Advances in Physical Organic Chemistry, Volume 31, First Edition

Advances in Physical Organic Chemistry

ADVISORY BOARD W. J. Albery, FRS University of Oxford A. L. J. Beckwith The Australian National University, Canberra R. Breslow Columbia University, New York L. Eberson Chemical Centre, Lund H. Iwamura Institute for Fundamental Research in Organic Chemistry, Fukuoka G. A. Olah University of Southern California, Los Angeles Z . Rappoport The Hebrew University of Jerusalem P. von R. Schleyer Universitiit Erlangen-Nurnberg G. B. Schuster University of Illinois at Urbana-Champaign

Advances in Physical Organic Chemistry Volume 31

Edited by

D. BETHELL Department of Chemistry University of Liverpool PO. Box 147 Liverpool L69 3BX

ACADEMIC PRESS San Diego London Boston New York Sydney Tokyo Toronto

This book is printed on acid-free paper. Copyright 0 1998 by ACADEMIC PRESS A11 Rights Reserved No part of this publication may be reproduced or transmitted in any form or by any

means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the Publisher. The appearance of the code at the bottom of the first page of a chapter in this book indicates the Publisher’s consent that copies of the chapter may be made for personal or internal use of specific clients. This consent is given on the condition, however, that the copier pay the stated per copy fee through the Copyright Clearance Center, Inc. (222 Rosewood Drive, Danvers, Massachusetts 01923), for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, €or creating new collective works, or for resale. Copy fees for pre-1998 chapters are as shown on the title pages If no fee code appears on the title page, the copy fee is the same as for current chapters 0065-3160198 $30.00

Academic Press 24-28 Oval Road, London NW17DX, UK

http:llwww.hbuk.co.uklapl

Academic Press a division of Harcourt Brace & Company 525 B Street, Suite 1900, San Diego, California 92101-4495, USA http:llwww.apnet.com

A catalogue record for this book is available from the British Library

ISBN 0-12-033531-X

qpeset by Keyset Composition, Colchester, Essex Printed and bound in Great Britain by MPG,Books Ltd, Bodmin, Cornwall 98 99 0001 02 03 MP 9 8 7 6 5 4 3 2 1

Contents

Preface

vii

Contributors to Volume 31

ix

Electrochemical Recognition of Charged and Neutral Guest Species by Redox-active Receptor Molecules

1

PAUL D. B E E R . P H I L I P A. G A L E

AND

ZHENG CHEN

Introduction 1 Electrochemical recognition of cationic guest species by redox-active receptor molecules 6 Electrochemical recognition of anionic guest species by redox-active receptor molecules 50 Towards electrochemical recognition of neutral guest species by redox-active receptor molecules 71 Conclusions 76 Acknowledgements 77 References 80 Appendix: Understanding cyclic voltammetry and square-wave voltammetry 84 Spin Trapping and Electron Transfer

LENNART EBERSON 1 Introduction 91 2 Redox mechanisms of spin trapping 93 3 Electron transfer theory 96 4 Spin trapping and electron transfer 101 5 Evidence for the ST'+-nucleophile mechanism under thermal conditions 105 6 Properties of the PBN and DMPO radical cations 114 7 Anodic spin trapping experiments 116 8 Photochemical spin trapping experiments 118 9 Example of problems in photo-initiated spin trapping 121 10 Ionizing radiation and spin trapping 126 11 Spin trapping of radicals generated by ultrasound (sonolysis) 126 12 Spin trapping in biochemicalhiological systems 127 13 Conclusions on the radical cation mechanism 129

91

CONTENTS

vi

14 Spin adduct formation via radical anions 129 15 The nucleophilic addition-oxidation mechanism 130 16 Bona Jide spin trappings: a recipe 136 References 137 Secondary Deuterium Kinetic Isotope Effects and Transition State

O L L E MATSSON

AND

143

K E N N E T H C. WESTAWAY

Introduction 144 Secondary a-deuterium KIEs in SN reactions 146 Secondary p-deuterium KIEs 197 Secondary deuterium KIEs and tunnelling 211 Remote secondary deuterium KIEs 231 New methods for the accurate determination of secondary deuterium KIEs 234 Conclusion 242 Acknowledgements 242 References 243 Catalytic Antibodies

249

G. M I CH A EL BLACKBURN, ANITA DATTA, H A Z E L D E N H A M A N D PAUL WENTWORTH J R 1 2 3 4 5 6 7 8 9 10 11

Glossary 250 Introduction 253 Approaches to hapten design 261 Spontaneous features of antibody catalysis 276 Performance analysis of catalytic antibodies 278 A case study: NPN43C9-an antibody anilidase 281 Rescheduling the regio- and stereo-chemistry of chemical reactions 285 Difficult processes 292 Reactive immunization 301 Medical potential of abzymes 304 Industrial potential of abzymes 309 Conclusions 311 Appendix: Catalogue of antibody-catalysed processes 313 References 385

Author Index

393

Cumulative Index of Authors

407

Cumulative Index of Titles

409

Physical organic chemistry (according to the liberal definition adopted in this series) continues to develop, and application of its methods and results continues apace in areas as diverse as biology and solid-state physics. Volume 31 of Advances in Physical Organic Chemistry has had a much longer period of gestation than is usual for the series. Indeed only one of the contributions originally commissioned met the manuscript deadline, and the present volume has a somewhat different balance of material from that originally envisaged. The hiatus has served, however, to remind the Editor of how much he asks of authors in terms of time, effort and organization in putting together their contributions. All those interested in following the development of physical organic chemistry in the diverse strands that make up the field owe much to their dedication. The present volume embodies several of the themes that have run through the series. The relationship of the structure of organic molecules to their properties measured quantitatively is represented in the first contribution on redox-switched ionophores, which find their application in analytical chemistry. Two contributions continue the theme of the transition state in physical organic chemistry, one concerned with obtaining structural information and the other with applying such knowledge to the design of models that can be used to develop new biological catalysts using modern methods of production, screening and isolation. Finally, the subject of spin-trapping of radicals is revisited in the interest of refining its application in radical detection not least in the realm of biological and medical research. As always, the Editor and his Advisory Board would be delighted to hear from readers abut their views on the series and suggestions for its future development, especially emerging topics and well-established ones where an up-to-date review would be timely.

vii

This Page Intentionally Left Blank

Contributors to Volume 31

Paul D. Beer Inorganic Chemistry Laboratory, University of Oxford, South Parks Road, Oxford OX1 3QR, UK

G. Michael Blackburn Krebs Institute, Department of Chemistry, University of Sheffield, Sheffield 53 7HF, UK Zheng Chen Department of Materials Science and Metallurgy, University of Cambridge, Pembroke Street, Cambridge CB2 3QZ, UK Anita Datta Scripps Research Institute and Skaggs Institute for Chemical Biology, 10550 North Torrey Pines Road, La Jolla, California CA92307, USA Hazel Denham . Krebs Institute, Department of Molecular Biology and Biotechnology, University of Sheffield, Sheffield S10 2TN, UK Lennart Eberson Chemical Physics, Department of Chemistry, University of Lund, PO Box 124, S-22100 Lund, Sweden Philip A. Gale Inorganic Chemistry Laboratory, University of Oxford, South Parks Road, Oxford OX1 3QR, UK Olle Matsson Department of Organic Chemistry, Uppsala University, Box 531, S-75121 Uppsala, Sweden Paul Wentworth Jr Scripps Research Institute and Skaggs Institute for Chemical Biology, 10550 North Torrey Pines Road, La Jolla, California CA92307, USA Kenneth C. Westaway Department of Chemistry and Biochemistry, Laurentian University, Sudbury, Ontario, Canada P3E 2C6

ix

This Page Intentionally Left Blank

Electrochemical Recognition of Charged and Neutral Guest Species by Redox-active Receptor Moleculest PAULD. BEER.PHILIP A. GALEAND ZHENGCHEN* Inorganic Chemistry Laboratory, University of Oxford, Oxford, UK *Department of Materials Science and Metallurgy, University of Cambridge, Cambridge, UK 1 Introduction The concept of a “coupled reaction system” Pathways for coupling electrochemicalkomplexation reactions 2 Electrochemical recognition of cationic guest species by redox-active receptor molecules Oxidizable cation sensors Water-soluble sensors for transition metals: ferrocene polyazamacrocycles Reducible cation sensors 3 Electrochemical recognition of anionic guest species by redox-active receptor molecules Anion recognition by cobaltocenium receptor molecules Porphyrin-based anion sensors Anion recognition by ruthenium(I1) bipyridyl receptors Receptors with multiple nonequivalent redox sites Anion binding by neutral ferrocene-amide receptors Ferroceneboronic acid Recognition of pairs of ions 4 Towards electrochemical recognition of neutral guest species by redox-active receptor molecules Calix[S]arenes Ferroceneboronic acid derivatives 5 Conclusions Acknowledgements References Appendix: Understanding cyclic voltammetry and square-wave voltammetry

1 2 4

6 6 28 35 50

50 58 62 66 66 69 70

71 71 75 76 77 80 84

1 Introduction

Stimulated by nature and in particular by the idea of modelling biotic “coupled reaction systems” such as ion transport and oxidative phosphorylation, recent attention has focused on a new generation of abiotic host ‘Manuscript received 9th October 1995 1 ADVANCES IN PHYSICAL ORGANIC CHEMISTRY VOLUME 3 I 0065-3160/Y8 $30.00

Copyright 0 1998 Academic Press All rights of reproduction in any form reserved

P D. BEER, !? A. GALE AND Z. CHEN

2

X = Electro/Photo-responsive Function G = Guest

Electrochemical (potentiaVcurrent) or optical response (visible colour change, fluorescence)

Fig. 1 An electrochemical or photochemical response i s generated by the receptor upon guest binding.

molecules that contain a responsive or signalling function appended to or as an integral part of a host receptor framework (Fig. 1). The detection and selective binding of cationic, anionic and neutral guests by such species is an area of intense current interest (Beer, 1992) of relevance to the advancement of chemical sensor technology. This review is concerned with receptor molecules that contain an electrochemical responsive signalling function, i.e. a redox-active centre, which is coupled to a host binding site (Beer, 1989, 1992; Kaifer and Echegoyen, 1990). Depending on the complementary nature of the host cavity, these systems can in principle be designed to recognize electrochemically the binding of any charged or neutral inorganic or organic guest species through a number of different possible coupling pathways (see below). Evidently selective binding of a particular guest species coupled with an electrochemical response is of paramount importance for future potential prototypes of new amperometric molecular sensory devices (Edmonds, 1988).

THE CONCEPT OF A “COUPLED REACTION SYSTEM’

The stability constant K of a hodguest (1:l) complex is defined by the equilibrium (l), kc

H+G

HG kd

where H, G and HG represent the host, guest and complex species.

REDOX-ACTIVE RECEPTOR MOLECULES

3

+e Hred

+G

Scheme 1 The scheme of one square for guest binding and electron transfer.

Studies in the area of electrochemical molecular recognition deal with bifunctional receptor molecules that contain not only binding sites but also one or more redox-active centres whose electron transfer reaction is coupled to the receptor’s complexation. Such systems can be described by the scheme of squares as shown in Scheme 1. In this scheme, H, G and HG in normal or subscript positions represent the host, guest and complex species respectively; subscripts “ox” and “red” indicate that the corresponding symbols or parameters refer to molecules in oxidized and reduced states; E o is the formal potential of the electron transfer reaction and K is the stability constant. According to thermodynamics, there are four relationships linking the concentrations of the four molecules at the four corners of the square. These are two Nernst equations for the upper (2) and lower (3) electron transfer reactions,

Combining (2) and (3):

4

I? D. BEER, F! A. GALE AND 2. CHEN

The two complexation equilibrium equations for the left and right complexation/decomplexation equilibria are (4)and (5) respectively.

Therefore,

Equation (6) links, in a simple way, the thermodynamically important stability constants KO,and Kredof a complex in different oxidation states with experimentally measurable redox potentials EH and EHG. Therefore it provides an easy way to obtain the ratio of K,,/Kred,which is a theoretically useful parameter known as the binding enhancement factor (BEF). We propose that a better description for this ratio would be the reaction coupling efficiency (RCE) since binding by so-called molecular switches may be reduced or enhanced, depending upon the particular system involved. Equation (6) also allows the calculation of KO, if Kred is known or vice versa. Receptors designed to recognize guest molecules electrochemically must couple the complexation process to the redox reaction, i.e. the two reactions must mutually influence each other. Electron insertion (reduction) or withdrawal (oxidation) from a host molecule will change the stability constant of the complex formed, leading to a change in the ratio of K0,/Kred.Equation ( 6 ) predicts that this change in stability constant will cause a change in the host’s redox potential. The magnitude and the direction of the potential change will depend primarily on the reaction coupling mechanism and the properties of the complexed guest molecule. The variations can be measured, for example, by voltammetric means.

PATHWAYS FOR COUPLING ELECTROCHEMICAL/COMPLEXATION REACTIONS

Ideally the electrochemical molecular recognition process should result in a large shift of the redox potential of the host species. The minimum magnitude of a potential shift is gauged by experimental error. For most voltammetric techniques, this error is about 5 5 mV. According to (6), the potential shift is This ratio reflects the influence of the redox determined by the ratio Kox/Kred. reaction upon complexation, in other words, the RCE. So far, the coupling has

REDOX-ACTIVE RECEPTOR MOLECULES

5

/

Through Space

(a)

,Bond linkage

/

Direct coordination

Change in conformation of redox centre

Fig. 2 The mechanisms for coupling electrochemical and complexation reactions.

been mainly realized through one or a combination of the following four pathways (Fig. 2): (a) Through-space electrostatic interaction between the redox centre(s) and the complexed guest molecule. Through-bond electrostatic communication provided typically by conjugated chemical bond linkage between the redox centre(s) and the binding cavity. Additional direct coordination bond formation between the redox centre and the complexed guest molecule. Conformationally induced perturbation of the redox centre(s) caused by the complexation of a guest molecule.

6

I? D. BEER, P A. GALE AND Z. CHEN

Examples of redox-active molecules which exhibit each of these mechanisms will be highlighted in the course of this review. The discussion of these types of systems is conveniently subdivided according to the nature of the target complexant, i.e. a cation (metal, ammonium), anion (halide, nitrate, hydrogensulfate, dihydrogen phosphate, etc.), ion pairs or neutral (organic or inorganic) guest. Later we shall see that a fifth “interference” coupling pathway can also be used, particularly in the detection of neutral species where there is little electrostatic interaction between guest and redox centre. 2 Electrochemical recognition of cationic guest species by redox-active receptor molecules

There are two distinct classes of redox-active cation sensors: those which can be oxidized and hence form less stable complexes (e.g. ferrocenecontaining receptors) and those which are reduced and hence form more stable complexes [e.g. quinone-, anthraquinone- and nitroaromatic-containing species; the latter two types of receptor have been covered in a recent review (Kaifer and Echegoyen, 1990) and will therefore not be referred to in great detail here]. Cyclic voltammetry (CV) is an electrochemical potential sweep technique commonly used for studying complexation reactions electrochemically. The newer electrochemical technique of square-wave voltammetry (SWV) has also been used successfully in more recent work as it provides a higher resolution of redox processes which are of similar potentials than is possible with CV owing to the elimination of the capacitative charging current of the electrochemical double layer (Osteryoung and O’Dea, 1987). More details are provided in the Appendix.

OXIDIZABLE CATION SENSORS

Ferrocene crown ether species The electrochemical properties of ferrocene have been utilized by many workers in the field of electrochemical molecular recognition. Saji (1986) showed that the previously synthesized (Biernat and Wilczewski, 1980) ferrocene crown ether molecule (Fig. 3; [l]),whose binding properties had previously been studied only by nmr and UVNis techniques (Akabori et af., 1983), could be used as an electrochemical sensor for alkali metal cations involving a combination of through-space and through-bond interactions. Initially, on addition of sodium cations to a solution of the ligand, two distinct CV waves were observed, corresponding to the uncomplexed and complexed compound [l](Fig. 4 ) (Charlot et af., 1962).The wave at the higher positive potential corresponds to the solution complexed species. The oxidized ferrocene crown ether has a lower binding constant with sodium than the

7

R EDOX-ACTIVE RECEPTOR MOLECULES

[11 Fig. 3 The structure of pentaoxa[l3]ferrocenophane [l].

+0.2

0 -0.2 -0.4 E N vs. FcIFc'

-0.6

Fig. 4 Cyclic voltammograms for 0.2 mmol dm-3 pentaoxa[l3]ferroceneophane (in the presence of 0.1 mol dm-3 Bu",PF, in CH,Cl,) in the absence of NaCIO, (a) and in the presence of 1 mmol dm-3 NaClO, (partially precipitated) in the course of stirring a solution for (b) 5 min and (c) 1h. Scan rate 40 mV s-'.

8

F! D. BEER, F! A. GALE AND Z. CHEN

W’, W2 = mini-grid platinum electrodes C’,C2= platinum plate counter electrodes R’ , R2 = saturated calomel reference electrodes Scheme 2 Transport of alkali metal cations across liquid membranes using [l]as a carrier.

unoxidized receptor owing to an electrostatic repulsion of the ferrocenium positive charge and the guest alkali metal cation. For sodium and lithium cations the RCEs (&)/Kh)) were 740 and 72 respectively. This repulsion can be used to switch off cation binding and this was utilized by Saji and Kinoshita (1986) to transport alkali metal cations across liquid membranes containing [l] as a carrier (Scheme 2). The evolution of a new set of electrochemical waves (as opposed to the gradual shifting of the redox couple) on addition of guest species may be due to a number of factors. If the complex formed has a particularly high stability constant and has a redox potential which is markedly different from that of the free ligand, a new set of waves may be observed. However, if the decomplexation kinetics of the complex formed is particularly slow on the electrochemical time scale then, as the potential is scanned between the vertex points during a cyclic voltammetric experiment, the solution complexed species will be stable over this time period and the two sets of waves will correspond to free ligand and complex. Therefore care should be taken to determine the cause of the evolution of a new set of electrochemical waves and

REDOX-ACTIVE RECEPTOR MOLECULES

9

Fig. 5 Ferrocene crown ether species.

it should not automatically be assumed that this phenomenon is due to a particularly high stability constant. In 1990 we reported the synthesis of new redox-responsive crown ether molecules that contain a conjugated link between the crown ether unit and a ferrocene redox-active centre (Beer et al., 1990a). Examples of some of the species synthesized are shown in Fig. 5. The electrochemical behaviour of these species was investigated and also the electrochemical behaviour of their analogues with a saturated link between the ferrocene unit and the crown ether. The changes in the CVs of [2a] upon addition of magnesium cations are shown in Fig. 6 . The metal cation-induced anodic shifts of [2a], [2b] and also their saturated analogue [3] and vinyl derivatives [4a], [4b] are shown in Table 1. These results show that significant anodic shifts in the ferrocene oxidation wave result if cations are added to the conjugated receptor systems where the welectron system links the heteroatoms of the ionophore to the redox centre.

F! D. BEER, I? A. GALE AND Z. CHEN

10

0.8

0.6

0.4 0.2 0.0 EN vs. SCE

Fig. 6 Cyclic voltammograms for 3 mmol dm-3 [2a] (in the presence of 0.2 mol dm-3 ButNBF, in CH,CI,): (a) in the absence of Mg2+and in the presence of (b) 0.75 equiv M g + , (c) 1.5 equiv MgZf.Scan rate 100 mV s-'.

