts to fall at higher ratios. At a given [OR]/[M2+]total ratio the plot of kobs vs. [M2+]total follows a square root dependence suggestive of a catalytically active monomer in equilibrium with an inactive dimer. The minimal process required to ﬁt the kobs vs. [OCH3]/[Zn2+]total and kobs vs. [M2+]total behavior for Zn2+-catalyzed methanolysis is shown in Scheme 1. The apparent dimerization constant (Kd) is a complex term that incorporates both dimerization of Zn2+(OCH3) and an ionization constant that connects Zn2+(OCH3)2 and (Zn2+(OCH3)2)2. The key point of this analysis is the requirement to limit the overall [Zn2+(OCH3)] through the formation of inactive dimeric species (Zn2+(OCH3))2 and (Zn2+(OCH3)2)2. Unfortunately, the dimerization constants for the 2(M2+(OR)) (M2+(OR))2+(M2+(OR)2)2 process are large (4 2 105 dm3 mol1) so the catalysis observed is not very effective even though the monomer is quite active. Given in Table 5 are the second-order rate constants for the Zn2+(OCH3)- and Cu2+(OCH3)-catalyzed methanolysis of paraoxon (1) and a closely related phosphorothionate pesticide fenitrothion (7) that are computed from the kobs vs. [M2+(OCH3)] data.36,37 The complexity exhibited by these transition metal ions can be reduced through the addition of complexing ligands such as phenanthroline (8) and 1,5,9-triazacyclododecane (9), a macrocyclic ligand, the Zn2+-complexes of which have been extensively investigated in water by Kimura.38 Detailed studies36,37 indicate that phenanthroline, when complexed to either Zn2+ or Cu2+ in the presence of 1 equivalent of added –OCH3, produces a kinetically active complex (8:M2+(OMe)) that still has an appreciable tendency to dimerize to an inactive dimer (10) as in

286

R.S. BROWN AND A.A. NEVEROV Ka1

Zn2+(HOCH3)n

Zn2+(HOCH3)n-1(-OCH3) + H+ [Zn2+(HOCH3)n-1(-OCH3)]2

2 Zn2+(HOCH3)n-1(-OCH3) Kd

[Zn2+(HOCH3)n-1(-OCH3)2]2 + 2H+

Scheme 1 A minimal scheme depicting ionization of a Zn2+(HOCH3)n species followed by formation of inactive Zn2+-dimers.

Table 5 Kinetic constants for the methanolysis of paraoxon 1 and fenitrothion 7 catalyzed by Zn2+ and Cu2+ in the absence and presence of ligands 8 and 9, T ¼ 25 1C Catalyst OCH3 Cu2+(OCH3)c Zn2+(OCH3)d 8:Cu2+:(OCH3)e 8:Zn2+:(OCH3)f 9:Cu2+:(OCH3)g 9:Zn2+:(OCH3)h i La3+ 2 ( OCH3)2

Kd (dm3 mol1)a

k2(1) (dm3 mol1 s1)a

k2 (7) (dm3 mol1 s1)a

Relative selectivityb

NA 42 105 42 105 42 105 42 105 NA NA NA

1.1 102 0.2270U02 1.2170.03 o0.2 2.0770.04 2.7670.17 0.8570.01 47.272.3

(7.270.2) 104 0.7970.03 0.1970.001 2.4470.06 0.3270.01 12.270.4 (4.870.2) 102 Non-reactive

1 55 2.4 186 2.35 67 0.86 0

a

Dimer association constant (Kd) and conditional second-order rate constant (k2(1) or k2(7)) for reaction of monomer with 1 or 7 deﬁned as in text. NA means non-applicable since there is no observable dimerization under the speciﬁc conditions. The Kd of 42 105 indicates very strong dimerization and is quoted as an upper limit based on an iterative ﬁtting procedure which provided the lowest standard deviations. Zn2+ results from ref. 36, Cu2+ results from ref. 37. b Deﬁned as (k2(7)/kOCH3 (7))/(k2(1)/kOCH3 (1)). c Based on ﬁts of kobs vs. [Cu2+]total data at [methoxide]/[Cu2+]total ratio of 0.5. d Based on NLLSQ ﬁts of kobs vs. [Zn2+]total data at [methoxide]/[Zn2+]total ratio of 0.3. e Based on ﬁts of kobs vs. [8:Cu2+]total data at [methoxide]/[Cu2+]total ratio of 0.5. f Based on NLLSQ ﬁts of kobs vs. [Zn2+:9]total data at [methoxide]/[Zn2+]total ratio of 0.5. g Based on linear ﬁts of kobs vs. [9:Cu2+:(OCH3)]total data at methoxide]/[Cu2+]total ratio of 0.5; ss pK a for ionization of 9:Cu2+ to generate 9:Cu2+(OCH3) is 8.75 70U1 by half neutralization. h Based on linear ﬁts of kobs vs. [9:Zn2+:(OCH3)]total data at [methoxide]/[Zn2+]total ¼ 1.0. The ﬁrst and second ss pK a values of 9.1 and 12.9 for 9:Zn2+ to generate 9:Zn2+ (OCH3) and 9:Zn2+(OCH3)2. i From ref. 17a.

Equation (12). However, in the case of complexation to 9 there is no dimerization and the catalytically active form is 9:M2+(OR). H N N

N 8

9

N N H

H

METAL-CATALYZED ALCOHOLYSIS

287

CH3 -O N Zn2+ N

OH3C

N Zn2+ N

CH3 -O k21 = 2.1 dm3mol-1s-1 N P Zn2+

2

N

5

Kd = >2x10 dm3mol-1

8:Zn2+(-OMe)

10 S

CH3

P

N PdII N CH3 CH OSO2CF3 3 11

(12)

O CH3 H

H2C O O O H3C H2C CH3 12

N N

S

N

P H2C O O N O H3C H2C CH3 13

Also included in Table 5 are the kinetic data for methanolysis of 1 and 7 catalyzed 2+ by the La3+ ( OCH3) and 9:M2+(OCH3) complexes and, for 2 ( OCH3)2, 8:M comparison purposes, the second-order rate constants for methoxide-catalyzed methanolysis. The ﬁnal column is a ‘relative selectivity (RS) parameter’ that describes the metal ions’ ability to select for a PQS species over the PQO species, deﬁned as (k2 (7)/kOCH3(7))/(k2(1)/kOCH3(1)). Of note is the fact that all the Cu2+-containing species have RS values of 55–186, meaning they strongly prefer the P ¼ S substrate but they will also catalyze methanolysis of the P ¼ O substrate. The Zn2+ complexes are less selective with RS values from 0.9 to 2.5 meaning they do not really discriminate well between the PQO and PQS classes. On the other hand, La3+ has a zero RS value and shows no propensity to react with the PQS substrates at all. In fact, none of the lanthanides we have studied promote the methanolysis or ethanolysis of the PQS derivatives, presumably due to the fact that these are considered ‘hard’ in the Pearson ‘hard/soft’ sense39 while the softer Cu2+ ion selects for the soft PQS derivatives. We have observed that cyclopalladated Pd complexes such as 11 rapidly methanolyze PQS pesticides like 7, diazinon (12) and quinalphos (13) but do not react with the PQO complexes nearly as fast which is also consistent with the ‘soft’ behavior of the Pd.40 These observations are also important from a mechanistic standpoint since they underscore the importance of transient metal ion:substrate binding which preceeds the catalytic events, and in cases where the hard/soft properties of the reaction partners do not match, one can safely predict little or no catalysis will be observed. While the above indicates that Ln3+ and transition metal ions in the presence of at least 1 equi. of –OR promote the alcoholysis of carboxylate and phosphate esters, sometimes by spectacular amounts, we have not presented evidence about the mechanism for the catalytic reactions. So far, the underlying theme is that the most active forms of the lanthanide ions are the Ln3+(OR) forms, either as a monomer (as in the case of Eu3+(OCH3)) or as a dimer (as in the case of La3+ 2 ( OCH3)2). For the transition metal ions the most active forms are those where one face of the

288

R.S. BROWN AND A.A. NEVEROV

metal ion is encapsulated with ligands such as phenanthroline (8) or 1,5,9-triazacyclododecane (9) along with a single alkoxide. MECHANISM OF ALCOHOLYSIS OF CARBOXYLATE ESTERS

The initial study of the La3+-catalyzed methanolysis of carboxylate esters16a re ported the apparent second-order rate constant for La3+ 2 ( OCH3)2-catalyzed methanolysis of some representative examples of aryl esters (2, 5 and 2,4-dinitrophenyl acetate (14)), phenyl benzoate (15) and three aliphatic esters, ethyl acetate, isopropyl acetate (16) and tert-butyl acetate (17). Given in Table 6 are the rate constants for the La3+ and methoxide-catalyzed methanolysis of these esters along with O

O H3C

NO2 O

14 O2N

O O

H3C

15

CH3 H3C O CH CH3 16

O

CH3 O C CH3 CH3 17

the catalytic rate acceleration relative to the background reaction at an essentially neutral ss pH of 8.5. There are a few features of note, some of which distinguish the methanolysis results from hydrolytic processes. First, simple metal ion-catalyzed hydrolysis of esters, including non-activated ones such as we have here is not, to our knowledge, a well-known phenomenon. Metal ion catalysis of hydrolysis is seen for esters having good leaving groups and particularly where there is an auxiliary binding site which places the metal ion in close proximity to the scissile CQO(LG) unit. Table 6 Maximal second-order rate constants for (La3+)2(CH3O)2-catalyzed methanolysis and second-order rate constants for methoxide attack on various esters, T ¼ 25 1C Ester 2 14 5 15 Ethyl acetate 16 17

3 1 1 c kester (dm3 mol1 s1)a,b kester Accelerationg s ) kester =kester 2 OMe (dm mol 2 OMe

72 2972 5873 3.070.06 0.143 0.0083 55.7 106

2207100c 410d 2.66e 0.33870.006f (5.770.2) 102 (7.070.1) 103 (2.070.1) 104

f f f

0.6–0.225 0.07 21 8.9 2.5 1.2 50.03

243,000 42,000 18,750,000 8,264,000 2,300,000 1,120,000 n.o.

a Methanolysis rates for esters 2, 14, 5, 15 determined by UV kinetics in CH3OH; ethyl acetate, 16, 17 determined by 1H NMR in d4-methanol. bs s pK a set at observed rate plateau for methanolysis of the esters between 8 and 9 through addition of 1:1 La3+/OCH3. c 27.4 oC; Schowen, R.L. and Behn, C.G. (1968). J. Am. Chem. Soc. 90, 5839. d Machacek, V., Mareckova, S. and Vojeslav, S. (1979). Collect. Czech. Chem. Commun. 44, 1779. e Milton, C.G., Gresser, M. and Schowen, R.L. (1969). J. Am. Chem. Soc. 91, 2047. f Ref. 16a. g Catalytic value at ss pH 8.5, in the presence of 5 103 mol dm3 (La3+)2(CH3O)2. Background rate s computed from kester OMe assuming value for solvent reaction at s pH 8.5 is entirely attributed to 16.77.h CH3O +ester; the autoprotolysis constant of methanol is 10 Acceleration computed at ss pH 8.7.

METAL-CATALYZED ALCOHOLYSIS H3C

O

H3C

-

O La3+

La3+ -O

289

+ R

R'

O

O

- O CH

-

O

La3+

La3+

H3C

CH3

O

O R'

3

R

La3+ + HOCH3 3+ O La + -H O -RC=O(OCH3) H3C R + HOCH 3

-HOR' -RC=O(OCH3) H3C

O

La3+

La3+

(poor LG) O-

O

O

R' R

-OR' CH3

-OR' H3C

O

(good LG)

La3+ O

H3C

O

La3+

La3+

O-

La3+

H3C

O -

O R'

H3C

O

O R'

R

R

Scheme 2 A proposed mechanism for La3+ (OCH3)2 catalyzed transesteriﬁcation of 2 carboxylate esters with good and poor leaving groups.

Second, for the three aryl esters (2, 5, 14) the greatest acceleration relative to the methoxide reaction is seen for the poorest leaving group (phenoxy 4 p-nitrophenoxy 4 2,4-dinitrophenoxy). Third, for the aliphatic esters (ethyl acetate, 16, 17) there is a large steric factor, perhaps superimposed on an electronic one which retards the metal-catalyzed reaction more than the methoxide-catalyzed reaction for the isopropyl and tert-butyl acetates. A proposed mechanism for the reaction is given in Scheme 2 where the catalytic entity (La3+ 2 ( OCH3)2) is shown as a bis-methoxy bridged dimer. According to the scheme, the CQO unit associates with the metal ion in a typical Lewis fashion for which the equilibrium constant should be greater in alcohol relative to water due to the lower dielectric constant. Since a methoxide bridged between the two metal ions should have limited nucleophilicity, we suggest that one of the methoxy bridges opens to reveal a strong Lewis acid activated CQO unit nearby a methoxide coordinated to the proximal La3+(OCH3) that subsequently generates a La3+-coordinated tetrahedral intermediate. In the case of good leaving groups the tetrahedral intermediate breaks down probably without assistance via coordination to the metal ion. However for poor leaving groups such as the aliphatic ones, microscopic reversibility suggests that if the formation of the tetrahedral intermediate results from attack of a metal-coordinated methoxy or ethoxy group, then its breakdown must proceed via metal ion assistance of OR0 departure. If the leaving group is very sterically demanding, as well as having a higher ss pK a for its conjugate acid such as in the case of tert-butoxy, then metal ion assistance of departure might be severely retarded to the point that the tetrahedral intermediate cannot break down productively and catalysis will be strongly reduced. The question of whether the reaction proceeds through an intermediate or whether it is a concerted displacement such as has been demonstrated by Ba-Saif and Williams41 for the displacement of the p-nitrophenoxy group from ester 2 (18d) by

290

R.S. BROWN AND A.A. NEVEROV

Table 7 Second-order rate constants for the methanolysis of esters 18, 19 promoted by methoxide, 9:Zn2+(OCH3) and La3+ 2 ( OCH3)2 at 25 1C Ester

18a(14) 18b 18c 18d(2) 18e 18f 18g(15) 18h 18i 19a 19b 19c (16) Methyl acetate

s s pK a

of ArOH in MeOH

kOCH3 (dm3 mol1 s1)

7.83b 8.84a 12.41b 11.30b 13.59b 14.7c 14.33b 15.04b 15.36b 15.78d 18.42d 19.79d 18.13

9:ZnðOCH Þ

9:ZnðOCH Þ

3 k cat (dm3 mol1 s1)h

3 k cat (dm3 mol1 s1)i

43.6 23.8 39.8 19.2 16.7 8.6 13.9 4.28 4.30 3.00 0.042 0.006 –

31 14.6 21.9 38.6g 27.7 18.0 29.33g – 4.33 10.0 0.071g 0.004g –

484.8 73.3 49.7 190e 7.04 1.58 2.01 (2.66)f 0.63 0.49 12.64 0.057g 0.007g 0.17e

a

Determined as the ss pH at half neutralization, this work. From refs. 6 and 42. pKa for 18i calculated according to the method provided in ref. (6a); MeOH s 2 O þ 3:56. ¼ 1:12pK H a s pK a c Schowen, R.L. and Lathan, K.S. (1967). J. Am. Chem. Soc. 89, 4677. d 2 O +6.53. Computed from ss pK MeOH ¼ 0.748pK H a a e From ref. 36. f From Milton, C.G., Gresser, M. and Schowen, R.L. (1969). J. Am. Chem. Soc. 91, 2047. g Data from ref. 16a for methoxide and lanthanum reactions; lanthanum reactions determined at pH ¼ 8.86 for 18d (2) and 18g (15); values are computed from the gradient of the plot of kobs vs. [La3+]total. h Computed from the gradient of kobs vs. [9:Zn2+(OCH3)] plots at ss pH 9.1 under half-neutralization conditions. i Computed from the gradient of kobs vs. [La3+] plots at ss pH 8.74 under buffered conditions. b

phenoxy and probably alkoxy nucleophiles has been addressed16f in the study of the 9:Zn2+(OCH3) and La3+ 2 ( OCH3)2-catalyzed methanolysis of an extensive series of carboxylate esters, 18 and 19. For the series of esters, the second-order rate constants for the reactions are presented in Table 7 along with the ss pK a values for the ArOH or HOR groups in methanol as reported in the literature6,42 or as determined by means identiﬁed in the study.16f From the autoprotolysis constant of methanol (1016.77 (mol dm3)2) the ss pK a of MeOH can be computed as 18.13 on the mol dm3 scale. O H3C

C

O O Ar

18 18a, Ar=2,4-dinitro 18b, Ar=pentafluoro 18c, Ar=3-nitro 18d, Ar=4-nitro 18e, Ar=4-chloro 18f, Ar=4-methoxy 18g, Ar=phenyl 18h, Ar=2,4-dimethyl 18i, Ar=2,3,5-trimethyl

H3C

C

O R

19 19a, R= CH2CF3 19b, R=CH2CH3 19c, R=CH(CH3)2

METAL-CATALYZED ALCOHOLYSIS

291

kcat9:Zn(OCH3)(dm3mol-1s-1)

100 10 1 0.1 0.01 0.001 7.5

10.0 12.5 15.0 17.5 pKa of ROH in methanol

20.0

9:ZnðOCH Þ

3 Fig. 4 A Brønsted plot of log pKa phenol in methanol vs. log kcat for methanolysis of aryl acetates promoted by 9:Zn2+(OCH3), T ¼ 251C, data in Table 7. Dashed line corresponds to NLLSQ ﬁt of data to Equation (15) encompassing all esters with b1 ¼ 0.02370.03 and b2 ¼ 0.69070.005 with a breakpoint of pKROH ¼ 14.8. Reproa duced from ref. 16f with permission.

kcatLa(OCH3)(dm3mol-1s-1)

1000 100 10 1 0.1 0.01 0.001 7.5

10.0 12.5 15.0 17.5 pKa of ROH in methanol

20.0

LaðOCH Þ

3 Fig. 5 A Brønsted plot of log pKa phenol in methanol vs. log kcat for methanolysis of aryl acetates promoted by La3+(OCH3), T ¼ 251C, data from Table 7. Dashed line corresponds to NLLSQ ﬁt of data to Equation (15) encompasses all esters with values of b1 ¼ 0.03 70.005 and b2 ¼ 0.71570.005 with a breakpoint of pKa ¼ 14.7. Reproduced from ref. 16f with permission.

The Brønsted plots shown in Figs 4 and 5 assist in visualizing the data of the metal-catalyzed reactions which show deﬁnite evidence of a break where the rate constants are quite sensitive to phenols or alcohols with high ss pK a values, but almost no sensitivity to the nature of the ROH groups with low ss pK a values. In these cases there is no obvious discrepancy between aliphatic and aryl esters, so all

292

R.S. BROWN AND A.A. NEVEROV

esters are used for the subsequent treatment which assumes that the overall metalcatalyzed processes follow the processes in Equations (13) and (14) involving a fast pre-equilibrium binding followed by reversible creation of an intermediate, the formation and breakdown of which can be rate-limiting depending on the nature of the leaving group. The general kinetic relationship in Equation (15) is derived from a steady state treatment where the

Kb X+ −

M ( OCH3) + RCO2R RCO2R':MX+(− OCH3)

(13)

RCO2R':MX+(− OCH3)

'

k1 k −1

(CH3O-To−):MX+

k2 P

(14)

pKa refers to the ss pK a of the ROH or ArOH leaving groups in methanol and the b terms refer to the binding and kinetic steps. Equations (16) and (17) are derived for limiting cases where formation and breakdown of the tetrahedral addition intermediate ((CH3 O TO ):Mx+) is rate-limiting. ðbbþb1þb2ÞpKa =ðC 1 10b1pKa þ C 2 10b2pKa Þ kobs 2 ¼ K b k1 k2 =ðk1 þ k 2 Þ ¼ C b C 1 C 2 10

(15) ðbbþb1ÞpKa kobs 2 ¼ K b k1 ¼ C b C 1 10

(16)

ðbbþb1b1þb2ÞpKa kobs 2 ¼ K b k1 k2 =k1 ¼ ðC b C 1 C 2 =C 1 Þ10

(17)

The dashed lines in Figs 4 and 5 are computed from the NLLSQ ﬁts of all the data to Equation (15): in Fig. 4 (bb+b1) ¼ 0.023 and (bb+b1b1+b2) ¼ 0.710 with a breakpoint at ss pK a ¼ 14.8; in Fig. 5 (bb+b1) ¼ 0.02 and (bb+b1b1+b2) ¼ 0.710 with a breakpoint at ss pK a ¼ 14.7. Overall the leaving group acidity ranges over 1011-fold while the kinetic data span respective ranges of 5000- and 9000-fold for the two metal ions. That each plot exhibits a pronounced break is consistent with a process that contains at least two steps where formation and breakdown of an intermediate is rate-limiting. In both the cases the descending and plateau b-values are experimentally the same with a plateau range that covers roughly 106-fold change in acidity of the leaving group with virtually no change in catalyzed rate constant. The data in Figs 4 and 5 are analyzed according to the simpliﬁed process presented in Scheme 3 where the ester substrate binds reversibly to the metal complex (9:Zn2+(OCH3) or La3+ 2 ( OCH3)2) followed by an intramolecular addition of the metal-coordinated methoxide to the bound CQO unit. Unlike the non-metalcatalyzed process of methoxide addition to the esters which must form a highly unstable, and perhaps kinetically unstable (in the case of aryl esters)41 anionic intermediate (TO ), the Mx+(OCH3)-catalyzed process forms an Mx+-stabilized tetrahedral intermediate (20:TO : Mx+) with a signiﬁcant lifetime that partitions between starting materials and products depending on the relative transition state

METAL-CATALYZED ALCOHOLYSIS

Ester

Kb

+ 2+ -

9:Zn ( OCH3)

R O

9:2+Zn

293

-

Zn2+:9 k2 O

k O C CH3 1 H3CO H3CO O OR OR k H3C -1 H3C 9:Zn2+(-OCH3) 20:TO-:Zn2+

Zn2+:9 O H3CO H3C

OR

Scheme 3 A simpliﬁed process for 9: Zn2+ (OCH3) or La3+ (OCH3)2-promoted meth2 anolysis of a carboxylate ester demonstrating the reversible formation of a metal-ion stabilized tetrahedral intermediate.

energies corresponding to k1 and k2. In the cases where methoxide addition to the ester is metal-ion assisted, the departure of poor leaving groups such as ethoxy or iso-propoxy must be metal ion assisted suggesting that the k2 step must be partially or entirely rate-limiting and Equation (17) applies. As its ss pK a decreases, eventually the leaving group is sufﬁciently good that it departs without metal-ion assistance and the rate-limiting step for the reaction must change from breakdown to formation of the intermediate. In both the La3+ and 9:Zn2+ cases this is when the leaving group s s pK a is about 14.7. In the plateau regions of Figs 4 and 5 the computed Brønsted b is 0 signifying little or no dependence on the nature of the leaving group. The observation is explained by the limiting case of Equation (16) as having the formation of the intermediate depending on the opposing effects of the leaving group on both the Kb and k1 steps of Scheme 3 and Equations (13) and (14), i.e. (bb+b1)-0. This is reminiscent of the known insensitivity of the rates of acid-catalyzed hydrolysis of esters to changes in the nature of the leaving group (r0) which Taft43 interpreted as resulting from a counterbalancing of the OR/OAr substituent’s electronic effect on the protonation equilibrium and subsequent nucleophilic attack of water on the protonated CQO. The descending wings of Figs 4 and 5 do have a dependence on the nature of the leaving group b ¼ 0.71 which, according to Equation (17), is consistent with signiﬁcant cleavage of the C–OR bond in the TS. Since the net effect of the substituent on the Kb and k1 steps cancel, this leads to the conclusion that the sum of (b1+b2) for the reversal of the intermediate and its forward cleavage is signiﬁcantly negative. Although the above concentrates mainly on the mechanistic aspects of the metalcatalyzed methanolysis process, it should be clear that there could be important applications for transesteriﬁcations in synthetic sequences where esters might be used as protecting groups for sensitive alcohols.10a Since the method works under essentially neutral conditions at ambient temperature, it should be particularly applicable to systems where there are acid or base sensitive structures. For example, as is shown in Scheme 4, deprotection of 6-exo-acetoxybicyclo[2.2.2]octan2-one using (La3+(OCH3))2 generated in situ by simple addition of 10 mM each of La(OTf)3 and NaOCH3 to a methanol solution, gave a 100% yield of the corresponding 6-exo-hydroxybicyclo[2.2.2]octan-2-one within 3 h with no observed epimerization at the 6-position.44 However, when the deacetylation is performed with NaOCH3 alone, there is rapidly established a 15/85 mixture of the exo- and

294

R.S. BROWN AND A.A. NEVEROV O H3CCO

HO

(La3+(-OCH3))2

O

O

CH3OH, 25oC exo NaOCH3 O

OH

H O

NaOCH3

O

endo

Scheme 4 A La3+-catalyzed deprotection reaction of 6-exo-acetoxybicyclo[2.2.2]octan-2one, where the mild nature of the transesteriﬁcation conditions does not promote rearrangement of the product (ref. 44).

endo-products formed by base-catalyzed equilibration through the aldehyde as shown in the scheme.

MECHANISM OF ALCOHOLYSIS OF NEUTRAL PHOSPHATE ESTERS

Our initial foray into metal-catalyzed alcoholysis of phosphate esters dealt with the methanolysis of the pesticide paraoxon17a where immediately it became evident that the log k2 vs. ss pH plot for La3+-catalyzed reaction showed a distorted bell-shaped proﬁle (given in Fig. 2 of ref. 17a). Preliminary analysis of the proﬁle indicated that the left ascending wing had a gradient of 1 which is consistent with the active species having a single methoxide, but the situation proved to be far more complex. Detailed analysis of the data required consideration of the La3+ speciation as outlined in section ‘Metal ion alcoholysis and titration in alcohol’, Table 3, which generated the same speciation plot as illustrated in Fig. 2. The kinetic data and speciations were ﬁtted to Equation (8), to compute the plot presented in Fig. 6 which shows the contribution of each of the La3+ 2 ( OCH3)n species to the observed curve (computed rate constants, along with those for the corresponding thiolate derivative 3 given in Table 8). The mechanism of the reaction was not investigated in detail at the early stages of these studies although it was clear that the acceleration of the decomposition of 1 and 3 was spectacular. For example, a 2 mM solution of total [La3+] ion, which gives a 1 mM solution of La3+ 2 , gives a t1/2 for decomposition of 1 of 20 sec at a ss pH of 8.3, corresponding to an acceleration of 109-fold over the background methoxide reaction at the same ss pH (t1/2 ¼ 600 years). With the La3+-catalyzed methanolysis of phenyl acetate (18g, 5) and p-nitrophenyl acetate (18d) carboxylate esters, the effect of the change in leaving group has only minor effect on the rate constant at ss pH 9 (38 and 29 dm3 mol1 s1, Table 7). However, preliminary

METAL-CATALYZED ALCOHOLYSIS

295

25

k2obs (dm3mol-1s-1)

20

15

10

5

0 5

6

7

8

9

10

11

12

s

pH s

Fig. 6 Plot of the predicted kobs vs. ss pH rate proﬁle for La3+-catalyzed methanolysis of 2 paraoxon (1) (solid line) based on the kinetic contributions of (left to right), La3+ 2 ( OCH3)1; 3+ 3+ 2:1 2:2 2:3 La3+ ( OCH ) ; La ( OCH ) and La ( OCH ) computed from the k , k 2 2 2 3 2 3 3 3 4 2 2 , k2 and 2:4 s k rate constants (Table 4), and their speciation as a function of s pH. Reproduced from ref. 10b with permission.