Table 1 The electrochemical anodic shifts of the ferrocene oxidation wave of [2a], [2b], [3], [4a] and [4b] upon addition of 4 equiv of cation. Compound

AE(Na+)/mV AE(K+)/mV AE(Mgz+)/mV

50 20 100

65 20 110

<5 <5 <20

30 20 70

30 20 60

REDOXACTIVE R ECEPTOR MOLEC ULES

11

.X

Fe X

X = CONH

PI

Fig. 7 Ferrocene bis(benz0)-crownether receptors [5]-[8].

Much smaller shifts are observed for the saturated analogue, suggesting a through-bond mode of coupling as the primary mechanism of electrochemical recognition with this type of system. The effect of the chargehadius ratio of the bound cation can be clearly seen in these data. The magnesium dication possesses the larger chargehadius ratio (i.e. polarizing ability) and thus produces anodic shifts approximately twice the size of the sodium monocation. Ferrocene bis-crown ether species Bis-crown ether receptor molecules, which consist of two crown ether subunits linked together by a hydrocarbon chain, are known to exhibit remarkable selectivity for group 1metal cations through the formation of 1:1 1igand:M’ intramolecular sandwich complexes (Bougoin et al., 1975; Kimura et al., 1985; Kibukawa et al., 1987). With this in mind, we focused on the syntheses of metallocene bis( benzo-15-crown-5 crown ether)s containing various linkages between the redox centre and the benzocrown ether (Fig. 7) (Beer, 1985a,b; Beer and Keefe, 1986; Beer et al., l989,1990b, 1993d). X-ray crystallographic analysis of the potassium complex of [5] confirmed, as expected, that these receptors form intramolecular 1:1 “sandwich” complexes with potassium cations (Fig. 8). Electrochemical complexation studies of [5],disappointingly, revealed that the reversible ferrocenoyl oxidation wave was not perturbed on addition of either sodium or potassium cations, implying that the complexed group 1

12

P D. BEER, P A. GALE AND Z. CHEN

Fig. 8 X-ray crystal structure of the potasssium complex of [5].

metal cation is too far away to influence the electron density at the ferrocene iron atom by through-space interactions. Also there is no through-bond pathway present to communicate the presence of the cation to the ferrocene centre; the amide linkage appears to be insulating. Studies on the Schiff base linked ferrocene bis-crown ether compound [6] have shown that it is not well-behaved electrochemically. The bis-benzo-15-crown-5ferrocene compound [7] containing two vinylic linkages was formed in a mixture of three isomeric components, the cis-cis, cis-trans and trans-trans isomers, which proved inseparable. However, the precedent of insignificant differences found between the magnitudes of the metal cation-induced anodic shifts in the ferrocenyl redox potentials of the respective separated cis and trans isomers [2a] and [2b] led us to use the same isomeric mixture of [7] throughout the subsequent FABMS and electrochemical group 1 and 2 metal cation complexation experiments. FABMS has been used as a semiquantitative indication of the selectivity of receptors for particular guest metal cations (Johnstone and Rose, 1983). The FABMS competition experiment on [7] with equimolar amounts of the nitrates of sodium, potassium, rubidium and caesium gave gas-phase complex ions of ([7] + K)' ion (mlz 809) and a minor peak ([7] + Rb)+ ion (mlz 855) exclusively. The relative peak intensities therefore suggested a selectivity order of K+ %- Rb+ B Na', Cs+, indicative of the bis-crown effect, the ability of bis-crown ether ligands to complex a metal cation of size larger than the cavity of a single crown ether unit, forming a sandwich structure. CVs of [7] were recorded after addition of calculated equivalents of Na+, K+ and Mg2' and equimolar mixtures of Na+/K+and Na'/K'/Mg2+. The results obtained are presented in Table 2. One-wave metal cation-induced anodic shifts of the ferrocenyl redox couple are observed (mediated by a throughbond coupling pathway), and interestingly the magnitudes of these are

REDOX-ACTIVE RECEPTOR MOLECULES

13

Table 2 Electrochemical data for ligand [7].

E V AE(Na+)/mVb AE(K+)/mVb AE (Mg2+)/mVb AE (Na+/K')/mV' AE (Na+/K'/Mg2')/mVc

0.340 55 35 110 35 40

Obtained in acetonitrile containing 0.2 mol dm-' BuZNBF, as supporting electrolyte. Solutions mol dm-3 in compound and potentials were determined with reference to SCE, were -3 X 0.2 V s-' scan rate. bOne-wave shift in oxidation potential produced by the presence of metal cations (4.0 equiv) added as perchlorate salts Positive values indicate anodic shift, negative values indicate cathodic shift. 'One-wave shift in oxidation potential produced by a mixture of metal cations (4.0 equiv) added in equimolar amounts.

Table 3 Electrochemical data for ligand [8]. EmN"

AE,/mVb AE, (Na+)/mV' AE,(K+)/mV' AE, (Na+/K+)/mVd

0.54 80 70 -60 - 60

"Obtained in acetonitrile containing 0.2 mol dm-3 BuVBF, as supporting electrolyte. Solutions were -2 X lo-' mol dm-3 in compound and potentials were determined with reference to SCE. *Separation between anodic and cathodic peak potentials of CVs; values for ferrocene under the same conditions ranged from 70 to 90 mV. 'One-wave shift in oxidation potential produced by the presence of metal cations added as NaPF, and KPF6 salts. Positive values indicate anodic shift, negative values indicate cathodic shift. dOne-wave shift in oxidation potential produced by a mixture of metal cations added as NAPF, and KPF, salts in equimolar amounts.

approximately double those induced by the same metal cations in the redox couple of the monosubstituted vinylic ionophore [4].When an equimolar mixture of Na+/K+ or Na+/K+Ng2+cations is added to solutions of [7], the ferrocene/ferrocenium redox couple shifts anodically by an amount approximately the same as that induced by the K+ cation alone, thus demonstrating the selectivity of the receptor for the potassium cation. Interestingly, the sulfur-linked bis-crown ligand [B] shows an unprecedented cathodic potential shift upon addition of K' cations to the electrochemical solution (Table 3). It is believed to be a conformational process that causes the anomalous shift of the ferrocene/ferrocenium redox couple and not a through-space or through-bond interaction, as these effects would produce the expected anodic potential shift of the ferrocene redox couple. The origin of the effect may be a redirection of the lone pairs of the sulfur donor atoms towards the iron centre upon complexation. This would increase the electron density

i? D. BEER, i? A. GALE AND Z. CHEN

14

n=l n=2

0

[9]n = 1 [lo] n = 2 Fe

1111 = M tl2j = Et [13] = CH&H2)20

Scheme 3 Compounds [9]-[13].

on the iron, causing a cathodic shift of the redox potential of the ferrocene group. The compounds [9] and [lo] were synthesized in an effort to increase the through-space interaction between the ferrocene moiety and the bound cation (Scheme 3). These ligands were prepared by reaction of two equivalents of the corresponding mono-azacrown ether with bis(chlorocarbony1)ferrocene. Significant one-wave anodic shifts were obtained on addition of group 1 metal cations to solutions of [9] and [lo] and the electrochemical results of group 1 cation binding on these and “model” compounds [11]-[13] (Scheme 3) are summarized in Table 4.The magnitude of the shift ( A E ) is dependent upon the polarizing power of the metal cation guest. Interestingly, Table 4 shows that although the model compounds [11]-[13] are electrochemically insensitive to the presence of Na’ and K + , the addition of Li+ to these simple acyclic ferrocene bis-tertiary amides results in a large anodic shift in the respective ferrocene oxidation wave and the appearance of a new redox couple associated with a lithium complex. 13CNmr titration experiments with [11]-[13] suggest that the lithium cation exclusively coordinates to the amide carbonyl oxygen donor atoms since there are no aza-crown ether moieties present. This coupling between the oxidation and the lithium cation complexation is therefore primarily a result of a mixture of through-space and through-bond pathways. Compounds [14] and [15] were produced by reducing [9] and [lo]

REDOX-ACTIVE RECEPTOR MOLECULES

15

Table 4 Electrochemical data and group I metal cation dependence of ferrocene amide aza crown ethers and "model" analogues. Compound

+0.67 40 20 70

EiD"

AE(NaC)lmVh AE(K+)/mVb AE(Li')lmV' ~

~

~~

+0.67 35 20 75

+0.68 <10 <10 360

+0.62

<10
+0.60 <10 <10 320

~~~

'Obtained in acetonitrile solution containing 0.2 mol dm-3 BuiNBF, as supporting electrolyte. Solutions were -2 X mol dm-3 in compound, and potentials were determined with reference to SCE. bOne-waveshift in oxidation potential produced by presence of metal cation (4 equiv) added as perchlorate salts 'New wave evolved.

respectively using LiAIH4 (Beer et al., 1994~).The single-crystal X-ray structure of the potassium complex of [9] is in Fig. 9 and shows that this ligand forms a 2: 1 complex with potassium cations with one cation bound in each aza-crown ether moiety. The electrochemical properties of [14] and [15] were studied using CV. Each compound exhibited a reversible one-electron oxidation wave. Anodic redox potential perturbations of the CV oxidation waves were observed (due to a through space interaction as the methylene linkage is insulating) on addition of excess amounts of Na+, K+, Mg2+and Ba2+cations. We have recently demonstrated (Beer et al., 1995b,c) that it is possible electrochemically to detect simultaneously the presence of two different cations bound in the redox-active ferrocene bis-crown ether receptor [15] as shown in Fig. 10. Figure 11 presents two SWV examples. In the first case (Fig. l l a ) , Mg(C104)2 is added first and then the solution is titrated with Ba(ClO,),. The voltammogram responds by showing first a new wave located between those of the bariumharium and magnesiumharium complexes. The bariumharium complex appears later and becomes dominant with further titration. The same changing pattern has been observed for adding barium cations first and then titrating with magnesium cations. When the two cations are added simultaneously at an equal amount, all the three waves appear, with the magnitude of the middle wave being obviously larger than that of the two pure complex waves. The potential of this new middle wave, 275 mV, is almost the average of those of the two pure complexes. This is again illustrated on the CVs (Fig. llc). As electrostatic through-space interactions contribute predominantly to the potential shifting of [15] upon binding of cations, these voltammetric results are strong evidence of the formation of an unsymmetrical and stable complex, [15]-Ba2+/Mg2+.Further evidence in

16

P D. BEER. P A. GALE AND 2. CHEN

n = 1 [14] n = 2 [151

Fig. 9 Compounds [14] n = 1 and [15] n = 2 (above) and crystal structure of the potassium complex of compound [15] (below).

Fig. 10 The mixed magnesiumhariurn complex of 1141.

support of the formation of the mixed complex was provided by FAB mass spectrometry. Recently, Plenio and co-workers have published a number of papers on ferrocenophanes, ferrocene cryptands and simple ferrocene amines (Plenio et al., 1993, 1994; Plenio and Diodone, 1995a). They have synthesized 22

Fig. 11 Square-wave (SW) voltammograms of [15] obtained (a) after adding about 1.6 equiv Mg(CIO,), first and then adding 1.0 (-), 1.6 (----) and 4.0 (------) equivalents of Ba(ClO& and (b) after adding simultaneously 1.0 (-), 2.0 (----) and 3.0 (------)equiv each of Ba(C104)2 and Mg(C10& In (c) are shown the cyclic voltammograms of [15] at the shown scan rates obtained after adding simultaneously 1.0 equiv of the two salts. Experimental conditions: SW voltammograms in (a) were recorded on a 7.0mm diameter glassy carbon disk working electrode. A 3.0mm diameter glassy carbon disk was used for recording (b) and (c).

t? D.BEER, P A. GALE AND Z. CHEN

18

400

350 -

s E Y

300. 250.

200 -

150100-

50.0 0.00 ~

1.6

1.8

2

2.2

2.4 2.6 1IFe-N distance [I O'pm"]

2.8

3

Fig. 12 A plot of the inverse Fe-N distances in four ferrocene amines against the differences of the redox couples of ligands in their free and protonated forms.

ferrocene nitrogen compounds, including protonated and methylated species, and studied their electrochemical properties by CV. Interestingly, they found a linear relationship between the inverse of the Fe-N distance and the difference in redox potentials between protonated and unprotonated ferrocene amines (Fig. 12). They attribute this behaviour to a Coulomb point charge model (i.e. a through-space interaction), according to equation (7).

Here W refers to the electrostatic energy of two charges, za, zb, separated by a distance r. Using this formula they generated two sets of AG values, one derived from electrochemical observations and one from X-ray crystal data. The two sets were found to be in good agreement. Electrochemical metal cation recognition studies on compounds [17] and [18] (Scheme 4)reveal that these receptors exhibit Li+ selectivity (Plenio and Diodone, 1995b). Upon addition of sodium or lithium cations to electrochemical solutions, significant anodic shifts of the redox potentials of the receptors are observed (Table 5). Interestingly, addition of cations to [18] caused the electrochemical behaviour to become reversible at room temperature. The highly selective coordination of Li+ vs Na+ or K+ can be used to detect Li+ electrochemically in the presence of large excesses of these other group 1 cations, making [18] a prototype amperometric Li+ sensor.

RE DOX-ACT1VE RECEPTOR MOLECULES

19

Fe

(a) H(N-12-C-4),Na2C03; (b) C2H4diyl(HN2-12-c-4)2,Li2C03.

Scheme 4 Synthesis of lithium-selective ferrocene-12-crown-4receptors. Table 5 Anodic shifts in the redox couples of [17] and [18] upon addition of Li+ or Na+ cations.

~

~~

AE (Li+)/mV AE(Na+)lmV

+100 +70

+ 140 +loo

Ammonium cation-selective redox-active receptors While examining the coordination properties of new di- and triaza-crown ether ligands containing multiple ferrocene moieties (Fig. 13) we discovered using 'H and 13Cnmr titration that these systems form selectively 1:l complexes with ammonium cations (Beer et al., 1993~). Significant one-wave anodic shifts of the ferrocene redox couple were observed using CV on addition of ammonium ions to solutions of compounds [19]-[22]; ligand [23], however, showed no response, suggesting that the amine nitrogen donor atoms are a prerequisite for ammonium binding (Table 6). Substantial anodic shifts of 220mV were observed with ligand [21], presumably due to a combination of through-space interactions and N-H+...O=C hydrogen bonds as illustrated in Fig. 14.

20

P D. BEER, t? A. GALE AND Z. CHEN

Fig. 13 Diaza and triaza crown ether species containing multiple ferrocene moieties.

REDOX-ACTIVE RECEPTOR MOLECULES

21

,o

0..

H

R Fig. 14 Proposed "lariat" coordination of NH: with [21]. Table 6 Electrochemical data and ammonium cation dependence for compounds [191-[23]. Compound

EN" AE,,lmVd AE(NH;)/mV' AE(K+)/mV' AE(CH,NH;)/mV' AE(PhCH,NHi)/mV'

+0.43' 90 30 20

+0.41" 90 50 40

-

-

+0.62' 100 220 50 <10
+0.54' 100 170 85 <10 <10

+0.67' 80
"Solutions were -2 X mol dm-' in compound, and potentials were determined with reference to the SCE. bThree-electron reversible oxidation process 'Tho-electron reversible oxidation process dSeparation between anodic and cathodic peak potentials; values for ferrocene under identical conditions ranged from 80 to 90 mV. 'Shift in respective ferrocenyl oxidation potential produced by presence of guest cation (2 equiv) added as their thiocyanate salts for potassium and ammonium, and their picrate salts for methylammonium and phenethylammonium.

Ferrocene cryptands A large number of ferrocene cryptand-type molecules have been reported in the literature in the last ten years. Hall and co-workers, the initial pioneers of these cryptand systems (Beer et af.,1984;Bell and Hall, 1980;Hammond et af., 1986;Bell et af.,1983;Hall and Sharpe, 1990,1991;Hall et af., 1990a,b, 1991a,b,

F? D.BEER,

22

P A. GALE AND Z. CHEN

1993), have used CV to investigate the coordination of alkaline earth and lanthanide metal cations by a series of ferrocene cryptands such as compound [24] (Fig. 15) (Hall and Chu, 1995). They noted that large anodic shifts of the ferrocenoyl redox couple are produced with these metal cations and that there exists a broad linear correlation between the AEIRvalue and the charge/radius ratio of the cationic guest species (Fig. 15). It has been proposed by Hall that this behaviour is indicative of a through-bond interaction (i.e. the cations are coordinating to the carbonyl group of the amide). There may also be a through-space contribution to the electrochemical shift. Alkali metal cations gave only small (<20 mV) anodic shifts with this cryptand. Gokel and co-workers have also published several papers on ferrocene cryptand species (Medina et al., 1992). Reduction of amide [24] with LiAlH, forms the amine cryptand [25] (Fig. 16). The electrochemical properties of [25] were studied by CV and it was found that there was a linear correlation between AElt2 values and the charge/(radius)* of the cation (Fig. 16). This relationship has been interpreted as being indicative of a through-space interaction between the bound cation and the redox centre (in contrast to the AEll20: chargehadius relationship found by Hall). This suggests that the cation is bound in the cavity of the cryptand since there are now no carbonyl groups for coordination of the cation. Indeed, this interpretation is borne out in the crystal structures of complexes of this ligand (Fig. 17). Interestingly Gokel has demonstrated the existence of a "direct coordination" coupling pathway between this ferrocene cryptand and a silver cation. Complexation studies were carried out with [24] and [25] (as well as other ferrocene cryptand-type species) by X-ray crystallography,FAB mass spectral analysis, nmr and UVNis spectroscopy.

250 -

200 :

Mg2+

2-

150

)

4

100

r

c!

/=

50 ;

Fig. 15 Plot of AEln versus chargehadius ratio for the complexation of [24] with various metal cations.

REDOX-ACTIVE RECEPTOR MOLECULES

23

z W-

Q

/ : 2 . o 0.15

O K '

Fig. 16 Plot of AEIn versus charge/(radius)' ratio for the complexation of Na+, K+ and CaZ+.

Compound [2S] was thus shown to have an unusual affinity for Ag+ cations. X-ray crystallographic determination of the structures of the free ligand, sodium and silver complexes were carried out and are shown in Fig. 17. The Ag-Fe distance in the silver complex of [2S] is only 3.37 A, whereas the Na-Fe distance in the sodium complex is 4.39 A. This evidence together with the FAB MS data and UV spectroscopic data suggests that there may be an interaction between the silver cation and the iron present in the ferrocene moiety. Electrochemical investigations by CV revealed that addition of substoichiometric amounts of NaClO, to the solution of the ligand caused the appearance of a new set of waves at 0.402V vs SSCE (Figs 18a and b). The currents for the new redox couple increase apparently linearly with the concentration of Na+ ion until a full equivalent is added, when the waves corresponding to free ligand have disappeared and the CV corresponds to the oxidation of the pure complex. This two-wave behaviour [which has been described previously (Charlot et al., 1962; Saji, 1986)] has been rationalized in terms of the high stability constant of [25].Na+ complex. Kinetic effects (i.e. slow decomplexation kinetics) were discounted on the basis that no change in the CV of a solution of [25] in the presence of 0.5 equiv Na+ cations was observed when the scan rate was varied between 0.02 and S V s-l. The CV of [25] has also been shown to be strongly affected by addition of Ag+ cations (Fig. 18c). The behaviour is similar to that observed on addition of sodium cations, but the magnitude of the AEIR value is much larger than that observed with Na' (Table 7). The AEln value with Ag+ is larger even than that with Ca2+. This is

F! D. BEER, I? A. GALE AND Z. CHEN

24

n

Fig. 17 The crystal structures of [25].H20 (left), [25].Na+ (right) and [25].Ag+ (facing page) illustrating the close proximity of the silver cation to the iron centre of the ferrocene.

REDOX-ACTIVE RECEPTOR MOLECULES

Fig. 17 (continued)

25

I? D. BEER, I? A. GALE AND Z.CHEN

26

-0.2

+0.8 EN vs SSCE

Fig. 18 Voltammetric response of acetonitrile solutions of 1.0 mmol dm-3 [25]: (a) stationary glassy carbon electrode (0.08 cm’), scan rate 100mV s-’; (b) same conditions as (a) with 0.5 equiv NaCIO., added; (c) same conditions as (a) with 0.5 equiv Ag+ added.

inconsistent with the relatively small charge-to-size ratio of the Ag+ cation. The fact that the Ag+ ion exerts a much larger effect on the half-wave potential of the ferrocenyl group than would be predicted in terms of its charge-to-size ratio suggests that the bound cation resides closer to the ferrocenyl subunit than the other cations studied. UVNis studies suggest this may be due to the ferrocene group acting as a donor to the d-orbital-bearing Ag+ cation. Electrochemical experiments were also conducted in an aqueous environment

REDOX-ACTIVE RECEPTOR MOLECULES

27

Table 7 Electrochemical data' for [25] in the absence and presence of several cations. Cation None Li Na+ K+ CaZ+ Ag+ +

Equiv. 0

0.5 0.5 0.5

0.5 0.5

E"

E:x

AE"

RCE(K/K+)"

0.216 0.210 0.214 0.224 0.214 0.214

0.402 0.348 0.488* 0.496

0.188 0.124 0.274 0.282

3 x 104 4x103 2 x 105 2x105

'K and K, represent the metal ion binding constants of the reduced and oxidized forms of the ligand respectively. The RCE (KK,) ratios given were obtained by optimizing the fit of experimental and simulated voltammograms bThisredox couple exhibited a marked degree of electrochemical irreversibility. 'E" and E& are the apparent half-wave potentialsof the free ligand and the specified metal ion complex respectively. The values are given in V vs SSCE. AEo is the difference between these two values

and it was found that compound [18] can selectively recognize silver cations in water.

Ferrocene crown receptors designed to show conformationally induced electrochemical perturbations

A new approach to the development of amperometric sensing devices of the future may centre on using conformational changes of a redox-active ligand induced by a bound guest to produce a modification of the redox centre's electronic environment and hence perturb its redox-responsive behaviour (Scheme 5) (Beer et al., 1995e). With this goal in mind the new polyferrocene bis(benzo-15-crown-5 ether) receptor molecules [26] and [27] were synthesized (Fig. 19). These species are designed to undergo a conformational change upon formation of an intramolecular potassium cation sandwich complex, bringing the respective redox centres into close proximity and consequently altering their redox properties. Proton nmr titration experiments of [26] and [27] with KPF, in acetonitrile revealed that in solution both compounds form 1: 1 intramolecular sandwich complexes with the potassium cation. A number of alkyl-, vinyl- and azo-linked bis(benzo-15-crown-5) ligands are well known to exhibit this mode of K + coordination. In the case of [26], a solid-state potassium complex was isolated whose elemental analysis and fast-atom bombardment mass spectrum ([26] OK' = 1083 complex ion) was in agreement with 1: 1 complex stoichiometry (Fig. 20). Cyclic voltammograms were recorded after progressively adding stoichiometric equivalents of K+, Na+ and Li' cations to solutions of the receptors. Disappointingly, with all metal cations only relatively small anodic

28

F! D. BEER, t? A. GALE AND Z. CHEN

Scheme 5 Conformational change in the bis-crown receptor results in a change in the

environment of the redox centres.

perturbations of the respective ferrocene/ferrocenium redox couple of [26] and [27] were observed. This result suggests that the binding of potassium cations, although resulting in a dramatic conformational change with the formation of an intramolecular sandwich complex, as evidenced from nmr and FABMS, does not significantly affect the electronic environments of the respective ferrocene redox centres.

WATER-SOLUBLE SENSORS FOR TRANSITION METALS : FERROCENE POLYAZAMACROCYCLES

The detection of ionic guests in an aqueous environment is a greater challenge than detection in organic media due to solvation effects. Cations and anions are strongly solvated in water via favourable electrostatic dipole-dipole and/or hydrogen-bonding interactions. Consequently, the host has to compete with these hydration spheres in order to bind the charged guest species successfully. We have synthesized new water-soluble polyaza and poly-

REDOX-ACTIVE RECEPTOR MOLECULES

0

29

0

Fig. 19 Ferrocene bis(benz0)-crownethers [26] and [27].

Fig. 20 The sandwich complex of [26] and K+.

30

F! D. BEER. !? A. GALE AND Z.CHEN

ammonium ferrocene macrocyclic ligands which are able electrochemically to recognize a variety of transition metal cations (Ni2+, Cu2+, Zn2+) and phosphate anions (HPOZ-, ATP) in an aqueous environment (Beer et al., 1993a, 1995b). Ligands [28]-[34] (Fig. 21) were synthesized by a variety of methods and their electrochemical properties were investigated using CV in acetonitrile, methanol and water; the results are summarized in Tables 8 and 9. In acetonitrile each ligand exhibited a one-electron oxidation wave in the range +0.4 to +0.6V (vs SCE). In water the reversibility and potential of the respective ferrocenyl redox couple of [28]-[34] was found to be dependent upon the pH of the aqueous solution (Vicek ef al., 1986). At pH values of 1-6 the ligands exhibit reversible one-electron oxidation waves at potentials of +0.5 to +0.4V (vs SCE). As the pH of the electrochemical solution is increased, the respective ferrocene oxidation wave becomes less anodic until at pH values 310.5 the CV wave reaches a constant minimum anodic potential and displays EC mechanistic electrochemical behaviour (Table 9). Macrocyclic polyamines which possess amine nitrogens separated by three of four methylene groups typically exhibit pK, values in the range of 7 and above (Kimura et al., 1982; Marcek and Burrows, 1986), whereas the smaller macrocycle 1,4,7-triazacyclononanedisplays pK, values of <2,6.82 and 10.42 (Zompa, 1978). Taking these literature pK, values and the pH-dependent electrochemical observations into account, it is highly likely that ligands [29] and [30] are fully protonated at pH values G6 and deprotonated at values 210.5, i.e. the respective pK, values of ligands [29] and [30] lie in the range 6-10.5. The pH-dependent electrochemical behaviour of ligands [28], [31] and [32] suggests that at pH values 210.5 they are fully deprotonated. In the pH range 1-6 the respective EIRvalues are independent of pH and attempts to investigate these ligands' electrochemical behaviour at pH values <1 led to their decomposition. The addition of stoichiometric amounts of Ni2+,Cu2+and Zn2+to solutions of [28]-[32] in acetonitrile led to large anodic shifts of the respective ferrocene/ferrocenium redox couple of up to 190 mV in the case of [29] and Cu2+(Table 8). Analogous experiments in water at pH values 10.5-12 revealed that [28]-[32] electrochemically recognize these transition metal cations in the aqueous environment (Table 9). Significant anodic perturbations were observed and, interestingly, the CV waves of the ligand-transition metal complexes were found to be reversible, implying that the nitrogen donor atoms of the corresponding macrocyclic ring linkage may be responsible for the EC mechanistic behaviour of the free ligands at high pH values. When an equimolar mixture of Ni2+,Cu2+and Zn2+was added to aqueous electrochemical solutions of [29] and [30]the ferrocene-ferrocenium redox couple shifted anodically by an amount approximately the same as that induced by the Cu2+cation alone. This result suggests that [29] and [30] are

/

\ 7-m

f? D. BEER. f? A. GALE AND Z. CHEN

32

Table 8 Electrochemical data for [28]-[34] in acetonitrile in the absence and presence of transition metal cations.' Compoundb

E,,(free)lmV' AEp,(Cuz')/mVd AEPa(Zn2+)lmVd AEp,(Ni2+)/mVd AEpa(Ag+)/mVd

I281

[29]

[30]

[31]

[32]

[33]'

[34]"

520 60 135 100

600 190

600 100 <10

480 20

180 110

135


410 90 85
-

-

-

-

f

45


-

-

-

'Data were obtained in acetonitrile solution containing 0.1 rnol dm-3 BuZNBF, as supporting rnol dm-3 in compound and potentials were determined electrolyte. Solutions were -3 X with reference to SCE at 21 2 1°C at 50 mV s-l scan rate. bThe CVs of [28], [29] and [31] consisted of a main current wave (reversible for [30] and [32] and EC mechanism for [28], [29] and [31]) corresponding to the Fc'/Fc couple and minor current waves (irreversible or quasi-reversible) from the oxidation of the amino groups 'Eparepresents the anodic current peak potential of the Fc+/Fc couple. dAnodic shifts of the anodic peak potential of the Fc'/Fc couple produced by the presence of metal cations (1 or 2 equiv added as their perchlorate salts). For [28], [29] and [31], after addition of cations, the current waves from the respective amino groups disappeared and that of the Fc+/Fc couple became reversible. 'Obtained in methanol. 'Instant oxidation by silver cations