3+ Table 8 Computed second-order rate constants (k2:n 2 ) for the La2 ( OCH3)n-catalyzed at each methanolysis of paraoxon 1 and its thioate derivative 3 obtained through ﬁts of kobs 2 a s s pH to Equation (8)

Paraoxonb 1

O,O-diethyl-p-nitrophenyl phosphorothioatec 3

a

La3+ 2 ( OCH3)n species

3 1 1 k2:n s 2 /dm mol

OCH3 3+ La2 ( OCH3)1 La3+ 2 ( OCH3)2 3+ La2 ( OCH3)3 La3+ 2 ( OCH3)4 OCH3 La3+ 2 ( OCH3)1 3+ La2 ( OCH3)2 La3+ 2 ( OCH3)4

kOCH3 ¼ 0:011 k2:1 2 ¼ 15.872.9 k2:2 2 ¼ 51.171.1 k2:3 2 ¼ 35.676.5 k2:4 2 ¼ 49.771.4 kOCH3 ¼ 0:12 k2:1 2 ¼ 11.675.3 k2:2 2 ¼ 28.471.2 k2:4 2 ¼ 16.171.6

Errors computed from the average % deviation in the ﬁtted numbers calculated by Equation (8) from the actual kinetic data. b Paraoxon data from ref. 17b. c DATA from ref. 17b

296

R.S. BROWN AND A.A. NEVEROV

study17b of the La3+-catalyzed methanolysis process for diethyl phenyl phosphate and 1 showed a very great sensitivity to the leaving group (3.5 103 and 51 dm3 mol1s1 respectively)17b suggesting the mechanisms for carboxylate and phosphate ester methanolysis might be different. Subsequent work on the La3+- and 9:Zn2+(OCH3)-catalyzed methanolysis of expanded sets of neutral OP substrates including phosphates and phosphorothioates,17c phosphonates17e and phosphonothioates17g led to the conclusion that all these materials probably react by a common mechanism without the formation of a kinetically detectable intermediate as is detailed below. Phosphates and phosphorothioates,17c phosphonates17e and phosphonothioates17g The metal ion promoted methanolysis of a series of aryl phosphates 1, 20a–f and aryl phosphorothioates 21a–f (in which 21b is the previously described 3) was studied in the presence of La3+ and 9:Zn2+(OCH3) in greater detail at 25 1C at ss pH values that correspond to the maximum for the metal-catalyzed reaction. In the case of La3+, the reaction ss pH was buffered by N-ethylmorpholine at 9.1 while for 9:Zn2+(OCH3) the operational ss pH was set at the ss pK a of 9.1 for ionization of 9:Zn2+(HOCH3) through half neutralization. The second-order rate constants for the metal-catalyzed reactions at this ss pH were determined from the gradients of the plots of the pseudo-ﬁrst-order rate constants for methanolysis of series 20 and 21 vs. [catalyst]. These are presented in Tables 9 and 10 along with those for the methoxide reactions at 25 oC.17c The Brønsted plots shown in Figs 7 and 8 and their gradients are computed from all data except for the lowest ss pK a member in each plot which can be seen to fall signiﬁcantly below the lines. O EtO

P

O OAr

OEt 20 a, Ar = phenyl b, Ar = pentafluorophenyl c, Ar = 4-chloro-2-nitrophenyl d, Ar = 3-nitrophenyl e, Ar = 4-chlorophenyl f, Ar = 4-methoxyphenyl (1) Ar = 4-nitrophenyl

EtO

P

SAr

OEt 21 a, Ar = phenyl b, (3) Ar = 4-nitrophenyl c, Ar = 4-chlorophenyl d, Ar = 3,5-dichlorophenyl e, Ar = 4-fluorophenyl f, Ar = 4-methoxyphenyl

The Brønsted plots of the log kOMe for methoxide reactions of the phosphate and 2 phosphorothioate esters vs. the ss pK a values45,46 (not shown here) provide reasonable linear correlations, the respective blg values being –0.7070.05 and –0.7670.08. As is the case for the metal-catalyzed reactions shown in Figs 7 and 8, the methoxide reactions also show that the derivatives having the lowest ss pK a values of the leaving phenol/thiophenol (20b and 21b (3)) are less reactive than predicted on the basis of the equilibrium ss pK a data. Cursory analysis might suggest a curvature in all the

METAL-CATALYZED ALCOHOLYSIS

297

Table 9 Second-order rate constants (k2) for the methanolysis of phosphates 20 and 1 promoted by methoxide, La3+ and 9:Zn2+(OCH3) in methanol solvent, T ¼ 25 1C k2 (OMe) (dm3 mol1 s1)

3 1 1 s ) kLa 2 (dm mol

Aryloxyphosphate

pHa a

b s s pHa

Pentaﬂuoro (20b) 4-chloro-2-nitro (20c) p-nitro (1) m-nitro (20d) p-chloro (20e) p-H (20a) p-methoxy (20f)

5.53 6.32

8.84 10.64

0.20170.002 (6.470.1) 102

1070740 18577

23.070.6 11.470.4

7.14 8.39 9.38 10.0 10.20

11.30 12.41 13.59 14.33 14.7

(1.0270.03) 102 (6.070.1) 103 (6.370.1) 104 1.4 104d (6.5070.01) 105e

23.270.9d 2.4270.07 (1.9870.07) 102 1.97 103d (2.270.1) 104e

1.3 0.5870.01 (8.0470.04)7103

k29:Zn(OMe) (dm3 mol1 s1)

(2.2570.05) 104e

a

Values of the pK a in water from ref. 42. 3 Values of the ss pK a in methanol from refs. 16f, 6a and 6b.cValues of kLa 2 determined in a 17 mmol dm N-ethylmorpholine buffer at ss pH 9.1. d Values of k2 from ref. 17b.eObtained from duplicate initial rate measurements monitored by 1H NMR in CD3OD as described in ref. 17c. b

Table 10 Second-order rate constants (k2) for the methanolysis of phosphorothioates 21 and 3 promoted by methoxide, La3+ and 9:Zn2+(OCH3) in methanol solvent, T ¼ 25 1C Aryl Sphosphorothioate

pKaa

b s s pK a

p-nitro (21b, 3) 3,5-dichloro (21d) p-chloro (21c) p-ﬂuoro (21e) p-H (21a) p-methoxy (21f)

4.61 5.07 5.97 6.54 6.68 6.95

8.4c 8.9 10.1 10.7 10.9c 11.2

a

kOMe 2 (dm3 mol1 s1) 0.12d 0.15270.001 (1.8870.03) 102 (1.1170.02) 102 4.8 103d (2.2270.03) 103

3 1 1 s ) kLa 2 (dm mol

12.4d 14.070.5 1.2370.05 0.4670.01 0.48d (9.370.2) 102

k29:Zn(OMe) (dm3 mol1 s1) 0.8470.01 0.9770.01 (11.670.01) 102 (5.3470.06) 102 (4.270.1) 102 (1.4670.04) 102

Aqueous pKa values from Hong, S.-B. and Rauchel, F.M. (1996). Biochemistry 35, 10904. Values of the ss pK a in methanol computed from 2-point linear regression pK(MeOH) ¼ 1.2 (pK(waa a

b

ter))+2.83. c

The experimental valuses ss pK a from Clare, B.W., Cook, D., Ko, E.C.F., Mac, Y.C. and Parker, A.J. (1966). J. Am. Chem. Soc. 88, 1911. d The value of kLa 2 from ref. 17b.

plots that is indicative of a change in the rate-limiting step commencing with the lowest ss pK a compounds. However, more detailed analysis indicates this is probably not so since non-linear Hammett and Brønsted behaviors have been observed before for the hydroxide reactions of substituted aryl benzoates47 and acetates.48 In addition Schowen49 reported that log k2 values for methoxide reactions of aryl acetates and carbonates in methanol are not linearly related to the ss pK a values for ionization of the corresponding phenols. This sort of curvature in the Brønsted plots does not result from a change in mechanism or rate-limiting step but from a greater importance of the resonance and inductive interactions in the equilibrium acid dissociation constants (on which the Hammett and ss pK a values are based) than in the kinetic processes where less charge development occurs in the rate-limiting TS.47–49 Since

298

R.S. BROWN AND A.A. NEVEROV

2.5 (dm3 mol-1s-1)

log k2La(OMe) or log k29:Zn(OMe)

5.0

La 0.0

Zn

-2.5

-5.0 8

9

10

11

12 13 ArOH

14

15

s s pKa of

Fig. 7 Brønsted plot of the log second-order rate constant for La3+- and 9:Zn2+(OCH3)catalyzed methanolysis of phosphates 20 (including 1) vs. the ss pK a values for the corresponding phenols; linear regressions through the La3+ and 9:Zn2+(OCH3) data (excluding 20b) give gradients of –(1.4370.08) (solid line, &) and (1.1270.13) (dashed line, J), respectively. Note that points on lower right for the p-methoxy derivative (20f) are coincident. Reproduced from ref. 17c with permission.

log k2La or log k29:Zn(OMe) (dm3mol-1s-1)

1 La 0 Zn -1

-2

8

9 s pK s a

10 of ArSH

11

Fig. 8 Brønsted plot of the log second-order rate constant for La3+- and 9:Zn2+(OCH3)catalyzed methanolysis of phosphorothioates 21 (including 3) vs. the ss pK a values for the corresponding thiophenols; linear regressions through the La3+ and 9:Zn2+(OCH3) data (excluding that for 3) give gradients of 0.8770.10 (solid line, J) and 0.7470.06 (dashed line, J), respectively. Reproduced from ref. 17c with permission.

METAL-CATALYZED ALCOHOLYSIS

299

the visible curvature in the Brønsted plots for methanolysis and metal-catalyzed methanolysis seems to occur at the same point, a plot of the log second-order rate constants for the two processes should be a straight line if the curvature is an artefact of the ss pK a scale. Indeed, as is shown in Fig. 12, these plots for the phosphates and phosphorothioates are linear, indicating that nucleophilic attack of the metal-coordinated alkoxide and methoxide behave the same way in response to varying the leaving group without a change in rate-limiting step over the series. The above gives a mechanism based justiﬁcation for excluding the lower ss pK a datum for linearization of the Brønsted plots for 20b and 21b for each of the methoxide and metal-catalyzed methanolysis reactions. The blg values of –0.70 and –0.76 for the methoxide reaction of the phosphate triesters and phosphorothioates, respectively, can be compared with the corresponding blg values of 0.43 and –0.44 for hydroxide attack on diethyl aryl phosphate triesters45,50 and –0.42 for hydroxide attack on diethyl S-aryl phosphorothioates.45 The relatively low blg values of 0.4 obtained with these nucleophiles is consistent with little cleavage of the P–OAr bond in the TS, and supports a two-step mechanism where there is a rate-limiting k1 step to form a pentacoordinate intermediate with preferential breakdown of the intermediate to product. In the case of the HO reacting with diethyl aryloxy phosphates Ba-Saif and Williams50 judged that HO displacement of aryloxy leaving groups was probably concerted although with little cleavage of the ArO–P bond. This conclusion is supported by the 18O-phenoxy kinetic isotope effect of 1.006 for hydroxide-promoted cleavage of paraoxon (1) which was interpreted51 as being consistent with a bond order of 0.75 for the P–OAr bond in the ‘SN2-like transition state of an associative mechanism with concerted, asynchronous departure of the leaving group’.

Methoxide reaction. For the methoxide reactions of the aryl phosphates 1 and 20 and the diethyl S-aryl phosphorothioates 21, the blg values are more negative by 0.3 unit than is the case for the hydroxide reactions.45,50 Such might suggest that there is more cleavage of the P–XAr bond in the transition state than is the case for the hydroxide reactions. Applying the ‘effective charge treatment’ described by Jencks52 and Williams53 to the Brønsted blg value suggests a process for the aryloxy phosphate triesters where the rate-limiting transition state has appreciable changes in the P–OAr bond. This could be result from a two-step process with CH3O attack being largely rate-limiting due to the fact that the methoxide nucleophile is a far poorer leaving group and better nucleophile than any of the aryloxy anions,54 or just as likely with a concerted process as shown in Scheme 5. It is instructive to consider the progress of the P–OAr bond cleavage in the TS in terms of the Lefﬂer parameter, a, which measures the ratio of the Brønsted blg for the TS relative to the beq for equilibrium transfers of acyl or phosphoryl groups between oxyanion nucleophiles. In the case of the transfer of the (EtO)2P ¼ O group, the beq value is 1.8750 which comes about because the O–Ar oxygen in the starting material has a net effective charge of +0.87. When methoxide is the nucleophile, the Lefﬂer parameter of blg/beq ¼ 0.37 suggests that the P–OAr bond change has progressed 37% of the way from starting material to product.

300

R.S. BROWN AND A.A. NEVEROV

CH3O- +

O α=0.37

O

-1

+0.87 EtO

P

CH3O

OAr βlg = −0.70

EtO

EtO

P

+ -+

OAr

OEt

O CH3O EtO

P

-1 + ArO OEt

Scheme 5 A proposed concerted mechanism for methoxide promoted methanolysis of carboxylate esters with displacement of aryloxy leaving groups.

For the methoxide reaction of the phosphorothioates, a similar sort of analysis for the reaction proceeding through a two-step or concerted process can be invoked. In this case there is no reported value for the effective charge on the S-atom in the ArS–P(QO) unit but based on comparison of the known effective charges on the S- and O-atoms of ArS–C(QO)CH3 and ArO–C(QO)CH3 of 0.4 and 0.7,53a one might expect that S is less positive than O in the case of the ArX–P(QO) unit. Assuming the effective charge on S is 0.5–0.6, the Lefﬂer a of 0.45–0.50 for a concerted P–SAr cleavage suggests that the change in the P–S bond has progressed about 50% of the way from starting material to product. Unfortunately, the data are not sufﬁcient that one can unambiguously tell the difference between a concerted reaction and one proceeding through a pentacoordinate intermediate. The situation might be diagrammatically shown by the MoreO’Ferrall Jencks diagram as in Fig. 9. This diagram is predicated on the expectation that the Q-corner, involving full cleavage of the P–XAr bond prior to methoxide attack, is sufﬁciently high in energy that any concerted process moves substantially toward the S-corner describing the pentacoordinate intermediate. If the intermediate is in fact formed, then either its formation (with a late transition state) or its breakdown (with an early transition state) could be rate-limiting, with the latter probably better ﬁtting Lefﬂer index data that shows the substantial progress in the P–XAr bonding. Metal– methoxide reactions of phosphates, phosphorothioates, phosphonates and phosphonothioates. The rate constant data for the (La3+(OCH3))2 and 9:Zn2+(OCH3)-catalyzed methanolysis of the phosphates and phosphorothioates are given in Tables 9 and 10, respectively, while those for the phosphonates (22a–f)17e and phosphonothioates (23a–e)17g are given in Tables 11 and 12 along with the corresponding methoxide data. The latter two series data are displayed in Figs 10 and 11 as Brønsted plots. O EtO

P

OAr

CH3 22 O EtO P SAr CH3 23

22 a Ar = (4-Cl-2-NO2)phenyl b Ar = (4-NO2)phenyl c Ar = (3-NO2)phenyl d Ar = (4-Cl)phenyl e Ar = phenyl f Ar = (4-OCH3)phenyl 23 a Ar = 3,5-dichlorophenyl b Ar = 4-chlorophenyl c Ar = 4-fluorophenyl d Ar = phenyl e Ar = 4-methoxyphenyl

METAL-CATALYZED ALCOHOLYSIS

301

Q CH3O- + (EtO)2P+(=O) +-XAr P---XAr dist.

P

CH3O-P(=O)(OEt)2 +-XAr

* *

CH3O- + (EtO)P(=O)XAr

* R

OCH3O-P(OEt)2

CH3O---P dist.

S

XAr

Fig. 9 A hypothetical More-O’Ferrall Jencks diagram for the attack of methoxide on O-aryl phosphate triesters (20) and S-aryl phosphorothioates (21). Note that the diagram for attack of a metal-coordinated methoxide would be similar, but Mx+-coordination would push the TS toward the S-corner, possibly stabilizing the pentacoordinated intermediate to the point that the reaction occurs stepwise with the likely rate-limiting step being breakdown. Table 11 Acid dissociation constants for the phenols in water and methanol, as well as second-order rate constants for the various methanolysis reactions of phosphonates 22a–e promoted by methoxide, La3+ and 9:Zn2+(OCH3) Aryloxyphosphonate 22a 22b 22c 22d 22e 22f a

pKa

s s pK a

kOMe (mol dm3 s1) 2

3 1 kLa s ) 2 (mol dm

k29:Zn(OMe) (mol dm3 s1)

6.32 7.14 8.39 9.38 10.00 10.20

10.64 11.30 12.41 13.59 14.33 14.70

14.370.1 (2171)a 2.7070.02 1.4470.02 0.11870.001 0.021770.0003 0.008470.0001

(2.6070.20) 104 (1.4570.03) 103 (2.6070.20) 102 3.9770.03 (4.070.1) 101 (1.4070.02) 101

51773 (510720)a 46.870.5 2271 0.4570.01 0.07270.002 0.01470.002

In DOCH3, determined at a 9:Zn2+(OCH3) ratio of 1:1:0.5.

The Brønsted plots for the (La3+(OCH3))2 and 9:Zn2+(OCH3)-catalyzed methanolyses of 20 and 21 shown in Figs 7 and 8 exhibit gradients that are much steeper than the corresponding Brønsted plots for the methoxide reactions. This is easily visualized in the log/log plot of the second-order rate constants of the metalcatalyzed reactions vs. the methoxide reaction which is shown in Fig. 12 where the gradient of the La3+ plot is 1.9470.10 while that for the 9:Zn2+(OMe) plot is 1.4970.11. The situation for the phosphorothioates is not nearly so pronounced since the similar plots for series 21 (not shown) has a gradient for the La3+ plot of 1.1570.10 while that for the 9:Zn2+(OMe) plot is 0.9970.06.17c δ− OAr

O M

P -

H3C

O

H3C

OAr P

OEt

OEt 24c

O MX+

X+

O

OEt

OEt 24i

302

R.S. BROWN AND A.A. NEVEROV

H

α = 0.60

N

OEt

O

N H

OAr

P OEt

Zn2+ -

OCH3

N H

25

The large negative blg values for the metal ion-catalyzed methanolysis of the phosphate esters suggests a process where there is considerable cleavage of the leaving group in the transition state, far more so than is the case for the methoxide reaction. In the cases of the carboxylate esters that we have reviewed in section ‘Mechanism of alcoholysis of carboxylate esters’, it appeared that the 9:Zn2+(OCH3)- and (La3+(OCH3))2-catalyzed reactions produce an intermediate whose formation and breakdown could be rate-limiting depending on whether the leaving OR/OAr group was actually stabilized by metal ion coordination. However, the situation with the present phosphate esters is clearly different because there is a large dependence of the rate on the leaving group in the ss pK a region where blg is zero for the reaction of the carboxylate esters. Accordingly we suggest that, for the metal-catalyzed methanolysis of these phosphates, there is little evidence for a

Table 12 Second-order rate constants for the methanolysis of phosphonates 23a–e catalyzed by methoxide, (La3+(OCH3))2, and 9:Zn2+(OCH3) in methanol at T ¼ 25oC Phosphonothioate 23a 23b 23c 23d 23e Et)(CH3)P( ¼ O)SCH2CH2NEt2

of thiol (methanol)a

kOMe 2 (mol1 dm3 s1)

k29:Zn(OMe) (mol1 dm3 s1)b

kLa 2 (mol1 dm3 s1)c

9.0870.04

2.1770.03 (1.9670.03)d 0.6070.01 0.1670.01 0.05470.002 0.04070.006

95.271.4 (75.774.6)d 7.8570.25 4.8370.11 2.6070.06 1.1670.03 38.0270.04f

(4.8470.09) 103

s s pK a

10.4770.01 11.0770.03 11.2870.14 11.9870.08 9.5470.04

e

(4.2570.12) 102 (1.6470.07) 102 (1.1070.02) 102 (3.5570.17) 101 (2.0670.09) 103f

a Experimental values from 2 mmol dm3 solutions titrated in methanol according to procedures in refs. 6 and 7. b Catalyst prepared in situ by adding 1 equivalent of each of Zn(OTf)2 and the triaza ligand along with 0.5 equivalent of tetrabutylammonium hydroxide in methanol to form solution at ss pH 9.1; kinetics determined in duplicate at 5 [catalyst] ranging from 0.4 to 2.0 mmol dm3. c Catalyst prepared in situ by adding a stock solution of La(OTf)3 in methanol to a 17 mmol dm3 Nethylmorpholine with perchloric acid in a 4:1 ratio to solution at ss pH 9.1; kinetics determined in duplicate at 5 [La3+]total from 0.4 to 2.4 mmol dm3. d In d1-methanol. e The value predicted is 2 103 mol1 dm3 s1 under highly basic conditions based upon a comparison with data for for O-ethyl S-ethyl methylphosphonothioate given in Yang, Y.-C., Berg, F.J., Szafraniec, L.L., Beaudry, W.T., Bunton, C.A. and Kumar, A. (1997). J. Chem. Soc., Perkin Trans. 2, 607. f Predicted kcatalyst computed using Equations (32) and (33). 2

METAL-CATALYZED ALCOHOLYSIS

303

log k2 (dm3mol-1 s-1)

5.0

2.5

0.0

-2.5 10

11

12 s s pKa

13

14

15

phenol

Fig. 10 Brønsted plots of the second-order rate constants for methoxide and metal ioncatalyzed methanolysis of 22a–f at 25 1C. (La3+(OCH3))2 (m, gradient ¼ 1.2670.06), 9:Zn2+(OCH3) (&, gradient ¼ +1.0670.09), OCH3 (., gradient ¼ 0.7670.06). Reproduced with permission from ref. 17e.

logk2 (dm3mol-1 s-1)

5.0

2.5

0.0

-2.5

9

10 s s pKa

11

12

of ArSH

Fig. 11 Brønsted plots for log kcatalyst vs. ss pK a of aryl thiol for the methanolysis of 23a–e, 2 3+ K, (La ( OCH3))2, gradient ¼ 0.7570.01; m, 9:Zn2+(OCH3), gradient ¼ 0.6670.04; ~, OCH3, gradient ¼ 0.6570.10. Redrawn from ref. 17g.

change in rate-limiting step for these aryloxy leaving groups. This leaves open the possibility that the reactions with the metal ions are proceeding via a hypothetical concerted TS 24c, but it cannot exclude the possibility that the reaction is stepwise with a metal stabilized pentacoordinated phosphorus intermediate (24i), the formation or breakdown of which could be rate-limiting throughout the series. This possibility is shown diagrammatically in Fig. 9. Although we have never observed evidence of saturation kinetics in any of the metal-catalyzed reactions of the neutral phosphorus esters,17 it is difﬁcult to envision

304

R.S. BROWN AND A.A. NEVEROV

3

log k2La or log k29:Zn(OMe) (dm3mol-1s-1)

2 1 0 -1 -2 -3 -4 -4

-3

-2

-1

0

log k2OMe (dm3mol-1s-1) 9:Zn(OMe) Fig. 12 Plots of log kOMe vs. log kLa (J) for the methanolysis of 2 2 (&) or log k2 phosphate esters (20, including paraoxon 1) at 25 1C: gradient for La3+ plot is 1.9470.10; slope for 9:Zn(OMe) plot is 1.4970.11 all data included. Note data points for the p-methoxy derivative (20f) in lower left corner are coincident. Reproduced with permission from ref. 17c.

a mechanism where the ion does not bind to the phosphate to provide Lewis activation toward attack. Indeed, phosphate complexation of lanthanides and actinides is well known55 and structures are reported for Zn2+-complexes of phosphine oxides56 and tritoluoyl phosphate57 where coordination does occur through the PQO unit. Two of the most likely mechanistic possibilities are shown in the equations below and involve a preequilibrium binding of the metal ion to the PQO unit, Equation (18), followed by a concerted displacement Equation (19), or a stepwise formation of a metal-coordinated pentacoordinated intermediate, Equation (20). In fact, derivation of the kinetic expression corresponding to the binding and two-step process of Equations (18) plus (20) gives the identical expressions as derived in Equations (15–17) for the carboxylate esters, while the concerted mechanism corresponding to Equations (18) plus (19) has the identical kinetic expression as given in Equation (16). In what follows we will deal with each of the possibilities. Kb MX+-(− OCH3):(EtO)2P(=O)XAr MX+-(−OCH3) + (EtO)2P(=O)-XAr (18) MX+-(−OCH3):(EtO)2P(=O)XAr

k1

P

(19)

METAL-CATALYZED ALCOHOLYSIS

MX+-(−OCH3):(EtO)2P(=O)XAr

305

k1 k-1

MX+- − OP(OCH3)(OEt)2XAr k2 MX+ + (EtO)2P(=O)(OCH3) + − XAr

(20) Concerted mechanism?. For the concerted mechanism, the kinetic expression (16) is used from which it can be seen that the overall Brønsted relationship is log kobs 2 ¼ flog C b þ log C 1 g þ ðbb þ b1 ÞpK a

(21)

where bb and b1 are the Brønsted b-values for the binding step and kinetic step and pKa represents the ss pK a values for the phenol or thiophenol in methanol. The actual linear regressions are: s HOAr log kLa 2 ðphosphates 20Þ ¼ ð17:60 1:07Þ ð1:43 0:05Þs pK a

(22)

s HOAr log k9:ZnðOMeÞ ðphosphates 20Þ ¼ ð13:05 1:59Þ ð1:12 0:13Þs pK a 2

(23)

s HSAr log kLa 2 ðphosphorothioates 21Þ ¼ ð8:93 1:01Þ ð0:87 0:10Þs pK a

(24)

s HSAr log k9:ZnðOMeÞ ðphosphorothioates 21Þ ¼ ð6:64 0:69Þ ð0:74 0:06Þs pK a 2

(25)

from which it can be seen that (bb+b1) assumes large negative values which correspond to a composite measure of the inﬂuence of the leaving group on the preequilibrium binding step and the kinetic step. It is difﬁcult to predict an exact value for the bb and what few data there are available in the literature58 suggest that this should be (+), but probably not large because electron donors are expected to promote the binding, but these are somewhat far away from the PQO binding site. Rackham reported that the europium shift reagent Eu3+(2,2,6,6-tetramethylheptane-3,5-dione)3 binds trimethyl phosphate and triphenyl phosphate with constants of 348 and 23.2 dm3 mol1, and suggested that the fall in the latter’s binding constant can be attributed to steric bulk and the inductive withdrawal of the phenoxy group relative to the methoxy group.59 Du Preez and Preston have reported that the extraction into toluene of ScIII, YtIII and the trivalent LnIII ions from aqueous nitrate solutions by coordination to neutral organophosphorus (P ¼ O) compounds correlates with the Taft s*-values of the substituents.55d These considerations indicate that for the La3+ or 9:Zn2+(OCH3) promoted methanolysis of the phosphates, the observed blg may be a lower limit measure of b1 because it will be offset by the positive bb. The general mechanism for the (La3+(OCH3))2-catalyzed concerted process is given in Scheme 6 in which, for the sake of visual clarity, we have omitted the methanols of solvation on each La3+ as well as any associated counterions. As was proposed for the carboxylate esters, a methoxy group bound between two La3+ ions will not be sufﬁciently nucleophilic to attack the phosphate,60 so we propose that the

306

R.S. BROWN AND A.A. NEVEROV CH3

CH3

O

O+

La3+

La3+

O-

Kb

O-

EtO

P

O-

EtO 20, 21

CH3

XAr

La3+

La3+

XAr

P

O CH3 complex

OEt

OEt

+HOCH3 -H+ -phosphate CH3

CH3 O-

La3+ O

La3+

O

P

-XAr-

O

k1

La3+ P

O H3C

CH3

La3+ α=0.76

O-

XAr

O-

La3+ O

La3+ O-

P

EtO OEt H3C

EtO TS

OEt

CH3

EtO

XAr

OEt

open complex

Scheme 6 A proposed mechanism for the concerted La3+ (OCH3)2-catalyzed methanolysis 2 reaction of phosphate triesters with XAr leaving groups.

pre-equilibrium binding of phosphate and metal ion to form the complex induces opening of one of the methoxy bridges to reveal a kinetically active open complex. Since the Lefﬂer parameter, a, for the La3+-catalyzed is blg/beq ¼ 1.43/ 1.87 ¼ 0.76, the transition state for this reaction (24c) has extensive cleavage of the P–OAr bond which would be consistent with a concerted reaction within the complex as drawn in Scheme 6. Catalytic turnover requires a ﬁnal dissociation of the diethyl methyl phosphate with the reformation of (La3+(OCH3))2. In the case of the 9:Zn2+(OCH3)-catalyzed reaction a similar concerted mechanism might be envisioned but this time the transition structure (shown above as structure 25) will be ﬁve-coordinate at Zn and P with a Lefﬂer a of blg/beq ¼ 1.12/1.87 ¼ 0.60. This also signiﬁes extensive dissociation of the P–OAr bond in the transition state, but less so than in the case of La3+ catalysis. In the case of La3+- and Zn2+-catalyzed methanolysis of the phosphorothioate esters the observed blg values of 0.87 and 0.74 also signify an associative mechanism with some departure of the leaving group, but it is difﬁcult to assign the extent of the bond cleavage since the beq value is not known for the phosphoryl transfer between thiol and oxygen nucleophiles. Stepwise mechanism?. The possibility exists that the metal-catalyzed reaction really proceeds as described in Equations (18) plus (20), where the actual kinetic steps involve formation of a pentacoordinate phosphorus intermediate, suggested to be akin to 24i. When viewed within the context of the More-O’Ferrall Jencks

METAL-CATALYZED ALCOHOLYSIS

307

diagram of Fig. 9, the S-corner of the diagram should be stabilized by association of the anionic intermediate with the electropositive metal ion driving the reaction further to a stepwise one. From the general steady state expression of Equation (15) can be deduced two limiting Brønsted relationships given in Equations (21) and (26) log kobs 2 ¼ flog C b þ log C 1 þ log C 2 log C 1 g þ ðbb þ b1 b1 þ b2 ÞpK a

ð26Þ

where the rate-limiting steps are, respectively, formation of the pentacoordinate intermediate and its breakdown. In the former case, all the arguments for blg ¼ bb+b1 are essentially the same as for the concerted mechanism, but with the large negative blg now referring largely to the kinetic step of formation of the metal-coordinated pentacoordinate intermediate. In this case the Lefﬂer index blg/beq ¼ 1.43/ 1.87 ¼ 0.76 is difﬁcult to rationalize as it implies a very extensive loosening of the P–XAr bond. However, under the likelihood that the rate-limiting step is the breakdown of the pentacoordinate intermediate, then blg ¼ (bb+b1b1+b2) and the Lefﬂer index makes more sense since the blg now explicitly includes a term relating to the step that involves cleavage of the intermediate. The (La3+(OCH3))2 and 9:Zn2+(OCH3) promoted methanolyses of the phosphonates (22)17e and phosphonothioates (23)17g generally follow the same sort of trends as the phosphates and phosphorothioates discussed above so they need not be discussed in great detail here. Analysis of the linear Brønsted plots for the phosphonates 22 gives the relationships shown in Equations (27)–(29) which shows the common trend that the blg observed for the metal-catalyzed reactions are greater than that of the methoxide reactions. Since the Lefﬂer parameter, a, for the La3+-catalyzed ¼ ð9:24 0:81Þ ð0:76 0:06Þss pK HOAr ; log kOMe 2 a s HOAr ; log kLa 2 ¼ ð17:78 0:84Þ ð1:26 0:06Þs pK a

r2 ¼ 0:9896; 6 data

(27)

r2 ¼ 0:9716; 6 data

(28)

¼ ð14:04 1:17Þ ð1:06 0:09Þss pK HOAr ; log k9:ZnðOMeÞ a 2 17e

r2 ¼ 0:9734; 6 data

(29)

process is blg/beq ¼ 1.26/1.5 ¼ 0.84, the transition state has extensive cleavage of the P–OAr bond which is best analyzed at this stage in terms of the concerted process or a two-step mechanism where breakdown of the Mx+-pentacoordinate intermediate is rate-limiting. In the case of the 9:Zn2+(OCH3)-catalyzed reaction a similar mechanism is envisioned but this time the transition structure will involve ﬁve-coordinate Zn2+ with a Lefﬂer a of blg/beq ¼ 1.06/1.5 ¼ 0.7. Catalysis of the cleavage of the phosphonothioates (23) also follows linear Brønsted relationships (Equations (30–32)) although the gradients of the metalcatalyzed processes are not nearly as steep as those of the phosphonates discussed above. This trend of steeper gradients for the OP esters bearing OAr leaving groups relative to SAr leaving groups seems to be a general one. The reasons for this are not immediately clear although it may be related to the fact that the O–Ar oxygen of the starting ester may be signiﬁcantly more positively charged than the S–Ar sulphur, i.e. beq (OAr)obeq (SAr) so that the lower Brønsted blg values still represent a very

308

R.S. BROWN AND A.A. NEVEROV

signiﬁcant cleavage of the P–Ar bond in the transition state. log kOMe ¼ ð6:39 1:12Þ ð0:65 0:10Þ ss pK HSAr 2 a s HSAr log kLa 2 ¼ ð10:51 0:10Þ ð0:75 0:01Þ s pK a

r2 ¼ 0:930

(30)

r2 ¼ 0:999

(31)

¼ ð7:89 0:41Þ ð0:66 0:04Þ ss pK HSAr log k9:ZnðOMeÞ a 2

r2 ¼ 0:990

(32)

Relative to the reactions of the corresponding O,O-diethyl O-aryl phosphate triesters with the same leaving groups,17c the phosphonates are about 100-fold more reactive with toward methanolysis promoted by methoxide, La3+ and the 9:Zn2+(OCH3) systems. Nevertheless, the reaction rates are quite spectacular considering that the catalysts are maximally operative at ss pH values near neutrality (8.34) in methanol. For the most reactive substrate (22a), a solution containing 2+ 1 mmol dm3 of catalyst (La3+ ( OCH3)) accelerates the meth2 ( OCH3)2 or 9:Zn anolysis relative to the background methoxide reaction at ss pH optimum values of 9.1 by 8.5 107 and 1.7 106 times, leading to t1/2 values of 0.026 and 1.33 seconds, respectively. Compounds of this general structure can be used as simulants for OP chemical warfare agents such as sarin (26), soman (27) and the V-agents such as VX (28). The fact that the metal-catalyzed methanolysis reactions of the neutral OP esters described above is so efﬁcient suggests that this methodology might prove effective in decomposition of OP CW agents, particularly in situations where neutral conditions and ambient temperature is desirable for decontamination of sensitive equipment. While we have not yet tested these systems on live agent, we can use the data here to predict the reactivity of the particularly noxious VX agent. To this end, we can use the Brønsted relationships of Equations (31) and (32) determined for the catalyzed reactions of the phosphonothioates (23) to predict the rate constants shown in O

O (CH3)2CHOP

F

CH3 26 3+

(CH3)2CHCHOP F CH3 27

CH3

O EtO

P

SCH2CH2N(CH(CH3)2)2

CH3 28

2+

Table 12 for the La and 9:Zn -catalyzed methanolysis of VX. Based on the experimental ﬁrst ss pK a of 9.54 for ionization of HSCH2CH2NEt2, which we feel is a good model for the HSCH2CH2N(CH(CH3)2)2 in 28, the predicted half-times for methanolysis of VX are 18 and 0.33 seconds in the presence of 1 mmol dm3 of 9:Zn2+(OCH3) or (La3+(OCH3))2.

6 Transition metal ion and La3+-catalysis of the alcoholysis of phosphate diesters Phosphate diesters of general structure 29 are among the most resistant species to hydrolysis due the reluctance of hydroxide or even water to attack an anionic

METAL-CATALYZED ALCOHOLYSIS

309

61 (RO)2PO On the other hand, hydrolysis is achievable under highly acidic 2 species. conditions once the phosphate is neutralized by protonation. In principle, metalcatalyzed hydrolysis should be an

O O-

O

O

ROPO

nucleobase

P OR'

RO

30 X=H, DNA 31 X=OH, RNA

-O

O

29 O

P -O

X OR

effective strategy if the role of the metal ion is to complex the phosphate in a Lewis acid/base fashion thereby neutralizing its negative charge and permitting attack of an anionic or metal-coordinated HO. In several reports where metal-catalyzed hydrolysis of phosphate diesters has been discussed, a dual role of the metal ion was proposed, acting both as a Lewis acid activator and deliverer of an intramolecularly bound OH group to the metal-bound OQP unit.62,63 Metal ion catalysis is not a universally effective for phosphate diester hydrolysis in water although a large number of studies have been reported where these do catalyze the hydrolysis of selected phosphate diesters (mainly p-nitrophenyl substituted).62 In aqueous solution, due to the solvation effects of H2O on both the metal ion (complex) and phosphate diester, their binding is not particularly effective unless special efforts are made to enhance this.64 Owing to their great functional group stability phosphodiesters make up the important link in the RNA and DNA biomolecules (general structures 31 and 30) that are responsible for the storage of genetic information.11,65–68 While the respective half-times for hydrolysis of RNA and DNA at pH 7 and 25 1C are 110 and up to 100 billion years, Nature provides enzymes that promote the cleavage by up to a factor of 1015–16 thus affording some of the most spectacular rate enhancements known. Many of these enzymes have active sites comprising two or more metal ions (usually Zn2+ and in some cases Mg2+, Ca2+ and Fe2+) as exempliﬁed by phosphodiesterases such as ribonuclease H from HIV reverse transcriptase,66 30 ,50 -exonuclease from DNA polymerase I,67 the P1 nucleases68 and phospholipase C.11,65 In order to cast some light on the metal ion-mediated biological phosphoryl transfer reactions, intense research was directed at understanding the origins of catalysis of phosphate diester cleavage provided by metal ion systems.69–71 From that earlier work and more recent reports72–78 it is seen that dinuclear complexes are usually more reactive than the mononuclear counterparts, and there are four proposed roles by which the dinuclear catalysts promote the phosphoryl transfer reactions: (1) by double Lewis acid activation of a phosphate diester through M2+–OP(OR)(OR)O–M2+ coordination; (2) through bifunctional catalysis whereby the metal ion activates the bound phosphate and delivers a metal-coordinated hydroxide, alkoxide or oxide that serves as a nucleophile or base; (3) by electrophilically assisting the departure of the phosphate’s leaving group through metal-coordination; and (4) as an

310

R.S. BROWN AND A.A. NEVEROV

electrostatic reservoir of (+)-charge to interact favorably with the anionic phosphate to stabilize the transition state for the phosphoryl transfer reaction. As important as all these effects are, there must be an additional factor which has heretofore not been demonstrated experimentally, as none of the Zn2+-based model systems achieves a catalysis in water that even remotely approaches enzymatic rates. However, as will be shown below, metal-catalyzed alcoholysis reactions shed important light on the effectiveness of a reduced dielectric constant/polarity medium as a possible way that Nature could accelerate these reactions. As will be seen, there are very profound increases in the strength of phosphate diester binding to metal ions in alcohols relative to water, and once bound the cleavage of these exhibits spectacular rate accelerations over the background reactions.

METAL-CATALYZED ALCOHOLYSIS OF AN RNA MODEL

The phosphate diester 2-hydroxypropyl-p-nitrophenyl phosphate79 (32, HPNPP) has been employed in numerous studies as an easily obtained and studied surrogate for RNA. The general mechanism for its reaction is depicted in Scheme 7 involves an intramolecular attack of the 2-hydroxy group on the phosphorus with expulsion of the leaving group to produce a cyclic phosphate (33) which undergoes subsequent ring opening to give 34a and 34b. The intramolecular cyclization confers a kinetic reactivity toward base promoted methanolysis that is some 3000-fold larger than a comparable phosphate diester that lacks this group (e.g. methyl p-nitrophenyl phosphate).80 The alcoholysis of this sort of phosphate can be directly compared with the situation in water since the cleavage step leading to the cyclic phosphate does not incorporate solvent during the departure of the leaving group which is the observable step in the catalyzed reaction. Our initial studies of the La3+-catalyzed methanolysis of HPNPP81 immediately indicated that there were important features of the reaction which are not seen in the O NO2

HO

O

P O O

HO

O

P

34a

O-

O

k1

O

P O

O

O + O-

33

32

NO2 O

+ OCH3 HO

k2(k2')

O 34b

P

OCH3

O-

Scheme 7 Generalized mechanism for the methanolysis reaction of 32, HPNPP proceeding via intramolecular cylcization to form a 5-membered cyclic phosphate which undergoes subsequent opening via attach of methoxide.

METAL-CATALYZED ALCOHOLYSIS

311

80

5

10 kobs, sec-1

60

40

20

0 0

1

2

3

104La(OTf)3,

4

5

M

Fig. 13 Plots of kobs vs. [La(OTf)3] for the transesteriﬁcation of 32 (2.02 105 M) in the low s s s pH regime at three values, T ¼ 25 1C. s pH ¼ 5.0 (&), 5.6 (’) and 6.4 (K). Lines through the data computed from ﬁts to Equation (5) of ref. 81. Reproduced from ref. 81 with permission.

aqueous studies of the La3+-catalyzed hydrolysis.82 At low ss pH between 4.6 and 7.6, one observes that plots of the kobs vs. [La3+] for the methanolysis of 32 exhibit very pronounced saturation curves as shown in Fig. 13 indicative of strong binding to the metal ion. Fitting of the data to a strong binding equation83 gives the pseudoﬁrst-order rate constants (kcat) at each ss pH as well as lower limits for the Kd dissociation constant of 106 mol dm3. The binding in methanol is at least 104-fold larger than the reported value of 1.4 102 mol dm3 for La3+ and 32 in aqueous solution at pH 6.8582 and is a consequence of the reduced dielectric constant of the alcoholic medium which strongly favors ion association. The plot of the kobs data vs. ss pH (not shown) is linear (slope ¼ 0.9670.07) and yields a value of 7 3 1 1 kOMe s for the second-order rate constant for OCH cat ¼ 2.65 10 dm mol 3 pro3+ moted cyclization of La -coordinated 32. As remarkable as this number seems, the actual mechanism is more complicated and, as is demonstrated by the data shown in Fig. 14, actually proceeds through dimers which are simply formulated as (La3+(32))2. Fitting of the kobs vs. [La3+:32] kinetic data to a standard one-site binding model84 gives values for the dimerization constant (Kdimer) and maximum 3 3 1 4 1 catalytic rate constant (kmax s cat ) of (9.170.5) 10 dm mol ; (3.3070.07) 10 3 3 1 2 1 at ss pH 5.0 and (7.170.9) 10 dm mol ; (1.0270.06) 10 s at ss pH 6.7. Interestingly, the calculated saturation curves for the dimers shown in Fig. 14 indicate that at very low concentrations, the rate constant becomes negligible which indicates that the La3+:32 monomer is far-less reactive (at least 100-fold) than the (La3+(32))2 dimer. The fact that the binding constant does not change with ss pH indicates

312

R.S. BROWN AND A.A. NEVEROV 80

5

70 4

50

3

40 2

30

104 kobs s-1

104 kobs s-1

60

20 1 10 0

0 0

1 2 3 104 ([NaHPNPP] = [La(OTf)3]), M

Fig. 14 Plots of observed pseudo-ﬁrst-order rate constants for the methanolysis of increasing and equimolar [La3+] ¼ [32, HPNPP] at 25 1C and ss pH 5.0 (N,N-dimethylaniline buffer, &, right axis) or ss pH 6.7 (2,6-lutidine buffer, ’, left axis). Lines through the data computed from ﬁts to a standard one-site binding model. Reproduced from ref. 81 with permission.

the dimerization does not require a methoxide, but since the log kmax value cat increases linearly with ss pH, the cyclization of the phosphate within the dimer requires 1 equivalent of (base) in a speciﬁc base-catalyzed process. The secondorder rate constant for the methoxide-induced cyclization within the dimer is 1.18 108 dm3 mol1 s1and relative to simple methoxide-promoted cyclization of 33 ((2.670.2) 103 dm3 mol1 s1) the dimer provides an acceleration of 4.6 1010-fold in the low ss pH regime. It is interesting that the rate enhancement we see in methanol is 3 million-fold larger than that for La3+:32 cyclization in water, for which a rate enhancement of 14,000-fold over the simple hydroxide-promoted cyclization has been reported.82 This points to a signiﬁcant solvent effect of methanol favoring formation of reactive dimers with two metal ions, the dinuclear metal ion core being a known motif employed by Nature in promoting such reactions in the active sites of enzymes. While the acceleration afforded to the cyclization of 32 by La3+ in methanol is certainly spectacular, this is not a biologically relevant metal ion and its charge exceeds that of the natural metal ion Zn2+. Very recent investigations of Zn2+catalysis of the methanolysis and ethanolysis of 32 indicated that there were indeed interesting catalytic effects, and that the situation in pure ethanol is quite different.85 Shown in Figs 15 and 16 are plots of the pseudo-ﬁrst-order rate constant (kobs) for ethanolysis and methanolysis of HPNPP (32) as a function of [Zn2+]total when the [–OR]/[Zn2+] ratio is 0.5. This ratio was chosen to buffer the system at the half neutralization ss pH of 7 in ethanol8 and 9.5 in methanol at [Zn2+]total ¼ 1–2 mM7

METAL-CATALYZED ALCOHOLYSIS

313

kobs, s-1

2.0×10-2

1.0×10-2

0 0

2.0×10-4

4.0×10-4 6.0×10-4 [Zn2+]t, M

8.0×10-4

Fig. 15 Plot of kobs vs. [Zn2+]total for the decomposition of 32 (8 106 mol dm3) in anhydrous ethanol at [OEt]/[Zn2+]total ¼ 0.5. Fitting the data to a universal binding equation gives an apparent dissociation constant Kd for Zn2+:32 of (6.270.1) 105 mol dm3 and a kmax of 1.88 102 s1; r2 ¼ 0.9811. Reproduced with permission from ref. 85.

3.5×10-3 3.0×10-3

kobs, s-1

2.5×10-3 2.0×10-3 1.5×10-3 1.0×10-3 5.0×10-4 0 0

1.0×10-3 2.0×10-3 3.0×10-3 4.0×10-3 5.0×10-3 [Zn2+], M

Fig. 16 A plot of kobs vs. [Zn2+]total for the cyclization of 32 (4 105 M) in anhydrous methanol at [OMe]/[Zn2+]total ¼ 0.5. Fitting the data to a universal binding equation gives an apparent dissociation constant Kd for Zn2+:32 of (4.2770.03) 104 mol dm3 and a kmax of 3.52 103 s1; r2 ¼ 0.9967. Reproduced with permission from ref. 85.

and in each case the active Zn2+-species contains 1 equivalent of alkoxide. The data appear to exhibit saturation behavior and can be ﬁt by the universal binding equation83,85 where Kd refers to the dissociation constant for Zn2+:32 Zn2++32. Although the ﬁts to the data shown on Figs 15 and 16 appear satisfactory, we know the situation is more complicated than can be assessed by this simple one-site binding model as we need to consider also the various equilibria of Zn2+-association with alkoxide and its oligomerization that produces dimers and higher-order

314

R.S. BROWN AND A.A. NEVEROV

aggregates. Nevertheless, what can be established is that at [Zn2+]total 4 0.4 mmol dm3 and 5 mmol dm3 in ethanol and methanol, respectively, essentially all the HPNPP is complexed to Zn2+. This binding is very much stronger than seen in water where it has been reported38b that Zn2+-complexes bind only weakly to phosphate diesters, the reported Kb values being o0.5 dm3 mol1. The Zn2+-catalyzed process in ethanol is more complicated than in methanol because there is evidence for a higher than ﬁrst-order dependence on [32]. For example, if we determine the kobs for the reaction starting with an initial [32] ¼ [Zn2+:0.5(NaOEt)] ¼ 0.4 mM, concentrations where all the phosphate is bound as a 1:1 complex, and then increase the concentration of each keeping the [32]/[Zn2+:0.5(NaOEt)] ratio at unity there is additional saturation shown in Fig. 17 indicative of a process where two (Zn2+:32) complexes come together to form a kinetically active dimeric species (Zn2+:32)2. This behavior is not seen in analogous experiments with 32 and Zn2+:0.5(NaOMe) in methanol, nor is it reported in water86 suggesting that those solvents do not engender the formation of kinetically active spontaneously formed dimers, at least at the concentrations employed. The similarity in the La3+-promoted cleavage of 32 in methanol that was reported above to what is observed with Zn2+ in ethanol suggests that the reactive dimers are probably doubly activated as (Zn2+:32)2 as shown in Scheme 8 where the actual catalytic role of the second phosphate in the dimer is to act as a template in stabilizing the dinuclear Zn2+-core. If the template role is operative, any phosphate

kobs, s-1

2

1

0 0.0000

0.0005

0.0010

0.0015

0.0020

0.0025

[Zn2+]bound (M)

Fig. 17 A plot of kobs vs. increasing [Zn2+]bound for the cyclization of 32 in anhydrous ethanol with [HPNPP]/[Zn2+]total/NaOEt ¼ 1:1:0.5. The [Zn2+]bound data used are corrected for the amount of unbound Zn2+ using the stability constant Kd of (6.270.1) 105 M as described in ref. The dissociation constant Kdimer for the presumed (32:Zn2+)2 complex is (6.270.5) 104 M and the maximum kmax for reaction of the dimeric complex is cat 2.9270.06 s1, r2 ¼ 0.9586. The measured ss pH at 2 mM is 7.1. Reproduced with permission from ref. 85.

METAL-CATALYZED ALCOHOLYSIS O OAr H O P OEt -O Zn2+ +

HO 32:Zn

2+

315

O

Kd

O O P OAr OEt -O 32:Zn+(-OEt) Zn2+ HO

H P OAr OEt O -O 2+ Zn Zn2+ EtO- -O O P ArO O OH HO

O

O-

H P OAr OEt O -O

H Zn2+ Zn2+ O -O O Et P ArO O - HOAr OH

kcatOEt O O O 2+

HOAr

H OEt

P -O

Zn

+ EtO- -O

Zn P O

O

O

2+

O kcat2OEt

O

P O

H O Et

2+

-O

Zn

Zn2+

-

-O O P ArO O

OEt OH

Scheme 8 Zn2+(OEt) promoted cleavage of 32 where two of the initially formed 32:Zn2+(OEt) complexes dimerize and subsequently undergo and intramolecular cyclization.

monoanion should be capable of performing this role which is veriﬁed by the observations85 that low concentrations of diphenyl phosphate (1 1062.5 104 M) exert a profound accelerating effect on the cyclization reaction of (Zn2+:32), the second-order rate constant being 790 dm3 mol1 s1 which is far too fast for any general base role. There are additional considerations that cast some important light on the chemical process proposed in Scheme 8. First, although the kinetics for appearance of p-nitrophenol are strictly ﬁrst order, each dimer releases two equivalents of p-nitrophenol in sequential steps, so the second release must be at least as fast or faster than OEt the ﬁrst, i.e. kOEt cat pkcat2. The phosphate products of the HPNPP cyclization and opening with ethoxide can also fulﬁll a template role so that in principle the product of the reaction actually serves to catalyze the loss of HPNPP by maintaining the dimeric form of the catalyst. Second, the observed pseudo-ﬁrst-order rate constant 1 (kmax at ss pH 7.1 at 2 mM [Zn2+:0.5(NaOEt)]. Consideration of the rates cat ) is 2.9 s excludes an external –OEt acting as a base since the observed process is at least 100 times faster than permitted by a diffusion limited abstraction of the 2-hydroxypropyl hydrogen by free (ethoxide) at the concentrations set at ss pH 7.1.87 The lack of general base assistance to the cyclization of 1 was noted before in processes not nearly as fast as that presented here on the bases of: (1) no observed general catalysis of cyclization of 32 by the strong base piperidine in the presence of metal ions;79 (2) no observed primary deuterium kinetic isotope effect on the cyclization of 32 when catalyzed by a dinuclear Zn2+-complex;75 and (3) no observed buffer catalysis of the cyclization of 32 promoted by a mononuclear Zn2+-complex where a speciﬁc base-catalyzed process was proposed.64 Thus, the proposed mechanism in Scheme 8

316

R.S. BROWN AND A.A. NEVEROV

involves two relatively non-reactive HPNPP:Zn2+ complexes (one suggested to have a coordinated ethoxide or the kinetic equivalent of an O-deprotonated hydroxypropyl group and another to have coordinated ethanols). These associate to form a reactive (HPNPP:Zn2+)2 dimer with a Zn2+-coordinated ethoxide (or its kinetic equivalent) which acts as an internal base to deprotonate the HPNPP. This process would be facilitated by the dimeric nature of the complex where the large positive charge on the Zn2+-core stabilizes the anionic deprotonated propanolate75 which subsequently cyclizes with the expulsion of p-nitrophenolate. The latter in turn acquires a proton to recreate the Zn2+-OEt in the nascent complex. The second equivalent of p-nitrophenol is released from the complex in a similar two-step pathway summarized as kOEt cat2. ZN2+ LIGAND MODELS FOR DINUCLEAR ENZYMES PROMOTING THE CLEAVAGE OF RNA

As stated above, Zn2+ in alcohol in the presence of alkoxide forms dimers and oligomers which can sometimes complicate the kinetic analysis, particularly in the cases of weakly binding substrates. However, complexation to the triaza ligand 938 simpliﬁes the speciation as only 9:Zn2+(HOR) and its deprotonated form 9:Zn2+(OR) (where RQCH3, CH2CH3) are present in solution. The cyclization of 32 mediated by 9:Zn2+(OH) in water has been reported88 to have a second-order rate constant of 0.018 dm3 mol1 s1which is about three times less than the rate constant for the hydroxide promoted cyclization (0.065 dm3 mol1 s1).72 However, when the reactions with 9:Zn2+(OR) are conducted in methanol and in ethanol, additional facets of the reaction are revealed. Presented in Fig. 1885 is a plot of the kobs for the cyclization reaction of 32 in methanol vs. [9:Zn2+]total under conditions where the [OCH3]/[Zn2+]total ¼ 0.5 which sets the measured ss pH at 9.2–9.3. The plot shows upward curvature consistent with a process bimolecular in [9:Zn2+]total. Non-linear least squares (NLLSQ) ﬁtting of the data to the expression kobs ¼ k1 [9:Zn2+]total+ 3 kobs [9:Zn2+]2total, gives k1 ¼ 18.9 dm3 mol1 s1 and kobs 3 3 ¼ (1.870.4) 10 3 1 2 1 (dm mol ) s , where the best ﬁt line through the data is shown in Fig. 18. Similar upward curving plots are found when the [OCH3]/[Zn2+]total ratio is 0.3 and 0.7, which is consistent with the participation of both the (CH3O)Zn2+:9 and (CH3OH)Zn2+:9 forms of the catalyst, but when the ratio is 1.0 (ss pH 11.09), the plot (not shown) is linear, indicative of the involvement of only one molecule of the 3 1 1 (CH3O)Zn2+:9 form, with a gradient kobs s . In no case do 2 ¼ 16.970.7 dm mol we observe evidence of saturation binding in methanol up to a [9:Zn2+]total of 5 103 mol dm3, so the Kdis for dissociation of any 9:Zn2+:32 complex must be at least ﬁve times higher. The situation in ethanol is similar but the upward curvature is manifested at lower [9:Zn2+] as demonstrated in Fig. 18 by the plot of kobs for the cyclization of 32 vs. [9:Zn2+]total under conditions where the [OEt]/[Zn2+]total ¼ 0.5 which sets the measured ss pH at 8.170.2. At progressively higher concentration this plot85 shows a downward curvature suggestive of a saturation phenomenon which is not seen in

METAL-CATALYZED ALCOHOLYSIS

317

0.15

0.30

0.25

0.15

0.05

kobs, s-1

0.20

kobs, s-1

0.10

0.10

0.05

0.00

0.00 0

1

2

3

4

5

6

[9:Zn2+]total

Fig. 18 Plot of kobs vs. total [9:Zn2+]total for the cyclization of 32 (8 105 dm mol1) in ethanol (’, left y-axis) and methanol (J, right y-axis) under conditions where the [OR]/ [Zn2+]total is 0.5. Dotted line corresponds to process ﬁrst-order in [9:Zn2+], while curved line is the ﬁt in methanol to the expression kobs ¼ k1 [9:Zn2+] + kobs [9:Zn2+]2. Heavy solid line 3 through the solid squares is the ﬁt of the data to Equation (34). Reproduced with permission from ref. 85.

methanol nor water at these concentrations. That this behavior does not come from an aggregation of the 9:Zn2+ species or from inhibition by triﬂate counterion is veriﬁed by the fact that the plot of the kobs for ethanolysis of the weak-binding substrate p-nitrophenyl acetate (2) at [OEt]/[Zn2+]total ¼ 0.5 is linear throughout the concentration range of 0.1 and 6 mmol dm3 with a gradient of 0.8170.01 dm3 mol1 s1. Additional studies indicated that the active species is (EtO)Zn2+:989 which is consistent the situation in methanol36 where the secondorder rate constant for (CH3O)Zn2+:9 promoted methanolysis of paraoxon is 0.84 dm3 mol1 s1. Overall the behavior in both methanol and ethanol is consistent with a process where there are two 9:Zn2+ complexes in the transition state for methanolysis of HPNPP given in Scheme 9. Notably, a sigmoidal plot was observed before for the hydrolysis of ethyl p-nitrophenyl phosphate catalyzed by Co(III)–cyclen (cyclen ¼ 1,4,7,10-tetraazacyclododecane) and analyzed in terms of a process analogous to that in Scheme 9.90 Since the methanol data show the bimolecular process involves a k1 process that is ﬁrst-order in 9:Zn2+, and another process dependent on [(CH3O)Zn2+:9] and [(CH3OH)Zn2+:9], we can speculate that, for the k2 process, one of the units binds the substrate, and the other serves to deprotonate it perhaps leading to a ternary complex of 32:[(CH3OH)Zn2+:9]2. For technical reasons relating to the inability to stabilize the ss pH at a [OEt]/[Zn2+]total ¼ 1.0, we have not speciﬁcally tested to see whether the bimolecular behavior in ethanol depends on the

318

R.S. BROWN AND A.A. NEVEROV

2+

2+

-

32 + (Zn :9(HOEt)) + (Zn :9( OEt))

Kb

32:(Zn2+:9(HOEt):(Zn2+:9(-OEt) 32-:(Zn2+:9(HOEt):(Zn2+:9(HOEt)

k1

kcat

P

P

Scheme 9 A minimal scheme for the reaction of 32 with 9: Zn2+(OEt) and 9: Zn2+(HOEt) accounting for the reaction kinetics which are both ﬁrst order and second order in [catalyst].

presence of both [(EtO)Zn2+:9] and [(EtO)Zn2+:9] but based on the results in methanol, there is no reason to suggest it does not. Fitting of the Fig. 18 data to the expression derived for the process of Scheme 9 in Equation (33) where [A] and [B] are [(EtO)Zn2+:9] and [(EtO)Zn2+:9] (equal concentrations) yields k1 ¼ 3.69 dm3 mol1 s1, Kb ¼ 2.85 108 (dm3 mol1)2 and kcat for the fully bound complex ¼ 0.13 s1. kobs ¼ kcat K b ½A½B=ð1 þ K b ½A½BÞ þ k1 ½A

(33)

EXHALTED CATALYSIS OF METHANOLYSIS OF HPNPP PROMOTED BY A DINUCLEAR COMPLEX IN METHANOL

The interesting ability of ethanol and methanol to recruit two independent 9:Zn2+ complexes together in the transition state for an enhanced rate of cleavage of HPNPP suggests that tethering the CH3OH and CH3O forms of the 9:Zn2+ moieties together would change a formally trimolecular process to a bimolecular one with a signiﬁcant rate enhancement anticipated if the two forms react cooperatively. The dinucleating ligand 3591 was prepared by Kim and Lim92 who studied the catalytic effect of its bis-Zn2+-complex (35:2Zn(II)) for hydrolysis of two phosphate diesters. Unfortunately, in water this complex is only 1.2 and 4.4 times more effective than 9:Zn2+ in promoting the hydrolysis of bis-p-nitrophenyl phosphate and p-nitrophenyl phosphate at pH 7.92 It appears that this complex is not particularly stable in water and may lose a Zn2+-ion in solution and cannot form a stable 2:2 ligand:Zn2+-complex. This behavior is akin to what is observed with the Cu2+- and Zn2+-complexes of the analogous bis-9[ane]N3 ligand 36.93,94 On the other hand, this complex displays a remarkable activity for the cyclization of HPNPP in methanol.95 We have determined recently the structure of the dinuclear complex 35:2Zn(II) by H

N 2+ N Zn

N 2+ N Zn

N

N

H

H 35:2Zn(II)

H N 2+ Cu N N

H 2 H

H -Cu2+

2+ N N Cu N

(36:Cu(II))2 H

36:2Cu(II)

X-ray diffraction.96 The structure, shown in Fig. 19, has the two [12]aneN3 rings facing each other and binding two Zn2+-ions which are bridged by a single

METAL-CATALYZED ALCOHOLYSIS

319

Fig. 19 The structure of 35:2Zn(II):OH(CF3SO 3 )3:CH3OH as determined by X-ray diffraction. For clarity, the triﬂate counter ions and methanol solvate are not shown.

hydroxide anion: also in the unit cell are three triﬂates for charge neutrality along with a single molecule of methanol. We assume that the lyoxide-bridged form is the one that is present when the catalytically active form is generated in situ through the sequential addition of 1 equivalent of 35 and NaOCH3, followed by 2 equivalent of Zn(OTf)2. It is important to note that the in situ formation of the active complex takes about 40–50 min, as judged by the fact that the catalytic activity of the solution continues to rise for that period of time, after which it assumes a constant value for several hours. Such an unusually slow complex formation can be attributed to initial formation of the mononuclear complex 35:Zn(II) where Zn2+ ion is ‘sandwiched’ between two triazamacrocycles within a single ligand molecule as was reported for the 36:Zn(II) complex.94 Intramolecular dissociation of such a complex and coordination of the second Zn2+-ion is probably responsible for the relatively slow formation of the thermodynamically stable and kinetically active di Zn(II)-complex which we formulate as 35:2Zn(II):(OCH3). Shown in Fig. 20 is a plot of kobs for the cyclization of 32 vs. [CH3O] ¼ [35: 2Zn(II)] determined at a measured ss pH ¼ 9.570.1. The primary data (dashed line) show downward curvature, although this not due to HPNPP binding but rather to a speciﬁc ion effect of the triﬂate anions that suppresses the rate (since each 35:2Zn(II)-complex carries with it 4 OTf counterions). The methanolysis of paraoxon (1) as a function of [35:2Zn(II):(OCH3)] which also exhibits a downward curvature as shown in Fig. 21. That the weakly binding substrate (1) exhibits the identical curvature in the plot as does the potentially strong binding 32 suggests that the curvature is independent of the substrate, and is more likely dependent on the presence of anions that accompany the Zn2+-complex. Indeed, we have observed

320

R.S. BROWN AND A.A. NEVEROV 300

kobs,s-1

200

100

0 0.00

0.25

0.50

0.75

1.00

1.25

1.50

[CH3O-]=[35:2Zn(II)], mmol dm-3

Fig. 20 A plot of the observed pseudo-ﬁrst-order rate constants (kobs) for the methanolysis of HPNPP (4 105 mol dm3) as a function of [35:2Zn(II)] in the presence of 1 equivalent of added CH3O per complex giving ss pH ¼ 9.5, T ¼ 2570.1 1C. Dotted line is presented as a visual aid directed through all actual data collected at 280 nm (&) or 320 nm (J) which are the wavelengths for disappearance of HPNPP and appearance of p-nitrophenol; solid line is a linear ﬁt of the data corrected for inhibition by triﬂate counterions at 280 nm (’) or 320 nm (K). Reproduced with permission from ref. 95.

4.5 4.0 104kobs, s-1

3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 0.0

0.5

1.0

1.5

2.0

[CH3O-]=[35:2Zn(II)], mmol dm-3

Fig. 21 A plot of the observed pseudo-ﬁrst-order rate constants (kobs) for methanolysis of 2.5 105 M paraoxon vs. [35:2Zn(II)] in the presence of 1 equivalent of added CH3O per complex, ss pH ¼ 9.5, T ¼ 2570.1 1C. Dotted line is presented as a visual aid directed through all actual data (J); solid line is a linear ﬁt through the data (K) corrected for inhibition by triﬂate counterions. Reproduced with permission from ref. 95.

that the catalytic activity toward methanolysis of paraoxon of a solution containing 0.4 mM of the [(CH3O):Zn2([12]aneN3)2] complex does decrease as a function of increasing [Bu4N+(OTf)]. The kinetic data for this inhibition, when ﬁt to a one-site binding model shown in Equation (34) give an inhibition constant of 14.9 mM. This

METAL-CATALYZED ALCOHOLYSIS

321

constant was used to calculate the free kobs ¼ kmax K inhib =ð½Bu4 NOTf þ K inhib Þ

(34)

[catalyst] under the kinetic conditions and when the kinetic data of Figs 20 and 21 are corrected for triﬂate inhibition by plotting the kobs vs. free [catalyst], linear correlations are observed, consistent with no saturation binding of either substrate. The gradient of the linear plot in Fig. 21 for paraoxon is 0.31470.006 dm3 mol1 s1, while that in Fig. 20 is (2.7570.10) 105 dm3 mol1 s1. It is important to note that the latter rate constant is the largest second-order rate constant reported for the cyclization of HPNPP by any catalytic system reported to date. It is evident that there is a very large cooperative effect for the two Zn2+-ions which is brought about by their complexation to 35 and the medium effect in methanol, since the catalyst is 1.1 108-fold more reactive toward HPNPP than is methoxide (kMeO ¼ 2.56 103 dm3 mol1s1).81 A 1 mM solution of the 2 35:2Zn(II):( OCH3) catalyst at ss pH of 9.5 provides a t1/2 for decomposition of HPNPP of 2.5 ms, providing a remarkable acceleration of 2 1012 over the background methoxide reaction at that ss pH. Furthermore, the exalted catalytic ability of 35:2Zn(II):(OCH3) is not limited to cyclization of HPNPP, but is also evident for the cleavage of a DNA model, methyl p-nitrophenyl phosphate (MNPP) The kinetic data are shown graphically in Fig. 22 which is analyzed95 to afford a Kb value of 0.37 mmol dm3 and a kmax value of 4.1 102 s1 for decomposition of the MNPP:35:2Zn(II):(OCH3) complex at ss pH 9.5. Relative to the background methoxide reaction with MNPP (kOMe ¼ (7.970.6) 107 dm3 mol1s1) this constitutes 2 12 an acceleration of 10 -fold at that ss pH. In terms of comparison of the second-order

102kobs, s-1

3

2

1

0 0.00

0.25

0.50

0.75

1.00

1.25

[35:2Zn(II):(-OCH3)], mmol dm-3

Fig. 22 A plot of kobs for methanolysis of 4 105 M methyl p-nitrophenyl phosphate (MNPP) vs. [35:2Zn(II)] in the presence of 1 equivalent of CH3O per ligand showing a saturation behavior, ss pH ¼ 9.5, T ¼ 2570.1 1C. Line through the data calculated by NLLSQ ﬁts to a Michaelis–Mentin equation corrected for complex dissociation95 giving a binding constant of KM ¼ 0.37 mmol dm3 and a maximum rate constant for reaction of the MNPP:[(CH3O–):35:2Zn(II)] complex of kmax ¼ (4.170.1) 102 s1. Reproduced with permission from ref. 95.