Table 9 Electrochemical data for [28]-[32] in water in the absence and presence of transition metal cations.'

E,,(free)/mV AEpa(Cu2' )/mV' AEp,(Znz+)/mVc AEpa(Ni2+)/mV'

215 25 105 60

230 70 4 0


250 80 20

210 30 40

<10

-

235 40 30


"Data were obtained in aqueous solution containing 0.2 rnol dm-3 KCI as supporting electrolyte. Solutions were -3 X mol dm-' in compound and potentials were determined with reference to SCE at 21 5 1°C at 50 mV s-l scan rate. The solution pH was adjusted with 0.5 rnol dm-3 KOH and 0.1 moldm-3 HCI. bFor [28], 1291, [30] and [31] the CVs were reversible one electron oxidations at pH < 6. At pH = 11, an EC mechanism was observed for [28], [29] and [31]. Minor oxidation waves of the amino groups appeared after that of the Fc+/Fc couple at slow scan rate. The CV of [32] was a one-electron reversible oxidation wave, less dependent on the solution pH, and showed no oxidation of the amino groups in the pH range explored. 'Anodic shifts of anodic current peak potential of the Fc+/Fc couple produced by the presence of metal cations (1 or 2 equiv added as their perchlorate salts).

REDOX-ACTIVE RECEPTOR MOLECULES

33

Fig. 22 Molybdenum and tungsten crown species.

first-generation prototype copper-selective amperometric sensors, capable of detecting the Cu2+ cation in the presence of Ni2+ and Zn2+ ions via through-space interactions.

Transition metal-crown ether systems Green and co-workers (Fu et al., 1988a,b) have prepared ionophoric ligands incorporating the oxidative, redox active organometallic d2-di-qcyclopentadienylmolybdenum(1V) or -tungsten(IV) fragments into a macrocyclic structural framework (Fig. 22). Compounds [35]-[37] were synthesized by the reaction of [M(r&HS),Cl2] (M = Mo or W) and the appropriate sodium thiolate, and [38] and 1391 from reaction of the same precursors with 4'-carboxybenzo-15-crown-5. Despite the fact that the respective reversible one-electron oxidation couples of [35], [38] and [39] proved insensitive to the presence of alkali metal cations, replacing the Et4NPF6supporting electrolyte with NaPF,, KPF, or LiPF, resulted in significant one-wave anodic shifts with compounds [36] and [37], of up to 110 mV with Na+ and [37]. Lowe and Garner (1993a,b) have synthesized three new dithiolene ligands and formed complexes with a variety of transition metals (Fig. 23) including molybdenum [40]-[42], tungsten [43] and [44] and nickel [45]-[48]. The electrochemical properties of the complexes and free ligands were studied by

I? D. BEER, I? A. GALE AND Z. CHEN

34

N i ( a " 7 OMe 2

Fig. 23 Benzodithiolene crown ether and veratrole complexes of transition metal cations.

CV and the results appear to indicate variations in the degree of delocalization in the metallodithiolene ring. FAB mass spectrometry and UVNis spectroscopy have been used to demonstrate the binding of the cations to the complexes, and CV studies have provided quantitative measures of the perturbations which result on cation binding. This behaviour is quite complex and readers are encouraged to consult the original papers. However, it does indicate that receptors [42] and [44]show most promise as sensors for alkali metal cations.

REDOX-ACTIVE RECEPTOR MOLECULES

35

REDUCIBLE CATION SENSORS

Nitroaromatic species

Reducible redox-active nitrobenzene macrocyclic polyether systems have been prepared by a number of groups (Kaifer and Echegoyen, 1990); in particular, Gokel and co-workers (Kaifer et al., 1983,1985)were arguably the first to demonstrate the electrochemical recognition of a sodium cation by such a system. For example, the introduction of sodium cations to an electrochemical solution of compound [48] causes the evolution of a new wave in the CV corresponding to solution complexed species. The redox-active nitro group is directly coordinated to the sodium cation (Scheme 6). However, the addition of sodium cations to electrochemical solutions of compound [49] has very little effect on the CV, presumably because the position of the nitro group on the aromatic ring allows no interaction between the sodium cation bound in the macrocycle and the nitro group (Scheme 7). The effects of simple

Scheme 6 A direct coordination pathway is possible between the binding site of [48] and the redox-active nitroaromatic moiety.

Scheme 7 A direct coordination coupling pathway is not possible between the redox centre of [49] and the binding site.

36

i? D. BEER, i? A. GALE AND Z. CHEN

through-space interactions between the bound cation and nitroaromatic group can therefore be dismissed in both these cases and the pathway for the coupling between the complexation and redox reactions with Na' and [48] is a direct coordination route.

Quinone- and anthraquinone-based receptors Quinone- and anthraquinone-containingspecies have also been utilized as electrochemical sensor species in another area of this chemistry pioneered by Gokel and co-workers (Gustowski et al., 1986; Delgado et al., 1988). These systems have been reviewed elsewhere (Kaifer and Echegoyen, 1990). Direct coordination between the redox centre and guest leads to generally large perturbations in the electrochemical behaviour of the host. A particularly elegant example of this type redodcoordination coupling is given by the quinone crown ether species synthesized by Cooper and co-workers (Wolf and Cooper, 1984; Delgado et al., 1992) (Fig. 24). A number of differently sized crown ethers were synthesized and the shift of the first reduction potential was found for each compound in the presence of excess alkali metal tosylate. The shifts were all between 60 and 70 mV for compound [50] but the larger crowns displayed larger shifts (Table 10). In contrast to the expected order of the magnitudes of the shifts from ion pairing effects alone, K+with compound [51] yields the largest potential shift followed by Rb+>Na+>Cs+>Li+. By oxidizing the free phenolic rings of various calixarenes it is pos-

n = 1,2,3,4

(501 (n = I ) [51] (n = 2) [52] (n = 3) [53] (n = 4)

Fig. 24 Redox-active crown ether n = 1 [50],n = 2 [51], n

=3

[52], n = 4 [53].

REDOX-ACTIVE RECEPTOR MOLECULES

37

Table 10 Anodic shifts (mV) in the formal reduction potentials of [50]-[53] upon addition of alkali metal cations.

AE(Li')/mV AE(Na')lmV AE(K+)/mV AE(Rbf)lmV AE(Cs+)lmV

66 68 68 67 60

d

56 130 162 138 117

38 68 106 114 132

33 67 74 87 91

\

R

[54] R = OEt [55]R = NEtz

1561

Fig. 25 Ester, amide and ether p-t-butylcalix[4]arenediquinonespecies.

sible to produce redox-active calixarenequinones (Casnati et af., 1993; Morita et al., 1989, 1992; Reddy et a/., 1992; Suga ef at., 1991). Thus the calixarene framework itself was rendered redox-active. It was therefore decided to exploit these types of systems by synthesizing ionophoric calix[4]arenediquinones for use as electrochemical sensors for cationic species (Beer et at., 1994b,d). Using the methodology of McKillop et af. (1970), which was applied to calixarenes by Gutsche (Reddy et af., 1992), the oxidation of p-t-butylcalix[4]arene-bis(ethyl ester) and the corresponding bis(diethy1 amide) (Collins et af.,1991) with thallium trifluoroacetate produced the new diester- and diarnide-~alix[4]arenediquinones[54]and [55] in 65% and 15% respective yields (Fig. 25).

38

F! D. BEER, I? A. GALE AND 2. CHEN

Fig. 26 The ammonium, potassium and strontium complexes of [54].

The p-t-butylcalix[4]arenediquinone-bis(methylether) [56],which had previously been synthesized by Gutsche and co-workers, was used as a model compound in these studies. The receptors [54]and [55] have been shown by nmr titration techniques to form 1:1solution state complexes with group 1and 2 metal, ammonium and alkylammonium guest species. The crystal structures of the NIX+,K+ and S?' complexes of [54]are shown in Fig. 26. UVNis spectroscopic titrations were used to determine association constant data for the calix[4]arenediquinone ligands with metal and ammonium cations and the results are summarized in Table 11. For both [54]and [55], Na+ and Ba2+ cations form very stable complexes. It is noteworthy that the diamide calix[4]arenediquinone [55]exhibits a larger association constant for BuNH: than for NH:, whereas with the bis(ethy1 ester) derivative [54]the reverse

REDOX-ACTIVE RECEPTOR MOLECULES

39

Table 11 Stability constants (dm3mol-') in acetonitrile calculated for [54], [55] and [56] (maximum error 215%). ~

~~

Compound ~

Na+

K+ Ba2+ NH; BuNH:

a

7.2 x 104 4.8 x 105 1.1x 104 6.6 X lo3

1.6 X 105 4.9 x 104 1.8 x 105 1.2 x 103 1.0x 104

~

b

1.9 x 104 5.3 x 103 b b

"Stability constant too large to be reliably calculated using curve-fitting method. bNoevidence of binding was seen.

trend is observed. This selectivity difference may be attributable to the comparatively greater degree of lipophilic character of the diamide which favours the alkylammonium guest species. The electrochemical properties of these molecules were investigated using cyclic and square-wave voltammetric techniques and it was found that they all exhibited similar electrochemical properties (Chen et al., 1994). A calixarenediquinone can accept a total of four electrons to become the tetra-anion. Figure 27 shows CVs of [55] in CHzClzat different scan rates. For the convenience of discussion, each of the reduction and the reoxidation or (") waves is labelled with a number and a number plus a superscript respectively. Compounds [54] and [56] exhibit CVs similar to those of [55] under the same conditions. CVs obtained in the mixed solvents for [54], [55] and [56] and pure CH3CN for [55]retain the main features of Fig. 27, although some modifications occur. For instance, for [55]wave 3" disappears, wave 2'is less obvious and the potential of wave 3 shifts cathodically. Wave couples 1/1' and 2/2' have previously been attributed to a oneelectron transfer to each of the quinone moieties present in the molecule. The third and fourth electron transfers are believed to form wave 3. Casnati et al. (1993) suggested that the irreversibility of this wave could be due to the formation of hydroquinone species insoluble in CH2C12.This was supported by an exhaustive electrolysis of a calix[4]arenediquinone solution which became non-electroactive after the experiment. The one-wave feature of wave 3 may be due to a minimized repulsive interaction between the reduced quinone groups upon protonation, leading to a smaller potential separation between consecutive electron transfer processes. This agrees with the observation that waves 1 and 2 merge into a single wave of typical EC shape after adding NH: or Bu"NH;, which may donate protons to the reduced quinone groups. Considering the possible kinetics involved in wave 3 (a two-electron and ( I )

F! D. BEER, f? A. GALE AND Z. CHEN

40

3

3

-0.4

1

-0.8

-1.2 EN vs. AgIAg’

-1.6

-2

Fig. 27 Cyclic (top) and square wave (bottom) voltammograms of [55] (1.4 X lob3mol dm-3) recorded in CH2C12 at different scan rates and frequencies (in order of decreasing amplitude of current: CVs 600, 400, 200, 100, 50,20 mV s-l; SWVs 80,40,20 and 10 Hz).

four-proton transfer process), the aprotic nature of CH2C12and the potential of wave 3”, incompletely protonated quinone dianion species may be responsible for wave 3”. If this is the case, the observed increase in wave 3” at fast scan rates (Fig. 27) can easily be explained by an EC mechanism. Its disappearance in the presence of acids or proton donors or in a less aprotic solvent like CH3CNor the mixture of CH3CN and CH2C12may be due to faster and more complete protonation processes. Combining the above discussions, the reduction of calix[4]arenediquinones may be thought of as an EEEC mechanism (Scheme 8). Compounds [54] and [55]have been shown to complex group 1 and 2 metal cations and also ammonium and alkylammonium cations by nmr and UVNis spectroscopies and also by a number of solid-state X-ray crystallographically determined structures. The quinone moieties in these molecules constitute not only the coordination site but also the redox-active centre. The complexation

REDOX-ACTIVE RECEPTOR MOLECULES

Q-Q

-

Q-

Q'-

41

; E",

171

Q, Q' -, Q 2 - , OH?-"- and QHJ2-k)- ( n = j + k; j = 0, 1, 2; k = 0, 1, 2; kand j may be different or equal) represent the neutral, radical anion, dianion and the protonated dianion forms of the quinone moiety in the molecule; Eoi( i = 1, 2 and 3) is the formal redox potential of the corresponding electron transfer reaction. Scheme 8

Electrochemical processes occurring in p-t-butylcalix[4]arenediquinones.

processes can therefore be followed by electrochemical means. Owing to the poor solubility of metal salts in CH2C12and of [56] and [55] in CH3CN, the experiments were conducted in mixtures of these solvents (or in CH3CN alone for [%I). The addition of one or more equivalents of NaC104 or KPF6 to electrochemical solutions of 1.551 resulted in the disappearance of waves 1 and 2 and the evolution of new reversible wave couples at more anodic potentials. Anodic potential perturbations were generally observed with [54], [55] and [56] with all cationic guests, and the results are summarized in Table 12. Interestingly, addition of one or more equivalents of NH4PF6 or Bu"NH3BF4 to electrochemical solutions of [55] resulted in EC mechanistic behaviour (Fig. 28) which was not affected by subsequent addition of equivalent amounts of Na' or K' cations. This finding is contrary to the respective association constant data calculated from UVNis titration results in which the ammonium association constants are smaller in magnitude compared to those of the alkali metal cations. The relatively strong interactions of these ammonium cations with the radical anions formed by the reduction of the respective quinone moieties of [55] may be responsible for this electrochemical observation and E C mechanistic behaviour. In the presence of more than one equivalent of Ba(ClO&, both of the CVs and SWVs of [S5] showed typical adsorption characteristics. The large shifts seen on addition of cations are indicative of a direct coordination pathway mediating between complexation and reduction processes. In an attempt to impart selectivity and greater thermodynamic cation stability to these types of calixarenediquinone systems, a calixarenediquinonecrown ether [57] (Fig. 29) was synthesized (Beer et al., 1994a). Electrochemical cationic recognition studies were carried out on [S7] using

F! D. BEER, F! A. GALE AND Z. CHEN

42

l

0

L

l

l

l

-0.32

l

l

.

l

1

I

1

I

I

-0.64 -0.96 EN vs. Ag/Ag'

,

I

1

-1.28

)

b

.

I

-1.6

mol dm-3) in Fig. 28 Cyclic (i) and square wave (ii) voltammograms of [55] (5 X acetonitrile in the presence of two equivalents of (a) m F 6 , (b) N&PF6, (c) Bu"NH3BF4and (d) Ba(ClO,),. The CVs were recorded at a scan rate of 100 rnV s-' and the SWVs were at 60 Hz. Ref Ag/Ag+.

43

R EDOX-ACTIVE RECEPTOR MOLEC U LES

Fig. 29 p-t-Butylcalix[4]arenediquinone-crown-5. Table 12 A summary of the changes in voltammetric properties of [54], [55] and [56]."

E,JV of each redox coupleb Free [54] Na'

+ + K+ + Ba2+ + NH: + Bu"NH;

(-0.75(s); -0.85(r)); -l.lO(r); -1.24(s) -0.70(r); -0.80(s) -0.75(s); -0.85(r); -0.93(r) -0.52(ec, ad) -0.67(ec) -0.68(ec)

Free [55] + Na+ + K+ + Ba2' + NH: + Bu"NH;

(-0.81(r); -0.91(r)); -l.ll(r); -1.26(q) -0.81(r); -0.89(r) -0.83(r); -0.92(r) -0.64(ec, ad) -0.69(ec) -0.78(ec)

Free [56] + Na'

-l.O5(r); - 1.09(r) -0.90(r); -0.99(s) -1.02(r); -l.O9(r) -0.52(ec, ad) -0.85(ec) -0.94(ec)

+ K+ + Ba2' + NH: + Bu"NH; ~~

~~

"Conducted in 10% CHZCl2-90% acetonitrile for compounds [54] and [56] and in acetonitrile [55] upon addition of 2 equiv of the respective cation; supporting electrolyte, 0.10 mol dm-3 TBABF,. *The potential of the reduction current peak: r, reversible; q, quasi-reversible;s, single reduction peak without corresponding reoxidation peak; ec, electron transfer followed by a chemical reaction; ec, ad, electron transfer followed by a chemical reaction with insoluble product which adsorbs on to the electrode surface. Prewaves are in parentheses.

44

P D. BEER, I? A. GALE AND 2. CHEN

Table 13 Reduction potentials of [57] and the anodic shifts in the presence of 1.0 or 2.0 equivalents of different cationic species." Solution

E J V of each redox couple (vs Ag/Ag+)

E,,*(free)N AE (K+)/mVb AE (Na+)/mVb AE (Ba2+)/rnVb AE (NH:)/rnV' AE(Bu"NH:)/rnV'

-1.155 210 255 555 405 355

-1.930 250 290 d d d

"Obtainedby both cyclic (100 mV s-l) and square-wave (10 Hz,Osteryoung-type)voltammetry in acetonitrile solution containing 0.1 mol dm-3 BuZNBF, as supporting electrolyte. Solutions were -1 X mol dm-3 in compound with reference to an Ag/Ag+ electrode (330 2 10 mV vs SCE) at 21 2 1°C. bAnodic shift of the reduction waves of [57] in the presence of 1.0 equiv of the respective cationic species added as their perchlorate or hexafluorophosphate salts. 'Anodic shift in the presence of 2.0 equiv of the respective cations. dThesecond reduction wave of [57] became obscure or disappeared in the presence of more than 1 equiv of the respective cations.

CV and SWV (Table 13). The receptor itself undergoes a reversible reduction at -1.15V and an irreversible redox process at -1.93V referenced to Ag/Ag+.Large anodic perturbations of the reduction waves were observed on addition of all the cationic guests with Ba2+, which possesses the largest charge-to-radius ratio, producing the greatest effect (AE = 555 mV). On addition of substoichiometric equivalents of Na+ cations the evolution of a new redox couple, substantially anodically shifted ( A E = 255 mV) results, until after 1equiv of Na+ the uncomplexed original wave has disappeared (Fig. 30). Very recently Bethel1 et al. (1995) have published the synthesis and electrochemical properties of ionophoric calix[4]arenes which are bridged at the lower rim by an anthraquinone group [58a-d] (Fig. 31). Cyclic voltammetry was used to study the changes in the electrochemical response of these ligands on addition of group 1metal cations. In all cases the addition of cations caused new reduction waves corresponding to solution complexed species at potentials anodically shifted from the reduction waves of the free ligand. With the addition of 1.0equiv of metal cation only the reduction waves corresponding to the complexed solution species were observed. On addition of 1.0 equiv of potassium perchlorate to [58a] the first one-electron reduction process splits into two waves at high scan rates (Fig. 32); this effect was not observed with sodium or lithium cations or with the R = CH2COOEt or CH2COONEt2 compounds. The authors attribute this effect to a particularly slow complexation process due to a change in the conformation of the ligand on complexation from a cone conformation in its free state to a partial cone in the potassium complex. Therefore, at high scan rates CV can distinguish between the complex and free ligand.

45

REDOX-ACTIVE RECEPTOR MOLECULES

I

I

I

I

I

I

0

-0.5

-1

-1.5

-2

-2.5

W vs AgJAg'

mol dm-3) in acetonitrile in the Fig. 30 Cyclic voltammograms of [57] (1.0 X absence (a) and the presence of 0.3 equiv (b) and 1.0 equiv (c) of sodium cations added as the perchlorate salt. Supporting electrolyte 0.1 mol dm-3 NBu;BF,. Scan rate: 100 mV s-l. Working electrode, glassy carbon.

Molybdenum nitrosyl receptors

A series of polyether macrocycles [59]-[66] (Fig. 33) that contain a coordinated reducible, redox-active 16-electron molybdenum nitrosyl (Mo(NO)}~+group have been prepared (Al-Obaidi et al., 1986; Beer e f al., 1987). Compounds [59]-[63] were synthesized from the reactions between [Mo(NO)LX,] (L- = tris(3,5-dimethylpyrazolyl)hydroborate;X = C1- or I-) and the appropriate amine substituted benzo-crown ether. Compounds

I? D. BEER, F! A. GALE AND Z. CHEN

46

[58a] R = M e [58b] R = CH,COOEt [ 5 8 ~R] = CHZCOONEtp [58d] R = H

0

Fig. 31 Anthraquinone bridged p-t-butylcalix[4]arene family [58a-d].

0.0

-0.5

-1.0

-1.5

-2.0

EN vs AgIAgCI

Fig. 32 Sweep rate dependence of the peak potentials for the [%a]: R = MeKClO, interaction: (a) 200 mV s-', (b) 1V s-', (c) 4 V s-', (d) 5 V s-I.

[64]-[66] were prepared from reactions between [Mo(NO)LI,] and tetra-, penta-, or hexa-ethylene glycol, respectively, in the presence of triethylamine. The electrochemical properties of [59]-[66] in the presence and absence of stoichiometric amounts of Na+ and K+ guest cations were investigated in

47

REDOX-ACTIVE RECEPTOR MOLECULES

1601 m = 1 n = 1 trans [611 m = 2 n = 1 cisbrans [621 rn = 2 n = 2 cishrans

H

/N L(N0)Mo

\

N

H

w

~ 3 1

L

=

/N-N H--B’-a

\

L(NO)Mo, /

[64] n = 1 [65]n = 2 [661n = 3

Fig. 33 Molybdenum nitrosyl receptor species.

48

P D. BEER, P A. GALE AND Z. CHEN

Table 14 Electrochemical cation dependence of reducible molybdenum macrocycles.

Compound

+ Na+ +- Na+ + K+ + Na'

-0.95

+ Na'

60

-0.89 -0.92

70 40

-0.855 -0.90

85 40

-0.835 -0.88

85 40

-1.25 - 1.32

90 40

- 1.oo

320

-1.00

280

-1.06

180

-0.94

+ K+

+ Na+ + K+ + Na' + K' + Na+ + Na'

-0.89 -0.96

-0.92 -1.36 -1.32

- 1.28 -1.24