322

R.S. BROWN AND A.A. NEVEROV

rate constants, deﬁned as kmax/Kb/kOMe , the 35:2Zn(II):(OCH3) catalyst is 2 8 1.4 10 -fold better than methoxide for promoting the methanolysis of MNPP. These accelerations exceed any previously reported catalysis of HPNPP cyclization or phosphate diester cleavage by several orders of magnitude, now approaching the realm of enzymatic catalysis. It is important to emphasize that this sort of catalysis is not seen with dinuclear Zn2+-containing catalysts in water, including 35:2Zn(II),92 and must be a consequence of a medium effect that optimizes substrate binding and catalytic transformation within a substrate/catalyst complex. Several dinucleating complexes76,97 of [9]aneN3 are known, but their Zn2+-complexes in water are, relative to the situation we observe here, only weakly catalytic toward the hydrolysis of phosphate diesters and toward the cyclization of HPNPP. Richard and Morrow76 have evaluated several of these and determined that only the di Zn2+-complex of 1,3-bis-N1(1,3,7-triazacyclononyl)propan-2-ol (37) had appreciable activity (120-fold) over its Zn:[9]aneN3 mononuclear analogue. Several other bis-complexes of the triazacyclononyl system showed only a 3–5-fold larger activity than the mononuclear complex. The di Zn2+-complex 38 is reported97 to promote the hydrolysis of bis-p-nitrophenyl phosphate (BNPP) with observed N Zn2+ N N

O

N Zn N

37

2+

N

N Zn2+ N N

N Zn2+ N N N N 38

H

saturation kinetics at pH 9.2, 35 1C and a Kb and kcat of 1.2 102 mol dm3 and 2.24 106 s1 for an overall second-order rate constant of kcat/Kb ¼ 1.87 104 dm3 mol1 s1. However, this value is only 7.5 times that of hydroxide98 indicating that very little exhalted catalysis is evident in water. Although HPNPP and MNPP are two closely related phosphate diesters in terms of charge, substitution and size, they exhibit far different appearances for the kobs vs. [35:2Zn(II)] plots as is evidenced by the data in Figs 20 and 22. Of course HPNPP is far more reactive than MNPP, but it should still require binding to the catalyst to achieve the very large accelerations observed for its cleavage by the dinuclear complex. The slower-reacting phosphate diester follows the expected pattern of saturation binding followed by a rate-limiting phosphate cleavage within the complex, but there is no visible saturation kinetics observed with HPNPP. The simplest explanation of the lack of observed saturation behavior in Fig. 20 is that the chemical step of cyclization of HPNPP is not rate-limiting, but some prior step is, such as productive binding. Shown in Scheme 10 is a working model for the binding and reaction of phosphate diesters promoted by 35:2Zn(II). Our current explanation is that there are two binding events, a rapid and reversible ﬁrst binding event, perhaps to form a transient complex with a single Zn2+-ion in the complex, followed by a rearrangement to form a doubly activated substrate, suggested to involve coordination of both Zn2+-ions to the phosphate. The latter mode of coordination has been shown to optimize metal ion catalysis of the cleavage of phosphate diesters2b and was also proposed in the La3+-catalyzed cleavage of (La3+:HPNPP)281 and the

METAL-CATALYZED ALCOHOLYSIS

S+

k1

Zn 35 Zn

S

k -1

Zn

Zn 35

323 k3

Zn

k2

35

S Zn

k-2

CH3O-

Product + Zn2:35

Scheme 10 A proposed simpliﬁed scheme for the three step reaction of HPNPP or MNPP with 35:2Zn(II) with two initial binding steps followed by a chemical step which releases p-nitrophenol. 60

+ H+

5.0

+ -OCH3

kobs, s-1

40 30

2.5

20

102kobs, s-1

50

10 0

0.0 0

1

2

[CH3O-]/[Zn2([12]aneN3)2 ratio

Fig. 23 A plot of the observed pseudo-ﬁrst-order rate constant for the methanolysis of 0.04 mM HPNPP (’, left axis) catalyzed by 0.2 mM 35:2Zn(II) or 0.04 mM methyl p-nitrophenyl phosphate (J, right axis) catalyzed by 0.4 mM 35:Zn(II) as a function of the [CH3O]/ [35:Zn(II)] ratio at 2570.1 1C. Experiments done by ‘pH jump’ method starting at a [CH3O]/ [35:Zn(II)] ratio of 1.0 (vertical dashed line, ss pH ¼ 9:5) and adding acid (left) or base (right). Reproduced with permission from ref. 95.

Zn2+-catalyzed cleavage of (Zn2+:HPNPP)2 in ethanol discussed above.85 For the slower reacting MNPP, the chemical cleavage step (represented by k3 and requiring a methoxide which is probably coordinated to one or both of the metal ions in 35:2Zn(II)) is relatively slow, so that both the pre-equilibrium steps are established and typical Michaelis–Menten behavior is observed with saturation at higher [35:2Zn(II)]. On the other hand, with the far more reactive HPNPP the chemical cyclization step, k3, is proposed to be faster than the k2 step in the concentration range of 35:2Zn(II) used here. In this event, the observed kinetics would be linear in [35:2Zn(II)] as is the case in Fig. 20, with kobs ¼ k1[35:2Zn(II)]k2/(k1+k2). This mechanism is also consistent with the base dependent behavior of the kinetics demonstrated in Fig. 23 for the reaction of MNPP and HPNPP catalyzed by 35:2Zn(II). The unusual increase in rate constant for the HPNPP reaction that accompanies addition of acid is not consistent with the cyclization step (k3) being rate-limiting. Rather, the increase in rate with added acid is more consistent with a process depending on binding the HPNPP to a greater amount of a complex devoid of an associated methoxide with its higher net positive charge attracting the negatively charged HPNPP. This requires that the chemical step of methoxide-dependent cyclization would be faster than the rearrangement step throughout most of the

324

R.S. BROWN AND A.A. NEVEROV

plot. At some point where the ss pH falls below a critical value and the complex contains little or no methoxide (or deprotonated HPNPP), the cyclization step slows and becomes rate-limiting which accounts for the discontinuity in the plot at the low [CH3O]/[ 35:Zn(II)] ratio. The simpliﬁed process described in Scheme 9 introduces another aspect of the medium effect that accelerates the chemical step of HPNPP cleavage to the extent that it is no longer rate-limiting with this catalyst. The reaction seems to be limited by an aspect of substrate binding, a phenomenon often observed in enzymatic systems where the catalytic steps are competitive with binding steps. While the secondorder rate constant of 2.75 105 dm3 mol1 s1 for the HPNPP reaction seems somewhat slow for a binding step involving simple ligand exchange on Zn2+ (known to be 107–108 dm3 mol1 s1), it is well within the range for more complex ions in which an internal rearrangement or breaking of intramolecular hydrogen bonds occurs, for example in salicylate (O2C–Ar–OH similar to O2P(OR0 )O–R–OH in the HPNPP anion) binding to Zn2+ where the rate constant for ligand exchange in water is reported to be 1.4 105 dm3 mol1 s1.99 Of course, this thesis remains to be tested and probably can be conﬁrmed or rejected through study of a series of close derivatives of HPNPP where the aryloxy group is varied which is currently underway in our laboratories.100

7

Conclusions

Metal ion-catalyzed hydrolytic processes have been studied for a long time, and many interesting systems have been explored which give valuable information about catalysis. However, with very few exceptions the catalysis afforded by these systems in water is disappointing when compared with enzymatic systems where a metal ion cofactor activates a substrate and a nucleophilic or basic group in an acyl or phosphoryl transfer process. It has been noted that bulk water may not be a good medium to approximate the medium inside the active site of an enzyme where it is now known that the effective dielectric constants resemble those of organic solvents rather than water. Our early studies of metal-catalyzed acyl transfer reactions were predicated on the idea the a reduced polarity/dielectric constant medium would allow one to better realize the catalytic potential of the metal ion by reducing the tightness of the solvation shell around the metal ion and its constituents, as well as allow a stronger interaction energy with substrate. The ﬁrst demonstrations of the pronounced catalysis of acyl transfers were from activated, and not particularly challenging, substrates such as acetyl imidazole which led to important understandings of the metal species in solution.10 These formed the basis for several additional studies of metal-catalyzed transesteriﬁcations of a wide variety of neutral and unactivated carboxylate esters and aryloxy phosphorus-based esters for which metal-catalyzed hydrolytic reactions are notoriously inefﬁcient. It now seems possible that this general area of study might have far-reaching implications for metal-catalyzed

METAL-CATALYZED ALCOHOLYSIS

325

transesteriﬁcation processes and biological ones which involve metal ions as cofactors for acyl and phosphoryl transfer reactions. Despite a great deal of effort that has led to an increased understanding of how enzymatic catalysis might occur, it is generally held that ‘none of several models so far described approaches the enormous catalytic efﬁciency of natural enzymes’.88 It is a venerable hypothesis that one mode by which enzymes could achieve exhalted efﬁciency is to utilize low polarity active sites that are specially tailored for the catalytic task at hand. The above account indicates that very simple systems comprising metal complexes and a medium effect engendered by the lower alcohols does give rate accelerations for acyl and phosphoryl transfers approaching those of enzymes. It remains to be seen whether these systems will work for more biologically relevant substrates. It is our hope that the ideas contained in this report will spur further work using the multiple effects of structure and medium to bring us closer to understanding the ways in which Nature performs such transformations.

Acknowledgements The authors are grateful to the myriad of undergraduate and graduate students, postdoctoral and summer researchers whose names appear on the publications from this laboratory. In addition, they acknowledge the ﬁnancial support of the Natural Sciences and Engineering Research Council of Canada, the Canada Foundation for Innovation, The United States Department of the Army, Army Research Ofﬁce, Grant No. W911NF-04-1-0057 and the Defense Threat Reduction Agency, Joint Science and Technology Ofﬁce (06012384BP) for ﬁnancial support of this work. Finally, they are indebted to Professor J. P. Guthrie (University of Western Ontario) and Professor Andrew Williams (retired from the University Chemical Laboratories, Canterbury, England) for helpful discussions on various aspects of this work.

References 1. (a) Przystas, T.J. and Fife, T.H. (1990). J. Chem. Soc., Perkin Trans. 2, 393 and references there in; (b) Suh, J. (1992). Acc. Chem. Res. 25, 273 and references therein; (c) Tecilla, P., Tonellato, U., Veronese, F., Fellugg, F. and Scrimin, P. (1992). J. Org. Chem. 62, 7621; (d) Chin, J., Jubian, V. and Mrejen, K. (1990). J. Chem. Soc., Chem. Commun. 1326; (e) Meriwether, L. and Westheimer, F. (1956). J. Am. Chem. Soc. 78, 5119; (f) Grant, T.J. and Hay, R.W. (1965). Aust. J. Chem. 18, 1189; (g) Groves, J.T. and Dias, R.M. (1979). J. Am. Chem. Soc. 101, 1033; (h) Groves, J.T. and Chambers, R.R. (1984). J. Am. Chem. Soc. 106, 630; (i) Groves, J.T. and Olson, J.R (1985). Inorg. Chem. 24, 2715; (j) Sayre, L., Reddy, K.V., Jacobson, A.R. and Tang, W. (1992). Inorg. Chem. 31, 935; (k) Fife, T.H. (1993). Acc. Chem. Res. 26, 325; (l) Fife, T.H. and Bembi, R. (1993). J. Am. Chem. Soc. 115, 11358; (m) Parac, T.N. and Kostic´, N.M. (1996). J. Am. Chem. Soc. 118, 51; (n) Singh, L. and Ram, R.N. (1994). J. Org. Chem. 59, 710; (o) Sutton, P.A. and Buckingham, D.A. (1987). Acc. Chem. Res. 20, 357; (p) Kimura, E., Nakamura, I., Koike, T., Shionoya, M., Kodama, Y., Ikeda, T. and Shiro, M. (1994). J. Am. Chem. Soc. 116, 4764; (q) Parac, T.N., Ullman, G.M. and

326

2.

3.

4.

5.

R.S. BROWN AND A.A. NEVEROV Kostic´, N.M. (1999). J. Am. Chem. Soc. 121, 3127 and references therein; (r) Ridder, A. M. and Kellogg, R. M. (1996). Models for zinc-containing enzymes. In Comprehensive Supramolecular Chemistry Vol. 4: Supramolecular Reactivity and Transport: Bioorganic Systems, Murakami, Y. (ed.), Chapter 11, Elsevier Science Ltd., Oxford; (s) Chin, J. (1991). Acc. Chem. Res. 24, 145 and references therein (a) Blasko, A. and Bruice, T.C. (1999). Acc. Chem. Res. 32, 475 and references therein; (b) Williams, N.H., Tasaki, B., Wall, M. and Chin, J. (1999). Acc. Chem. Res. 32, 485 and references therein; (c) Moss, R.A., Park, B.D., Scrimin, P. and Ghirlanda, G. (1995). J. Chem. Soc., Chem. Comm. 1627; (d) Moss, R.A., Zhang, J. and Bracken, K. (1997). J. Chem. Soc., Chem. Comm. 1639; (e) Sumaoka, J., Miyama, S. and Komiyama, M. (1994). J. Chem. Soc., Chem. Comm. 1755; (f) Morrow, J.R., Buttrey, L.A. and Berback, K.A. (1992). Inorg. Chem. 31, 16; (g) Morrow, J.R., Buttrey, L.A., Shelton, V.M. and Berback, K.A. (1992). J. Am. Chem. Soc. 114, 1903; (h) Breslow, R. and Zhang, B. (1994). J. Am. Chem. Soc. 116, 7893; (i) Takeda, N., Irisawa, M. and Komiyama, M. (1994). J. Chem. Soc., Chem. Comm. 2773; (j) Hay, R.W. and Govan, N. (1990). J. Chem. Soc., Chem. Commun. 714; (k) Schneider, H.-J., Rammo, J. and Hettich, R. (1993). Angew. Chem. Int. Ed. Engl. 32, 1716; (l) Ragunathan, K.G. and Schneider, H.-J. (1996). Angew. Chem. Int. Ed. Engl. 35, 1219; (m) Go´mez-Tagle, P. and A.K. Yatsimirsky (1998). J. Chem. Soc., Dalton Trans. 2957; (n) Roigk, A., Hettich, R. and Schneider, H.-J. (1998). Inorg. Chem. 37, 751 and references therein; (o) Go´mez-Tagle, P. and Yatsimirski, A.K. (2001). J. Chem. Soc., Dalton Trans. 2663; (p) Jurek, P.E., Jurek, A.M. and Martell, A.E. (2000). Inorg. Chem. 39, 1016 (a) Parshall, G.W. and Ittel, S.D. (1992). Homogeneous Catalysis: The Applications and Chemistry of Catalysis by Soluble Transition Metal Complexes (2nd edn), pp. 269–296. Wiley Interscience, New York; (b) Stanton, M.G., Allen, C.B., Kissling, R.M., Lincoln, A.L. and Gagne´, M.R. (1998). J. Am. Chem. Soc. 120, 5981 and references therein; (c) Otera, J. (1993). Chem. Rev. 93, 1449; (d) Okano, T., Miyamoto, K. and Kiji, J. (1995). Chem. Lett. 246; (e) Ranu, B.C., Dutta, P. and Sarkar, A. (1998). J. Org. Chem. 63, 6027; (f) Baldini, L., Bracchini, C., Cacciapaglia, R., Casnati, A., Mandolini, L. and Ungaro, R. (2000). Chem. Eur. J. 6, 1322; (g) Cacciapaglia, R., Di Stephano, S., Kelderman, E. and Mandolini, L. (1999). Angew. Chem. Int. Ed. 38, 349; (h) Cacciapaglia, R., Mandolini, L., Arnecke, R., Bo¨hmer, V. and Vogt, W. (1998). J. Chem. Soc., Perkin Trans. 2, 419; (i) Breccia, P., Cacciapaglia, R., Mandolini, L. and Scorsini, C. (1998). J. Chem. Soc., Perkin Trans. 2, 1257; (j) Stevels, W.M., Ankone´, M.J.K., Dijkstra, P.J. and Feijen, J. (1996). Macromolecules 29, 8296; (k) Hinde, N.J. and Hall, C.D. (1998). J. Chem. Soc., Perkin Trans. 2, 1249 and references therein; (l) Otton, J., Ratton, S., Vasnev, V.A., Markova, G.D., Nametov, K.M., Bakhmutov, V.I., Komarova, L.I., Vinogradova, S.V. and Korshak, V.V. (1988). J. Polym. Sci., Part A: Polym. Chem. 26, 2199; (m) Cacciapaglia, R., Casnati, A., Mandolini, L., Reinhoudt, D.N., Salvio, R., Sartori, A. and Ungaro, R. (2005). J. Org. Chem. 70, 5398; (n) Cacciapaglia, R., Di Stefano, S., Fahrenkrug, F., Luening, U. and Mandolini, L. (2004). J. Phys. Org. Chem. 17, 350 (a) Balakrishnan, V.K., Dust, J.M., vanLoon, G.W. and Buncel, E. (2001). Can. J. Chem. 79, 157; (b) Nagelkerke, R., Pregel, M.J., Dunn, E.J., Thatcher, G.R.J. and Buncel, E. (1995). Org. React. 29, 11; (c) Dunn, E.J. and Buncel, E. (1989). Can. J. Chem. 67, 1440; (d) Dunn, E.J., Moir, R.Y., Buncel, E., Purdon, J.G. and Bannard, R.A.B. (1990). Can. J. Chem. 68, 1837; (e) Pregel, M.J. and Buncel, E. (1991). J. Chem. Soc., Perkin Trans. 2, 307; (f) Pregel, M.J., Dunn, E.J. and Buncel, E. (1991). J. Am. Chem. Soc. 113, 3545; (g) Pregel, M.J., Dunn, E.J., Nagelkerke, R., Thatcher, G.R.J. and Buncel, E. (1995). Chem. Soc. Rev. 449. Bates, R.G. (1969). Medium effects and pH in non-aqueous solvents. In Solute– Solvent Interactions, Coetzee, J.F. and Ritchie, C.D. (eds), 45. Marcel Dekker, New York

METAL-CATALYZED ALCOHOLYSIS

327

6. (a) Bosch, E., Rived, F., Roses, M. and Sales, J. (1999). J. Chem. Soc., Perkin Trans. 2, 1953; (b) Rived, F., Rose´s, M. and Bosch, E. (1998). Anal. Chim. Acta 374, 309; (c) Bosch, E., Bou, P., Allemann, H. and Rose´s, M. (1996). Anal. Chem. 68, 3651; (d) Canals, I., Portal, J.A., Bosch, E. and Rose´s, M. (2000). Anal. Chem. 72, 1802 7. Gibson, G.T.T., Neverov, A.A. and Brown, R.S. (2003). Can. J. Chem. 81, 495 8. Gibson, G.T.T., Mohamed, M.F., Neverov, A.A. and Brown, R.S. (2006). Inorg. Chem. 45, 7891 9. Harned, H.S. and Owen, B.B. (1957). The Physical Chemistry of Electrolytic Solutions. ACS Monograph Series 137 (3rd edn), p. 161. Reinhold Publishing, New York. 10. (a) Brown, R.S. and Neverov, A. (2002). J. Chem. Soc., Perkin Trans. 2, 1039; (b) Brown, R.S., Neverov, A.A., Tsang, J.S.W., Gibson, G.T.T. and Montoya-Pela´ez, P. (2004). J. Can. J. Chem. 82, 1791 11. (a) Lipscomb, W. and Stra¨ter, N. (1996). Chem. Rev. 35, 2024; (b) Wilcox, D.E. (1996). Chem. Rev. 96, 2435; (c) Coleman, J.E. (1998). Curr. Opin. Chem. Biol. 2, 222; (d) Parkin, G. (2004). Chem. Rev. 104, 699; (e) Cowan, J.A. (1998). Chem. Rev. 98, 1067; (f) Weston, J. (2005). Chem. Rev. 105, 2151 12. (a) Cleland, W.W., Frey, P.A. and Gerlt, J.A. (1998). J. Biol. Chem. 273, 25529; (b) Simonson, T., Carlsson, F. and Case, D.A. (2004). J. Am. Chem. Soc. 126, 4167 and references therein 13. (a) Crosby, J. and Lienhard, G.E. (1970). J. Am. Chem. Soc. 92, 5707; (b) Crosby, J., Stone, R. and Lienhard, G.E. (1970). J. Am. Chem. Soc. 92, 2891 14. Lowry, T.H. and Richardson, K.S. (1987). Mechanism and Theory in Organic Chemistry (3rd edn), pp. 361–367, 589–590, 600. Harper and Row, New York 15. Levine, I.N. (1978). Physical Chemistry (4th edn). pp. 276–281, McGraw-Hill, USA. 16. (a) Neverov, A.A., McDonald, T., Gibson, G. and Brown, R.S. (2001). Can. J. Chem. 79, 1704; (b) Neverov, A.A. and Brown, R.S. (2001). Inorg. Chem. 40, 3588; (c) Neverov, A.A. and Brown, R.S. (2000). Can. J. Chem. 78, 1247; (d) MontoyaPelaez, P.J. and Brown, R.S. (2002). Inorg. Chem. 41, 309; (e) Neverov, A.A., Gibson, G. and Brown, R.S. (2003). Inorg. Chem. 42, 228; (f) Neverov, A.A., Sunderland, N.E. and Brown, R.S. (2005). Org. Biomol. Chem. 3, 65 17. (a) Tsang, J.S., Neverov, A.A. and Brown, R.S. (2003). J. Am. Chem. Soc. 125, 7602; (b) Tsang, J.S.W., Neverov, A.A. and Brown, R.S. (2004). Org. Biomol. Chem. 2, 3457; (c) Liu, T., Neverov, A.A., Tsang, J.S.W. and Brown, R.S. (2005). Org. Biomol. Chem. 3, 1525; (d) Lu, Z.-L., Neverov, A.A. and Brown, R.S. (2005). Org. Biomol. Chem. 3, 3379; (e) Lewis, R.E., Neverov, A.A. and Brown, R.S. (2005). Org. Biomol. Chem. 3, 4082; (f) Maxwell, C., Neverov, A.A. and Brown, R.S. (2005). Org. Biomol. Chem. 3, 4329; (g) Melnychuk, S.A., Neverov, A.A. and Brown, R.S. (2006). Angew. Chem. Int. Ed. 45, 1767 18. (a) Amis, E.S. (1955). Anal. Chem. 27, 1672; (b) Amis, E.S. (1953). J. Chem. Educ. 30, 351 19. (a) Landskroener, P.A. and Laidler, K.J. (1954). Effects of Solvation of Hydrolysis. Ph.D Dissertation, Catholic University of America, Catholic University of America Press: Washington; (b) Laidler, K.J. and Landskroener, P.A. (1956). Trans. Faraday Soc. 52, 200 20. Martell, A.E. and Motekaitis, R.J. (1988). Determination and Use of Stability Constants. VCH Publishers, New York 21. (a) Moeller, T., Martin, D.F., Thompson, L.C., Ferru´s, R., Feistel, G.R. and Randall, W. (1965). J. Chem. Rev. 65, 1; (b) Burgess, J. (1978). Metal Ions in Solution. John Wiley and Sons, New York 22. (a) Fritz, J.S. (1973). Acid– Base Titrations in Nonaqueous Solvents. Allyn and Bacon, Boston, MA; (b) Huber, W. (1967). Titrations in Nonaqueous Solvents. Academic Press, New York; (c) Gyenes, I. (1967). Titration in Non-Aqueous Media. D. Van Nostrand Company, New Jersey

328

R.S. BROWN AND A.A. NEVEROV

23. deLigny, C.L. and Rehbach, M. (1960). Recl. Trav. Chim. 79, 727 24. Bates, R.G., Paabo, M. and Robinson, R.A. (1963). J. Phys. Chem. 67, 1833 25. (a) Grunwald, E. and Berkowitz, B.J. (1951). J. Am. Chem. Soc. 73, 4939; (b) Gutbezahl, B. and Grunwald, E. (1953). J. Am. Chem. Soc. 75, 559–565 26. (a) Bates, R.G. (1964). Determination of pH: Theory and Practice (2nd edn). p. 244, John Wiley and Sons, New York; (b) This value was calculated by Bates26a using the equation d ¼ E¯ j logm gH taking values for logm gH and E¯ j (termed log l) from Grunwald25 27. Gans, P., Sabatini, A. and Vacca, A. (1996). Talanta 43, 1739 28. (a) Moeller, T., Martin, D.F., Thompson, L.C., Ferru´s, R., Feistel, G.R. and Randall, W.J. (1965). Chem. Rev. 65, 1; (b) Burgess, J. (1978). Metal Ions in Solution. John Wiley and Sons, New York 29. (a) Baes Jr., C.F. and Mesmer, R.E. (1976). The Hydrolysis of Cations. John Wiley and Sons, New York; (b) Martell, A.E. and Motekaitis, R.J. (1988). Determination and Use of Stability Constants. VCH Publishers, New York; (c) Martell, A.E. and Smith, R.M. (1974). Critical Stability Constants, Vols 1–6, Plenum Press, New York 30. Simms, H.S. (1926). J. Am. Chem. Soc. 48, 1239 31. Neverov, A.A., Montoya-Pela´ez, P. and Brown, R.S. (2001). J. Am. Chem. Soc. 123, 210 32. Montoya-Pela´ez, P., Gibson, G.T.T., Neverov, A.A. and Brown, R.S. (2003). Inorg. Chem. 42, 8624 33. Liu, T., Neverov, A.A. and Brown, R.S. unpublished observations 34. Given that the autoprotolysis constants of ethanol and methanol are 1019.1 and 1016.77, respectively, the background kobs values for the lyoxide-catalyzed ethanolysis and methanolysis reactions of paraoxon at ss pH 7.3 (ethanol) and 8.3 (methanol) are 8.1 1015 and 3.5 1011 s–1, respectively 35. Okano, T., Miyamoto, K. and Kiji, J. (1995). Chem. Lett. 246 reported the attempted transesteriﬁcation of various esters promoted by Ln(OiPr)3 alkoxides which showed exceedingly poor activity even at higher temperatures and prolonged reaction times. From our work, the trialkoxy forms would be the least active, while the monoalkoxy forms would be the most active 36. Desloges, W., Neverov, A.A. and Brown, R.S. (2004). Inorg. Chem. 43, 6752 37. Neverov, A.A. and Brown, R.S. (2004). Org. Biomol. Chem. 2, 2245 38. (a) Kimura, E., Shiota, T., Koike, T., Shiro, M. and Kodama, M. (1990). J. Am. Chem. Soc. 112, 5805; (b) Koike, T. and Kimura, E. (1991). J. Am. Chem. Soc. 113, 8935 39. Smith, M.B. and March, J. Advanced Organic Chemistry (5th edn), pp. 338–342. Wiley Interscience, New York, and references therein 40. Lu, Z.-L., Neverov, A.A. and Brown, R.S. (2005). Org. Biomol. Chem. 3, 3379 41. Ba-Saif, S., Luthra, A.K. and Williams, A. (1987). J. Am. Chem. Soc. 109, 6362 42. (a) Palm, V.A. (1975). Tables of Rate and Equilibrium Constants of Heterocyclic Reactions (Suppl. Vol. 1 (1)), Proizbodstvenno-Izdatelckii Kombinat Biniti, Moscow; (b) Palm, V.A. (1985). Tables of Rate and Equilibrium Constants of Heterocyclic Reactions (Vol. 1, Issue 3). Tartuskii gosudarsvennii Universitet, Tartu 43. (a) Taft Jr., R.W. (1952). J. Am. Chem. Soc. 74, 2729–3120; (b) Taft Jr., R.W. (1953). J. Am. Chem. Soc. 75, 4231 44. Di Stephano, S., Leonelli, F., Garofolo, B., Mandolini, L., Bettolo, B. and Migneco, M. (2002). Org. Lett. 4, 2783 45. Hong, S.-B. and Rauchel, F.M. (1996). Biochemistry 35, 10904 aquenous pKa values for Table 10 46. Clare, B.W., Cook, D., Ko, E.C.F., Mac, Y.C. and Parker, A.J. (1966). J. Am. Chem. Soc. 88, 1911 experimental pKa values for Table 10 47. (a) Chaw, Z.S., Fischer, A. and Happer, D.A.R. (1971). J. Chem. Soc. (B) 1818; (b) Kirsch, J., Clewell, A. and Simon, A. (1968). J. Org. Chem. 33, 127; (c) Humffray, A.A. and Ryan, J.J. (1967). J. Chem. Soc. (B) 468

METAL-CATALYZED ALCOHOLYSIS

329

48. (a) Ryan, J.J. and Humffray, A.A. (1966). J. Chem. Soc. (B) 842; (b) Bruice, T.C. and Mayahi, M. (1960). J. Am. Chem. Soc. 82, 3067; (c) Kirsch, J.F. and Jencks, W.P. (1964). J. Am. Chem. Soc. 86, 837 49. Mitton, C.G., Schowen, R.L., Gresser, M. and Shapley, J. (1969). J. Am. Chem. Soc. 91, 2036 50. Ba-Saif, S.A. and Williams, A. (1988). J. Org. Chem. 53, 2204 51. Caldwell, S.R., Raushel, F.M., Weiss, P.M. and Cleland, W.W. (1991). Biochemistry 30, 7444 52. (a) Jencks, W.P. and Gilchrist, M. (1968). J. Am. Chem. Soc. 90, 2622; (b) Hupe, D.J. and Jencks, W.P. (1977). J. Am. Chem. Soc. 99, 451; (c) Sayer, J.M. and Jencks, W.P. (1977). J. Am. Chem. Soc. 99, 464; (d) Gresser, M.J. and Jencks, W.P. (1977). J. Am. Chem. Soc. 99, 6963 53. (a) Thea, S. and Williams, A. (1986). Chem. Soc. Rev. 15, 125; (b) Williams, A. (1984). Acc. Chem. Res. 17, 425; (c) Williams, A. (2000). Concerted Organic and Bio-Organic Mechanisms. CRC Press, Boca Raton 54. Given that the autoprotolysis constant of methanol is 1016.77 (mol dm–3)2 (ref. 6) one can compute a ss pK a of methanol of 18.13 on the mol dm3 scale 55. (a) Petrova, J., Momchilova, S., Haupt, E.T.K., Kopf, J. and Eggers, G. (2002). Phosphorus, Sulfur Silicon Relat. Elem. 177, 1337; (b) Peterman, D.R., Fox, R.V., and Rollins, H.W. (2002). Abstract of Papers, 223rd ACS National Meeting, Orlando, FL, April 7–11, NUCL-131; (c) Ferraro, J.R., Herlinger, A.W. and Chiarizia, R. (1998). Sol. Extract Ion Exch. 16, 775; (d) Du Preez, R. and Preston, J.S.S. (1986). S. Afr. J. Chem. 39, 89–137; (e) Peppard, D.F., Mason, G.W., Driscoll, W.J. and McCarty, S.J. (1959). J. Inorg. Nucl. Chem. 12, 141; (f) Lebedeva, E.N., Zaitseva, M.G., Galaktionova, O.V., Bystrov, L.V. and Korovin, S.S. (1981). Koordinatsionnaya Khimiya 7, 870 CAN 95:87058; (g) Galaktionova, O.V., Lebedeva, E.N., Yastrebov, V.V. and Korovin, S.S. (1980). Zh. Neorg. Khim. 25, 2660 CAN 93:226562; (h) Pyartman, A.K., Kopyrin, A.A., Puzikov, E.A. and Bogatov, K.B. (1996). Zh. Neorg. Khim. 41, 347 CAN 125:124905; (i) Pyartman, A.K., Kopyrin, A.A., Puzikov, E.A. and Bogatov, K.B. (1996). Zh. Neorg. Khim. 41, 686 CAN 126:11927; (j) Pyartman, A.K., Keskinov, V.A., Kovalev, S.V. and Kopyrin, A.A. (1997). Radiochemistry (Moscow) (Translation of Radiokhimiya) 39, 142 CAN 127:210889; AN 1997:471873 56. Kosky, C.A., Gayda, J.-P., Gibson, J.F., Jones, S.F. and Williams, D.J. (1982). Inorg. Chem. 21, 3173 57. Mikulski, C.M., Pytlewski, L.L. and Karagannis, N.M. (1979). Inorg. Chim. Acta 32, 263 58. Schurhammer, R., Erhart, V., Troxler, L. and Wipf, G. (1999). J. Chem. Soc., Perkin Trans. 2, 2423 and references therein. 59. Rackham, D.M. (1980). Spectrosc. Lett. 13, 513 60. (a) Williams, N.H., Cheung, W. and Chin, J. (1998). J. Am. Chem. Soc. 120, 8079; (b) Wahnon, D., Lebuis, A.-M. and Chin, J. (1995). Angew. Chem. Int. Ed. Engl. 34, 2412 61. Williams, N.H. and Wyman, P. (2001). J. Chem. Soc., Chem. Commun. 1268 62. (a) Blasko, A. and Bruice, T.C. (1999). Acc. Chem. Res. 32, 475 and references therein; (b) Williams, N.H., Tasaki, B., Wall, M. and Chin, J. (1999). Acc. Chem. Res. 32, 485 and references therein; (c) Moss, R.A., Park, B.D., Scrimin, P. and Ghirlanda, G. (1995). J. Chem. Soc., Chem. Commun. 1627; (d) Moss, R.A., Zhang, J. and Bracken, K. (1997). J. Chem. Soc., Chem. Commun. 1639; (e) Sumaoka, J., Miyama, S. and Komiyama, M. (1994). J. Chem. Soc., Chem. Commun. 1755; (f) Morrow, J.R., Buttrey, L.A. and Berback, K.A. (1992). Inorg. Chem. 31, 16; (g) Morrow, J.R., Buttrey, L.A., Shelton, V.M. and Berback, K.A. (1992). J. Am. Chem. Soc. 114, 1903; (h) Breslow, R. and Zhang, B. (1994). J. Am. Chem. Soc. 116, 7893; (i) Takeda, N., Irisawa, M. and Komiyama, M. (1994). J. Chem. Soc., Chem. Commun. 2773; (j) Hay, R.W. and

330

63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80.