"Obtained in MeCN solution containing 0.2 mol dm-' BuzNBF4 as supporting electrolyte. Solutions were -2 X mol dm-3 in complex, and potentials were determined with reference to ferrocene as internal standard but are quoted relative to SCE. bShift in reduction potential produced by the presence of Na' or K+ added as their BPh4salts, in solution, and in aliquots to provide up to 2 molar equivalents with less than 5% volume change in the solution. Broadening of the CV trace was observed when between 0.2 and 1.0 equivalentsof Na+ or K+had been added. After -1.2 equivalents had been added the trace returned to its original shape and no further changes were observed on adding an additional 0.8 equivalent to give a cation/complex molar ratio of 2. The shifts may be compared with the effects of adding aliquots of NaBPh, to solutions containing [Mo(N0)LCl(NHC6&-3.4-(OMe)zJ1 or [Mo(NO)L-(NHC6&-3,4-(OMe),),] for which shifts in reduction potential of less than 10 mV were found under similar conditions

acetonitrile solution by CV. Table 14 shows that addition of alkali metal salt in 1:1molar ratio produces anodic shifts (AE) in the original redox couple of 40-320 mV in the reduction potentials of the respective host's molybdenum redox centre. It can be concluded from Table 14 that the polarizing power of the metal cation guest is again of great importance in determining the magnitude of the anodic shift (see Hall and Chu, 1995); the larger K+ cation having the smaller chargehadius ratio produces approximately half the shift of the Na+ guest cation.

49

R EDOX-ACTIVE RECEPTOR MOLECULES

Me

Fig. 34 The redox-active bipyridinium bis(benz0)-crown ether [67].

Fig. 35 The crystal structure of the 1 : l intramolecular sandwich complex barium complex of [67].

Bipyridinium crown receptors designed to exhibit conformationally induced electrochemical perturbations on cation complexation

Cation-induced conformational changes in redox-active molecules may perturb their electrochemical behaviour (Gourdon, 1992). The bipyridinium bis(benzo) crown (Fig. 34) is one example (Beer et af., 1994e). Our earlier work on ferrocene bis-crown ethers has shown that the amide bond linkage is insulating (Beer et al., 1990a) so that any perturbation of the redox behaviour of [67] will not be caused by through-bond interactions. It was c o n h e d by nmr, UVNis techniques and X-ray crystallography (Fig. 35) that this molecule binds groups 1 and 2 metals and ammonium cations forming 1 : l intramolecular sandwich complexes with Ba2+,K+ and NH: and 2:l complexes with Mg2+ and Na' with a cation in each crown ether moiety which necessitates a significant twisting of the 4,4'-bipyridinium redox moiety. However, the formation of the Na+ complex does not require such a dramatic change in conformation for the 4,4'-bipyridinium unit.

F! D. BEER, f? A. GALE AND Z.CHEN

50

Table 15 Electrochemical data for [67] in acetonitrile containing 0.2 mol dm-3 BuWBF, as supporting electrolyte (4 equiv cation salt added). Redox couple 2+/1+ ElnlV AEdmV AE(Ba2+)/mV AE (K+)/mV AE (NH:)/mV AE(Na+)/mV

-0.73 70 45 10 10 -10

1+/o -0.87 70

- 10 -40

-40

30

Upon addition of Ba2+ cations, the 2+/1+ bipyridinium redox couple is shifted anodically by 45 mV and the 1+/0 couple is shifted cathodically by 10 mV. K+ and N).I+ produce similar effects (Table 15). However, addition of Na+ cations causes a small cathodic shift to the 2+/1+couple and an anodic perturbation to the 1+/0 couple. This is in agreement with the proposed conformational change pathway for coupling the complexation and redox reactions. 3 Electrochemical recognition of anionic guest species by redox-active receptor molecules

Stimulated by how nature utilizes negatively charged species for numerous biochemically important pathways, structural roles in proteins and polyanions for the storage and transmission of genetic information, there is intense current interest in the molecular recognition of anion guest molecules by positively charged or electron-deficient neutral abiotic organic receptor molecules (Dietrich, 1993;Katz, 1991; Schmidtchen, 1988). Chemically, anions are utilized in many chemical reactions acting as nucleophiles (CN-), bases (-OR)and redox-active centres (S20i-). The environmental impact of anionic pollutants such as excess nitrates from agricultural fertilizers leads to eutrophication of rivers. Surprisingly, then, the design and synthesis of specific ligands that have the capability of detecting anions optically (Czarnik, 1994) and/or electrochemically (Beer, 1994) in aqueous and nonaqueous media are extremely rare. ANION RECOGNITION BY COBALTOCENIUM RECEPTOR MOLECULES

A first venture into this field was to utilize the pH-independent, positively charged, redox-active cobaltocenium moiety (Beer and Keefe, 1989; Beer et

REDOX-ACTIVE RECEPTOR MOLECULES

51

0

Fig. 36 Tripodal cobaltocenium anion receptors.

al., 1992). The tripodal receptors [68] and [69] (Fig. 36) were synthesized in 65% and 60% yields respectively (Beer et al., 19938). The addition of tetrabutylammonium chloride to 'H solutions of [68] and [69] in deuteriated acetonitrile resulted in remarkable nmr shifts of the respective proton signals of both receptors. Of particular note were the substantial downfield shifts of the amide protons (A8 = 1.28 ppm for [68] and 1.52ppm for [69]) on addition of one equivalent of chloride. These results suggest that a significant -CO-NH.-.CIhydrogen-bonding interaction contributes to the overall anion complexation process. Subsequent 'H nmr titration curves suggesting 1:1stoichiometry with anion complexes of [68] and [69] were found in all cases. Negligible shifts were observed under identi-

52

t? D. BEER. I? A. GALE AND Z. CHEN

0

do,,,

8

pF6

Fig. 37 Model cobaltocenium monosubstituted ester and amide species.

H

co

+

H

Fig. 38 Tertiary amides [72] and [73].

cal experimental conditions with cobaltocenium hexafluorophosphate ester derivative [70] (Fig. 37). However, the simple mono-amide-substituted cobaltocenium compound [71] did exhibit some significant solution interactions with halide anions (Table 16). These highlight the importance of the (CONH) amide group in anion binding. To test this hypothesis further, simple acyclic cobaltocenium derivatives (Fig. 38) containing tertiary amide groups ([72] and [73]) were prepared and complexation with C1- and Br- anions was investigated in solution by nmr spectroscopy. No shifts of the host's proton resonances were observed, emphasizing again the importance of the amide N-H group in anion binding. The cathodic shifts of the redox potentials of compounds [68], [69] and [71] on addition of halide anions are due to the stabilization of the cobaltocenium cation by the bound anion which causes the redox couple to shift to a more

REDOX-ACTIVE RECEPTOR MOLECULES

53

Table 16 Electrochemical data for compounds [68]-[73],

-0.74'

(V)" AE(F-)lmV' AE(CI-)ImV' AE(Br-)/mV EL12

55d

30 -

-0.75' 60d 40

-

-0.45

-0.74

<5 <5 <5

-

30 40

-0.50 <5 <5 <5

-0.60 <5 <5 <5

"Obtained in MeCN solution containing 0.2 mol dm-3 BuiNBF, as supporting electrolyte. Solutions were -2 X lW3 mol dm-3 in ligand, and potentials were determined with reference to SCE. bThree-electron reduction process as determined by coulometric experiments. 'Cathodic shift in reduction potential produced by the presence of anions (4 equiv) added as their ammonium of butylammonium salts. dValues obtained in DMSO solution.

0

co

co

+

Q

pF6-

+

[75] pF6-

[77]X = OMe [78]X = H

Fig. 39 Simple acyclic cobaltocenium amide derivatives [74]-[78].

negative potential. The receptors not containing the amide CONH group showed no electrochemical response to the addition of anions, ruling out the possibility of the cathodic shift being caused by ion-pairing effects. New monosubstituted cobaltocenium derivatives functionalized with hydrogen-bonding amine groups (Fig. 39) have been prepared and the

54

I? D. BEER, I? A. GALE AND Z. CHEN

Fig. 40 The solid-stateX-ray crystal structure of the bromide complex of [78].

electrochemical response of these receptor species has been investigated (Beer ef al., 1995f). The crystal structure of the bromide complex of [78] has been elucidated and clearly shows bromide anions hydrogen-bonded to the amide N-H groups and, interestingly, also to the cyclopentadienyl hydrogen atoms (Fig. 40). Significant one-wave cathodic shifts of the cobaltoceniudcobaltocene redox couples of receptors [74]-[76] (Table 17) are observed with all anionic guest species (Fig. 41), primarily owing to through-space interactions between the anion and the cobaltocenium redox-centre. The complexed anion effectively stabilizes the positive charge of the cobaltocenium unit. It is noteworthy that the largest cathodic perturbations are observed with the H,PO; anion guest, 240mV with [77] and [78]. The stability constant data obtained by nmr titration methods are in Table 18, which shows the highest K-values are obtained for [78] with H2PO;.

REDOX-ACTIVE RECEPTOR MOLECULES

55

Fig. 41. Cyclic voltammograms in CH3CNof [76] in the absence (a) and presence (b) of excess chloride ion.

Table 17 Electrochemical data for compounds [74]-[78].

Compound

c1c1c1-

-1.10 -1.08 - 1.03 -0.80 -0.80 -0.80 -0.83 -0.83 ~

AEI,2/mVb

Anion

EII2N"

35 30 55

c1-

90

BrHZPO;

40 240 85 240

c1-

H2PO; ~

~~

~

~

~

Obtained in acetonitrile solution containing 0.2 mol dm-3 Bu$NBF4 as supporting electrolyte. Solutions were -1 X lo-' mol dm-' in ligand and potentials were obtained with reference to Ag/Ag+ electrode. bCathodicshift in reduction potential produced by presence of anions (up to 10 equiv) added as their tetrabutylammoniumsalts.

In an effort to impart selectivity and enhance complex stability for this class of anion receptor novel ditopic biscobaltocenium receptor molecules (Fig. 42) have been synthesized and their coordination and electrochemical properties have been studied (Beer et al., 19935 1995h). The two positively charged metallocene centres linked by various alkyl, aryl and calix[4]arene spacers

? ! D. BEER, P A. GALE AND 2. CHEN

56

Table 18 Stability constant data determined by nmr titration techniques.

Compound

Anion

Kldm3mol-'

c1-

24"

c1c1c1Brc1BrH2PO; ~

~

~

~

770" 630" 30' 25' 35' 25 320'

'

Solvent CD3CN CDSCN CD&N 4-DMSO &-DMSO d,j-DMSO d,yDMSO 4-DMSO

~

"Errors estimated to be <5%. bErrorsestimated to be s10%.

R = Alkyl, Aryl, calix[4]arene spacer R' = COzEt, H

Fig. 42 Schematic representation of anion complexation by a di-topic biscobaltocenium receptor.

may cooperate in the molecular recognition of mono- or dianionic guest substrates. Proton nmr halide anion titrations reveal that the ethyl- [79], propyl[80] and butyl- [81] linked derivatives (Fig. 43) form complexes of 1:l stoichiometry in acetonitrile solution. Stability constant determinations suggest that the ethyl derivative [79] exhibits selectivity for the chloride anion in preference to bromide or iodide. As the chain length increases, so the selectivity for chloride decreases and also the magnitude of the stability constant which is evidence for an anionic chelate effect with the chloride anion. Receptors containing larger aryl [HI, [83], [84] and alkylamino spacers [85] (Fig. 43) form complexes of 2: 1 halide anion:receptor stoichiometry.

57

REDOX-ACTIVE RECEPTOR MOLECULES

0

0

[82] \

0

~841

R =

Fig. 43 Biscobaltocenium derivatives.

58

F1 D. BEER, f? A. GALE AND Z. CHEN

Receptor [86] (Fig. 44) forms extremely stable 1:l anion complexes with chloride, bromide and H2PO; in dimethyl sulfoxide solutions and with the adipate anion in acetone solutions. Interestingly, this receptor displays selectivity for chloride over dihydrogenphosphate. The results of electrochemical investigations are summarized in Table 19. It is noteworthy that chloride, by virtue of its higher charge density, causes relatively larger cathodic perturbations than bromide. Interestingly, as observed with monosubstituted cobaltocenium derivatives, the dihydrogen phosphate anion produces the largest cathodic shifts. The addition of tetrabutylammonium adipate to electrochemical solutions of compound [86] led to a cathodic shift of 50 mV, suggesting that this receptor can electrochemically recognize this dianionic guest in acetone solution. Similar electrochemical experiments with other biscobaltocenium receptors gave inconclusive results because of solubility problems. Although not strictly relevant to amperometric sensor technology, various metalloporphyrins [Co(III), Mn(III), Fe(II1); Fig. 451 have been shown to sense anions pofenfiometricufly with selectivity sequences dependent on the centrally bound metal (Amman et al., 1986; De ef al., 1994). For example the anti-Hofmeister selectivity sequence SCN- > I- > C10; > NO; > Br- > C1- > NO; was exhibited by PVC membrane electrodes containing [87]. Kadish et ul. (1989) have described the effect of axially bound anions on the electroreduction of tris(1V) porphyrins in THF. Cyclic voltammetric investigations of zinc tetraphenylporphyrin in dichloromethane in the presence of background electrolyte anions reveal significant perturbations of the metalloporphyrin's first one-electron oxidation, ranging from 0.86 V for TBAPF6 to 0.50 V for TBACl (Seely et af.,1994). Expanded porphyrins (sapphyrins) in their protonated forms can bind anions such as phosphate and nucleotides, but no electrochemical investigations have been reported to date (Sessler ef al., 1991,1992).

PORF'HYRIN-BASED ANION SENSORS

Recently a porphyrin unit has been incorporated into an anion receptor (Fig. 46) (Beer ef al., 19958). 'H nmr titration experiments with this compound demonstrated the formation of 1:1 stoichiometric complexes with tetrabutylammonium halides, nitrate, hydrogensulfate and dihydrogenphosphate. The electrochemical properties were investigated by CV and SWV. The effects of adding anions to solutions of the porphyrin are summarized in Table 20. Interestingly,the porphyrin reduction waves are not significantlyperturbed by any anionic guest, but the cobaltocenium moieties do show cathodic shifts of up to 225 mV with the dihydrogenphosphate anions.

REDOX-ACTIVE RECEPTOR MOLECULES

59

Fig. 44 Compound [86] and its X-ray crystal structure (including two CH&N solvent molecules).

Table 19 Electrochemical data for compounds [79]-[86]. Receptor

E W

AEdCl-)lmVb

AEdBr-)/mVb

AEdHSO;)/mVb

-0.88 -0.89 -0.89 -0.85 -0.79 -0.90 -1.04 -0.8Sd

60 45

40 30 30

-

-

65 35

165 200

40 70 65 70 50 5Sd

c

3od

115 Sod

AEdH2PO;)/mVb

C

250 C

in receptor and “Obtained in acetonitrile solution containing 0.2 mol dm-3 BGNFJF, as supporting electrolyte. Solutions were -1 X W3mol potentials were obtained with reference to an Ag/Ag’ electrode. Coulornetricinvestigationssuggest Ef values represent a two-electron reduction process Cathodic shift in reduction potential produced by presence of anions (up to 4 equiv) added as their tetrabutylammonium salts “Precipitation of complex observed; no CV could be obtained. dObtained in acetonitrile solution.

REDOX-ACTIVE R ECEPTOR MOLECULES

61

~ 7 1 Fig. 45 Metalloporphyrins for potentiometric anion setting.

Fig. 46 cisJ,10,15,20-Mesotetrakis(R-substituted)porphyrinreceptor [87].

F! D. BEER, I? A. GALE AND Z. CHEN

62

Table 20 Electrochemical data" for [87] and its electrochemical anion recognition

properties. Porphyrin oxidation

Porphyrin reduction

+0.75' 15 10 5 50 75

-1.24, -1.60 <5, <5 4, <5 <5, <5 4, <5 <5, <5

Free ligandN AE(C1-)/mVd AE(Br-)/mVd AE (NO;)/rnVd AE (HSO;)/rnVd AE (H2PO;)/mVd ~______

Cobaltoceniurn reduction - 1.5w

40 35

5 75

225

~

"Obtainedin acetonitrile solution containing 0.2 mol dm-3 BuiNBF, as supporting electrolyte. Solutions were -1 X lo-' mol dm-3 in receptor and potentials were determined with reference to an Ag/Ag+ electrode. b'Ibo-electron process 'Four-electron process dCathodic shift in redox wave produced by the presence of anions (up to 5.0 equiv) added as their tetrabutylammonium salts

ANION RECOGNITION BY RUTHENIUM(I1) BIPYRIDYL RECEPTORS

In the light of the crucial importance of hydrogen-bonding to the anion recognition process in metallocene organometallic receptors, it can be reasoned that, in principle, any Lewis acidic binding site in close proximity to one or more amide groups may lead to the successful molecular recognition of an anionic guest species. With this in mind, the Lewis acidic and photoactive ruthenium(I1) bipyridyl moiety in combination with a secondary amide group is an attractive building block to utilize for the design of innovative spectral and electrochemical sensory agents for anions (Beer et al., 1993e). The novel acyclic, cyclic and calixarene ruthenium(I1) bipyridyl receptors [88]-[92] (Figs 47 and 48) have therefore been synthesized (Szemes et al., 1996). Single-crystalX-ray structures of the chloride complex of [88] (Fig. 49) and the dihydrogenphosphate complex of [92] (Fig. 50) again illustrate the importance of hydrogen-bonding in the complexation of anions. In the former complex six hydrogen bonds (two amide and four C-H groups) stabilize the C1- anion and three hydrogen bonds (two amide and one calix[4]arene hydroxyl) effect H2PO; complexation with [92]. 'H nmr titration investigations in d6-DMS0 show that these receptors form strong, and, in the case of the macrocyclic ligand [91] and calix[4]arene containing receptor [92], highly selective complexes with H2PO; having stability constants of 8 X 103 and 2.8 X lo4dm3mol-' respectively. CV and SWV were used to investigate the electrochemical anion recognition properties of these species in acetonitrile and the results are summarized in Tables 21 and 22. With reference to the known electrochemical properties of [Ru(bipy),] [PF6I2,the respective reversible oxidation and reduction redox couples exhibited by the receptors can be assigned to the metal-centred

63

REDOX-ACTIVE RECEPTOR MOLECULES

[88] R =

[89] R = -(CH&OMe

Fig. 47 Acyclic and macrocyclic ruthenium(I1)bipyridylamides [88]-[91].

Fig. 48 A p-t-butylcalix[4]arene bridged at the lower rim with a ruthenium(I1)bipyridyl amide unit [92].

64

F! D. BEER, I? A. GALE AND 2. CHEN

Fig. 49 The X-ray crystal structure of the chloride complex of [MI.

oxidation (1.3 to 1.2 V) and the three bipyridyl-centred reductions in the range (-1.2 to -2.0 V). Because of the electron-withdrawing nature of the carbonyl moieties, the least cathodic ligand-centred reduction couple can be assigned to the amide-substituted bipyridyl group present in each receptor. Interestingly, it is this redox couple that exclusively undergoes significant cathodic perturbations on addition of anionic guest species (Table 22). This is in agreement with the crystal structures of [88] binding C1- and also [92] binding H2P04and with 'H nmr titration data that anion recognition takes place in the amide-bipyridyl vicinity of the respective receptor. Analogous electrochemical anion recognition experiments with [R~(bipy)~][PF,],gave no evidence of anion complexation. The novel results of electrochemical experiments with receptor [92] are particularly interesting. When an equimolar mixture of H2P0;, HSO; and C1- (each 5 X mol dm-3) was added to an acetonitrile solution of [92] (5 X mol dm-3), the ligand-centred reduction couple shifted cathodically by 175 mV (a shift similar to that caused by the presence of H2PO; alone).

R EDOX-ACTIVE R ECEPTOR MOLECULES

65

b

Fig. 50 The X-ray crystal structure of the dihydrogenphosphate complex of [92]. (Note the dimerization of the dihydrogenphosphate anions.)

Table 21 Electrochemical data' for compounds [88]-[92].

Receptor [881 1891 1901 ~911 ~921

RuII/III

E(+2/+1)/V

E(+lIO)N

E(0I-l)V

1.10 1.12 1.01 1.02 1.12

-1.40 -1.44 -1.48 -1.56 -1.40

-1.78 -1.80 -1.84 -1.87 -1.80

-2.02 -2.01 -2.06 -2.07 -2.01

"Obtained in acetonitrile solution containing 0.1 mol dm-3 BU~NPF,as supporting electrolyte. mol dm-3 in receptor and potentials were determined with reference to Solutions were -5 X an Ag'/Ag electrode (330 V t 5 mV vs SCE) at 21°C at 50 mV sC1 scan rate.

66

P D. BEER, P A. GALE AND Z. CHEN

Table 22 Cathodicperturbations of first ligand centred reduction couple observed on addition of various anions"

Receptor

AE(H,PO;)/mV

AE(HSO;)/mV b

15 -

20 15

AE(C1-)/mV

AE(Br-)/mV

40 65 70 65 70

30 60 10 -

60

~~

"Obtainedin acetonitrile solution containing 0.1 mol dm-3 Bu!JVPF, as supporting electrolyte. Cathodic shifts of reduction potential produced by presence of anions (up to 10 equiv) added as their tetrabutylammonium salts bPrecipitationof complex prevented a AE value from being determined.

This was found still to be the case when the other anions were in a tenfold excess over H2P0;, showing conclusively that [92] is acting as a selective H2PO; amperometric sensor.

RECEPTORS WITH MULTIPLE NONEQUIVALENTREDOX SITES