81. 82. 83.

84.

85.

R.S. BROWN AND A.A. NEVEROV Govan, N. (1990). J. Chem. Soc., Chem. Commun. 714; (k) Schneider, H.-J., Rammo, J. and Hettich, R. (1993). Angew. Chem. Int. Ed. Engl. 32, 1716; (l) Ragunathan, K.G. and Schneider, H.-J. (1996). Angew. Chem. Int. Ed. Engl. 35, 1219; (m) Go´mez-Tagle, P. and Yatsimirsky, A.K. (1998). J. Chem. Soc., Dalton Trans. 2957; (n) Roigk, A., Hettich, R. and Schneider, H.-J. (1998). Inorg. Chem. 37, 751 and references therein; (o) Go´mez-Tagle, P. and Yatsimirski, A.K. (2001). J. Chem. Soc., Dalton Trans. 2663; (p) Jurek, P.E., Jurek, A.M. and Martell, A.E. (2000). Inorg. Chem. 39, 1016 (a) Hegg, E.L. and Burstyn, J.N. (1998). Coord. Chem. Rev. 173, 133; (b) Jancso´, A., Mikkola, S., Lo¨nnberg, H., Hegetschweiler, K. and Gadja, T. (2003). Chem. Eur. J. 9, 5404 Feng, G., Mareque-Rivas, J.C., Martin de Rosale´s, R.T. and Williams, N.H. (2006). J. Am. Chem. Soc. 127, 13470 Stra¨ter, N., Lipscomb, W.N., Klabunde, T. and Krebs, B. (1996). Angew. Chem. Int. Ed. Engl. 35, 2024 Davies, J.F., Hostomska, Z., Hostomsky, Z., Jordan, S.R. and Mathews, D.A. (1991). Science 252, 88 Beese, L.S. and Steitz, T.A. (1991). EMBO J. 10, 25 Lahm, A., Volbeda, S. and Suck, D. (1990). J. Mol. Biol. 215, 207 Molenveld, P., Engbertsen, J.F.J. and Reinhoudt, D.N. (2000). Chem. Soc. Rev. 29, 75 Mancin, F., Scrimin, P., Tecilla, P. and Tonellato, U. (2005). Chem. Commun. 2540 Morrow, J.R. and Iranzo, O. (2004). Curr. Opin. Chem. Biol. 8, 192 Yamada, K., Takahashi, Y.-i., Yamamura, H., Araki, A., Saito, K. and Kawai, M. (2000). Chem. Commun. 1315 Feng, G., Mareque-Rivas, J.C. and Williams, N.H. Chem. Commun. 1845 Mancin, F., Rampazzo, E., Tecilla, P. and Tonellato, U. Eur. J. Chem. 281 Yang, M.-Y., Iranzo, O., Richard, J.P. and Morrow, J.R. (2005). J. Am. Chem. Soc. 127, 1064 Iranzo, O., Elmer, T., Richard, J.P. and Morrow, J.R. (2003). Inorg. Chem. 42, 7737 Iranzo, O., Richard, J.P. and Morrow, J.R. (2004). Inorg. Chem. 43, 1743 Iranzo, O., Kovalevsky, A.Y., Morrow, J.R. and Richard, J.P. (2003). J. Am. Chem. Soc. 125, 1988 Brown, D.M. and Usher, D.A. J. Chem. Soc. 6558 The methoxide reaction of 33 in pure methanol has a second-order rate constant for methoxide promoted methanolysis of (2.5670.16) 103 M1s1, while the secondorder rate constant for attack of methoxide on methyl p-nitrophenyl phosphate ((7.970.6) 107 M1s1 at 25 1C), Neverov, A.A., Brown, R.S. (2001). Inorg. Chem. 40, 3588 Tsang, J.S., Neverov, A.A. and Brown, R.S. (2003). J. Am. Chem. Soc. 125, 1559 Morrow, J.R., Buttrey, L.A. and Berback, K.A. (1992). Inorg. Chem. 31, 16 A universal expression for both strong and weak binding scenarios is given below: kobs ¼ kcat ð1 þ K d ½33 þ ½La3þ K d X Þ=ð½33 ð2K d ÞÞ where Kd refers to the dissociation constant for La3+:33 La3++33; [33] and [La3+] are total concentrations and X is given as: X ¼ fð1 þ 2K d ½1 þ 2 ½Zn2þ K d þ K 2d ½12 2 K 2d ½Zn2þ ½1 þ ½Zn2þ 2 K 2d g0:5 . The equation was derived from the equations for equilibrium binding and for conservation of mass by using the commercially available MAPLE software, Maple V Release 5, Waterloo Maple Inc., Waterloo, Ontario, Canada At ﬁrst glance the process described in Equation (6) is bimolecular in [La3+:33], but the kinetics strictly adhere to a ﬁrst-order process for the loss of starting material. However, the subsequent observation that the products of the reaction actually catalyze the decomposition of starting material allow us to treat the kinetics at each ss pH according to 3+ a simple one-site binding model: kobs ¼ kmax :33]init/(Kd+[La3+:33]init) cat [La Liu, C.T., Neverov, A.A. and Brown, R.S. (2007). Inorg. Chem. 46, 1778

METAL-CATALYZED ALCOHOLYSIS

331

86. Breslow, R., Berger, D. and Huang, D.-L. (1990). J. Am. Chem. Soc. 112, 3686 report that 0.5 mM Zn2+ at pH 7 in water at 37 oC produces a pseudo-ﬁrst-order rate constant of 1.71 102 h1 (4.75 106 s1) from which a second-order rate constant of 9.5 103 dm3 mol1 s1 can be calculated at that pH. 87. A diffusion limited proton abstraction would occur at kdif ¼ 1010 M1 s1, so at ss pH 7.1 where [OEt] ¼ 1012 the maximum pseudo-ﬁrst-order rate constant would be 2 1 kmax s cat ¼ 10 88. Bonfa´, L., Gatos, M., Mancin, F., Tecilla, P. and Tonellato, U. (2003). Inorg. Chem. 42, 3943 89. Since the active species for the ethanolysis of 2 is 9: Z n2+(OEt), the true-second-order rate constant for the reaction would be twice the gradient of the plot, or 1.62 dm3 mol1 s1 90. Rawlings, J., Cleland, W.W. and Hengge, A.C. (2003). Inorg. Biochem. 93, 61 91. Ligand 35 was originally prepared by Weisman, G.R., Vachon, D.J., Johnson, V. B., Gronbeck, D.A. (1987). J. C. S. Chem. Commun. 886, along with numerous others containing the [12]aneN3 and [9]aneN3 binding group 92. Kim, J. and Lim, H. (1999). Bull. Korean. Chem. Soc. 20, 491 93. Haidar, R.H., Ipek, M., Dasgupta, B., Yousef, M. and Zompa, L. (1997). J. Inorg. Chem. 36, 3125 94. Dasgupta, B., Haidar, R., Hsieh, W.-Y. and Zompa, L. (2000). J. Inorg. Chim. Acta 306, 78 95. Neverov, A.A., Lu, Z.L., Maxwell, C.I., Mohamed, M.F., White, C.J., Tsang, J.S.W. and Brown, R.S. (2006). J. Am. Chem. Soc. 128, 16398 96. Lu, Z.-L and Brown, R.S. To be published 97. Vichard, C. and Kaden, T.A. (2002). Inorg. Chim. Acta 337, 173 98. The second-order rate constant for hydrolysis of BNPP at 35 1C is reported to be 2.4 104 s1 38b 99. Diebler, H., Secco, F. and Vetorini, M. (1989). J. Phys. Chem. 93, 1691 100. Liu, C.T., Bunn, S., Neverov, A.A. and Brown, R.S. To be published

This page intentionally left blank

AUTHOR INDEX

Abou-Hamdan, A. 205–208 Abramovitch, R.A. 39 Achiwa, T. 36 Adam, K.R. 62, 64 Adam, W. 226, 243 Adams, A. 186, 194–196 Adams, M. 85–86 Adelfang, J.L. 51 Aebisher, D. 248 Aguirre, G. 225, 261 Ahlberg, P. 154 Ahmad, A. 38 Aihara, S. 253 Akaba, R. 253 Akasaka, K. 186, 191 A˚kerman, B. 186, 203 Akimoto, H. 253 Alabugin, I.V. 1, 4, 8, 10–12, 17–22, 31 Alkorta, I. 36 Allen, C.B. 272, 280, 283 Allen, F.H. 49 Almgren, M. 177 Al-Souﬁ, W. 179–180, 205, 213 Alvaro, M. 225, 231, 261 Amat, A.M. 231 Ameen, S. 173 Ames, B.N. 97 Amis, E.S. 275–276 Anderson, M.W. 229 Andrews, A.W. 105 Andrews, L.E. 37, 39, 54, 67, 70, 72, 74–75, 80, 83, 94, 97, 106, 109–110, 112–113 Anslyn, E.V. 168 Apeloig, Y. 139, 151, 158 Araki, A. 309, 316 Araki, Y. 247 Arends, I.W.C.E. 261 Arques, A. 231 Assa-Munt, N. 186, 196–197 Atwell, G.J. 186, 194–196 Atwood, J.L. 167

Aubard, J. 186 Auer, A.A. 159 Auner, N. 149 Avramenko, V.I. 37, 42, 54, 56, 89, 94 Badjic, J.D. 168 Baes, C.F. 278 Bag, S.S. 3 Baggott, J.E. 261 Baguley, B.C. 186, 195–196 Baik, M. 16 Bailly, C. 186, 199–200 Bain, A.D. 36 Balakrishnan, V.K. 272, 280, 283 Baliga, R. 186, 201–203 Balko, B. 171 Ballew, M. 173 Ballinger, T.H. 231 Balster, A. 22 Banks, T.M. 37, 39, 52–54, 70–71, 74–75, 81, 85, 97–99, 104–107, 109–110, 112–113 Baraldi, P.G. 105 Barich, D.H. 128, 155–156, 236, 238 Barnes, P. 227 Barra, M. 205 Barros, T.C. 192, 205, 215 Bartlett, R.J. 127, 148 Barton, D.H.R. 56 Bartsch, H. 37 Ba-Saif, S.A. 289, 292, 299 Basak, A. 3, 38, 70 Bashir-Heshemi, A. 156 Bashore, C.G. 36 Baskakis, C. 243 Basu, P. 231 Bates, R.G. 272, 277, 280 Bauch, C. 149 Bauer, L. 59 Bauman, J.D. 186 Bausch, J. 157 Bauta, W.E. 12 333

334 Beauchamp, J.L. 241–242 Beck, L.W. 155 Becker, A.R. 70 Becker, W. 179 Beckett, J.L. 205, 213 Beese, L.S. 309 Beitz, J.V. 173 Beland, F.A. 37 Bellamy, L.J. 51 Benish, M.A. 154 Bennet, A.J. 58, 205, 215 Bent, H.A. 21 Berback, K.A. 311–312 Berger, D. 314 Berger, H. 4 Berger, R.L. 171 Bergman, R.G. 2 Berkowitz, B.J. 277, 328 Bernardi, F. 43 Bernasconi, C.F. 170 Berndt, A. 139 Bettolo, B. 293 Bhattacharyya, S. 16 Bills, W.D. 36, 40 Bilyarska, L. 248 Birch, D.J.S. 176 Birks, J.B. 177 Birney, D.M. 27 Bittman, R. 186 Biver, T. 186, 192, 203 Blackburn, G.M. 36–37, 71, 97 Blasko`, A. 186, 202, 271, 280, 283, 290, 309 Blatter, F. 253–258 Blau, L. 186 Bloor, D.M. 174–175, 205, 210–212 Boatz, J.A. 156 Boche, G. 37, 113 Boger, D.L. 105, 186, 200 Boggs, J.E. 126 Bohne, C. 167–168, 176–178, 186, 192, 196, 205, 208, 214–216 Bolt, J.D. 205 Bolte, M. 149 Bonfa´, L. 316, 325 Bonin, A.M. 37, 39, 52–54, 65, 67–71, 74, 80–82, 94, 97–99, 104–110, 112–113 Bonnet, G. 178–179 Borcherdt, W. 171 Borden, W.T. 22

AUTHOR INDEX Bordner, J. 36 Bortolini, R. 186, 202 Bosch, E. 272–273, 276–277, 290, 302 Bosold, F. 37 Bossmann, S.H. 231 Boston, R.C. 186, 197 Bourdelande, J.L. 233 Boyce, C.W. 105 Boyd, M.R. 115 Boykin, D.W. 186 Brana, M.F. 186, 199–200 Brassett, C. 186 Braude, V. 139 Braun, A.M. 231 Breck, D.W 229 Breiner, B. 1, 10 Bremer, M. 150, 154 Breneman, C. 35, 50 Bresloff, J.L. 186, 188–189 Breslow, R. 314 Breusegem, S.Y. 172, 186, 203 Brinker, U.H. 232 Britten, J.F. 36 Brixner, S. 135–136 Broene, R.D. 16 Brooks, M.E. 38 Brown, D.A. 59 Brown, D.M. 310, 315 Brown, G.R. 205, 212 Brown, H.C. 65 Brown, K. 39 Brown, P. 205, 211 Brown, R.D. 126 Brown, R.S. 36, 58, 271, 273, 275–280, 282, 284–285, 287, 302–303, 309–310, 312, 315–318, 321–324 Browne, K.A. 186, 202 Bruice, T.C. 186, 202, 271, 280, 283, 290, 297, 309 Brun, R. 186 Brunelle, P. 149 Bryant, M.S. 37 Bubnovskaya, L.N. 20 Buccigross, J.M. 67, 72–73, 83–84, 94 Buchholz, H.A. 151 Buchwald, S.L. 16 Buck, M. 22 Bugnon, P. 205–208 Bu¨hl, M. 126

AUTHOR INDEX Bui, B.H. 6 Buncel, E. 272, 280, 283 Bunn, S. 324 Burrichter, A. 153, 156–158 Burstyn, J.N. 309 Buttrey, L.A. 311–312 Buzek, P. 126–129, 135, 138, 143–144, 146, 148–149, 151 Bytheway, I. 56 Cacciari, B. 105 Cacho, M. 186, 199–200 Cain, B.F. 195 Caldin, E.F. 173 Caldwell, J.W. 241 Caldwell, S.R. 299 Calin, M. 157 Callender, R. 173 Campbell, J.J. 37, 39, 44, 52–53, 60, 65, 67, 70–75, 77–79, 81, 83, 85, 94, 97–98, 104–105, 110 Cann, J.R. 185 Cano, M.L. 232 Cantrill, S.J. 168 Carlsson, C. 203 Carmichael, I. 187 Carneiro, J.W. de M. 128–129, 142–143 Carpenter, B.K. 27 Carr, S.W. 227 Carrasco, C. 186, 199–200 Carrol, F.A. 74 Carter, P. 12 Casanova, J. 133 Casida, M.E. 131–132 Cativiela, C.C. 36 Catlow, C.R.A. 229 Cecil, T.L. 15 Cera, C. 71, 97 Chaires, J.B. 186, 197 Challis, B. 36 Challis, J. 36 Chambers, C.C. 77 Chan, J. 232, 248, 250 Chan, Y.-Y. 255 Chandrasekhar, J. 232, 235, 244 Chaw, Z.S. 297 Cheeseman, J.R. 151 Chen, F.M. 186, 201–203 Chen, H. 36

335 Chen, P. 20–21, 27 Chen, Y.Z. 246–247 Cheng, B. 38 Cheng, P. 225 Cherry, W.R. 43 Chervin, I.I. 58–59 Cheung, S.T. 177 Cheung, W. 305 Chiapperino, D. 39 Chin, J. 305 Chisholm, J.D. 186 Chou, D.T.H. 205, 215 Christe, K.O. 156 Christoff, M. 205, 215–216 Chung, C.J. 177, 205 Cioslowski, J. 144 Clare, B.W. 296 Clark, A.E. 5 Clarke, R.J. 22, 205, 208–210 Clay, S.F. 37, 54, 70, 97, 106, 108–109, 112–113 Clegg, R.M. 172, 186, 188–189, 203 Cleland, W.W. 299, 317 Clements, B. 42, 54 Clennan, E.L. 225, 231–239, 240–241, 247–248, 250–251, 253–254, 259 Clewell, A. 297 Coates, J.H. 205, 208–210, 213 Coe, J.W. 36 Collis, M. 54 Collyer, C.A. 186, 194–195 Colson, P. 186 Colvin, M.E. 37, 97 Connors, K.A. 205 Cook, D. 296 Cooley, J.H. 36, 40 Corbin, D.R. 228 Corin, A.F. 186, 192–194 Corma, A. 225–226, 231, 261 Cormier, W.E. 229 Cory, M. 186, 196 Cousins, G.R.L. 168 Cow, C.C. 36 Cow, C.N. 36 Cox, K.K. 186, 202–203 Cox, P.A. 226, 229 Cozens, F.L. 232 Cramer, C.J. 39, 77 Cramer, F. 205–206, 208

336 Cremer, D. 12, 20, 27, 126, 154 Crippen, G.M. 103, 105 Crofton, M.W. 133 Cromwell, N.H. 51 Crooks, J.E. 172–173 Crosby, J. 274 Crothers, D.M. 186, 188–192, 194, 197, 201–203 Csicsery, S.M. 225 Cundy, C.S. 226, 229 Czarny, A. 186 Czerlinski, G.H. 170 D0 Alessio, R. 186, 202 Danovich, D. 139 Dapprich, J. 181 Dasgupta, B. 318–319 Dattagupta, N. 186, 197 Davidse, P.A. 38 Davidson, E.R. 5 Davies, J.E.D. 167, 309 Davies, M.C. 183–184 De Almeida, M.V. 56 De Biasi, A. 186, 203 De Fetyer, S. 168, 205, 214 De Kimpe, N. 80, 113 De Schryver, F.C. 168, 176, 205, 214 De Vos, D.E. 258 Debnath, A.K. 105 Demont, P.M. 205, 208 Deng, F. 155, 232 Denny, W.A. 105, 186, 194–197 Desloges, W. 285, 317 Di Stephano, S. 293 Dicks, A.P. 38 Diebler, H. 324 Ding, D. 186 Dobrowolski, D.C. 247 Dodin, G. 186 Donovan, D.J. 146 Dougherty, D.A. 236, 238 Douglas, K.T. 172, 186, 203 Douhal, A. 176 Dourlent, M. 186, 192 Doyle, M.P. 18 Dunard, F.A. 205, 208 Dunitz, J.D. 44–45 Duspara, P.A. 36 Dust, J.M. 272, 280, 283

AUTHOR INDEX Dutta, P.K. 227 Dwyer, T.J. 181–182 Dyck, A.S.M. 205, 208 Dye, D.F. 16 Dyer, A. 226, 231 Dyer, R.B. 173 Eastman, N.L. 205, 213 Eaton, D.F. 228, 233 Eggers, F. 175 Eggers, G. 304 Ehrenberg, M. 186, 193 Eiger, E. 38 Elguero, J. 36 Eliel, E.L. 18, 242 Elliott, J. 12 Elmer, T. 309, 322 Elson, E.S. 178, 186–188 Enderlein, J. 178–180 Engbertsen, J.F.J. 309 Engel, B. 11 Engelhardt, G. 113 Englander, S.W. 186–188 Englebienne, P. 183–185 Epiotis, N.D. 43 Ercolani, G. 168 Erdinger, L. 113 Erhart, V. 305 Eriksson, M. 186, 203 Ervin, J. 173 Eschenmoser, A. 43 Eshdat, L. 4 Evans, C.H. 205, 214 Evans, D.F. 253 Exner, O. 59 Eyring, E.M. 205, 210–211 Facompre, M. 186 Falkenberg-Andersen, C. 131 Falvey, D.E. 39 Famulok, M. 37 Fan, M.-C. 107 Farcasiu, D. 128 Feeder, N. 36, 55, 57 Feichtinger, D. 20–21, 27 Feigon, J. 186, 195, 197 Feistel, G.R. 276, 278 Felekyan, S. 179–180, 205, 213

AUTHOR INDEX Felton, J.S. 37, 97 Feng, G. 309, 315 Feng, K. 246 Feng, L. 27 Ferguson, D.B. 155 Ferguson, L.R. 105 Ferna´ndez, L. 155, 226 Ferrer, B. 231 Ferriera, J.A. 56 Ferru´s, R. 276, 278 Field, J.P. 173 Fife, T.H. 271, 283 Fingerle, R.E. 186, 197 Finlay, G.J. 195 Fischer, A. 297 Fishbein, J.C. 38 Fisher, J.J. 55 Fitter, J. 178–180 Flanagan, M.F. 15 Fleischer, U. 127, 150, 154 Flynn, G.H. 173 Fokin, A.A. 22 Foote, C.S. 233, 247 Ford, G.P. 71 Forne´s, V. 231 Forster, W. 74, 78 Forsyth, D.A. 129 Fortt, S. 12 Foster, R. 256 Fouad, F.S. 3, 6 Fox, K.R. 186 Fox, M.A. 126, 235 Fox, R.V. 304 Frank, J. 178, 205, 215 Fransson, L. 37, 39, 54, 70, 74, 80, 97, 109–110, 112 Frazier, R.A. 183–184 Freeman, R. 181 Frei, H. 231, 253–258 Freier, S.M. 173 French, H.T. 62 Friauf, W. 171 Friesen, M. 37 Frimer, A.A. 257 Frisch, M.J. 151 Fritz, J.S. 276 Froudakis, G.E. 235, 238, 240, 246 FuX, M. 126, 147 Fu, P.P. 37

337 Fuess, H. 235 Fujimoto, H. 240 Fujimoto, M. 205–208 Fujita, Y. 205, 208 Fukahori, T. 205, 210–212 Fukui, K. 240 Fullenkamp, D.E. 186 Funck, T. 175 Funk, R.L. 15 Furusawa, C.-I. 253 Fyfe, C.A. 229 Gabbay, E.J. 186, 197 Gadosy, T.A. 38 Gagne´, M.R. 272, 280, 283 Galasso, V. 144 Galbraith, J.M. 7, 13, 22 Galembeck, S.E. 154 Gallaso, V. 145 Galletero, M.S. 233 Galm, U. 3, 6 Gans, P. 277, 279–280 Garcı´ a, H. 155, 225–226, 231–233, 261 Garcia, P. 90, 93 Gardiner, W.C. 173, 186, 190 Garofolo, B. 293 Gatos, M. 316, 325 Gauss, J. 125–126, 130, 134–135, 138, 147, 151, 154, 159 Gayda, J.-P. 304 Geacintov, N.E. 71, 186, 193–194 Gehlen, M.H. 176 George, H. 157 Georgiadia, R.M. 186 Gerdes, R.G. 37, 39, 52, 70, 97–98 Ghirlanda, G. 271, 280, 283, 290, 309 Ghose, A.K. 103, 105 Gibson, G.T.T. 273, 275–280, 282, 284, 302, 309, 312, 324 Gibson, J.F. 304 Gil, A.M. 36 Gilbert, A. 261 Gilchrist, M. 299 Gillson, A.-M.E. 36–37, 39, 44, 47–50, 54, 67, 70, 72, 74, 80, 83, 94, 97, 106, 109–113 Glass, W.K. 59 Glende, C. 113

338 Glover, S.A. 35–39, 42, 44, 47–56, 59–60, 65, 67–75, 77–87, 90, 92–94, 97–99, 104–113 Gnann, R. 156 Gniazdowski, M. 71, 97 Gob, S. 231 Gobbi, G.C. 229 Gold, V. 62–63 Goosen, A. 38, 70, 92 Gordon, M.S. 64 Go¨sch, M. 178 Goto, Y. 186 Goyne, T.E. 247 Goze, C. 225 Grandinetti, F. 39 Grassian, V.H. 230–231, 254–255, 257–258 Greatrex, R.A. 126 Green, R.J. 183–184 Greenberg, A. 35–36, 50 Greer, A. 248 Gregor, I. 178–180 Greif, P.C. 186, 197 Gresser, M. 297 Grey, C.P. 230, 236 Grier, D. 186, 197 Grieser, F. 177 Gruebele, M. 173 Grunenberg, J. 142 Grunewald, E. 79 Grunwald, E. 277, 328 Guengerich, F.P. 37 Guss, J.M. 186, 194–195 Haack, T. 113 Hachoumy, M. 158 Hadzialic, G. 38 Hagen, E.L. 129 Hager, M.H. 3, 6 Haidar, R.H. 318–319 Hajji, L. 225 Hall, D.G. 205, 210–212 Hall, D.R. 185 Hall, M.B. 56, 73, 84 Haller, J. 205, 210 Halvorsen, A. 74 Hamelberg, D. 186 Hamilton, A.D. 168 Hammond, B.L. 142

AUTHOR INDEX Hammond, G.A. 67, 72, 83, 94 Hammond, G.P. 37, 39, 44, 52–53, 60, 65, 67–70, 73, 82, 84, 94, 97–98, 104–105 Haˆncu, D. 128 Hanrath, M. 11 Hansch, C. 105–106, 116 Hanson, D.C. 247 Happer, D.A.R. 297 Harling, R. 12 Harned, H.S. 273, 275 Harris, J.M. 70 Harris, N.R. 7, 13, 22 Harrison, P.H.M. 36 Hartz, N. 128 Hartzell, C.J. 205, 213 Harvey, R.G. 71 Hashimoto, S. 230–231, 253, 259 Hatch, F.T. 37, 97 Haupt, E.T.K. 304 Haw, J.F. 128, 155, 232, 236, 238 Hawkins, G.D. 77 Hayashi, K. 205, 208 Head, N.J. 151, 156 Hegg, E.L. 309 Hehre, W.J. 27, 46 Heiliger, L. 156 Heiner, T. 157 Heinrich, J.L. 38 Helmick, J.S. 38 Heneghan, C.S. 155, 232 Hengge, A.C. 317 Hepp, A. 142 Herrmann, W. 205, 212 Hersey, A. 205, 208 Hess, P.H. 51 Hilderbrandt, R. 126 Hinton, J.F. 126 Hiromi, K. 186, 191 Hoekman, D. 106 Hoffmann, R. 27, 145 Hoffner, J. 20–21, 27 Hoge, B. 156 Hogrel, J.F. 186, 192 Hollenstein, S. 154 Holzwarth, J.F. 173, 178, 186, 188, 192–193, 205, 215 Homma, A. 36 Hong, S.-B. 296, 299 Hopf, H. 3–4

AUTHOR INDEX Hoppe, R. 235 Horn, C. 3 Hosseini, M.W. 168 Hostomska, Z. 309 Hostomsky, Z. 309 Houk, K.N. 39, 240 Houssier, C. 186 Hrovat, D.A. 22 Hrusak, J. 39 Hsieh, W.-Y. 318–319 Huang, D.-L. 314 Huang, T.-B. 107 Huffmann, J.C. 151 Hug, G.L. 187 Hughes, T.S. 27 Humffray, A.A. 297 Humphreys, W.G. 37 Hunt, T.A. 258 Huskens, J. 168 Ibata, T. 66 Ignatov, S.M. 58–59 Ileva, A. 248 Iliev, V. 248 Imamura, A. 27 Imhof, R.E. 176 Inagaki, S. 36, 240 Ingold, K.U. 92 Inoue, Y. 205 Ipek, M. 318 Iranzo, O. 309, 315–316, 322 Isaacs, N.S. 61–62, 64, 70, 74, 79 Ishikawa, K. 205, 210 Itoh, M. 253 Ittel, S.D. 272, 280, 283 Ivonin, S.P. 37, 42, 54, 56, 89, 94 Jacobs, P.A. 258 Jaffe´, H.H. 79 Jafri, J.A. 14 Jagod, M.-F. 133 Jakokby, W.B. 86 James, T.G. 38 Jasko´lski, W. 225 Jemmis, E.D. 158 Jencks, W.P. 70, 299 Jenkins, S.I. 36 Jenkins, T.C. 186, 198–199 Jenks, W.S. 259