The anion coordination properties of receptors such as compound [93] (Fig. 51) are currently under investigation. This molecule contains both a redoxactive ruthenium bipyridyl moiety and also a cobaltocenium unit. This type of host has already been shown by 'H nmr and fluorescence emission spectroscopy to exhibit remarkable selectivity for the chloride anion in preference to dihydrogenphosphate (Beer and Szemes, 1995).

ANION BINDING BY NEUTRAL FERROCENE-AMIDERECEPTORS

Nature uses globular protein domains to bind sulfate and phosphate anions using respectively 7 and 12 complementary anion-hydrogen bond arrangements. With this in mind we decided to construct new, neutral ferrocene derivatives that contain various hydrogen bond donor and acceptor sites for anion recognition. The first indication that neutral amide-containing ferrocene derivatives may electrochemically recognize anions in their own right (Beer et al., 1993b) came from the electrochemical anion coordination properties of the novel multimetallocene compound [94] (Fig. 52). 'H Nmr titration investigations with tetrabutylammonium dihydrogenphosphate, hydrogensulfate and chloride in acetonitrile solution indicated that the receptor formed 1:1 complexes with each of the anionic guests. CV anion recognition investigations

REDOX-ACTIVE RECEPTOR MOLECULES

67

(PFi)2.2H20

P31

Fig. 51 The chloride complex of compound [93].

=aN

[95]R = "Bu [g61

Fig. 52 Ferrocenyl anion receptors 1941-[97].

f? D. BEER, I? A. GALE AND Z. CHEN

68

Table 23 Electrochemicaldata for [94].

Ferrocenyl redox couple/mV

Cobaltocenium redox couplelmV

240 180 90 60 40

-810 -870 230

90 70

"Obtained in MeCN solution containing 0.1 mol dm-3 BuiNBF4 as supporting electrolyte. Solutions were -1 X mol dm-3 in compound and potentials were determined with reference to an Ag/Ag+ electrode at 21 2 1"C,50 mV s-' scan rate. EpBand EF represent the anodic and cathodic peak potentials 'Cathodic shifts in the metallocene redox couples produced by the presence of anion ( 5 equiv) added as their tetrabutylammoniumsalts dAsthe concentration of the anion increased, the cathodic current peak potential of the ferrocenelferrocenium redox couple began to exhibit the features of an EC mechanism.

[981

Fig. 53 Lower-rim-bridged ferrocenylcalix[4]areneanion receptor [98].

revealed not only that the cobaltoceniudcobaltocene redox couple undergoes cathodic perturbations of up to 230 m V with HzPOQ but also that the ferrocene/ferrocenium couple shifts cathodically by up to 90 mV in the presence of H2P0; (Table 23). As a consequence of these electrochemical findings the neutral ferrocene amide derivatives [95]-[97] (Figs 52 and 53) were prepared and their electrochemical interactions with anions were investigated. Remarkable downfield shifts of the respective amide protons, ranging from A8 0.9 to 2.5ppm, are exhibited by receptors [96]-[98] on addition of tetrabutylammonium dihydrogenphosphate, hydrogensulfate and chloride salts to deuteriated acetonitrile-DMSO solutions. The resulting

REDOX-ACTIVE RECEPTOR MOLECULES

69

Table 24 Electrochemical data for [95]-[98].” Compound

[951

1961

[971

[981

Epa(*ee)lmVb

290 265 60 20 30

360 270 100 40

320 220 180 50 20

450 380 110

Epc(free)lmVb

AE(H,PO;)/mWd A E(HSO;)lmV‘*d AE( C1-)/mVc,d

<5

<5

40

“Refer to footnote a Table 23. b E p , and EPerepresent the anodic and cathodic current peak potentials of the ferrocenelferrocenium redox couple of the free ligand. ‘Cathodic shifts in the ferrocene redox couples produced by the presence of anion (5 equiv) added as the tetrabutylammonium salts. dAs the concentration of the anion increased, the ferrocene/ferrocenium redox couple began to exhibit the features of an EC mechanism.

titration curves suggested that [96]-[98] in solution form complexes of 1:1 stoichiometry with all three anionic guests. Relatively smaller amide perturbations are observed with [95], which contains only one amide moiety, and it is not possible therefore to elucidate the [95] :anion stoichiometries from the nmr titration results. The potential values of the reversible redox couple and the results of electrochemical anion recognition experiments are given in Table 24. Significant cathodic perturbations of the current peak potentials are observed for the respective ferrocenyl oxidations in [96]-[98] upon addition of anions. Interestingly, as the concentration of the anionic guest is increased, the shape of the oxidation wave changes from that characteristic of a reversible redox process to one for an EC mechanism. It is noteworthy that [97], which contains a tetrahedral cavity, exhibits the largest cathodic shift (180 mV) with H2P0,, an anionic guest having a complementary shape. Electrochemical competition experiments demonstrated that when an equimolar mixture of H2PO;, HSO; and C1- is used the redox couple shifts by the same amount as when H2P04 is the only anion present. The same result can even be obtained when HSO; and C1- are in tenfold excess concentration over H2P04. Analogous competition experiments with the two anions HSO; and C1- suggest an overall selectivity order for [96], [97] and [98] of H2P04 >HSO; > C1-. These results suggest that [96], [97] and [98] are neutral prototype dihydrogenphosphate-selectiveamperometric sensors, capable of detecting the H2P0: anion in the presence of tenfold amounts of HSOi and C1- anions.

FERROCENEBORONIC ACID

Recently Shinkai and co-workers (Dusemund et al., 1995) have reported the electrochemical properties of ferroceneboronic acid (Fig. 54) in the presence

70

F1 D. BEER, F! A. GALE AND Z. CHEN

Fe

Fig. 54 Ferroceneboronic acid [99].

of anions. This simple molecule has excellent selectivity in the recognition of fluoride ions in the presence of halides and other anions including SCN-, SO:-, and H2PO;. They found a KO,-valueof 1000dm3mol-’ in MeOH-H20 (1:9) for fluoride compared to values of less than 2 for chloride and bromide. The interaction between the boronic acid group and fluoride is attributed to the “hardness” of the boron atom, which strongly interacts with the fluoride anion which is a hard base. On oxidation the ferrocene group becomes more electron-withdrawing, so decreasing the electron density on the boron atom and therefore increasing the strength of the interaction with fluoride. RECOGNITION OF PAIRS OF IONS

We have seen that both cations and anions can be successfully detected electrochemically. The detection of a pair of ions, that is a cation and an anion simultaneously, is a new challenge in this area of chemistry. A ligand containing binding sites for both cations and anions is required for studying the recognition of such entities. Several such difunctional ligands have been reported (Arafa et al., 1992; Flack et al., 1993; Rudkevich et al., 1994; Savage et al., 1994) but none contains redox active groups. Recently reported electrochemical and proton nmr studies of a ferrocenecarboxamidesubstituted diaza-18-crowd receptor [22] have shown, for the first time, that it complexes and electrochemicallyrecognizes a cation-anion pair as indicated in Fig. 55 (Beer et al., 1995d) via subtle changes in the electrochemistry of this species. In summary, the CV shape of compound [22] in the presence of both anion and cation is more reversible than it is in the presence of only the anion and is located at a less positive potential than in the presence of the cation alone. The SWV titration curves, plotting the SWV wave potential against equivalents of anion added, show that the anion is more strongly complexed by the ligand-cation complex than by the free ligand.

REDOX-ACT1VE RECEPTOR MOLECULES

71

Fig. 55 The cation-anion complex of compound [22].

Compound [22] forms thermodynamically stable complexes with Ba2+and

IS+.The binding of anions by free [22] is relatively weak; the metal cation therefore cooperates in the anion complexation process. Electrochemical studies using CV and SWV have shown that [22] is capable of electrochemically recognizing both cations and anions simultaneously and readers are urged to consult the original publication for a detailed explanation of this effect. 4 Towards electrochemical recognition of neutral guest species by redox-active receptor molecules

The design of redox-active sensors for neutral guests is more challenging than that of sensors for charged species because the electrochemical properties of such a host will undergo only small electrostatic perturbations on binding. Therefore, alternative strategies must be employed when designing an electrochemical sensor for a neutral guest. CALIX[S]ARENES

As part of an ongoing research programme aimed at the construction of redox-active calixarene-based host molecules designed to sense neutral

F? D. BEER, !? A. GALE AND Z. CHEN

72

. ' I /

\

I

\

0

'/

0

\

'I

gJ$ &0 (!j

OH

0

Fig. 56 p-t-Butylcalix[5]arene trisferrocenoyl ester [lOo].

guests electrochemically (Beer, 1992), the inclusion properties of modified calix[5]arenes have been examined. Calix[S]arenes can only be synthesized in relatively small yields (Stewart and Gutsche, 1993), but, owing to their larger cavity size, they may possess a greater propensity completely to include small organic molecules than an analogous calix[4]arene (Gutsche and Alam, 1988). The chemistry of the calix[5]arenes is still relatively unexplored and there are few papers detailing modification via lower rim substitution (Stewart et af., 1995). p-t-Butylcalix[S]arene trisferrocene ester [lo01 (Fig. 56) has therefore been synthesized and its electrochemical properties have been examined (Beer et al., 1995a). Crystals of [loo] suitable for X-ray structure determination were grown from a dichloromethane-ethanol solvent mixture. The structure of the resultant [100].ethanol inclusion complex is shown in Fig. 57 together with the atomic numbering scheme of the oxygen atoms. Compound [lo01 contains three ferrocene ester groups appended to the lower rim of the calixarene. Overall the geometry is similar to that of calix[5]arenes which are unsubstituted at the lower rim, taking up the cone conformation. In the complex, hydrogen bonds are formed to a totally included ethanol molecule. The oxygen atom [0(800)] is lying within hydrogen-bonding distance of both unsubstituted oxygen atoms at the bottom of the cone [e.g. O(350) 2.81 A, O(550) 2.82 A] and at larger distances from the other oxygen atoms O(150) 3.51 A, O(250) 3.13 A, O(450) 3.72 A. Electrochemical studies of the behaviour of [lo01 (5 X mol dmV3in dichloromethane solution containing 0.1 mol dm-3 [NBu!]BF4 as supporting electrolyte) have been carried out using cyclic and square-wave voltammetric techniques. The receptor itself undergoes two quasi-reversible oxidations at Epl= +350 mV and Ep2= +450 mV referenced to Ag/Ag+. Rotating disk

REDOX-ACTIVE RECEPTOR MOLECULES

73

Fig. 57 Side view with oxygen numbering scheme of the molecular structure of the ethanol inclusion complex of [loo].

electrode electrochemistry was used to resolve the two oxidation processes. It is possible to express the instantaneous current in terms of the limiting current and the potential applied. Figure 58 illustrates this for [loo]. It can be seen that the limiting current of the first oxidation process is half that of the second. This corresponds to one ferrocene being oxidized at 350 mV and the other two at the larger anodic potential of 450 mV. Preliminary electrochemical investigations on the effects of addition of potential neutral guests (DMF, DMSO, ethanol) to a solution of [loo] in CH2C12show an interesting effect. For example Fig. 59 shows the CV of [loo] in CHzClz background solution upon addition of progressive amounts of DMF. The dielectric constant of CH2C12is 8.9. Addition of DMF ( E = 36.7) shows the peaks merging, but the addition of toluene ( E = 2.4) caused little change in the CV. It may therefore be inferred that the splitting of the peak in CHzClz alone is at least partiany due to the interaction between the ferrocene moieties. The decrease in the interaction upon addition of the solvent of higher dielectric constant implies that the polar solvent is interposing itself between the ferrocene moieties. As the dielectric constant of the interposed solvent increases, the shielding between the ferrocene centres increases owing to the guest’s higher polarizability (Fig. 60). The inclusion of the more highly polarizable guest therefore interferes with the electrochemical interactions

P D. BEER, I? A. GALE AND 2. CHEN

74

0 0

0

50 -

-

X

Current experimental) Current [simulated) 1st electron transfer (calculated) 2nd and 3rd electron transfers (calculated)

-

40

U

:

3 30 c

c

a

5

0

20

1

10 -

0 200

250

300

350 400 450 E ( mV ) vs Ag'/Ag

500

550

600

Fig. 58 Computer fit of the Nernst equation to the rotating disk electrode electrochemistry, at 121 rpm, of [loo] (5 X mol dmP3) in CHZClzwith Bu;NBF4 (0.1 mol dm-3) as supporting electrolyte.

-

(100)in 5ml CH2C12

+ lOOpl DMF + 400pl DMF + 700p1 DMF + 10OOpl DMF

Fig. 59 The two redox processes become more equivalent upon addition of DMF to the solution of [lo01 (5 X mol dm-7 in CHzClzwith Bu!NBF4 (0.1 mol dm-3) as supporting electrolyte.

75

REDOX-ACTIVE RECEPTOR MOLECULES

--

F

Solvent Insertion in lower rim

FCt

Fig. 60 Insertion of polar guest species into the lower rim of [lo01causing a decrease in the interaction between the ferrocene groups.

bTb -676 \ ’+, ’,,, ”,.\+‘ .\‘“

Through Space Electrochemical Interaction Between Redox Centres

”‘

\

Presence of Guest Disrupts Interaction

Scheme 9 The “interference pathway” for coupling complexation and redox reactions.

present in the host, so producing an electrochemical response. It is tentatively suggested that this is an example of a fifth “pathway” (see Scheme l ) , coupling an electrochemical reaction to, in this case, the inclusion of a neutral guest (Scheme 9). This new pathway relies on an interaction between redox centres present in the host molecule. In this case the presence of nonequivalent multiple ferrocene moieties in close proximity allows such an interaction to take place. FERROCENEBORONIC ACID DERIVATIVES

Shinkai and co-workers (Ori and Shinkai, 1995) have recently extended their work using ferroceneboronic acid derivatives to the detection of neutral molecules. A chiral ferroceneboronic acid [loll bearing a tertiary amino group

76

I? D. BEER, I? A. GALE AND Z. CHEN

&

,NHMe,

& (

\

+e-

H

+ )-[lo1lH+

&

NHMe2

-8'

25' H

(+)-[101]'H'

-a:+aoH : I

OH Kmd

(

+ )-[lo1]*saccharide

(

+

Sugar

KO.

-H+

+ )-[lo1]+*saccharide

Scheme 10 Scheme of one square for saccharide binding by [101]H+.

has been shown to bind saccharides at -pH 7 (Scheme 10). Values of Kredand KO,were determined using NPV (normal pulse voltammetry) and the results are shown in Table 25. The saccharide-induced electrochemical potential shift was only observed at pH 7, suggesting that the process actually observed is the conversion of [101]H+ into [loll .saccharide. The KO, values are generally greater than the Kredvalues by an order of magnitude. This is due to the increased acidity of the boronic acid functionality in [loll+. It is also shown that chiral discrimination occurs with linear sugars but not with cyclic ones. This may be due to an OH--.Feinteraction which CPK models have shown is possible for the linear saccharides but not for the cyclic sugars. 5 Conclusions

This review has been concerned with illustrating the concept of electrochemical recognition of guest species by redox-active host systems and the pathways by which the complexation and electrochemical reactions are coupled together. Initially, four pathways for the communication of complexation and redox processes were introduced and the examples presented in the review

REDOX-ACTIVE RECEPTOR MOLECULES

77

Table 25 Association constants (Kredand KO,).” KO,

Kred

Saccharide D-Fructose D-Glucose D-Mannitol D-Sorbitol L-Sorbitol L-Iditol Pentaerythritol

(+)-[1011

(-)-[loll

(+)-[loll

(-)-[1011

1552

1452

28 2 3 110-clO 76 5 7

27 I 4 7027 120 2 10

190 ? 20 622 170 2 20 720 2 50 500 2 40 500 5 50

220 5 20 652 150 5 20 500 -t 40 710 5 50 390 5 40 4453

b

b

13?2

“25”C,pH 7.0 with 0.1 mol dm-3 phosphate buffer. *No change in NPV.

were put into the context of these mechanisms (Fig. 61). It is hoped that these routes for communication have been clearly illustrated so that the future design of electrochemical sensor species may proceed along rational lines. A challenge for the chemist will be the continued design of new receptor species which can sense charged guest species electrochemically or optically in an aqueous environment. As this area of chemistry progresses the production of “real” devices that can be used as diagnostic tools in the field for the detection of various chemical species will become more widespread. One route to such a device is electropolymerization (Beer et al., 1993h). This technique requires the incorporation of an electropolymerizable group into the receptor. However, once this is achieved the fabrication of the device - a modified electrode in this case - simply involves the immersion of the electrode into a solution of the receptor (and perhaps a simple electropolymerizable monomer spacer) and then growing a polymer film by conventional chronoamperometric (constant potential electrolysis) or cyclic voltammetric (potential cycling) methods. Examples of this design strategy include the pyrrole-substituted calixarene [lo21 (Chen et al., 1995) shown in Fig. 62. The inclusion properties of this first calix[4]arene modified electrode with various guest molecules are currently under investigation. The detection of neutral guests by redox-active hosts will be another challenge for the future of this area of chemistry. We have suggested one way in which this could be achieved via a proposed “interference” pathway. Time will tell whether these goals are met. Acknowledgements

We thank the postgraduate and postdoctoral research co-workers for their motivation and determination to succeed in this multidisciplinary research

Q

U Through Space

Through Bond

10 .

"

1 Q '.

o+

,O,,"' 1541

EtO

OEt

Direct Coordination

Conformational Change

Interference

Fig. 61 Examples of complexes, each displaying a different electrochemical-complexation communication pathway.

80

I? D. BEER, f? A. GALE AND Z. CHEN

Fig. 62 Electropolymerizable calix[4]arene [102].

field; their names appear in the references. Special thanks go to Dr M. G. B. Drew at the University of Reading for his many crystal structure determinations and molecular mechanics calculations and to Dr H. R. Powell at the Cambridge Crystallographic Data Centre. We gratefully acknowledge the financial support of the EPSRC, BP (Sunbury), MediSense, Ministry of Defence, Kodak, Unilever, ICI, The Royal Society and NATO.

References Akabori, S., Habata, Y., Sakamoto, Y., Sato, M. and Ebine, S. (1983). Bull. Chem. SOC. Jpn. 56,537 Al-Obaidi, N., Beer, P. D., Bright, J. F'., Jones, C. J., McCleverty, J. A. and Salam, S. S. (1986). J. Chem. SOC., Chem Commun. 239 Amman, D., Huser, M., Krautler, B., Rusterholz, B., Schulthess, l?, Lindemann, B., Halder, E. and Simon, W. (1986). Helv. Chim. Acta 69,849 Arafa, E. A., Kinnear, K. I. and Lockhart, J. C. (1992). J. Chern. SOC., Chem. Commun. 61 Bard, A. J. and Faulkner, L. R. (1980). Electrochemical Methods. Wiley, New York Beer, F! D. (1985a). J. Organornet. Chem. 297,313 Beer, l? D. (1985b). J. Chem. Soc., Chem. Commun.1115 Beer, P.D. (1989). Chem. SOC.Rev. 18,409 Beer, I? D. (1992). Adv. Znorg. Chem. 39,79 [and references cited therein] Beer, P. D. (1994). Adv. Muter. 6,607 Beer, P. D. and Keefe, A. D. (1986). J. Organornet. Chem. 306,C10 Beer, P. D. and Keefe, A. D. (1989). J. Orgartomet. Chem. 375, C40

REDOX-ACTIVE RECEPTOR MOLECULES

81

Beer, P. D. and Szemes, F. (1995). J. Chem. SOC., Chem. Commun. 2245 Beer, P. D., Hammond, P. J., Dudman, C. and Hall, C. D. (1984). J. Organomet. Chem. 263. C37 Beer, k. D., Jones, C. J., McCleverty, J. A. and Salam, S. S. (1987). J. Inclusion. Phenom. 5.504 Beer, P. D., Sikanyika, H., Slawin, A. M. Z. and Williams, D. J. (1989). Polyhedron 8, 879 Beer, P. D., Blackburn, C., McAleer, J. F. and Sikanyika, H. (1990a). Znorg. Chem. 29. 378 Beer; P. D., Keefe, A. D., Sikanyika, H., Blackburn, C. and McAleer, J. F. (1990b). J. Chem. SOC.,Dalton Trans. 3289 Beer, P. D., Hesek, D., Hodacova, J. and Stokes, S. E. (1992). J. Chem. SOC., Chem. Commun. 270 Beer, P. D., Chen, Z., Drew, M. G. B., Kingston, J. E., Ogden, M. I. and Spencer, P. (1993a).J. Chem. SOC.Chem. Cornmun. 1046 Beer, P. D., Chen, Z., Goulden, A. J., Graydon, A., Stokes, S. E. and Wear, T. (1993b). J. Chem. SOC.,Chem. Commun. 1834 Beer, F? D., Crowe, D. B., Ogden, M. I., Drew, M. G. B. and Main, B. (1993~).J. Chem. SOC., Dalton Tram. 2107 Beer, P. D., Danks, J. P., Hesek, D. and McAleer, J. F. (1993d). J. Chem. Soc., Chem. Commun. 1735 Beer, I? D., Dickson, C. A. P., Fletcher, N., Goulden, A. J., Grieve, A., Hodacova, J. and Wear, T. (1993e). J. Chem. SOC.,Chem. Commun. 828 Beer, P. D., Drew, M. G. B., Hazlewood, C., Hesek, D. and Stokes, S. E. (1993f). J. Chem. SOC., Chem. Commun. 229 Beer, P. D., Hazlewood, C., Hesek, D., Hodacova, J. and Stokes, S. E. (19938).J. Chem. SOC.,Dalton Trans. 1327 Beer, P. D., Kocian, O., Mortimer, R. J. and Ridgway, C. (1993h). J. Chem. SOC.,Dalton Trans. 2629 Beer, P. D., Chen, Z., Drew, M. G. B. and Gale, P. A. (1994a). J. Chem. SOC.,Chem. Commun. 2207 Beer, €? D., Chen, Z., Drew, M. G. B., Gale, P. A., Heath, J. A., Knubley, R. J. and Ogden, M. I. (1994b). J. Znc. Phenom. Mol. Recognit. Chem. 19,343 Beer, P. D., Chen, Z., Drew, M. G. B. and Pilgrim, A. J. (1994~).Znorg. Chim. Acta 225, 137 Beer, P. D., Chen, Z. and Gale, P. A. (1994d). Tetrahedron 50,931 Beer, P. D., Chen, Z., Grieve, A. and Haggitt, J. (1994e). J. Chem. Soc., Chem. Commun. 2413 Beer, P. D., Chen, Z., Drew, M. G. B. and Gale, P. A. (1995a). J. Chem. SOC., Chem. Commun. 1851 Beer, P. D., Chen, Z. and Pilgrim, A. J. (1995b). J. Chem. SOC., Faraday Trans. 4331. Beer, P. D., Chen, Z. and Pilgrim, A. J. (199%). J. Electroanal. Chem. 444,209. Beer, P. D., Chen, Z. and Ogden, M. I. (1995d). J. Chem. SOC., Faraday Trans. 295 Beer, P. D., Crane, C. G., Danks, J. I?, Gale, P. A. and McAleer, J. F. (1995e). J. Organomet. Chem. 490,143 Beer, P. D., Drew, M. G. B., Graydon, A. R., Smith, D. K. and Stokes, S. E. (1995f). J. Chem. SOC., Dalton Trans. 403 Beer, P. D., Drew, M. G. B., Hesek, D. and Jagessar, R. (1995g). J. Chem. SOC.,Chem. Comrnun. 1187 Beer, P. D., Hesek, D., Kingston, J. E., Smith, D., Stokes, S. E. and Drew, M. G. B. (1995h). Organometallics 7 , 3288 Bell, A. P. and Hall, C. D. (1980). J. Chem. SOC., Chem. Commun. 163

82

P D. BEER, I? A. GALE AND Z. CHEN

Bell, A. P., Hammond, P. J. and Hall, C. D. (1983). J. Chem. SOC., Perkin Trans. 1, 707 Bethell, D., Dougherty, G. and Cupertino, D. C. (1995). J. Chem. Soc., Chem. Commun. 675 Biernat, J. F. and Wilczewski, T. (1980). Tetrahedron 36,2521 Bougoin, M., Wong, K. H., Hui, J. Y. and Smid, J. (1975). J. Am. Chem. SOC. 97, 3462 Casnati, A., Comelli, E., Fabbi, M., Bocchi, V., Mori, G., Ugozzoli, F., Lanfredi, A. M. M., Pochini, A. and Ungaro, R. (1993). Red. Trav. Chim. Pays-Bas 112,384 Charlot, G., Badoz-Lambling, J. and Bhillion, B. (1962). In Electrochemical Reactions (ed. G. Charlot). Elsevier, Amsterdam, p. 46 Chen, Z., Gale, P.A., Heath, J. A. and Beer, F! D. (1994). J. Chem. SOC.,Faraday Tram. 2931 Chen, Z., Gale, F! A. and Beer, I? D. (1995). J. Electroanal. Chem. 393,113 Collins, E. M., McKervey, M. A., Madigan, E., Moran, M. B., Owens, M., Ferguson, G. and Hams, S. J. (1991). J. Chem. SOC., Perkin Trans. 1, 3137 Czarnik, A. W. (1994). Acc. Chem. Res. 21,302 De, G., Li, J.-Z., Yu, R.-Q. and Zheng, G. D. (1994). Anal. Chem. 66,2245 Delgado, M., Gustowski, D. A., Yoo, H., Gokel, G. W. and Echegoyen, L. (1988). J. Am. Chem. SOC. 110,119 Delgado, M. R. E., Wolc J., Hartman, J. R., McCafferty, G., Yagbasan, R., Rawle, S. C., Watkin, D. J. and Cooper, S. R. (1992). J. Am. Chem. SOC. 114,8983 Dietrich, B. (1993). Pure Appl. Chem. 65,1457 Dusemund, C., Sandanayake, K. R. A. S. and Shinkai, S. (1995). J. Chem. SOC., Chem. Commun. 333 Edmonds, T. E. (1988). In Chemical Sensors (ed. T. E. Edmonds). Blackie, London, p. 193 Flack, S. S., Chaumette, J. L., Kilburn, J. D., Langley, G. J. and Webster, M. (1993). J. Chem. SOC., Chem. Commun. 399 Fu, E., Granell, J., Green, M. L. H., Lowe, V. J., Marder, S. R., Saunders, G. C. and Tbddenham, M. (1988a). J. Organomet. Chem. 355,205 Fu, E., Green, M. L. H., Lowe, V. J. and Marder, S. R. (1988b). J. Organomet. Chem. 341, c39 Gourdon, A. (1992). New J. Chem. 16,953 Gustowski, D. A., Delgado, M., Gatto, V. J., Echegoyen, L. and Gokel, G. W. (1986). J. Am. Chem. SOC.108,7553 Gutsche, C. D. and Alam, I. (1988). Tetrahedron 44,4689 Hall, C. D. and Chu, S. Y. F. (1995).J. Organomet. Chem. 498,221 Hall, C. D. and Sharpe, N. W. (1990). Organometallics 9, 952 Hall, C. D, and Sharpe, N. W. (1991). J. Organomet. Chem. 405,365 Hall, C. D., Danks, I. F!, Lubienski, M. and Sharpe, N. W. (1990a).J. Organomet. Chem. 384,139 Hall, C. D., Danks, I. P., Nyburg, S. C., Parkins, A. W. and Sharpe, N. W. (1990b). Organometallics 9, 1602 Hall, C. D., Parkins, A. W., Nyburg, S. C. and Sharpe, N. W. (1991a). J. Organomet. Chem. 407,107 Hall, C. D., Tucker, J. H. R. and Sharpe, N. W. (1991b). Organometallics 10, 1727 Hall, C. D., Tbcker, J. H. R., Chu, S. Y. E, Parkins, A. W. and Nyburg, S. C. (1993). J. Chem. SOC., Chem. Commun. 1505 Hammond, I? J., Beer, P. D., Dudman, C., Danks, I. l? and Hall, C. D. (1986). J. Organomet. Chem. 306,367 Johnstone, R. A. W. and Rose, M. E. (1983). J. Chem. SOC.,Chem. Commun. 1268

REDOX-ACTIVE RECEPTOR MOLECULES

83

Kadish, K. M., Xu, Q. Y. Y., Bhaskar Maiya, G., Barbe, J.-M. and Guilard, R. (1989). J. Chem. SOC.,Dalton Trans. 1531 Kaifer, A. E. and Echegoyen, L. (1990). Redox control of cation binding in macrocyclic systems. In Cation Binding by Macrocyclic Systems (ed. Y. Inoue and G. W. Gokel). Marcel Dekker, New York, p. 363 Kaifer, A., Echegoyen, L., Gustowski, D. A., Goli, D. M. and Gokel, G. W. (1983). J. Am. Chem. SOC. 105,7168 Kaifer, A., Gustowski, D. A., Echegoyen, L., Gatto, V.J., Schultz, R. A., Cleary, T. P., Morgan, C. R., Goli, D. M., Rios, A. M. and Gokel, G. W. (1985).J. Am. Chem. SOC. 107,1958 Katz, H. E. (1991). In Znclusion Compounds (ed. J. L. Atwood, J. E. D. Davies and D. D. MacNicol). Academic Press, New York, p, 391 Kibukawa, K., He, G. H., Abe, A., Goto, T., Arata, R., Ikeda, T., Wada, F. and Matsuda, T. (1987). J. Chem. SOC., Perkin Trans. 2, 135 Kimura, E., Kodama, M. and Yatsunami, T. (1982). J. Am. Chem. SOC. 104, 3182 Kimura, K., Sakamoto, H., Koseki, Y. and Shono, Y. (1985). Chem. Lett. 1241 Lowe, N. D. and Garner, C. D. (1993a). J. Chem. SOC., Dalton Trans. 2197 Lowe, N. D. and Gamer, C. D. (1993b). J. Chem. Soc., Dalton Trans. 3333 Marcek, J. F. and Burrows, C. J. (1986). Tetrahedron Lett. 29, 6231 McKillop, A., Swann, B. P. and Taylor, E. C. (1970). Tetrahedron 26,4031 Medina, J. C., Goodnow, T.T., Rojas, M. T., Atwood, J. L., Lynn, B. C., Kaifer, A. E. and Gokel, G. W. (1992). J. Am. Chem. SOC. 114,10583 Morita, Y., Agawa, T., Kai, Y., Kanehisa, N., Kasai, N., Nomura, E. and Taniguchi, H. (1989). Chem. Lett. 1349 Morita, Y., Agawa, T., Nomura, E. and Taniguchi, H. (1992). J. Org. Chem. 57,3658 Ori, A. and Shinkai, S. (1995). J. Chem. SOC., Chem. Commun. 1771 Osteryoung, J. and O’Dea, J. J. (1987). In Electroanalytical Chemistry 14, 209 Plenio, H. and Diodone, R. (1995a). J. Organomet. Chem. 492,73 Plenio, H. and Diodone, R. (1995b). Znorg. Chem. 34, 3964 Plenio, H., El-Desoky, H. and Heinze, J. (1993). Chem. Ber. 126,2403 Plenio, H., Yang, J., Diodone, R. and Heinze, J. (1994). Znorg. Chem. 33,4098 Reddy, P. A., Kashyap, R. P., Watson, W. H. and Gutsche, C. D. (1992). Israel J. Chem. 32,89 Rudkevich, D. M., Brzozka, Z., Palys, M., Visser, H. C., Verboom, W. and Reinhoudt, D. N. (1994). Angew. Chem. Znt. Ed. Engl. 33,467 Saji, T. (1986). Chem.. Lett. 275 Saji, T. and Kinoshita, I. (1986). J. Chem. SOC., Chem. Cornmun. 716 Savage, P. B., Holmgren, S. K. and Gellman, S. H. (1994). J. Am. Chem. SOC. 116, 4069 Schmidtchen, F. P. (1988). Nachr. Chem. Tech. Lab. 8, 10 Seely, G. R., Gust,’D., Moore, T. A. and Moore, A. L. (1994). J. Phys. Chem. 98, 10 659 Sessler, J. L., Furuta, H. and Cyr, M. J. (1991). J. Am. Chem. Soc. 113,6677 Sessler, J. L., Moody, J. D., Ford, D. A. and Lynch, V. (1992). Angew. Chem. Znt. Ed. Engl. 31,452 Southampton Electrochemical Group (1985). Instrumental Methods in Electrochemistry. Ellis Horwood, Chichester Stewart, D. R. and Gutsche, C. D. (1993). Org. Prep. Proc. Znt. 25, 137 Stewart, D. R., Krawiec, M., Kashyap, R. P., Watson, W. H. and Gutsche, C. D. (1995). J. Am. Chem. SOC. 117, 586 [and references cited therein] Suga, K., Fujihira, M., Monta, Y. and Agawa, T. (1991). J. Chem. SOC.,Faraday Trans. 1575

I? D. BEER, F! A. GALE AND Z. CHEN

84

Szemes, F., Hesek, D., Chen, Z., Dent, S. W., Drew, M. G. B., Goulden, A. J., Graydon, A. R., Grieve, A., Mortimer, R. J., Wear, T,,Weightman, J. S. and Beer, €! D. (1996). Inorg. Chem. 35,5868 Vicek, A. A., Volke, J., Pospisil, L. and Kalvoda, R. (1986). In Physical Methods of Chemistry (ed. B. W. Rossiter and J. F. Hamilton). Wiley, New York, p. 797 Wolc R. E. J. and Cooper, S. R. (1984). J. Am. Chem. SOC.106,4646 Zompa, L. (1978). Inorg. Chem. 17,2531

Appendix: Understanding cyclic voltammetry and square-wave voltammetry FUNDAMENTALS (Bard and Faulkner, 1980 Southampton Electrochemistry Group, 1985)

Voltammetry is the study of electrochemical processes occurring at an electrode-electrolyte interface with respect to the electrode potential (E) and current ( I ) , and also the time (r) if non-steady-state conditions are required. A voltammogram plots I against E and specifies the experimental time scale. Voltammetry is often conducted by applying different potential or current perturbations, by using electrodes of various physical and chemical properties and geometric structures, and by forcing a type of controllable convection in the electrochemical solution. Let us consider a solution containing the supporting electrolyte and only the ligand 0 which is in its stable oxidized state and can undergo the following electrode reduction to give rise to the product R:

If the electrode is flat with a typical surface area, A, of the order of square millimetres, the I-E relation of reaction (11) is then governed by (12): ( E - E " ) ) - c::"""" exp(

(1 - n)nF RT

and also by Fick's first law, which can be converted to equations (13) and (14): dcSdace I = -nF-= -nFD,dt

deb""" dr

(A.3)

where F and E" are the Faraday constant and the standard electrode potential; and cgmc.are the concentrations of 0 and R at the electrode surface;

R EDOX-ACTIVE RECEPTOR MOLECU LES

85

Do and D Rare the respective diffusion coefficients;k" and a are known as the standard (electron transfer) rate constant and electron transfer coefficient respectively, and both are kinetic parameters characterizing the feasibility of the electron transfer; x is the distance away from the electrode surface. After the electrode reaction starts at a potential close to E", the concentrations of both 0 and R in a thin layer of solution next to the electrode become different from those in the bulk, co and cR. This layer is known as the diffusion layer. Beyond the diffusion layer, the solution is maintained uniform by natural or forced convection. When the reaction continues, the diffusion layer's thickness, I, increases with time until it reaches a steady-state value. This behaviour is also known as the relaxation process and accounts for many features of a voltammogram. Besides the electrode potential, equations (A.3) and (A.4) show that the electrode current output is proportional to the concentration gradient dc&""""/dx or dc2"""ld.x. If the concentration distribution in the diffusion layer is almost linear, which is true under a steady state, these gradients can be qualitatively approximated by equation (AS).

This approximation is very useful for qualitatively understanding, for example, cyclic voltammetry.

CYCLIC VOLTAMMETRY

The most popular voltammetric technique is probably cyclic voltammetry (CV), partly because of its early development in theory and the availability of the corresponding commercial equipment. In this technique, the electrode potential is first scanned linearly with time from a starting potential, where no reaction occurs, passing E", towards another potential, and then reversed back to the starting potential. In this case, the time variable can be conveniently represented by the scan rate, v. A CV voltammogram can be recorded under either a dynamic or a steady state depending on the electrode design and solution convection mode. In a stationary solution with a conventional disk electrode, if the scan rate is sufficiently high to ensure a non-steady state, the current will respond differently to the forward and backward potential scan. Figure 63 shows a typical CV for a reversible reduction.' 'In this review, wherever electrochemistry is concerned, the reversibility of a reaction refers firstly to the chemical reversibility.It also requires that the electron transfer reaction occurs at such a rate that the rate of the whole electrodic process, which is measured by the output current of the electrode, is controlled by the diffusion of the redox species towards the electrode surface. Furthermore, the surface concentrations of 0 and R at a given potential should be governed by the Nernst equation.

86

t? D. BEER, t? A. GALE AND Z.CHEN

0.4

0.6

-E/V

4

Reverse scan current base line

Fig. 63 A cyclic voltammogram often observed for a highly reversible one-electron reduction in a non-aqueous solution

The CV shows an asymmetrical peak-shaped current wave or wave couple. The current output increases upon increasing the potential scan rate and it is the potential scan rate that contributes dominantly to the peak shape of the

cv.2

When the potential scan is very slow, the relaxation of the diffusion layer can always follow the change in cgrtace and a steady-state voltammogram (SSV) is recorded. At a higher scan rate, the change in cgrtacebecomes faster than the relaxation of the diffusion layer; at a given potential a larger current output is found owing to a thinner diffusion layer according to the approximation (AS). This accounts for the scan rate dependence of the CV current. On the other hand, when c&u;Lrface effectively approaches zero, increasing the potential further will no longer affect the current output. However, the diffusion layer is, at this moment, far from its stable state but continues to relax, which further increases the layer's thickness. According to the approximation (AS), this will result in a decrease in the current output, giving the peak shape of the CV. A similar discussion can be applied to the re-oxidation process when the potential scan is reversed. However, it should be noted that when the scan is reversed, the diffusion layer contains not only R but also 0.The reduction of 0 continues, although the scan is reversed, and will dominate the current *Theoretically,the CV can be strictly derived from the diffusion behaviour of the reactant and product.

REDOX-ACTIVE RECEPTOR MOLECULES

87

output until the scan approaches the potential where the oxidation of R dominates. This accounts for the positive starting current of the oxidation wave. The smaller peak current of the re-oxidation wave results from some of the product diffusing into the bulk solution and so not being re-oxidized on the voltammetric time scale. The key parameters from a CV measurement include the wave shape, the peak potential(s), E,, and E,,, and, more importantly, their dependence on the scan rate. For reversible and many quasi-reversible systems, the average of EpB and E,, equals or closely approximates El12.For judging the reversibility of an electrode reduction like reaction (A.l) at 25"C, the useful criteria are: (1) AE, = EPa- Epc= 59/n mV; (2) I I p J ~ p c I= 1; (3) both Zpa and I,, are proportional to v112; (4) both Epaand Epcare independent of the scan rate; ( 5 ) at potentials beyond EPaor E,,, Z is proportional to t-'".

Any deviation from the above criteria is indicative of kinetic complications and should be treated individually. However, one case is worthy of note. In non-aqueous solutions, it is commonly observed that AE,, for example, has typical values between 70 and 100 mV owing to the so-called ZR drop resulting from the uncompensated and relatively large solution resistance. While ZR compensation techniques are available, they are not always reliable, and it is more convenient to compare the measured AE, with that of a known reversible reaction measured under similar conditions. In electrochemical recognition studies, an often encountered kinetic case is the so-called EC mechanism, when a chemical reaction occurs following the electron transfer. One type of subsequent chemical reaction originates from complexation or decomplexation upon changing the oxidation state of the ligand. Another type is direct reaction, such as ion-pair formation, between the guest species and the redox centre after the electron transfer. In both cases, because the electron transfer reaction product is consumed by the chemical reaction, the electron transfer becomes faster but its reverse process becomes slower. The corresponding changes on the CV are that the forward scan wave shifts in opposition to the scan direction and the reverse scan wave becomes smaller or disappears. For a completely irreversible reaction, there will be only a single current wave whose peak potential is the only potential parameter available from a CV. In this case, because the peak potential depends strongly on scan rate, it should not be used to evaluate which can only be obtained by SSV. However, reporting the peak potential and its dependence on scan rate is by no means less important than reporting and a more comprehensive knowledge should be obtained by reporting both and investigating their difference upon changing the scan rate and other experimental conditions.

88

I? D. BEER, P A. GALE AND 2. CHEN

SQUARE-WAVE VOLTAMMETRY (SWV) (Osteryoung and ODea, 1987)

In a CV measurement, the current output always contains two components: the Faradaic current, IF,due to the reaction of the redox species and the capacitive charging current, Ic, which results from the charging of the electrode double layer and the diffusion layer. (This diffusion layer contains all charged and polar species in the solution and therefore differs from that of the redox species.) While IF changes linearly with vl'* as determined by diffusion, Zc is directly proportional to v as shown below, where C , is the total electrode capacitance and q the added capacitance charge: dq d(C,E) - c,-dE = C D V I --=-dt dt dt

'-

This means that with increasing scan rate or lowering the solution concentration, the effect of Zc will increase. Because a peak-shaped CV can only be obtained at a sufficiently high scan rate, the effect of electrode capacitance charging limits the CV application in low-concentration solutions. SWV has been developed to overcome this problem and to increase the quantitative accuracy of voltammetric techniques. The concentration for recording a SW voltammogram can be as low as a hundredth of that for recording a CV. The potential perturbation in SWV is a stair-like square wave as shown in Fig. 64. Taking an average over time, the potential change in SWV can be considered as linear and the concept of scan rate still applies. In practice, the scan rate of an SW voltammogram is often described in terms of the frequency of the square wave, f,which can be converted to the scan rate by the following equation (see Fig. 64):

Unlike in CV, pairs of current measurements are made on each period of the square wave. These are at time tfonvard late in the forward pulse, named Iforward, and t,,,,,, in the reverse pulse, named I,,,,,,. Both ?forward and t,,,,,, are much greater than the time for fully charging the electrode capacitance, so that only the Faradaic current is recorded. With calculation of Inet,the difference and Ireverse, SWV presents three types of peak-shaped I-E between Iforward relations. Figure 65 displays the SWV of a reversible one-electron reduction process. The unsymmetrical shapes of the forward and reverse components of an SW voltammogram have similar origins as those of the CV. However, unlike the reverse scan wave in CV, the reverse SWV wave is measured almost simultaneously with the forward component. Therefore, there is much less accumulation of the reaction product at the electrode surface during the potential scan. This feature of SWV makes it very useful for understanding

89

REDOX-ACTIVE RECEPTOR MOLECULES

E

*

V

l/f

.......

-

s!

%l-a$

'

EEnarmsd

t

>

Fig. 64 The potential-time wave for square-wave voltammetry. This wave may be and obtained by superimposing a square wave with constant pulse height (EqUme) width (1/(2f);f is the wave frequency) on a staircase wave with constant increment ( E i n m e m e d and width [1/(2f)l.

Fig. 65 A square-wave voltammogram often observed for a highly reversible one-electron reduction in a non-aqueous solution.

90

I? D. BEER, F! A. GALE AND Z. CHEN

cyclic voltammograms distorted by product precipitation onto the electrode surface and by other slow subsequent chemical reactions. The symmetrical shape of I,,, vs E plots results mainly from the diminution of the diffusioncontrolled limiting current which, in theory, is the same on both forward and reverse components and also the fact that Znetreaches maximum at E". Therefore, while the value of E" or may be directly read from the peak potential, Ep,of the Znet vs E curve for reversible or quasi-reversible reactions, the peak current, Zp, serves as a sensitive parameter proportional to concentration for quantitative analysis. Another advantage of SWV over CV can be seen when dealing with a separate multi-electron transfer reaction. The CV current wave of each or each group of electrons always contains the contribution from the previous electron transfer, particularly the diffusion-controlled current. Separating currents from different electron transfers can be tedious, if not impossible. It can be even worse when we have to take into account the capacitive charging current. Since both capacitive and diffusion-controlled currents are absent or at least minimized on the Z,, vs E curve of an SW voltammogram, current waves from each electron transfer are much better resolved and more accurate information can be obtained.