339 Jensen, F. 248 Jeschek, G. 51, 55, 57 Jeuell, C.L. 131 Jiang, Z.Q. 247 Jime´nez, A.I. 36 Jobe, D.J. 210 Jockusch, S. 234 Johns, J. 90, 93 Johnson, C.K. 48–49 Johnson, I. 203 Johnson, J.E. 36 Johnston, L.J. 236, 259 Jones, D.J. 225 Jones, G.B. 3, 6, 12, 22 Jones, K.M. 70 Jones, S.F. 304 Jonsson, M. 203 Jordan, S.R. 309 Joubert, A. 186, 199–200 Jovin, T.M. 186, 188–189, 192–194 Joy, A. 244 Ju, J. 3, 6 Jung, H. 37 Kaanumalle, L.S. 243–244 Kaatze, U. 205, 210 Kaden, T.A. 322 Kadlubar, F.F. 37 Kahley, M.J. 38 Kaila, M. 186, 196 Kallenbach, N.R. 186–188 Kalyanasundaram, K. 176 Kao, H.-M. 230 Karagannis, N.M. 304 Karlsson, J. 186, 203 Karplus, M. 36 Kates, M.H. 148 Kato, S. 210 Katsumata, H. 36 Kaufmann, F.P. 139 Kaupp, M. 126, 156 Kausch, M. 150 Kawai, M. 309, 316 Kawase, M. 38, 70 Kawatkar, S.P. 4 Kayser, K.J. 38 Kearns, D.R. 186, 195–197 Keith, T.A. 151

340 Kellard, B. 37, 71, 97 Keller, B. 205, 212 Kelly, D.P. 131 Kelly, G. 231 Kelly, H.C. 205, 208 Kennedy, S.A. 38 Kerr, G.T. 229 Kerwin, S.M. 1, 27 Khenkin, A.M. 261 Kierzek, R. 173 Kiji, J. 284 Kikugawa, Y. 38, 70 Kim, C.-S. 17 Kim, J. 318, 322 Kimura, E. 285, 316 Kirby, A.J. 36, 50, 55, 57 Kirchen, R.P. 143–144 Kirsch, J. 297 Kirschhock, C. 235 Kisiel, U. 205, 208 Kissling, R.M. 272, 280, 283 Kitagawa, T. 173 Kitamura, T. 38, 70 Kitano, H. 205, 208 Klabunde, T. 309 Klein, M. 3, 113 Kleinman, M.H. 168, 176–177, 214–215 Kleinpeter, E. 57 Klevit, R.E. 186, 202 Klinowski, J. 227, 229 Klots, E.A. 37, 42, 47, 54, 56, 89, 94 Knoche, W. 210 Ko, E.C.F. 296 Kobori, A. 186 Koch, A. 57 Koch, W. 127–129, 142, 154 Kodama, M. 285, 316 Koehli, T.-P. 143 Koga, N. 11, 13 Koike, T. 285, 316 Kojima, M. 253, 259 Kolb, B.A. 38 Kollman, P.A. 241 Kolubayev, T. 186, 193–194 Komarov, I.V. 36, 50, 55, 57 Kondo, M. 205, 210–211 Ko¨nig, B. 3 Konkoli, Z. 154 Kopf, J. 304

AUTHOR INDEX Kornyushyna, O.S. 203 Kos, A.J. 158 Koschinsky, R. 143 Kosky, C.A. 304 Kosmas, G. 244 Kostyanovsky, R.G 37, 42–43, 47, 54, 56, 58–59, 89, 94 Kovalenko, S.V. 4, 17–20, 31 Kovalevsky, A.Y. 309 Kowski, K. 36, 50, 55 Kraka, E. 12, 20, 27, 154 Krawietz, T.R. 232 Krebs, B. 309 Krebs, O. 243 Krichevsky, O. 178–179 Kronja, O. 129, 143–144 Krossing, I. 139 Kruger, J.D. 142 Kudryavtsev, V. 179 Ku¨hnemuth, R. 179–180, 205, 213 Kumar, A. 186 Kumar, C.V. 186 Kumar, D. 27 Kumasawa, H. 186 Kunz, Y.K. 36 Kupfer, R. 232 Kuramoto, N. 205, 210–212 Kuriyama, Y. 259 Kuroda, M. 205 Kuroda, Y. 205 Kusmierek, J.T. 97 Kutzelnigg, W. 154 Kuzmin, V.A. 186, 193–194 Laali, K.K. 154, 156 Lacy, E.R. 186, 202–203 Lahm, A. 309 Laidig, K.E. 146, 149 Laidler, K.J. 276 Laird, R.M. 74, 78 Lakshminarasimhan, P.H. 230, 236, 259 Lamberson, C.R. 186, 196 Lambert, J.B. 151, 155 Landrie, C.L. 38, 70 Landskroener, P.A. 276 Lang, N.P. 37 Larsen, R.G. 254, 257–258 Larsen, S.C. 230–231, 254–255, 257–258

AUTHOR INDEX Larsson, A. 186, 203 Lauder, I. 62, 64 Lavigne, J.J. 168 Le, N.M. 186, 202–203 Ledney, M. 227 Lee, K.J. 247 Lee, M.P.H. 186, 202–203 Lefﬂer, J.E. 7, 79 Lehn, J.M. 167–168 Lenhard, J.R. 203 Lennartz, R. 135, 138, 159 Lenoir, D. 126, 151 Leo, A.J. 105–106, 116 Leonelli, F. 293 Leung, H.-K. 255 Leupin, W. 186, 195–197 Levine, I.N. 274, 309 Levitin, I.Y. 20 Levy, G.C. 58 Levy, J.B. 151 Levy, R. 186, 197 Lewis, F. 10 Lewis, K.D. 4 Lewis, R. 12 Li, G. 258 Li, H.J. 27–28, 186, 190–192 Li, P. 230 Li, X.Y. 39, 186, 196, 151, 233–236, 255 Liang, G. 126, 151, 156 Liao, Y. 178, 205, 215 Licence, V.E. 38 Lichter, R.L. 58 Liebman, J.F. 35 Lienhard, G.E. 274 Lim, H. 318, 322 Lim, K. 126 Limbach, H.H. 142 Lin, D.X. 37 Lin, J. 38–39 Lin, L. 155 Lincoln, A.L. 272, 280, 283 Lincoln, P. 186, 195, 203 Lincoln, S.F. 205, 208–210, 213 Linton, B. 168 Lipscomb, W.N. 274, 309 Liu, B. 127 Liu, C.T. 315–316, 323–324 Liu, R.H. 157 Liu, T. 282

341 Liu, W. 56 Liu, X. 255 Lok, S.M. 142 Lombardy, R.L. 186 Long, F.A. 62 Longfellow, C.E. 173 Loontiens, F.G. 172, 186, 203 Lopez de Compadre, R.L. 105 Lopez, X. 36 Losytskyy, M.Y. 203 Lowry, T.H. 274 Lu, Z.-L. 287, 318, 321 Lukashov, S.S. 203 Luthra, A.K. 289, 292 Lye, P.G. 205–208 Ma, H. 173 Ma, J.C. 236, 238 Mac, Y.C. 296 Macgregor, R.B. 188–189 Macho, V. 142 MacNicol, D.D. 167 MacRae, A.I. 186, 196 Madsen, E.M. 186 Maeda, Y. 205, 208 Maerker, C. 126–127, 129, 143–144, 146, 148–149 Maestre, N. 186, 199–200 Magde, D. 178, 186–188 Mageswaran, R. 59 Magnus, P. 12 Magonski, J. 38 Maireles-Torres, P. 225 Maleveille, C. 37 Malkin, V.G. 126, 131–132, 156 Malkina, O.L. 131–132, 152–153, 156 Mancin, F. 309, 316, 325 Mandal, C. 186–188 Mandal, S. 3 Mandolini, L. 168, 293 Mann, J.E. 258 Manoharan, M. 1, 8, 10–12, 17–22 Marcandalli, B. 173, 186, 188, 192–193 March, J. 287 Marcus, B.K. 229 Marcus, D.M. 232 Mareque-Rivas, J.C. 309, 315 Maron, D.M. 97 Ma´rquez, F. 226

342 Martell, A.E. 276 Martı´ , V. 155, 226, 232 Martin de Rosale´s, R.T. 309, 315 Martin, D.F. 276, 278 Martin, K.A. 38 Marx, D. 126 Mary, C. 52–53, 90, 93 Matar, M. 142 Mathew, T. 156 Mathews, D.A. 309 Matsubara, C. 259 Matsui, Y. 205, 208 Matta, C.F. 36 Matzger, A.J. 4 Maxwell, C.I. 318, 321 Mayahi, M. 297 Mayr, H. 143 Mazepa, A.V. 37, 42, 47, 54 Mazzini, S. 186, 202 McCleland, C.W. 38, 70, 92 McClelland, R.A. 38–39 McDonald, T. 275, 309 McGarvey, D.J. 255 McGhee, J.D. 186 McLennan, D.J. 74 McMahon, T.B. 55 McNee, I.R. 70 Mehmedovic, M. 186, 203 Mencarelli, P. 168 Menozzi, M. 186, 197 Mente, S.R. 205, 213 Merbach, A.E. 205–208 Mercero, J.M. 36 Mesmer, R.E. 278 Meyer, R. 151 Meyer-Almes, F.J. 186, 188–190 Miecznik, P. 205, 210 Migneco, M. 293 Mihalic, Z. 126 Mikulski, C.M. 304 Millar, S.P. 16 Miller, E.C. 37, 97 Miller, J.A. 37, 97 Mills, I.M. 126 Mills, N.S. 154 Miranda, M.A. 231 Mison, P. 126, 151 Mitchell, P.A. 240 Mitchell, S.C. 86

AUTHOR INDEX Mitra, A. 156 Mitton, C.G. 297 Miyahara, Y. 210 Miyamoto, K. 284 Miyasato, M. 4 Miyazawa, E. 38, 70 Mizutani, Y. 173 Mo, G. 37, 56, 67, 72–74, 83–84, 86, 94, 99, 113 Mochida, K. 205, 208 Moeller, T. 276, 278 Moffatt, J.R. 70 Mohamed, M.F. 273, 276–278, 280, 282, 312, 318, 321 Mohammed, S.A. 59 Molenveld, P. 309 Momchilova, S. 304 Monaco, R.R. 173, 186, 190 Monahan, S.L. 186, 196 Mongelli, N. 186, 202 Montgomery, M.N. 36 Montoya-Pela´ez, P. 273, 275, 279, 282, 284, 324 Moreau, F. 253–254 Morgan, K.L. 36 Morokuma, K. 11, 13 Morrow, J.R. 309, 311–312, 315–316, 322 Mortelmans, K. 97 Moss, R.A. 271, 280, 283, 290, 309 Motekaitis, R.J. 276 Mpourmpakis, G. 235, 240 Muchall, H.M. 148 Mujika, J.I. 36 Mulder, A. 168 Mu¨ller, B. 152–153 Mu¨ller, K. 43 Mu¨ller, T. 135, 138, 149, 151, 159 Mu¨ller, W. 186, 194 Mulliken, R.S. 253 Mulzer, J. 86 Murov, S.L. 187 Mwakı´ bete, H. 205, 211 Myhre, P.C. 142, 146 Myli, K.B. 255, 257 Naﬁsi, K. 203 Nagae, O. 36 Nakajoh, M. 259 Nakamura, Y. 154

AUTHOR INDEX Nakatani, K. 186 Nalley, E.A. 36 Naruse, Y. 36 Nau, W.M. 176, 205, 214 Navarro-Va´zquez, A. 3, 5, 8 Neidle, S. 186, 198–199 Nelson, A. 168 Nencka, R. 243 Netzel, T.L. 203 Neumann, R. 261 Neuvonen, H. 57 Neuvonen, K. 57 Neverov, A.A. 271, 273, 275–280, 282, 284–285, 287, 302–303, 309–310, 312, 315–318, 321–324 Newton, M.D. 14 Nguyen, B. 186 Nicholas, J.B. 128, 155–156, 232 Nicolaou, K.C. 11 Nikitin, E. 126 Nishikawa, S. 205, 210–212 Nishimura, J. 154 Noguchi, E. 36 Nolet, M.-C. 177 Nomura, H. 210 Norde´n, B. 186, 195 Novak, M. 38–39 Novo, M. 179–180, 205, 213 Nyquist, R.A. 51, 55, 57 O’Brien, D.H. 157 O’Connor, D.V. 176 Ogul’Chansky, T.Y. 203 Ohno, M. 36 Ohwada, T. 36 Oka, T. 133 Okamoto, I. 36 Okamoto, Y. 65 Okano, L.T. 205, 215–216, 284 Okubo, T. 177, 205, 208 Olah, G.A. 126–128, 131, 133, 146, 150–151, 153, 156–158 Onodera, K. 253 O¨rstan, A. 205 Ortiz, J.A. 38, 70 O’Shannessy, D.J. 185 Osinsky, S.P. 20 Ostermeier, M. 149 Otani, Y. 36

343 Otto, A.H. 126, 154 Ovans, R. 177 Overman, L.E. 4 Owen, B.B. 273, 275 Paabo, M. 277 Pace, A. 231, 234, 241, 248 Pace, T.C.S. 167, 186, 196 Pa´linko´, I. 157 Palm, V.A. 290, 297 Palmer, L.C. 181 Palmer, M.H. 126 Palomares, E. 226 Palomino, E. 247 Pan, G.L. 231, 251, 259 Panda, M. 38 Panov, A.G. 254, 257 Panthananickal, A. 105 Paoletti, C. 186 Park, B.D. 271, 280, 283, 290, 309 Park, J.W. 208 Parker, A.J. 296 Parrish, D.A. 12 Parshall, G.W. 272, 280, 283 Pasanen, P. 57 Patra, D. 178–180 Patrick, D.A. 186 Paull, K.D. 35, 115 Peabody, S. 21 Pearce, S.W. 186, 197 Pelton, J.G. 186, 202 Peng, M.L. 246 Perera, S.A. 127, 148 Perlstein, D.L. 230 Perrin, C.L. 181–182 Peterman, D.R. 304 Peters, H.E. 38 Petersen, J.L. 27–28 Petrova, J. 304 Pettit, T.L. 235 Phillips, D.R. 176, 186, 197 Pidko, E.A. 254 Pincock, J.A. 6 Pink, M. 16 Pitchumani, K. 244 Pitterna, T. 12 Planelles, J.H. 225 Pleshkova, A.P. 37, 42, 54, 56, 89, 94 Plourde, G.W. 12

344 Poliks, M.D. 232 Pople, J.A. 46, 158 Porschke, D. 173, 186, 188–190 Prakash, A.S. 37, 39, 52–53, 70–71, 81, 97–99, 104–105, 107, 186, 194–195 Prakash, G.K.S. 126–128, 133, 146, 150–151, 153, 156–158 Prall, M. 3, 5, 8–9, 11, 22 Prasad, G. 247 Pregel, M.J. 272, 280, 283 Preston, J.F. 186 Price, C.A. 186, 202–203 Price, H. 52–53 Primo, J. 231 Prins, L.J. 168 Pritchard, J.G. 62 Pritchett, A. 103, 105 Pross, A. 70, 74 Prusik, T. 186, 194 Przystas, T.J. 271, 283 Pulay, P. 126, 154 Pullman, A. 71, 97 Pullman, B. 71, 97 Pulzonetti, M. 186, 203 Pytlewski, L.L. 304 Qian, X. 107 Raabe, I. 139 Rabalakos, C. 235, 240, 243 Rabinovitz, M. 4 Rachdi, F. 225 Rackham, D.M. 305 Rademacher, P. 36, 50, 55 Radom, L. 46 Ragg, E. 186, 202 Raghavachari, K. 129, 142 Rajagopal, S. 38 Rajski, S.R. 97 Ramachandran, K. 186 Ramamurthy, V. 225, 228, 230, 232–236, 243–244, 259 Ramlall, P. 38 Rampazzo, E. 309 Ramstein, J. 186, 193 Randall, W.J. 276, 278 Ranganayakulu, K. 131, 143 Rao, V.J. 230 Raptis, C. 243

AUTHOR INDEX Rassing, J. 175 Rassolov, V. 155 Rasul, G. 126–128, 133, 150–151, 153, 156–158 Rauchel, F.M. 296, 299 Rauh, S. 210 Rauk, A. 36–37, 47, 51, 56, 67, 72–74, 83–84, 94, 143 Raushel, F.M. 299 Rawat, D.S. 3, 16 Rawlings, J. 317 Rawlins, M.L.M. 36 Ray, S. 233 Rebek, J. 168, 181 Reddy, V.P. 133, 156 Redmond, R.W. 176 Reed, E.C. 39 Regenfuss, P. 186, 203 Rehfuss, B.D. 133 Rehm, D. 261 Reich, R.M. 172 Reichardt, C. 254 Reichardt, P.B. 4 Reichel, F. 126, 154 Reid, B.R. 186, 202 Reija, B. 179–180, 205, 213 Reimer, R. 186, 197 Reinhoudt, D.N. 168, 309 Reinsborough, V.C. 205, 208, 210 Rekharsky, M.V. 205 Ren, D. 38 Rest, A.J. 253 Rewcastle, G.W. 186, 194–196 Richard, J.P. 309, 315–316, 322 Richardson, K.S. 274 Rigler, R. 178, 181, 186, 193 Rinco, O. 177 Ritchie, C.D. 70 Rived, F. 272–273, 276–277, 290, 302 Rizzo, V. 186, 197 Robbins, R.J. 230, 232, 235–236, 243–244 Roberts, C.J. 183–184 Roberts, J.D. 146 Robinson, B.H. 171–173, 205, 208 Robinson, R.A. 277 Roche, C.J. 186, 197 Rodriguez, L.J. 205, 210–211 Roelens, S. 168 Rohrbach, R.P. 205, 210–211

AUTHOR INDEX Rollins, H.W. 304 Romagnoli, R. 105 Rosenfeld, J. 129 Rose´s, M. 272–273, 276–277, 290, 302 Roth, W.R. 3 Rowan, S.J. 168 Rowbottom, C.A. 37–39, 44, 52–53, 60, 65, 67, 70, 94, 97–98, 104–105 Rowe, J.E. 67 Rozie´re, J. 225 Ruane, P.H. 39 Rudchenko, V.F. 43, 58–59 Russell, K.C. 17 Ryan, J.J. 297 Sabater, M.-J. 231 Sabatini, A. 277, 279–280 Sabelko, J. 173 Sacchi, N. 186, 197 Sadat-Ebrahimi, S.E. 172, 186, 203 Saenger, W. 205–206, 208 Saionz, K.W. 186, 200 Saito, I. 186 Saito, K. 309, 316 Sakamoto, T. 38, 70 Sakoda, M. 186, 191 Sakuragi, H. 253 Saladino, A.C. 258 Salahub, D.R. 131–132 Sales, J. 272–273, 276–277, 290, 302 Salisbury, K. 253 Samardjiev, I.J. 36 Sander, W. 22 Sanders, J.K.M. 168 Sanderson, D.R. 233 Sandhagen, C. 179 Sandison, M. 247 Sandvick, P.E. 4 Sanjua´n, A. 261 Santarsiero, B.D. 58 Sartorius, J. 107 Sastre, G. 225 Saudan, C. 205–208 Saunders, M. 129, 143, 146, 148–149 Sauvage, J.-P. 168 Savin, A. 126 Scaiano, J.C. 176, 205, 225, 231–232, 261 Schaap, A.P. 247 Schiller, R.L. 205, 208–210, 213

345 Schindler, M. 126, 133, 143, 149–150, 154 Schleyer, P.v.R. 126–129, 135, 138, 142–144, 146, 148–151, 154, 158 Schmidt, A. 173 Schmidt, M.W. 64 Schmitt, H. 113 Schmitz, L.R. 131 Schneider, H.-J. 107 Schoonraad, J.L. 38, 70, 92 Schottelius, J. 20–21, 27 Schowen, R.L. 297 Schreiner, P.R. 3–9, 11–13, 22 Schroder, D. 39 Schroder, S. 37 Scho¨tz, K. 150, 154 Schuchardt, U. 261 Schuck, P. 183–185 Schulz-Ekloff, G. 235 Schumacher, R.R 52–53, 75, 81 Schurhammer, R. 305 Schwaller, M.-A. 186 Schwarz, H. 39 Scott, A.P. 38–39, 46, 70 Scrimin, P. 271, 280, 283, 290, 309 Secco, F. 186, 192, 203, 324 Seidel, C.A.M. 179–180, 205, 213 Seiyama, A. 205–208 Sekino, H. 127, 148 Sen, S.E. 225 Sha, F. 186, 201–203 Shafer, S.G. 70 Shaik, S. 7, 13, 22, 43 Shaikh, A.U. 37 Shailaja, J. 232, 235, 243–244 Shakesheff, K.M. 183–184 Shamma, T. 153 Shapley, J. 297 Sheehy, J.A. 156 Sheldon, R.A. 261 Shen, B. 3, 6 Shimada, M. 38, 70 Shimkin, M. 105 Shiota, T. 285, 316 Shirai, T. 205, 208 Shiro, M. 4, 285, 316 Shishkin, O.V. 37, 42, 47, 54 Shtamburg, V.G. 37, 42, 47, 54, 56, 73, 84, 89, 94

346 Shudo, K. 36 Shusterman, A.J. 105 Shustov, G.V. 43 Siddiqui, S. 247 Sieber, S. 126–129, 143–144, 146, 148–149, 151, 154 Siehl, H.-U. 125, 126, 129, 134–136, 138–139, 143–144, 146–147, 151–153, 159 Sigan, A.L. 20 Silverstein, R.M. 51 Simms, H.S. 279 Simon, A. 297 Singer, B. 97 Singleton, J. 235 Sivaguru, J. 232, 234–235 Skeels, G.W. 229 Skloss, T.W. 232 Smith, A.L. 11 Smith, C.T. 71 Smith, M.B. 287 Smith, S.M. 225 Snyder, J.P. 12 Sodeau, J.R. 253 Soﬁkiti, N. 240, 243 Somayaji, V. 58 Song, W. 155, 232 Songstad, J. 74 Sorensen, T.S. 131, 143–145, 149, 156 Spackman, M.A. 230 Spalluto, G. 105 Spatz, H.-C. 205–206, 208 S Prakash, G.K. 156 Springﬁeld, J.R. 36 Sram, J.P. 232, 236–239, 241 Srivastava, S. 39 Staley, R.H. 241–242 Stamatis, N. 139 Stange, G. 186, 193 Stanton, J.F. 125, 134–135, 159 Stanton, M.G. 272, 280, 283 Staral, J.S. 146 Stefaniak, K. 192, 205 Steitz, T.A. 309 Stephenson, L.M. 240, 257 Stern, C.L. 151 Stimson, V.R. 62, 64 Stoddart, J.F. 168 Stratakis, M. 235, 238, 240, 243–244, 246 Stra¨ter, N. 274, 309

AUTHOR INDEX Streck, R. 51, 55, 57 Strehlow, H. 173, 181–182 Subramanian, G. 158 Suck, D. 309 Suh, J. 271, 283 Sukhai, P. 38 Sullivan, K.A. 225 Sullivan, M.B. 39 Sun, F. 145 Sun, H. 253–258 Sun, S. 36 Sun, W. 225 Sunko, D.S. 144 Sunoj, R.B. 232, 235, 244 Sutherland, I.O. 86 Sutin, N. 173 Sutter, J.R. 172 Svensson, P. 154 Szejtli, J. 204 Tabushi, I. 205 Taeschler, C. 145, 149 Taft, R.W. 293 Taguchi, K. 205 Takahashi, Y.-i. 309, 316 Takeya, H. 259 Takisawa, N. 205, 211 Tamao, K. 4 Tang, S.L.Y. 255 Tanious, F.A. 186, 198–200 Tantillo, D.J. 145 Tardy, C. 186, 199–200 Tarling, S.E. 227 Tawarah, K. 205, 210–212 Taylor, D.K. 56 Tecilla, P. 309, 316, 325 Ten Have, J.F. 37, 39, 52, 70, 97–98 Tendler, S.J.B. 183–184 Thea, S. 299 Theiss, J. 105 Thomas, J.K. 176–177 Thomas, K.J. 259 Thomason, M.A. 205, 211 Thompson, L.C. 276, 278 Thompson, N.L. 178–180 Thomson, J.A. 186, 197 Thomson, L.M. 56, 73, 84 Thornton, E.R. 70 Thorpe, J.W. 80

AUTHOR INDEX Thorson, J.S. 3, 6 Throckmorton, J.R. 36, 40 Thurn, T. 174–175, 205 Tidwell, R.R. 186 Tidwell, T.T. 157 Timmerman, P. 168 Tine`, M.R. 186, 192 Tokumaru, K. 253 Tonellato, U. 309, 316, 325 To¨ro¨k, B. 157 Torres, P.D. 128, 155–156 Toscano, J.P. 39 Totah, N.I. 254, 257 Toutchkine, A. 248, 251 Troxler, L. 305 Trucks, G.W. 151 Truhlar, D.G. 39, 77 Truong, P.N. 64 Tsang, J.S.W. 273, 275, 284, 303, 310, 318, 321–322, 324 Tsubomura, H. 253 Tsuchiya, M. 253 Tsuno, Y. 66 Tsygankov, A.V. 37, 42, 47, 54, 56, 89, 94 Tsykalova, M.V. 20 Tu, B. 246 Tucker, D.J. 36, 37, 39, 44, 47–50, 56, 73–74, 85–86, 111, 113 Tung, C.H. 246–247 Tuppurainen, K. 105 Turnbull, W.B. 168 Turner, D.H. 173 Turner, P. 36–37, 39, 44, 47–50, 56, 73–74, 111, 113 Turro, N.J. 177, 205, 234 Uchiyama, M. 36 Ugawa, T. 205, 210–212 Uhl, W. 142 Usher, D.A. 310, 315 Vacca, A. 277, 279–280 Van Hoonacker, A. 183–185 Van Lanen, S.G. 3, 6 Van Losenoord, A. 90, 93 van Santen, R.A. 254 van Stam, J. 205, 214 Van Vranken, D.L. 186

347 Vancik, H. 126, 144 Vankar, Y.D. 156 vanLoon, G.W. 272, 280, 283 Vanoppen, D.L. 258 Vasenkov, S. 231, 253–255, 258 Vasudevan, S. 233 Veal, J.M. 186, 196 Venturini, M. 186, 192, 203 Verhas, M. 183–185 Verhe´, R. 80, 113 Verrall, R.E. 174–175, 210 Vetorini, M. 271, 324 Vichard, C. 322 Vieth, H.M. 142 Villani, R.P. 205, 208 Vina, J. 86 Vincer, K. 241 Vinkovic-Vrcek, I. 144 Vinogradov, A. 142 Virtanen, P.O.I. 70 Viruela, P.M. 225 Vo¨gtle, F. 167–168 Voigtmann, U. 113 Volbeda, S. 309 Volk, R. 173 von Hippel, P.H. 186 v.R. Schleyer, P. 46 Vrcek, V. 126, 129, 143–144 Wagnerova, 233 Wagstaff, K. 144 Wakelin, L.P.G. 186, 188–190, 194–196 Waldmeyer, J. 186, 193–194 Walenzyk, T. 3 Wallau, M. 261 Walton, J.C. 92 Wan, C. 173 Wan, Z. 259 Wandel, H. 12 Wang, H. 247 Wang, K.K. 27–28 Wang, Q. 156 Wang, Z. 155, 232 Ward, A.D. 67 Wardrop, D.J. 38, 70 Ware, W.R. 177 Waring, M.J. 186, 188–190 Warkentin, J. 80 Warner, B.P. 16

348 Warner, P.M. 12, 22 Webb, G. 146 Webb, W.W. 178, 186–188 Weber, H.-P. 230 Webster, F.X. 51 Wei, D.-Z. 107 Wei, W. 7, 13, 22 Weinhold, F. 21 Weiss, P.M. 299 Weller, A. 261 Wemmer, D.E. 186, 202 Wenk, H.H. 22 Wenthold, P.G. 5 Wenz, G. 205, 212 Wenzler, D.L. 4 Werner, K. 151 Werstiuk, N.H. 148 Westerlund, F. 186, 195, 203 Westman, G. 186, 203 White, A.M. 157 White, C.J. 318, 321 White, S. 241 Whitlock, H.W. 4 Wiberg, K.B. 35–36, 47, 50, 146, 149 Widengren, J. 178–179, 181 Wiest, O. 12 Wilen, S.H. 18, 242 Wilhelmsson, L.M. 186, 195 Wilkinson, F. 231 Williams, A.P. 173, 289, 292, 299 Williams, D.J. 304 Williams, J.C. 240 Williams, N.H. 305, 309, 315 Williams, R.M. 15, 97 Wilson, W.D. 184–186, 196–200, 202–203 Winkler, F.K. 44–45 Winzek, C. 173, 186, 188, 192 Winzor, D.J. 185 Wipf, G. 305 Wittkopp, A. 7, 9, 13, 22 Wo¨hrle, D. 235 Wojcik, J.F. 205, 210–211 Wolf, L.K. 186 Wolff, H. 173 Wolinski, K. 126 Wothers, P.D. 36, 55 Wu, L.Z. 246–247 Wyman, P. 309 Wyn-Jones, E. 174–175, 205, 210–212

AUTHOR INDEX Xiang, Y. 230–231, 257 Xu, M. 258 Xu, T. 128, 155–156, 232, 236, 238 Yamada, K. 309, 316 Yamada, S. 36, 58 Yamaguchi, K. 36, 205, 210–212, 240 Yamaguchi, S. 4, 205, 210, 212 Yamamoto, G. 36 Yamamoto, K. 173 Yamamura, H. 309, 316 Yang, M.-Y. 309, 315–316 Yao, W. 107, 142, 146 Yarmoluk, S.M. 186, 203 Yates, J.T. 231 Yates, Y.L. 43 Yavari, I. 146 Ybarra, C. 154 Yen, S.-F. 186, 199–200 Yim, C.T. 205, 212 Ying, Y.-M. 247 Yokoo, N. 205, 210–212 Yoon, K.B. 254 York, C. 157 Yoshida, N. 205–208 Young, E.R.R. 15 Yousef, M. 318 Yudin, A.K. 156 Yue, E.W. 11 Yukawa, Y. 66 Zagorujko, L.I. 20 Zaleski, J.M. 3, 5, 16 Zeidan, T. 19–20 Zeiger, E. 97 Zewail, A.H. 173 Zhang, D. 235 Zhang, J. 232, 236, 238 Zhang, L.P. 246–247 Zhang, L.Z. 225 Zhang, R.Y. 246 Zhang, S. 151 Zhang, X. 176, 205, 214 Zhang, Y. 255 Zhao, D. 246 Zhao, M. 203 Zhao, Y. 151 Zholobenko, V.L. 255

AUTHOR INDEX Zhou, W. 232, 248, 250, 253 Zhu, C. 255 Zhu, X.X. 205, 212 Zicovich-Wilson, C.M. 225–226 Zimmerman, H.E. 6

349 Zimmerman, S.C. 186, 196 Zolotoi, A.B. 43 Zompa, L. 318–319 Zon, G. 186, 196 Zubatyuk, R.I. 37, 42, 47, 54

This page intentionally left blank

Cumulative Index of Authors Abboud, J.-L.M., 37, 57 Ahlberg, P., 19, 223 Alabugin, I., 42, 1 Albery, W.J., 16, 87; 28, 139 Alden, J.A., 32, 1 Alkorta, I., 37, 57 Allinger, N.I., 13, 1 Amyes, T.L., 35, 67; 39, 1 Anbar, M., 7, 115 Antoniou, D., 41, 317 Arnett, E.M., 13, 83; 28, 45 Ballester, M., 25, 267 Bard, A.J., 13, 155 Basner, J., 41, 317 Baumgarten, M., 28, 1 Beer, P.D., 31, I Bell, R.P., 4, 1 Bennett, J.E., 8, 1 Bentley, T.W., 8, 151; 14, 1 Berg, U., 25, 1 Berger, S., 16, 239 Bernasconi, C.F., 27, 119; 37, 137 Berreau, L.M., 41, 81 Berti, P.J., 37, 239 Bethell, D., 7, 153; 10, 53 Blackburn, G.M., 31, 249 Blandamer, M.J., 14, 203 Bohne, C., 42, 167 Bond, A.M., 32, 1 Bowden, K., 28, 171 Brand, J.C.D., 1, 365 Bra¨ndstro¨m, A., 15, 267 Braun-Sand, S., 40, 201 Breiner, B., 42, 1 Brinker, U.H., 40, 1 Brinkman, M.R., 10, 53 Brown, H.C., 1, 35 Brown, R.S., 42, 271 Buncel, E., 14, 133 Bunton, C.A., 21, 213 Cabell-Whiting, P.W., 10, 129 Cacace, F., 8, 79

Capon, B., 21, 37 Carter, R.E., 10, 1 Chen, Z., 31, 1 Clennan, E.L., 42, 225 Collins, C.J., 2, 1 Compton, R.G., 32, 1 Cornelisse, J., 11, 225 Cox, R.A., 35, 1 Crampton, M.R., 7, 211 Datta, A., 31, 249 Da´valos. J.Z., 37, 57 Davidson, R.S., 19, 1; 20,191 de Gunst, G.P., 11, 225 de Jong, F., 17, 279 Denham, H., 31, 249 Desvergne, J.P., 15, 63 Detty, M.R., 39, 79 Dosunmu, M.I., 21, 37 Drechsler, U., 37, 315 Eberson, K., 12, 1; 18, 79; 31, 91 Eberson, U., 36, 59 Ekland, J.C., 32, 1 Eldik, R.V., 41, 1 Emsley, J., 26, 255 Engdahl, C., 19, 223 Farnum, D.G., 11. 123 Fendler, E.J., 8, 271 Fendler, J.H., 8, 271; 13, 279 Ferguson, G., 1, 203 Fields, E.K., 6, 1 Fife, T.H., 11, 1 Fleischmann, M., 10, 155 Frey, H.M., 4, 147 Fujio, M., 32, 267 Gale, P.A., 31, 1 Gao, J., 38, 161 Garcia-Viloca, M., 38, 161 Gilbert, B.C., 5, 53 Gillespie, R.J., 9, 1 Glover, S.A., 42, 35 Gold, V., 7, 259 Goodin, J.W., 20, 191 351

Gould, I.R., 20, 1 Greenwood, H.H., 4, 73 Gritsan, N.P., 36, 255 Hamilton, T.D., 40, 109 Hammerich, O., 20, 55 Harvey. N.G., 28, 45 Hasegawa, M., 30, 117 Havjnga, E., 11, 225 Henderson, R.A., 23, 1 Henderson, S., 23, 1 Hengge, A.C., 40, 49 Hibbert, F., 22, 113; 26, 255 Hine, J., 15, 1 Hogen-Esch, T.E., 15, 153 Hogeveen, H., 10, 29, 129 Horenstein, N.A., 41, 277 Hubbard, C.D., 41, 1 Huber, W., 28, 1 Ireland, J.F., 12, 131 Iwamura, H., 26, 179 Johnson, S.L., 5, 237 Johnstone, R.A.W., 8, 151 Jonsa¨ll, G., 19, 223 Jose´, S.M., 21, 197 Kemp, G., 20, 191 Kice, J.L., 17, 65 Kirby, A.J., 17, 183; 29, 87 Kitagawa, T., 30, 173 Kluger, R.H., 25, 99 Kochi, J.K., 29, 185; 35, 193 Kohnstam, G., 5, 121 Korolev, V.A., 30, 1 Korth, H.-G., 26, 131 Kramer, G.M., 11, 177 Kreevoy, M.M., 6, 63; 16, 87 Kunitake, T., 17, 435 Kurtz, H.A., 29, 273 Le Fe`vre, R.J.W., 3, 1 Ledwith, A., 13, 155 Lee, I., 27, 57 Lee, J.K., 38, 183 Liler, M., 11, 267 Lin, S.-S., 35, 67

352 Lodder, G., 37, 1 Logan, M.E., 39, 79 Long, F.A., 1, 1 Lu¨ning, U., 30, 63 Maccoll, A., 3, 91 MacGillivray, L.R., 40, 109 Mandolini, L., 22, 1 Manoharan, M., 42, 1 Maran, F., 36, 85 Matsson, O., 31, 143 McWeeny, R., 4, 73 Melander, L., 10, 1 Mile, B., 8, 1 Miller, S.I., 6, 185 Mo, Y., 38, 161 Modena. G., 9, 185 More O’Ferrall, R.A., 5, 331 Morsi, S.E., 15, 63 Mu¨llen, K., 28, 1 Mu¨ller, P., 37, 57 Nefedov, O.M., 30, 1 Nelsen, S.F., 41, 185 Neta, P., 12, 223 Neverov, A.A., 42, 271 Nibbering, N.M.M., 24, 1 Norman, R.O.C., 5, 33 Novak, M, 36, 167 Nu´n˜ez, S., 41, 317 Nyberg, K., 12, 1 O’Donoghue, A.M.C., 35, 67 Okamoto, K., 30, 173 Okuyama, T., 37, 1 Olah, G.A., 4, 305 Olsson, M.H.M., 40, 201 Oxgaard, J., 38, 87 Paddon-Row, M.N., 38, 1 Page, M.I., 23, 165 Parker, A.J., 5, 173 Parker. V.D., 19, 131; 20, 55 Peel, T.E., 9, 1 Perkampus, H.H., 4, 195 Perkins, M.J., 17, 1 Pittman, C.U., Jr., 4, 305 Platz, M.S., 36, 255 Pletcher, D., 10, 155 Poulsen, T.D., 38, 161 Pross, A., 14, 69; 21, 99 Quintanilla, E., 37, 57 Rajagopal, S., 36. 167

CUMULATIVE INDEX OF AUTHORS Rajca, A., 40, 153 Ramirez, F., 9, 25 Rappoport, Z., 7, 1; 27, 239 Rathore, R., 35, 193 Reeves, L.W., 3, 187 Reinboudt, D.N., 17, 279 Richard, J.P., 35, 67; 39, 1 Ridd, J.H., 16, 1 Riveros, J.M., 21, 197 Robertson, J.M., 1, 203 Romesberg, F.E., 39, 27 Rose, P.L., 28, 45 Rosenberg, M.G., 40, 1 Rosenthal, S.N., 13, 279 Rotello, V.M., 37, 3l5 Ruasse, M.-F., 28, 207 Russell, G.A., 23, 271 Saettel, N.j., 38, 87 Samuel, D., 3, 123 Sanchez, M. de N. de M., 21, 37 Sandstro¨m, J., 25, 1 Save´ant, J.-M., 26, 1; 35, 117 Savelli, G., 22, 213 Schaleger, L.L., 1, 1 Scheraga, H.A., 6, 103 Schleyer, P., von R., 14, 1 Schmidt, S.P., 18, 187 Schowen, R.L., 39, 27 Schuster, G.B., 18, 187; 22, 311 Schwartz, S.D., 41, 317 Scorrano, G., 13, 83 Shatenshtein, A.I., 1, 156 Shine, H.J., 13, 155 Shinkai, S., 17. 435 Siehl, H.-U., 23, 63 Siehl, H-U., 42, 125 Silver, B.L., 3, 123 Simonyi, M., 9, 127 Sinnott, M.L., 24, 113 Speranza, M., 39, 147 Stock, L.M., 1, 35 Strassner, T., 38, 131 Sugawara, T., 32, 219 Sustmann, R., 26. 131 Symons, M.C.R., 1, 284 Takashima, K., 21, 197 Takasu, I., 32, 219

Takeuchi, K., 30, 173 Tamara, C.S. Pace, 42, 167 Tanaka, K.S.E., 37, 239 Tantillo, D.J., 38, 183 Ta-Shma, R., 27, 239 Tedder, J.M., 16, 51 Tee, O.S., 29, 1 Thatcher, G.R.J., 25, 99 Thomas, A., 8, 1 Thomas, J.M., 15, 63 Tidwell. T.T., 36, 1 Tonellato, U., 9 185 Toteva, M.M., 35, 67; 39, 1 Toullec, J., 18, 1 Tsuji, Y., 35, 67; 39, 1 Tsuno, Y., 32, 267 Tu¨do¨s, F., 9, 127 Turner, D.W., 4, 31 Turro, N.J., 20, 1 Ugi, I., 9, 25 Walton, J.C., 16, 51 Ward, B., 8, 1 Warshel, A., 40, 201 Watt, C.I.F., 24, 57 Wayner, D.D.M., 36, 85 Wentworth, P., 31, 249 Westaway, K.C., 31, 143; 41, 219 Westheimer, F.H., 21, 1 Whalen, D.L., 40, 247 Whalley, E., 2, 93 Wiest, O., 38, 87 Williams, A., 27, 1 Williams, D.L.H., 19, 381 Williams, J.M., Jr., 6, 63 Williams, J.O., 16, 159 Williams, K.B., 35, 67 Williams, R.V., 29, 273 Williamson, D.G., 1, 365 Wilson, H., 14, 133 Wolf, A.P., 2, 201 Wolff, J.J., 32, 121 Workentin, M.S., 36, 85 Wortmaan, R., 32, 121 Wyatt, P.A.H., 12, 131 Zimmt, M.B., 20, 1 Zipse, H., 38, 111 Zollinger, H., 2, 163 Zuman, P., 5, 1

Cumulative Index of Titles Abstraction, hydrogen atom, from O—H bonds, 9, 127 Acid–base behaviour macroeycles and other concave structures, 30, 63 Acid–base properties of electronically excited states of organic molecules, 12, 131 Acid solutions, strong, spectroscopic observation of alkylcarbonium ions in, 4, 305 Acids, reactions of aliphatic diazo compounds with, 5, 331 Acids, strong aqueous, protonation and solvation in, 13, 83 Acids and bases, oxygen and nitrogen in aqueous solution, mechanisms of proton transfer between, 22, 113 Activation, entropies of, and mechanisms of reactions in solution, 1, 1 Activation, heat capacities of, and their uses in mechanistic studies, 5, 121 Activation, volumes of, use for determining reaction mechanisms, 2, 93 Addition reactions, gas-phase radical directive effects in, 16, 51 Aliphatic diazo compounds, reactions with acids, 5, 331 Alkene oxidation reactions by metal-oxo compounds, 38, 131 Alkyl and analogous groups, static and dynamic stereochemistry of, 25, 1 Alkylcarbonium ions, spectroscopic observation in strong acid solutions, 4, 305 Ambident conjugated systems, alternative protonation sites in, 11, 267 Ammonia liquid, isotope exchange reactions of organic compounds in, 1, S56 Anions, organic, gas-phase reactions of, 24, 1 Antibiotics, b-lactam, the mechanisms of reactions of, 23, 165 Aqueous mixtures, kinetics of organic reactions in water and, 14, 203 Aromatic photosubstitution, nucleophilic, 11, 225 Aromatic substitution, a quantitative treatment of directive effects in, 1, 35 Aromatic substitution reactions, hydrogen isotope effects in, 2, 163 Aromatic systems, planar and non-planar, 1, 203 N-Arylnitrenium ions, 36, 167 Aryl halides and related compounds, photochemistry of, 20, 191 Arynes, mechanisms of formation and reactions at high temperatures, 6, 1 A-SE2 reactions, developments In the study of, 6, 63 Base catalysis, general, of ester hydrolysis and related reactions, 5, 237 Basicity of unsaturated compounds, 4, 195 Bimolecular substitution reactions in protic and dipolar aprotic solvents, 5, 173 Bond breaking, 35, 117 Bond formation, 35, 117 Bromination, electrophilic, of carbon–carbon double bonds: structure, solvent and mechanisms, 28, 207 I3

C NMR spectroscopy in macromolecular systems of biochemical interest, 13, 279 Captodative effect, the, 26, 131 Carbanion reactions, ion-pairing effects in, 15,153 Carbene chemistry, structure and mechanism in, 7, 163 Carbenes generated within cyclodextrins and zeolites, 40, 1 353

354

CUMULATIVE INDEX OF TITLES

Carbenes having aryl substituents, structure and reactivity of, 22, 311 Carbocation rearrangements, degenerate, 19, 223 Carbocationic systems, the Yukawa–Tsuno relationship in, 32, 267 Carbocations, partitioning between addition of nucleophiles and deprotonation, 35, 67 Carbocations, thermodynamic stabilities of, 37, 57 Carbon atoms, energetic, reactions with organic compounds, 3, 201 Carbon monoxide, reactivity of carbonium ions towards, 10, 29 Carbonium ions, gaseous, from the decay of tritiated molecules, 8, 79 Carbonium ions, photochemistry of, 10, 129 Carbonium ions, reactivity towards carbon monoxide, 10, 29 Carbonium ions (alkyl), spectroscopic observation in strong acid solutions, 4, 305 Carbonyl compounds, reversible hydration of, 4, 1 Carbonyl compounds, simple, enolisation and related reactions of, 18, 1 Carboxylic acids, tetrahedral intermediates derived from, spectroscopic detection and investigation of their properties, 21, 37 Catalysis, by micelles, membranes and other aqueous aggregates as models of enzyme action, 17, 435 Catalysis, enzymatic, physical organic model systems and the problem of, 11, 1 Catalysis, general base and nucleophilic, of ester hydrolysis and related reactions, 5, 237 Catalysis, micellar, in organic reactions; kinetic and mechanistic implications, 8, 271 Catalysis, phase-transfer by quaternary ammonium salts, 15, 267 Catalytic antibodies, 31, 249 Cation radicals, in solution, formation, properties and reactions of, 13, 155 Cation radicals, organic, in solution, and mechanisms of reactions of, 20, 55 Cations, vinyl, 9, 135 Chain molecules, intramolecular reactions of, 22, 1 Chain processes, free radical, in aliphatic systems involving an electron transfer reaction, 23, 271 Charge density-NMR chemical shift correlation in organic ions, 11, 125 Charge distribution and charge separation in radical rearrangement reactions, 38, 111 Chemically induced dynamic nuclear spin polarization and its applications, 10, 53 Chemiluminesance of organic compounds, 18, 187 Chiral clusters in the gas phase, 39, 147 Chirality and molecular recognition in monolayers at the air–water interface, 28, 45 CIDNP and its applications, 10, 53 Computer modeling of enzyme catalysis and its relationship to concepts in physical organic chemistry, 40, 201 Computational studies of alkene oxidation reactions by metal-oxo compounds, 38, 131 Computational studies on the mechanism of orotidine monophosphate decarboxylase, 38, 183 Conduction, electrical, in organic solids, 16, 159 Conﬁguration mixing model: a general approach to organic reactivity, 21, 99 Conformations of polypeptides, calculations of, 6, 103 Conjugated molecules, reactivity indices, in, 4, 73 Cross-interaction constants and transition-state structure in solution, 27, 57 Crown-ether complexes, stability and reactivity of, 17, 279 Crystalographic approaches to transition state structures, 29, 87 Cycloaromatization reactions: the testing ground for theory and experiment, 42, 1 Cyclodextrins and other catalysts, the stabilisation of transition states by, 29, 1

CUMULATIVE INDEX OF TITLES

355

D2O—H2O mixtures, protolytic processes in, 7, 259 Degenerate carbocation rearrangements, 19, 223 Deuterium kinetic isotope effects, secondary, and transition state structure, 31, 143 Diazo compounds, aliphatic, reactions with acids, 5, 331 Diffusion control and pre-association in nitrosation, nitration, and halogenation, 16, 1 Dimethyl sulphoxide, physical organic chemistry of reactions, in, 14, 133 Dioleﬁn crystals, photodimerization and photopolymerization of, 30, 117 Dipolar aptotic and protic solvents, rates of bimolecular substitution reactions in, 5, 173 Directive effects, in aromatic substitution, a quantitative treatment of, 1, 35 Directive effects, in gas-phase radical addition reactions, 16, 51 Discovery of mechanisms of enzyme action 1947–1963, 21, 1 Displacement reactions, gas-phase nucleophilic, 21, 197 Donor/acceptor organizations, 35, 193 Double bonds, carbon–carbon, electrophilic bromination of: structure, solvent and mechanism, 28, 171 Dynamics for the reactions of ion pair intermediates of solvolysis, 39, 1 Dynamics of guest binding to supramolecular systems: techniques and selected examples, 42, 167 Effect of enzyme dynamics on catalytic activity, 41, 317 Effective charge and transition-state structure in solution, 27, 1 Effective molarities of intramolecular reactions, 17, 183 Electrical conduction in organic solids, 16, 159 Electrochemical methods, study of reactive intermediates by, 19, 131 Electrochemical recognition of charged and neutral guest species by redox-active receptor molecules, 31, 1 Electrochemistry, organic, structure and mechanism in, 12, 1 Electrode processes, physical parameters for the control of, 10, 155 Electron donor–acceptor complexes, electron transfer in the thermal and photochemical activation of, in organic and organometallic reactions. 29, 185 Electron spin resonance, identiﬁcation of organic free radicals, 1, 284 Electron spin resonance, studies of short-lived organic radicals, 5, 23 Electron storage and transfer in organic redox systems with multiple electrophores, 28, 1 Electron transfer, 35, 117 Electron transfer, in thermal and photochemical activation of electron donor-acceptor complexes in organic and organometallic reactions, 29, 185 Electron transfer, long range and orbital interactions, 38, 1 Electron transfer reactions within s- and p-bridged nitrogen-centered intervalence radical ions, 41, 185 Electron-transfer, single, and nucleophilic substitution, 26, 1 Electron-transfer, spin trapping and, 31, 91 Electron-transfer paradigm for organic reactivity, 35,193 Electron-transfer reaction, free radical chain processes in aliphatic systems involving an, 23, 271 Electron-transfer reactions, in organic chemistry, 18, 79 Electronically excited molecules, structure of, 1, 365 Electronically excited states of organic molecules, acid-base properties of, 12, 131 Energetic tritium and carbon atoms, reactions of, with organic compounds, 2, 201

356

CUMULATIVE INDEX OF TITLES

Enolisation of simple carbonyl compounds and related reactions, 18, 1 Entropies of activation and mechanisms of reactions in solution, 1, 1 Enzymatic catalysis, physical organic model systems and the problem of, 11, 1 Enzyme action, catalysis of micelles, membranes and other aqueous aggregates as models of, 17, 435 Enzyme action, discovery of the mechanisms of, 1947–1963, 21, 1 Equilibrating systems, isotope effects in NMR spectra of, 23, 63 Equilibrium constants, NMR measurements of, as a function of temperature, 3, 187 Ester hydrolysis, general base and nucleophitic catalysis, 5, 237 Ester hydrolysis, neighbouring group participation by carbonyl groups in, 28, 171 Excess acidities, 35, 1 Exchange reactions, hydrogen isotope, of organic compounds in liquid ammonia, 1, 156 Exchange reactions, oxygen isotope, of organic compounds, 2, 123 Excited complexes, chemistry of, 19, 1 Excited molecular, structure of electronically, 3, 365 Finite molecular assemblies in the organic solid state: toward engineering properties of solids, 40, 109 Fischer carbene complexes, 37, 137 Force-ﬁeld methods, calculation of molecular structure and energy by, 13, 1 Free radical chain processes in aliphatic systems involving an electron-transfer reaction, 23, 271 Free Radicals 1900–2000, The Gomberg Century, 36, 1 Free radicals, and their reactions at low temperature using a rotating cryostat, study of, 8, 1 Free radicals, identiﬁcation by electron spin resonance, 1, 284 Gas-phase heterolysis, 3, 91 Gas-phase nucleophilic displacement reactions, 21, 197 Gas-phase pyrolysis of small-ring hydrocarbons, 4, 147 Gas-phase reactions of organic anions, 24, 1 Gaseous carbonium ions from the decay of tritiated molecules, 8, 79 General base and nucleophilic catalysis of ester hydrolysis and related reactions, 5, 237 The Gomberg Century: Free Radicals 1900–2000, 36, 1 Gomberg and the Nobel Prize. 36, 59 H2O—D2O mixtures, protolytic processes in, 7, 259 Halides, aryl, and related compounds, photochemistry of, 20, 191 Halogenation, nitrosation, and nitration, diffusion control and pre-association in, 16, 1 Heat capacities of activation and their uses in mechanistic studies, 5, 121 Heterolysis, gas-phase, 3, 91 High-spin organic molecules and spin alignment in organic molecular assemblies, 26, 179 Homoaromaticity, 29, 273 How does structure determine organic reactivity, 35, 67 Hydrated electrons, reactions of, with organic compounds, 7, 115 Hydration, reversible, of carbonyl compounds, 4, 1 Hydride shifts and transfers, 24, 57 Hydrocarbon radical cations, structure and reactivity of, 38, 87 Hydrocarbons, small-ring, gas-phase pyrolysis of, 4, 147

CUMULATIVE INDEX OF TITLES

357

Hydrogen atom abstraction from 0—H bonds, 9, 127 Hydrogen bonding and chemical reactivity, 26, 255 Hydrogen isotope effects in aromatic substitution reactions, 2, 163 Hydrogen isotope exchange reactions of organic compounds in liquid ammonia, 1, 156 Hydrolysis, ester, and related reactions, general base and nucleophilic catalysis of, 5, 237 Interface, the air-water, chirality and molecular recognition in monolayers at, 28, 45 Intermediates, reactive, study of, by electrochemical methods, 19, 131 Intermediates, tetrahedral, derived from carboxylic acids, spectroscopic detection and investigation of their properties, 21, 37 Intramolecular reactions, effective molarities for, 17, 183 Intramolecular reactions, of chain molecules, 22, 1 Ionic dissociation of carbon-carbon a-bonds in hydrocarbons and the formation of authentic hydrocarbon salts, 30, 173 Ionization potentials, 4, 31 Ion-pairing effects in carbanion reactions, 15, 153 Ions, organic, charge density-NMR chemical shift correlations, 11, 125 Isomerization, permutational, of pentavalent phosphorus compounds, 9, 25 Isotope effects and quantum tunneling in enzyme-catalyzed hydrogen transfer. Part I. The experimental basis, 39, 27 Isotope effects, hydrogen, in aromatic substitution reactions, 2, 163 Isotope effects, magnetic, magnetic ﬁeld effects and, on the products of organic reactions, 20, 1 Isotope effects, on NMR spectra of equilibrating systems, 23, 63 Isotope effects, steric, experiments on the nature of, 10, 1 Isotope exchange reactions, hydrogen, of organic compounds in liquid ammonia, 1, 150 Isotope exchange reactions, oxygen, of organic compounds, 3, 123 Isotopes and organic reaction mechanisms, 2, 1 Kinetics, and mechanisms of reactions of organic cation radicals in solution, 20, 55 Kinetics and mechanism of the dissociative reduction of C—X and X—X bonds (X ¼ O, S), 36, 85 Kinetic and mechanistic studies of the reactivity Zn–Ohn (n = 1 or 2) species in small molecule analogs of zinc-containing metalloenzymes, 41, 81 Kinetics and spectroscopy of substituted phenylnitrenes, 36, 255 Kinetics, of organic reactions in water and aqueous mixtures, 14, 203 Kinetics, reaction, polarography and, 5, 1 b-Lactam antibiotics, mechanisms of reactions, 23, 165 Least nuclear motion, principle of, 15, 1 Macrocyles and other concave structures, acid-base behaviour in, 30, 63 Macromolecular systems of biochemical interest, 13C NMR spectroscopy in, 13, 279 Magnetic ﬁeld and magnetic isotope effects on the products of organic reactions, 20, 1 Mass spectrometry, mechanisms and structure in: a comparison with other chemical processes, 8, 152 Matrix infrared spectroscopy of intermediates with low coordinated carbon silicon and germanium atoms, 30, 1

358

CUMULATIVE INDEX OF TITLES

Mechanism and reactivity in reactions of organic oxyacids of sulphur and their anhydrides, 17, 65 Mechanism and structure, in carbene chemistry, 7, 153 Mechanism and structure, in mass spectrometry: a comparison with other chemical processes, 8, 152 Mechanism and structure, in organic electrochemistry, 12, 1 Mechanism of the dissociative reduction of C—X and X—X bonds (XQO, S), kinetics and, 36, 85 Mechanisms for nucleophilic aliphatic substitution at glycosides, 41, 277 Mechanisms of hydrolysis and rearrangements of epoxides, 40, 247 Mechanisms of oxygenations in zeolites, 42, 225 Mechanisms, nitrosation, 19, 381 Mechanisms, of proton transfer between oxygen and nitrogen acids and bases in aqueous solutions, 22, 113 Mechanisms, organic reaction, isotopes and, 2, 1 Mechanisms of reaction, in solution, entropies of activation and, 1, 1 Mechanisms of reaction, of b-lactam antibiotics, 23, 165 Mechanisms of solvolytic reactions, medium effects on the rates and, 14, 10 Mechanistic analysis, perspectives in modern voltammeter: basic concepts and, 32, 1 Mechanistic applications of the reactivity–selectivity principle, 14, 69 Mechanistic studies, heat capacities of activation and their use, 5, 121 Mechanistic studies on enzyme-catalyzed phosphoryl transfer, 40, 49 Medium effects on the rates and mechanisms of solvolytic reactions, 14, 1 Meisenheimer complexes, 7, 211 Metal-catalyzed alcoholysis reactions of carboxylate and organophosphorus esters, 42, 271 Metal complexes, the nucleophilicity of towards organic molecules, 23, 1 Methyl transfer reactions, 16, 87 Micellar catalysis in organic reactions: kinetic and mechanistic implications, 8, 271 Micelles, aqueous, and similar assemblies, organic reactivity in, 22, 213 Micelles, membranes and other aqueous aggregates, catalysis by, as models of enzyme action, 17, 435 Molecular recognition, chirality and, in monolayers at the air-water interface, 28, 45 Molecular structure and energy, calculation of, by force-ﬁeld methods, 13, 1 N-Acyloxy-N-alkoxyamides – structure, properties, reactivity and biological activity, 42, 35 N-Arylnitrinium ions, 36, 167 Neighbouring group participation by carbonyl groups in ester hydrolysis, 28, 171 Nitration, nitrosation, and halogenation, diffusion control and pre-association in, 16, 1 Nitrosation, mechanisms, 19, 381 Nitrosation, nitration, and halogenation, diffusion control and pre-association in, 16, 1 NMR chemical shift-charge density correlations, 11, 125 NMR measurements of reaction velocities and equilibrium constants as a function of temperature, 3, 187 NMR spectra of equilibriating systems, isotope effects on, 23, 63 NMR spectroscopy, 13C, in macromolecular systems of biochemical interest, 13, 279 Nobel Prize, Gomberg and the, 36, 59 Non-linear optics, organic materials for second-order, 32, 121 Non-planar and planar aromatic systems, 1, 203

CUMULATIVE INDEX OF TITLES

359

Norbornyl cation: reappraisal of structure, 11, 179 Nuclear magnetic relaxation, recent problems and progress, 16, 239 Nuclear magnetic resonance see NMR Nuclear motion, principle of least, 15, 1 Nuclear motion, the principle of least, and the theory of stereoelectronic control, 24, 113 Nucleophiles, partitioning of carbocations between addition and deprotonation, 35, 67 Nucleophili aromatic photolabstitution, 11, 225 Nucleophilic catalysis of ester hydrolysis and related reactions, 5, 237 Nucleophilic displacement reactions, gas-phase, 21, 197 Nucleophili substitution, in phosphate esters, mechanism and catalysis of, 25, 99 Nucleophilic substitution, single electron transfer and, 26, 1 Nucleophilic substitution reactions in aqueous solution, 38, 161 Nuckophilic vinylic substitution, 7, 1 Nucleophilic vinylic substitution and vinyl cation intermediates in the reactions of vinyl iodonium salts. 37, 1 Nucleophilicity of metal complexes towards organic molecules, 23, 1 O—H bonds, hydrogen atom abstraction from, 9, 127 One- and two-electron oxidations and reductions of organoselenium and organotellurium compounds, 39, 79 Orbital interactions and long-range electron transfer, 38, 1 Organic materials for second-order non-linear optics, 32, 121 Organic reactivity, electron-transfer paradigm for, 35, 193 Organic reactivity, structure determination of, 35, 67 Orotidine monophosphate decarboxylase, the mechanism of, 38, 183 Oxyacids of sulphur and their anhydrides, mechanisms and reactivity in reactions of organic, 17, 65 Oxygen isotope exchange reactions of organic compounds, 3, 123 Partitioning of carbocations between addition of nucleophiles and deprotonation, 35, 67 Perchloro-organic chemistry: structure, spectroscopy and reaction pathways, 25, 267 Permutations isomerization of pentavalent phosphorus compounds, 9, 25 Phase-transfer catalysis by quaternary ammonium salts, 15, 267 Phenylnitrenes, Kinetics and spectroscopy of substituted, 36, 255 Phosphate esters, mechanism and catalysis of nuclcophilic substitution in, 25, 99 Phosphorus compounds, pentavalent, turnstile rearrangement and pseudoration in permutational isomerization, 9, 25 Photochemistry, of aryl halides and related compounds, 20, 191 Photochemistry, of carbonium ions, 9, 129 Photodimerization and photopolymerization of dioleﬁn crystals, 30, 117 Photosubstitution, nucleophilic aromatic, 11, 225 Planar and non-planar aromatic systems, 1, 203 Polarizability, molecular refractivity and, 3, 1 Polarography and reaction kinetics, 5, 1 Polypeptides, calculations of conformations of, 6, 103 Pre-association, diffusion control and, in nitrosation, nitration, and halogenation, 16, 1 Principle of non-perfect synchronization, 27, 119 Products of organic reactions, magnetic ﬁeld and magnetic isotope effects on, 30, 1

360

CUMULATIVE INDEX OF TITLES

Protic and dipolar aprotic solvents, rates of bimolecular substitution reactions in, 5, 173 Protolytic processes in H2O—D2O mixtures, 7, 259 Proton transfer between oxygen and nitrogen acids and bases in aqueous solution, mechanisms of, 22, 113 Protonation and solvation in strong aqueous acids, 13, 83 Protonation sites in ambident conjugated systems, 11, 267 Pseudorotation in isomerization of pentavalent phosphorus compounds, 9, 25 Pyrolysis, gas-phase, of small-ring hydrocarbons, 4, 147 Radiation techniques, application to the study of organic radicals, 12, 223 Radical addition reactions, gas-phase, directive effects in, 16, 51 Radical rearrangement reactions, charge distribution and charge separation in, 38, 111 Radicals, cation in solution, formation, properties and reactions of, 13, 155 Radicals, organic application of radiation techniques, 12, 223 Radicals, organic cation, in solution kinetics and mechanisms of reaction of, 20, 55 Radicals, organic free, identiﬁcation by electron spin resonance, 1, 284 Radicals, short-lived organic, electron spin resonance studios of, 5, 53 Rates and mechanisms of solvolytic reactions, medium effects on, 14, 1 Reaction kinetics, polarography and, 5, 1 Reaction mechanisms, in solution, entropies of activation and, 1, 1 Reaction mechanisms, use of volumes of activation for determining, 2, 93 Reaction velocities and equilibrium constants, NMR measurements of, as a function of temperature, 3, 187 Reactions, in dimethyl sulphoxide, physical organic chemistry of, 14, 133 Reactions, of hydrated electrons with organic compounds, 7, 115 Reactive intermediates, study of, by electrochemical methods, 19, 131 Reactivity, organic, a general approach to; she conﬁguration mixing model, 21, 99 Reactivity indices in conjugated molecules, 4, 73 Reactivity-selectivity principle and its mechanistic applications, 14, 69 Rearrangements, degenerate carbocation, 19, 223 Receptor molecules, redox-active, electrochemical recognition of charged and neutral guest species by, 31, 1 Redox and recognition processes, interplay between, 37, 315 Redox systems, organic, with multiple electrophores, electron storage and transfer in, 28, 1 Reduction.of C—X and X—X bonds (XQO, S), kinetics and mechanism of the dissociative, 36, 85 Refractivity, molecular, and polarizability, 3, 1 Relaxation, nuclear magnetic, recent problems and progress, 16, 239 Selectivity of solvolyses and aqueous alcohols and related mixtures, solvent-induced changes in, 27, 239 Short-lived organic radicals, electron spin resonance studies of, 5, 53 Small-ring hydrocarbons, gas-phase pyrolysis of, 4, 147 Solid state, tautomerism in the 32, 129 Solid-state chemistry, topochemical phenomena in, 15, 63 Solids, organic, electrical conduction in, 16, 159 Solutions, reactions in, entropies of activation and mechanisms, 1, 1 Solvation and protonation in strong aqueous acids, 13, 83

CUMULATIVE INDEX OF TITLES

361

Solvent effects, reaction coordinates, and reorganization energies on nucleophilic substitution reactions in aqueous solution, 38, 161 Solvent, protic and dipolar aprotic, rates of bimolecular substitution-reactions in, 5, 173 Solvent-induced changes in the selectivity of solvolyses in aqueous alcohols and related mixtures, 27, 239 Solvolytic reactions, medium effects on the rates and mechanisms of, 14, 1 Spectroscopic detection of tetrahedral intermediates derived from carboxylic acids and the investigation of their properties, 21, 37 Spectroscopic observations ofalkylcarbonium ions in strong acid solutions, 4, 305 Spectroscopy, 13C NMR. in macromolecular systems of biochemical interest, 13, 279 Spectroscopy of substituted phenylnitrenes, kinetics and 36, 255 Spin alignment, in organic molecular assemblies, high-spin organic molecules and 26, 179 Spin trapping, 17, 1 Spin trapping, and electron transfer, 31, 91 Stability and reactivity of crown-ether complexes, 17, 279 Stereochemistry, static and dynamic, of alkyl and analogous groups, 25, 1 Stereoelectronic control, the principle of least nuclear motion and the theory of, 24, 113 Stereoselection in elementary steps of organic reactions, 6, 185 Steric isotope effects, experiments on the nature of, 10, 1 Structure, determination of organic reactivity, 35, 67 Structure and mechanism, in curbene chemistry, 7, 153 Structure and mechanism, in organic electrochemistry, 12, 1 Structure and reactivity of carbencs having aryl substitutents, 22, 311 Structure and reactivity of hydrocarbon radical cations, 38, 87 Structure of electronically excited molecules, 1, 365 Substitution, aromatic, a quantitative treatment of directive effects in, 1, 35 Substitution, nueleophilic vinylic, 7, 1 Substitution reactions, aromatic, hydrogen isotope effects in, 2, 163 Substitution reactions, bimolecular, in protic and dipolar aprotic solvents, 5, 173 Sulphur, organic oxyacids of, and their anhydrides, mechanisms and reactivity in reactions of, 17, 65 Superacid systems, 9, 1 Tautomerism in the solid state, 32, 219 Temperature, NMR measurements of reaction velocities and equilibrium constants as a function of, 3, 187 Tetrahedral intermediates, derived from carboxylic acids, spectroscopic detection and the investigation of their properties, 21, 37 The interplay between experiment and theory: computational NMR spectroscopy of carbocations, 42, 125 The interpretation and mechanistic signiﬁcance of activation volumes for organometallic reactions, 41, 1 The physical organic chemistry of very high-spin polyradicals, 40, 153 Thermodynamic stabilities of carbocations, 37, 57 Topochemical phenomena in solid-slate chemistry, 15, 63 Transition state analysis using multiple kinetic isotope effects, 37, 239 Transition state structure, crystallographic approaches to, 29, 87 Transition state structure, in solution, effective charge and 27, 1

362

CUMULATIVE INDEX OF TITLES

Transition stale structure, secondary deuterium isotope effects and, 31, 143 Transition states, structure in solution, cross-interaction constants and, 27, 57 Transition states, the stabilization of by cyclodextrins and other catalysts, 29, 1 Transition states, theory revisited, 28, 139 Tritiated molecules, gaseous carbonium ions from the decay of, 8, 79 Tritium atoms, energetic reactions with organic compounds, 2, 201 Turnstile rearrangements in isomerization of pentavalent phosphorus compounds, 9, 25 Unsaturated compounds, basicity of, 4, 195 Using kinetic isotope effects to determine the structure of the transition states of SN2 reactions, 41, 219 Vinyl cation intermediates, 37, 1 Vinyl cations, 9, 185 Vinyl iodonium salts, 37, 1 Vinylic substitution, nuclephilic, 7, 1; 37, 1 Voltammetry, perspectives in modern: basic concepts and mechanistic analysis, 32, 1 Volumes of activation, use of, for determining reaction mechanisms, 2, 93 Water and aqueous mixtures, kinetics of organic reactions in, 14, 203 Yukawa–Tsuno relationship in carborationic systems, the, 32, 267

SUBJECT INDEX

thermal decomposition reactions, see Thermal decomposition reactions N-acyloxy-N-alkoxyamides, structure, 43–47 anomeric effects, 44, 58 infrared spectroscopy, 51–56 NMR spectroscopy, see Nuclear magnetic resonance spectroscopy sp2 hybridisation, 43 sp3 hybridisation, 44, 49–50 X-ray structures, 47–51 N-acyloxy-N-alkoxybenzamides, solvolysis reactions, 64–65 N-acyloxy-N-alkoxycarbamate, structure, 42, 47–48, 49f N-acyloxy-N-alkoxynitrenium ions, 63, 68, 99–100 N-acyloxy-N-alkoxyurea, structure, 43, 47–48, 49f, 56 Alcohol, potentiometric titration in, 276–278 with metal ions, 278–284 Alcoholysis, with transition metal ion and Ln3+ catalysts of carboxylate esters, 288–294 of neutral phosphate esters, 294–308 concerted mechanism, 305–308 metal–methoxide reactions, 300–305 methoxide reactions, 299–300 phosphates and phosphorothioates, phosphonates and phosphonothioates, 296–299 of phosphate diesters, 308–324 exalted catalysis of methanolysis of 2-hydroxypropyl-p-nitrophenyl phosphate (HPNPP) in methanol, 318–324 metal-catalyzed alcoholysis of an RNA model, 310–316 Zn2+ ligand models for dinuclear enzymes promoting the cleavage of RNA, 316–318

N-acetoxy-N-acetyl arylamines, 39 N-acetoxy-N-alkoxybenzamides, 59 acid-catalysed solvolysis of, 65t acid independent and uncatalysed solvolysis of, 61t SN2 reaction of N-methylaniline with, 79t, 100 N-acetoxy-N-butoxybenzamides, 60, 66–67, 115 N-acetoxy-N-tert-butoxybenzamide, 55 activity levels of, 100 nucleophilic substitution reactions, 71, 78, 86 solvolysis reactions, 60, 66–67, 68 Acridine derivatives, binding with DNA, 190–194 N-acyloxy-N-alkoxyamides, 37–39 biological activity, see N-acyloxyN-alkoxyamides, biological reactivity chemical reactivity, see N-acyloxyN-alkoxyamides, chemical reactivity N-alkoxy-N-acyl nitrenium ions, 38 resonance forms in twisted or pyramidal, 55f spectroscopic properties, see Nuclear magnetic resonance spectroscopy structure, see N-acyloxy-N-alkoxyamides, structure synthesis, 39–43 N-acyloxy-N-alkoxyamides, biological activity, 97–115 anticancer activity of, 115 mutagenicity of, in Ames Salmonella/ microsome assay, 97–115 N-acyloxy-N-alkoxyamides, chemical reactivity, 59–96 factors contributing, 59–60 nucleophilic substitution reactions, see Nucleophilic substitution reactions solvolysis studies, see Solvolysis 363

364 N-alkoxy-N-acyl nitrenium ions, 38 Alkoxynitrenium ions, see N-AcyloxyN-alkoxynitrenium ions Alkyl and cycloalkylmethyl cations, NMR spectroscopy, 126–133 bridged cyclobutyl cations, 145–146 2-butyl cation, 129 cyclopropylethyl cations, 131 cyclopropylmethyl cation, 131, 145, 146, 147 diprotonated methane, 126–127 E-1-cyclopropyl-2-(triisopropylsilyl)ethyl cation, 131, 132f, 134f E-1-cyclopropyl-2-(trimethylsilyl)ethyl cation, 131, 132, 134f ethyl cation, 127 GIAO-DFT methods, 129–131 isopropyl cations, 128 2-methyl-2-butyl cation, 129 2-methyl-1-triisopropylsilylpropyl2-cation, 128 propyl cations, 127 protonated methane, 126 tert-butyl cation, 128 Amides, 35–36, see also N-acyloxyN-alkoxyamides anomeric, 37 linkages, 35 resonance and HOMO–LUMO interaction, 36f twisted, 36, 36f 9-Aminoacridine carboxamide derivatives binding with DNA, 194, 195 Anomeric amides, 37, see also N-acyloxyN-alkoxyamides Antiaromatic region, in enediynes, 14, 15 Aromatic amines, 38 metabolism, 37 reaction of N-acyloxy-N-alkoxyamides with, see HERON Arylnitrenium ions, 37–38 1-Azaadamantan-2-one, 36, 37f Azo dyes, for binding dynamics study of guests cyclodextrins (CD), 205–208 Bergman cyclization, 3, 16, 25 acid-catalyzed, 20 activation energy, 10–11 deceleration of, 12, 12f

SUBJECT INDEX effect of protonation of amino enediynes, 20–21 electronic effects, 6–10 enediyne, 3, 18, 25 MO correlation diagrams for, 8f photochemical, 15, 28–29 radical-anionic, 10f, 25, 26f rehybridization effects in, 21–22 stabilizing p–p* interactions, 13–14 syn-OMe groups and ortho substituents role, 17–19 Bicyclic and polycyclic carbocations, NMR spectroscopy, 145–150 1, 2-dimethyl-2-norbornyl cation, 149 bicyclobutonium ions, 145–146, 146f 3-endo-trialkylsilylbicyclobutonium ions, 147–148 1-methylbicyclobutonium ion, 147 2-methyl-2-norbornyl cation, 149 2-norbornyl cation, hypercoordinated, 148 silanorbornyl cations, 149 1-silylcyclobutyl cation, 147 Bicyclobutonium ions, 131, 145–146, 147–148 Bisoxo substitution, 44, 48, 58 Bridged cyclobutyl cations, 146, see also Bicyclobutonium ions "Camel through the eye of a needle" syntheses, in zeolites, 231, 232f Carboxylate esters, alcoholysis of , with transition metal ion and Ln3+ catalysts, 288–294 Catalysts, alcoholysis of transition metal ion and Ln3+, see Transition metal ion and Ln3+ catalysts, alcoholysis C1–C5 cyclization, 9, 10f enediyne radical-anions, 25, 25f, 26f of enediynes, 5f Charge transfer (CT) intrazeolite photooxygenation, 253–254, 255, 257 Cis-effect, in singlet oxygen ene reaction, 240, 243 Concerted mechanism, of neutral phosphate esters, 305–308 general mechanism, 306 Cyanine dyes, for binding dynamics study of guests to DNA, 203, 204

SUBJECT INDEX Cyclic enediynes, see under Enediynes Cycloalkyl cations, NMR spectroscopy, 142–144 cyclopentyl cation, 142 methylcyclohexyl cation, 143 nonamethylcyclopentyl cation, 143 Cycloalkylmethyl cations, NMR spectroscopy, see Alkyl and cycloalkylmethyl cations Cycloaromatization reactions diversity of, 3–6 electronic effects, 7–10 of enediynes, 3–6, 4f MO correlation diagrams for, 7–9, 8f p-bond, 2–3 effects in early stage, 7–10 in product stability, 30–31 on reactant stability, 22–23 product stabilization, 27–31 rehybridization effects in, 21–22 s-bond, 2–3 effects in early stage, 7–10 interaction of non-bonding electrons, 27 on reactant stability, 10–22 schematic representation of, 2f steric assistance mechanism for ortho substituents in, 17–20 steric repulsion of ortho substituents of enediynes in, 20–21 strain effect on, 28–29 zwitterionic products in, 27–28 Cyclodextrins (CD), study of binding dynamics of guest molecules to, 204–205 using azo dyes, 205–208 ﬂuorescence correlation spectroscopy, 213 hydroxyphenylazo arylsulfonate derivatives binding, 205–207 laser ﬂash photolysis, 215–216 NMR experiments, 212–213 stopped-ﬂow experiments, 205, 207, 208 temperature jump experiments, 208–210 time-resolved ﬂuorescence experiments, 214 ultrasonic relaxation experiments, 210–212

365 Daunomycin, binding dynamics of DNA with, 197–198 DNA, study of binding dynamics of guests molecules to, 186 acridine derivatives binding with, 190–194 9-aminoacridine carboxamide derivatives binding with, 194, 195 cyanine dyes, 203, 204 daunomycin binding, 197–198 distamycin binding with, 201–203, 204 ethidium bromide binding with, 186–187, 188t, 189–190 ﬂuorescence correlation spectroscopy, 187–188, 188t groove binding guests, 201–204 Hoechst dyes, 203–204 intercalative guest molecules, 186–201 laser ﬂash photolysis, 193–194 naphthalene diimide with alkylamino substituents binding with, 199–200 NMR study, 196–197, 202 proﬂavine binding with, 190–194 stopped-ﬂow methodology, 187, 189–191, 194, 195–196 surface plasmon resonance studies, 199–200, 202–203 temperature jump experiments, 188–191, 192, 193, 197, 203 time-resolved ﬂuorescence, 201 Dynamic 1H NMR spectroscopy, for N-acyloxy-N-alkoxyamides, 59 Enediynes amino, 12t, 20, 21t antiaromatic region in, 14–15 Bergman cyclization, 3, 18, 25 C1–C5 cyclization of, 5f, 25 cyclic, strain and antiaromaticity in, 11–16 cyclization of, with terminal ﬂuorosubstituents, 22 dianionic cyclization, 4, 6f ligand–metal coordination to control strain, 16–17 9-membered, 11–12, 12f 10-membered, 11 11-membered, 12 steric assistance mechanism for ortho substituents for strain control, 17–20

366 steric repulsion of ortho substituents for strain control, 20–21 tetradentate enediyne ligands, for activation, 16–17, 18f Esters, see Carboxylate esters; Phosphate diesters; Phosphate esters Ethidium bromide, binding with DNA, 186–187, 188t, 189–190 Faujasite (FAU), 226–227 ball and stick structure of, 226–227, 227f Fluorescence correlation spectroscopy, 178–181 antibunching phenomenon, 179 cyclodextrins (CD), binding dynamics of guests binding to, 213 DNA, binding dynamics of guests binding to, 187–188, 188t N-Formyloxy-N-methoxyformamide, 111 HERON migration transition states, 95, 96t HERON reactions of, 75–76, 77t reaction of azide with, 85f structure, 44, 45f, 46f, 47 Frei intrazeolite photooxygenation, 253f, 255, 258, 261, see also Wagnerova Class II intrazeolite photooxygenation GIAO-DFT methods, 126 alkyl and cycloalkylmethyl cations, 129–131 GIAO method (Gauge Independent Atomic Orbitals), 126 GIAO-MP2/TZP approach, 126 vinyl cations, 137 Groove binding guests, binding dynamics of DNA with, 201–204 cyanine dyes, 203, 204 distamycin, 201–203, 204 Heteroatom rearrangements on nitrogen (HERON) reaction, 67, 70–74, 83, 94 bonding in, 72 in thermal decomposition reactions, 93–96 Heteroatom stabilized carbocations, NMR spectroscopy, 156–158 halomethyl cation, 156

SUBJECT INDEX o-, m- and p-phenylene bis(1,3-dioxolanium)dications, 156 protonated carbonic acid, 157 Hoechst dyes, for binding dynamics study of guests to DNA, 203–204 HPNPP (2-Hydroxypropyl-p-nitrophenyl phosphate), metal-ion catalyzed methanolysis, 310–316 with dinuclear complex in methanol, 318–324 m-Hydrido-bridged carbocations, NMR spectroscopy, 144–145 carbodication, 145 cyclooctyl cation, 144 Hydrogen abstraction, intrazeolite photooxygenations, 236–240, 243, 247, 248, 253 Hydroperoxysulfonium ylide (HPSY), 248, 249f, 251, 251f IGLO method, 125–126 Infrared spectroscopy, for N-acyloxyN-alkoxyamides structure, 51, 52t– 54t, 55–56 N-acetoxy-N-tert-butoxybenzamide, 55 N-acyloxy-N-alkoxyalkylamides, 55 N-acyloxy-N-alkoxyureas, 56 N-chlorohydroxamic esters, 56 N,N-dialkoxyamides, 56 Intercalative guest molecules, binding dynamics of DNA with, 186–201 acridine derivatives, 190–194 9-aminoacridine carboxamide derivatives, 194, 195 daunomycin, 197–198 ethidium bromide, 186–187, 188t, 189–190 naphthalene diimide with alkylamino substituents, 199–200 proﬂavine, 190–194 Lanthanide ion catalysts, alcoholysis with, see Transition metal ion and Ln3+ catalysts, alcoholysis Laser ﬂash photolysis (LFP), 170, 175–178 cyclodextrins (CD), binding dynamics of guests binding to, 215–216 DNA, binding dynamics of guests binding to, 193–194 Lithium naphthalamide, 4

SUBJECT INDEX Mesitylene, 94 Metal-catalyzed alcoholysis reactions, 272–276 of carboxylate esters, 288–294 of neutral phosphate esters, 294–308 of phosphate diesters, 308–324 potentiometric titrations in alcohol and, 276–284 Metal coordination, in enediynes, 16–17 N-Methoxyformamide, structure, 44, 46–47, 46f N-Methoxy-N-benzoylnitrenium ion, 64, 66 Methyl-p-nitrophenyl phosphate (MNPP), 321–323 Molecular orbital (MO) correlation diagrams, for cycloaromatization reactions, 7–9, 8f Mutagenic activity of N-acyloxy-N-alkoxyamides, 81 in Ames Salmonella/microsome assay, 97–115 hydrophobicity and, 105 intercalation effects, 106–109 mutagen stability and, 104–105 mutagen structure and, 104 quantitative structure activity relationships governing, 106 steric effects on, 105–106, 109–115 Myers-Saito cyclizations, 4f, 11, 30, 30f Nitrenium ions, 38, 70, see also N-AcyloxyN-alkoxynitrenium ions; Arylnitrenium ions formation, 99 phenylnitrenium ions, 39 substituted with heteroatoms, 39 NMR spectroscopy, for carbocations, 125 alkyl and cycloalkylmethyl cations, 126–133, see also individual entry bicyclic and polycyclic carbocations, 145–150 cycloalkyl cations, 142–144 heteroatom stabilized carbocations, 156–158 m-hydrido-bridged carbocations, 144–145 p-stabilized carbocations, 150–155 vinyl cations, 133–142 Nuclear magnetic resonance (NMR) study, 181–182

367 cyclodextrins (CD), binding dynamics of guests binding to, 212–213 DNA, binding dynamics of guests binding to, 196–197, 202 Nuclear magnetic resonance spectroscopy for N-acyloxy-N-alkoxyamides 13 C NMR spectroscopy, 56–58 15 N NMR spectroscopy, 58–59 dynamic 1H NMR spectroscopy, 59 Nucleophilic substitution (SN2) reactions, N-acyloxy-N-alkoxyamides, 70–90 alcoholysis reactions, 89–90 with aromatic amines, HERON reactions, 70–74 with azide ions, 83–86 electronic effects, 78–79 with hydroxide ions, 82–83 kinetic studies, 74–77 steric effects, 79–81 with thiols, 86–89 Paraoxon, 284 ethanolysis of, 282, 286t transesteriﬁcations of, 282 p-bond, in cycloaromatization reactions, 2–3 effects in early stage, 7–10 effects in product stability in, 30–31 Persulfoxide (PS), 248, 249f stabilization, 250–251 Phenylnitrenium ions, 39 Phosphate diesters, alcoholysis of, 308–324 exhalted catalysis of methanolysis of 2-hydroxypropyl-p-nitrophenyl phosphate (HPNPP) in methanol, 318–324 metal-catalyzed alcoholysis of an RNA model, 310–316 methyl-p-nitrophenyl phosphate (MNPP), 321–323 Zn2+ ligand models for dinuclear enzymes promoting the cleavage of RNA, 316–318 Phosphate esters, alcoholysis of neutral, 294–308 concerted mechanism, 305–308 metal-methoxide reactions of phosphates, phosphonates, phosphonothioates, and phosphorothioates, 300–305 methoxide reactions of aryl esters, 299–300

368 phosphates and phosphorothioates, phosphonates and phosphonothioates, 296–299 Polycyclic carbocations, see Bicyclic and polycyclic carbocations, Potentiometric titrations, in alcohol, 276–278 with metal ions, 278–284 solvation effects, 277–278 Proﬂavine, binding with DNA, 190–194 p-systems, orthogonal, 3 in-plane and out-of-plane, 7–10 Radical-anionic Cope rearrangement, 5 Rehybridization effects, in cycloaromatization reactions, 21–22 Relative selectivity (RS) parameter, 287 s-bond, in cycloaromatization reactions, 2–3 effects in early stage, 7–10 effects on reactant stability, 10–26 s- effects interaction of non-bonding electrons, 27 Schmittel cyclization, 4f, 30, see also Cycloaromatization reactions Schreiner cyclization, 4f, 30, see also Cycloaromatization reactions Sensitizers, singlet oxygen, 233–234, 235–236 "Ship-in-a-bottle" syntheses, in zeolites, 231, 232f Singlet oxygen sensitizers, 233–234, 235–236 Solvolysis reactions, of N-acyloxyN-alkoxyamides, 60–70 AAl1 mechanism, 61–62 AAl2 mechanism, 61, 62 N-acetoxy-N-butoxybenzamide, 60–63, 66–67, 68, see also individual entry N-acyloxy-N-alkoxybenzamides, 64–65 p-Stabilized carbocations, NMR spectroscopy, 150–155 arenium cations, 151 benzylic mono- and dications, substituted, 153–154 b-silyl substituted benzyl cations, 151–152 cyclobutadienyl dication, 150

SUBJECT INDEX cyclobutenyl cation, 154 1,3-dimethyl cyclopentenyl cation, 155 disilylated arenium ions, 151 homotropylium cation, 154 pentamethylcyclopentadienyl cation, 154–155 silylated arenium ions, 151 1,2,3-trimethylcyclopentenyl cations, 155 Steric assistance mechanism for ortho substituents in enediynes, 17–20 Steric effects, of N-acyloxy-N-alkoxyamides mutagenic activity and, 105–106, 109–115 in nucleophilic substitution (SN2) reactions, 79–81 Steric repulsion of ortho substituents in enediynes, 20–21 Stopped-ﬂow experiments, 171–172 cyclodextrins (CD), binding dynamics of guests binding to, 205, 207, 208 DNA, binding dynamics of guests binding to, 187, 189–191, 194, 195–196 Supramolecular dynamics, techniques for ﬂuorescence correlation spectroscopy, 178–181 laser ﬂash photolysis, 170, 175–178 nuclear magnetic resonance (NMR) experiments, 181–182 stopped-ﬂow experiments, 171–172 surface plasmon resonance (SPR) instruments, 183–185 temperature jump experiments, 172–174 time-resolved ﬂuorescence, 175–178 ultrasonic relaxation measurements, 174–175 Supramolecular systems, 167, 168 cyclodextrins (CD), study of binding dynamics of guests to, 204–216 DNA, study of binding dynamics of guests to, 186–204 dynamics, techniques to study, see Supramolecular dynamics, techniques for relaxation kinetics measurement, 168, 169–172 Surface plasmon resonance (SPR) experiments, 183–185 binding dynamics of guests binding to DNA, 199–200, 202–203

SUBJECT INDEX Temperature jump experiments, 172–174 cyclodextrins (CD), binding dynamics of guests binding to, 208–210 DNA, binding dynamics of guests binding to, 188–191, 192, 193, 197, 203 heating by laser light absorption, 172, 173 Joule heating in, 172–173, 174 microwave heating in, 172, 173 Tetradentate enediyne ligands, for enediynes, 16–17, 18f Thermal decomposition reactions, of N-acyloxy-N-alkoxyamides, 90–96 free radical decomposition, 91–93 HERON reactions, 93–96 Thiazine dyes, 233 Time-resolved ﬂuorescence experiments, 175–178 cyclodextrins (CD), binding dynamics of guests binding to, 214 DNA, binding dynamics of guests binding to, 201 single photon counting detection method in, 176, 178 Transesteriﬁcations, of neutral carboxylate and organophosphate esters with transition metal ion and Ln3+ catalysts, 284–288 alcoholysis of carboxylate esters, 288–294 alcoholysis of neutral phosphate esters, 294–308 Transition metal ion and Ln3+ catalysts, alcoholysis of carboxylate esters with, 288–294 of neutral phosphate esters with, 294–308 of phosphate diesters with, 308–324 Triple-zeta (TZP) basis, 135, 136 Twisted amides, 36, 36f Ultrasonic relaxation measurements, 174–175 acoustic resonator cavity technique, 175 cyclodextrins (CD), binding dynamics of guests binding to, 210–212 pulsed technique, 175

369 Vinyl cations, NMR spectroscopy, 133–142 a-vinyl substituted vinyl cations, 136 a-cyclopropylcyclopropylidenmethyl cation, 137, 140f cyclopropylidene substituted dienyl cation, 137 E-1,3-dienyl-2-cation, 135–136 GIAO-MP2/TZP approach, 137 b-silyl stabilized vinyl cation, 139–140 structure, 135 Z-1,3-dienyl-2-cation, 135–136 Wagnerova Class I intrazeolite photooxygenation, 233–253 of alkanes, 234–235 of alkenes, 235–243 regiochemistry, 236, 237, 243, 244 steric conﬁnement effects, 237f, 246–247 Wagnerova Class II intrazeolite photooxygenation, 253–261 of alkanes, 256f, 258–259 of alkenes, 253–257, 253f, 256f charge-transfer (CT) complexes in, 253–254, 255, 257 intrazeolite electrostatic ﬁeld, 254–255, 256 Wagnerova oxygenation classiﬁcation, 233f X-ray structures of N-acyloxyN-alkoxyamides, 47–51 N-acyloxy-N-alkoxycarbamate, 47–48, 49f N-acyloxy-N-alkoxyurea, 47–48, 49f N-formyloxy-N-methoxyformamide, 44, 45f, 46f, 47 N-methoxyformamide, 44, 45f, 46f Zeolites, 226 acid and base properties, 230 "Camel through the eye of a needle" syntheses in, 231, 232f faujasite (FAU), 226–227 intrazeolite products recovery, 232 mordenite, 228, 229 oxygenations in, see Zeolites, oxygenations in properties, 229–230

370 reactant and reagent loading, 230–231 reaction monitoring, 231–232 "Ship-in-a-bottle" syntheses, 233– 234 structures, 226–229 synthesis, 229 zeolite A, 228f, 229 zeolite L, 228 zeolite X, 227–228 zeolite Y, 227–228

SUBJECT INDEX ZSM-5, 228, 229, 246, 247–248 Zeolites, oxygenations in Wagnerova Class I intrazeolite photooxygenation, 233–253 Wagnerova Class II intrazeolite photooxygenation, 253–261 ZSM-5, 228, 229, 246, 247–248 Zwitterionic products, in cycloaromatization reactions, 27–28

Advances in Physical Organic Chemistry. Volume 20

Read more

Advances in Physical Organic Chemistry, Volume 30

Read more

Advances in Physical Organic Chemistry, Volume 11

Read more

Advances in Physical Organic Chemistry, Volume 22

Read more

Advances in Physical Organic Chemistry, Volume 07

Read more

Advances in Physical Organic Chemistry, Volume 30

Read more

Advances in Physical Organic Chemistry, Volume 26

Read more

Advances in Physical Organic Chemistry, Volume 44

Read more

Advances in Physical Organic Chemistry, Volume 35

Read more

Advances in Physical Organic Chemistry, Volume 06

Read more

Advances in Physical Organic Chemistry, Volume 37

Read more

Advances in Physical Organic Chemistry, Volume 04

Read more

Advances in Physical Organic Chemistry, Volume 24

Read more

Advances in Physical Organic Chemistry, Volume 05

Read more

Advances in Physical Organic Chemistry, Volume 10

Read more

Advances in Physical Organic Chemistry, Volume 35

Read more

Advances in Physical Organic Chemistry, Volume 26

Read more

Advances in Physical Organic Chemistry, Volume 38

Read more

Advances in Physical Organic Chemistry, Volume 09

Read more

Advances in Physical Organic Chemistry, Volume 07

Read more

Advances in Physical Organic Chemistry, Volume 03

Read more

Advances in Physical Organic Chemistry, Volume 41

Read more

Advances in Physical Organic Chemistry, Volume 12

Read more

Advances in Physical Organic Chemistry, Volume 43

Read more

Advances in Physical Organic Chemistry, Volume 28

Read more

Advances in Physical Organic Chemistry

Read more

Advances in Physical Organic Chemistry

Read more

Advances in Physical Organic Chemistry, Volume 31, First Edition

Read more

Advances in Physical Organic Chemistry, Volume 31, First Edition

Read more

Progress in Physical Organic Chemistry (Volume 13)

Read more

Recommend Documents

Advances in Physical Organic Chemistry. Volume 20

Contents Contributors t o Volume 20 V Magnetic Field and Magnetic Isotope Effects on the Products of Organic Reaction...

Advances in Physical Organic Chemistry, Volume 30

Advances in Physical Organic Chemistry ADVISORY BOARD W. J. Albery, FRS University of Oxford A. L. J. Beckwith The Au...

Advances in Physical Organic Chemistry, Volume 11

Advances in Physical Organic Chemistry This Page Intentionally Left Blank Advances in Physical Organic Chemistry Vo...

Advances in Physical Organic Chemistry, Volume 22

dvances in hysical Organic Chemistry In Memoriam Professor Victor Gold, FRS 29 June 1922 - 29 September 1985 Foundin...

Advances in Physical Organic Chemistry, Volume 07

Advances in Physical Organic Chemistry This Page Intentionally Left Blank Advances in Physical Organic Chemistry ...

Advances in Physical Organic Chemistry, Volume 30

Advances in Physical Organic Chemistry ADVISORY BOARD W. J. Albery, FRS University of Oxford A. L. J. Beckwith The Au...

Advances in Physical Organic Chemistry, Volume 26

Advances in Physical Organic Chemistry ADVISORY BOARD W. J. Albery, FRS University of Oxford, Oxford A. L. J. Beckwit...

Advances in Physical Organic Chemistry, Volume 44

Advances in Physical Organic Chemistry Volume 44 Editor JOHN P. RICHARD Department of Chemistry University at Buffalo, ...

Advances in Physical Organic Chemistry, Volume 35

Editor's preface This volume is the first since Donald Bethell served as Editor of the series, and he set a very high ...

Advances in Physical Organic Chemistry, Volume 06

Advances in Physical Organic Chemistry This Page Intentionally Left Blank Advances in Physical Organic Chemistry ...