Spin Trapping and Electron Transfer LENNART EBERSON

Department of Chemistry, Lund University, Lund, Sweden

1 2 3 4 5

6 7 8 9

10 11 12 13 14 15 16

Introduction Redox mechanisms of spin trapping Electron transfer theory Spin trapping and electron transfer Kinetic scheme for spin adduct formation Evidence for the ST+-nucleophile mechanism under thermal conditions Oxidation of PBN by hexachloroosmate(V) ion in the presence of nucleophiles Oxidation of PBN by tris(6bromophenyl)aminium ion (TBPA+) in the presence of nucleophiles Properties of the PBN and DMPO radical cations Anodic spin trapping experiments Photochemical spin trapping experiments Examples of problems in photo-initiated spin trapping Formation of chloro spin adducts Formation of cyano and thiocyano spin adducts Formation of the trinitromethyl spin adduct Formation of imidyl spin adducts Trapping of aroyloxyl radicals Ionizing radiation and spin trapping Spin trapping of radicals generated by ultrasound (sonolysis) Spin trapping in biochemicalhiological systems Conclusions on the radical cation mechanism Spin adduct formation via radical anions The nucleophdic addition-oxidation mechanism Trapping of the hydroxyl radical Bona fide spin trappings: a recipe References

91 93 96 101 104 105 106 110 114 116 118 121 121 122 123 123 124 126 126 127 129 129 130 132 136 137

1 Introduction

Spin-trapping is a much-used technique for the detection and identification of transient radicals (for reviews on fundamental aspects, see Janzen, 1971; Evans, 1979; Perkins, 1980; Janzen and Haire, 1990). A spin trap (ST) is a compound containing a functional group to which a transient radical X' can 91 ADVANCES IN PHYSICAL ORGANIC CHEMISTRY 0065-3160/98 $30.00 VOLUME 31

Copyright 0 1998 Academic Press A / / rights of reproduction in any form rcscrved

L. EBERSON

92

add with formation of a new, persistent radical X-ST' with easily observable epr spectral characteristics [equation (l)].IIfipical spin traps contain a nitroso group, as in 2-methyl-2-nitrosopropane (MNP) [11, 2,4,6-tri-t-butylnitrosobenzene (TBN) [2], or a nitrone function, as in N-t-butyl-a-phenylnitrone (PBN) [3] or 5,5-dimethyl-l-pyrroline-l-oxide (DMPO) [4].Compounds with a carbon-carbon double bond can also act as spin traps, e.g., 1,l-di-tbutylethylene [5]. X

111 MNP

+ ST

+

X-ST

[21 TBN

(1)

[31 PEN

0 141 DMPO

I51

The typical spin trapping experiment is designed in the following way. The spin trap is added in low concentration to a solution containing the components of a suspected radical reaction, the reaction is initiated (by addition of a critical component, by raising the temperature of the sample, by photolysis, by electrolysis, by radiation, etc.) and monitored by epr spectroscopy. The appearance and identification of a spin adduct epr spectrum of X-ST' is then considered evidence that X was an intermediate in the reaction. It was soon realized (Perkins, 1980) that such a conclusion is valid only if the spin trap is devoid of all reactivity other than the capability of reacting with X to form a persistent adduct X-ST'. This ideal is not fulfilled by any of the spin traps so far used; the spin trap is seldom just an innocent collector of radicals but can participate in the reaction under study in various ways. The most important of these are connected with the redox properties of spin traps and species derived from them, and the common theme to be discussed here is the

93

SPIN TRAPPING AND ELECTRON TRANSFER

electron transfer (ET) reactivity of spin trap systems and the risk that a particular spin adduct has been formed in a way different from equation (1). It should be pointed out from the beginning that even rather weak redox agents like dioxygen can be suspected to initiate pathways to X-ST' which do not involve the formation of X . Thus the redox properties of the components of almost any spin trapping experiment should be a matter of concern in its execution and in the interpretation of the observations.

2 Redox mechanisms of spin trapping It was pointed out at an early stage (Forrester and Hepburn, 1971) that both nitroso compounds and nitrones can add nucleophilic species X:- in equilibrium processes to give anions of hydroxylamines, as exemplified by [l] and [3] in equations (2) and (3). Such anions or their protonated forms are easily oxidized, for example by dioxygen or, in the case of equation (2) even the original nitroso compound, and then produce the same spin adduct as if X had been added to [l]or [3] [see equations (4)and (5)].

PhCH=N(O)Bu' [31

+ X:-

PhCH(X)-N(O-)Bu'

+

0 2 -+

PhCH(X)-N(O-)Bu'

PhCH(X)-N(O)Bu'

+ 0;

(3)

(5)

For [3], anions such as acetate, cyanide and nitromethanide were shown to form spin adducts with traces of adventitious dioxygen as the oxidant. A different version of equation (3) - addition of an organolithium (RLi) or Grignard reagent (RMgX) to [3] followed by dioxygen treatment - had actually been used earlier for the verification of the structure of alkyl and aryl radical adducts to [3] (Janzen and Blackburn, 1969). Little study has been devoted to the nucleophilic addition-oxidation mechanism since its inception, in spite of the fact that it could be critically important to avoid in many situations, not least in experiments purporting to demonstrate the dangers of radicals generated in biological systems. Thus, for example, reports of trapping of the hydroxyl radical in biological systems would sound a far less ominous note if the addition-xidation mechanism were involved. Recently, interest in it has been revived after the finding that many heteroaromatic bases

94

L. EBERSON

containing the N-H group, such as lH-1,2,3-triazole, benzotriazole, pyrazole, imidazole, benzimidazole and purine, undergo addition to the nitrone functionality with great ease, thereby generating a hydroxylamine intermediate which can be oxidized by mild oxidants to persistent spin adducts with characteristic epr spectral parameters (Carloni et al., 1996; Alberti et al., 1997). The second ET mechanism of importance in spin trapping may prevail in reactions taking place under strongly oxidizing or reducing conditions. In the treatment to follow, we will deal mainly with oxidation reactions involving radical cations (for reviews on radical cation reactions, see Bard et al., 1976; Hammerich and Parker, 1984; Yoshida, 1984) since these are by far the most commonly encountered and studied ones; it is to be stressed, however, that analogous considerations apply to reductive processes (see Section 14). The main feature of this mechanism originates from the fact that spin traps have redox reactivity, as exemplified by PBN [3], which is oxidized to its radical cation at Epa= 1.47 V vs SCE in acetonitrile (McIntire et al., 1980). Thus a strong one-electron oxidant (Ox) is capable of oxidizing PBN to its radical cation, and the latter can react with a nucleophile Nu- present to give the spin adduct [equations (6)-(7)], provided the nucleophile is resistant towards the oxidant. If the intention of the study is to show that Ox reacted with Nu-, forming Nu’, the appearance of Nu-PBN will not immediately be interpretable as evidence for the latter process. Ox+PBN

-

Red+PBN+

(6)

PBN+ + NU- + NU-PBN

(7)

In a reducing environment, conditions may allow for the same type of mechanism to occur, but with the radical anion of the spin trap as the intermediate. Actually, the possibility of radical ion-mediated “spin trapping” was first discussed in a study of a reductive system, namely in the search for radical intermediates in the reaction between alkanethiolates and alkyl halides conducted in the presence of TBN [2] (Crozet et al., 1975). TBN is known to trap primary radicals with formation of nitroxides (attack of R’at N), and it was therefore anomalous to find alkoxyaminyl radicals (attack of R’ at 0) in the above reaction. It was suggested that the alkanethiolate or some other reductant reduces TBN to its radical anion, which attacks the alkyl halide via oxygen in an SN2fashion, as in equations (8) and (9) (see p. 129). RS-

+ R-N=O

R-N-0-

+ R’X

* RS’

+ R-N-O-

SNZ

+

R-N-OR’

+ X-

(8)

(9)

SPIN TRAPPING AND ELECTRON TRANSFER

95

This finding initiated programmes aimed at elucidating the redox properties of spin traps (as first defined for PBN by Bard et al., 1974) and to define their “potential windows”, i.e., the potential range within which the spin trap would not undergo electron transfer in any direction (McIntire et al., 1980; Sosonkin ef al., 1982; Gronchi et al., 1983; Cerri et al., 1989; Gronchi and Tordo, 1993). A few examples of spin adduct formation via radical cations were described, such as the formation of the hydroxyl adduct of PBN when photolysed in an aqueous solution of sodium peroxydisulfate (10) (Janzen and Coulter, 1984), formation of the chloro adduct of PBN upon photooxidation in dichloromethane in the presence of an arylonium chloride [Ar,X+Cl-; X = N, P, S, I; equation (11)](Baumann et al., 1985), and formation of the hydroxyl adduct of DMPO from DMPO+,generated by y-irradiation and characterized by its epr spectral properties at 77 K and allowed to react with water at 270 K (12) (Chandra and Symons, 1986). By 1986, the radical cation mechanism was explicitly utilized to produce ‘70-labelled hydroxyl and superoxide radical adducts of DMPO and PBN (Mottley et al., 1986).

hv

PBN

-

PBN+

acceptor (Ar,X’.)

-

M)

~

CI-PBN

(11)

HO-DMPO

(12)

270 K

CFCI,, 77 K

DMPO

CI

electron

PBN*

DMPO+

H20

Co Y

Extended studies later showed that the radical cation mechanism of equations ( 6 )and (7) is prevalent in strongly oxidizing systems (Eberson, 1992; Eberson and Nilsson, 1993), especially under photolytic conditions where excited states are formed (Eberson, 1994). The latter normally can act as strong oxidants or reductants, as in equation ( l l ) , and thus can create radical ions under seemingly mild conditions. Since this type of mechanism was judged to be much more common than thought, the name “inverted spin trapping” was coined for it in view of the inverted situation of electron demand which appears when an electron is formally transferred between the spin trap and the nucleophile or electrophile; see equations (13)-(16). True spin trapping: Inverted spin trapping: True spin trapping: Inverted spin trapping:

ST + Nu’- NU-ST’

ST”

+ Nu-

ST + E

-+

-

Nu-ST’

E-ST’

ST-’ + E+ + E-ST‘

(13)

(14)

(15) (16)

96

L. EBERSON

3 Electron transfer theory

Before we can enter a discussion of the redox processes involved in the two mechanisms defined above, we need a simple theoretical background which provides relevant insights into the phenomenon of ET. The Marcus theory of outer-sphere ET provides such a framework for the delineation of mechanistic domains, thanks to its origin in a simple model and its classical nature (Marcus, 1964; Marcus and Sutin, 1985; for applications in organic chemistry, see Eberson, 1982b, 1987). The Marcus model of intermolecular ET refers to reaction (17) and is detailed in Fig. 1. D +A+ D ++A -

(17) The two reacting species, D (after donor, the reductant) and A (after acceptor, the oxidant) are approximated as two spheres of charges Z1 and 2, and radii rl and r2, D being symbolized by the larger sphere in which the arrow represents the electron to be moved. The two spheres first have to diffuse

A

D

Precursor complex

Successor complex Fig. 1 The Marcus model of intramolecular electron transfer, involving two colliding spheres of charges Z1 and Z2 and radii rl and r2.

SPIN TRAPPING AND ELECTRON TRANSFER

97

4

I

b

Reaction coordinate Fig. 2 Potential energy curves of the (D A) and ( D + A - ) states.

together with rate constant kd to form the precursor complex, usually assumed to be at collision distance between the two centres, r12= rl + r,. An electron is transferred under Franck-Condon restrictions, i.e., so rapidly that the positions of all nuclei are frozen during the event. From this it follows that the energy levels between which the electron jumps must be equal to within +RT (0.6 kcal mol-' at room temperature). This requirement is met by increasing the energy of the two species until the desired matching of the energy levels is reached; at this point the transition state is reached (Fig. 2). The increase in energy is achieved by bond and solvent reorganization, associated with bond and solvent reorganization energies hi and A,, symbolized in Fig. 1 by changes in the radii of the spheres in the transition state. The sum of these energy terms is A, the reorganization energy. Bond reorganization hi involves bond stretching and compression, angle deformation and torsional movement, whereas solvent reorganization A, involves movement of solvent molecules around the reactants to adapt to the changing electrostatic situation in connection with the electron jump. The change in bond reorganization is symbolized in Fig. 1 by a change in the original radii of the spheres (dotted lines to solid lines). Both terms can be calculated by classical procedures, and hi is also accessible by quantumchemical calculations (Eberson et al., 1993). A special case of ET reactions is the simple self-exchange reaction shown in equations (18) and (19) for D and A, respectively; here only an electron is exchanged and no net chemical change occurs. D+D+

+Df+D

(18)

A+A-

_1

A-+A

(19)

98

L. EBERSON

The reorganization energy of a self-exchange reaction is denoted h(0) (from the fact that AGO = 0) and is an important quantity in the Marcus theory, where it can be shown that the activation free energy of a self-exchange reaction, AG(O)*,is equal to h(0)/4. It is also possible to measure rate constants of self-exchange processes experimentally and thus get access to X(0) via this relationship. It is important to remember that the Marcus model refers to a weakly electronically coupled model, as embodied in the term “outer-sphere” ET. Thus it must be assumed that the electronic overlap between the two reactants is so small that no quantum-chemical effects ensue, yet that there must be enough overlap for the transmission coefficient K of the Eyring equation to be equal to 1 (the reaction must be adiabatic). Usually, this minimum overlap requirement is put at a fairly low level, around 0.1 kcal mol-l, which causes no problems for most reactions involving at least one organic species. Some discussion has centred around what should be regarded as the maximally allowed overlap energy for the reaction still to be classified as involving outer-sphere ET. The problem is not likely to be solved easily by any kind of experimentation, since the concept of outer-sphere ET is a theoretical construction and thus instead needs some kind of agreement as to what should actually constitute the borderline. This author, inspired by an early, excellent account of the Marcus theory (Reynolds and Lumry, 1966), has long put the outer-sphere limit at S1kcal mol-’. However, this delimitation immediately suggests that reactions with an electronic overlap energy in the transition state of 31 kcal mol-’ should be of inner-sphere type and not able to be included within the confines of the Marcus theory. The ET reduction of alkyl halides (for reviews, see SavCant, 1990; Lund et al., 1997) exemplifies this problem; it is possible to show that the transition state interactions in several cases are around 8.5 kcal mol-l, clearly indicating that such reactions are not strictly of outer-sphere ET type according to the 1kcal mol-’ limit (Eberson and Shaik, 1990). Yet such reactions obey the formalism of the Marcus theory, but with smaller reorganization energies the more the reactions are removed from ideal Marcusian behaviour (Eberson, 1987). Thus we can state with certainty that there must be a grey zone between outer- and inner-sphere behaviour which cannot easily be defined. The model of Fig. 1, in combination with the assumptions mentioned in the text, led Marcus to derive a quadratic expression for the free energy of activation AG* of the ET reaction (17), namely (20), where the “effective” standard free energy change AGO’ for the ET step is given by (21).

A G O ’

= AGO

+ (21 - Z2-1)- e’f

Dr12

SPIN TRAPPING AND ELECTRON TRANSFER

99

The first term of equation (20) describes the change in electrostatic free energy of the precursor complex as the two spheres move from an infinite distance to the distance r12in the precursor complex, e being the electronic charge, D the dielectric constant and f the usual factor describing the effect of ionic strength. The second term of equation (20) is the parabolic one, with the reorganization energy X as defined above and AGO' defined by equation (21). The "effective" standard free energy change, AGO', is composed of the standard free energy change of the ET step under the prevailing conditions, AGO, to which is added a term which describes the change in the electrostatic situation upon ET in the transition state. With r12in 181, the value of 2 = 331.2 gives the electrostatic terms expressed in kcalmol-I. As an example, if we oxidize a neutral organic compound RH ( Z 2 = 0, r2 = 2.5 A) with Fe(CN)i- (Zl = -3, rl = 3.5 181) in acetonitrile (D= 36.2) at zero ionic strength ( f = l), the electrostatic term in equation (20) becomes 0, and that of equation (21) is equal to (-3 - 0 - 1)X 331.2/(38 X 6) = -6.1 kcal mol-' or 0.26 eV. This should thus be subtracted from AGO, so lowering the free energy change and thus AG*. We can also express this situation as an increase in the effective standard potential of the Fe(CN)i-/Fe(CN):- redox couple by 0.26 V under these conditions, i.e., as long as neutral compounds are oxidized. The electrostatic terms can be reasonably well handled in solvents of high dielectric constant, but problems are raised by some solvents of widespread use in spin trapping, for example dichloromethane ( D = 8.9), chloroform ( D = 4.8) and benzene ( D = 2.3), in which the electrostatic terms calculated as above for acetonitrile become -24.8, -46 and -96 kcal mol-', respectively. Already in dichloromethane the effective standard potential of Fe(CN)i-/Fe(CN):- is increased by 1.08 V and in benzene by an absurdly high 4.2 V ! These problems aside, the Marcus equation is a useful guide for judging the feasibility of ET steps in solvents possessing D 3 10. The usefulness is increased vastly by the fact that the reorganization energy A for any D-A reaction as a good approximation can be calculated as the mean value of the reorganization energies of the contributing self-exchange reactions (22). = A(O)D.+/D +

(22)

Pictorial representations of AG* vs AGO' according to equation (20), using different values of A, are shown in Fig. 3. Conversion of AG$ to logk values with the aid of the Eyring equation gives Marcus plots of the type shown in Fig. 4. This global view shows clearly the demarcation line between reactions following Marcus behaviour and others. Reactions represented by AG' - AGO' points located well into the empty space to the right of the Marcus curve family can hardly correspond to electron transfer reactions, as exemplified by alkyl halide/Nu reactions (Eberson, 1982a, 1987; Lund et al., 1997) in Fig. 5.

L. EBERSON

100

30 20 10

0

-40

-20 0 20 AG"'/kcal mol-'

40

Fig. 3 Plots of AG* vs AGO' according to equation (20), the electrostatic term being put equal to 0, for A = ( a ) 10, ( b ) 20, ( c ) 30, ( d ) 40 and ( e ) 50 kcal mol-'.

10

log k 0

-10

-30

-10 10 AG"'/kcal rno1-l

30

Fig. 4 Family of Marcus parabolas (log (k/dm3mol-' s-') vs AGO') in the A range of 5-59 kcal mol-', drawn for steps of A of 2 kcal mol-'.

SPIN TRAPPING AND ELECTRON TRANSFER

12

101

I

log k 4

-12

-100

-60

-20

20

60

I00

AG"'kca1 mol-' Fig. 5 Marcus plot of electron transfer reductions of alkyl halides, as contrasted with archetypal SN2 substitution processes (Finkelstein reactions, circled: see Eberson, 1982a, 1987).

In conclusion, the free energy change of an ET step is already a good indicator of the feasibility of the reaction. A highly endergonic reaction, with, say, AGO' > 20 kcal mol-', corresponds to a rather slow ET reaction that is not likely to compete with other reactions of polar nature. In the region where AGO' lies between 20 and - A kcal mol-', we need to apply the Marcus approach in order to get an approximate value of the ET rate constant, whereas at AGO' < -A kcal mol-' most intermolecular ET reactions appear to be diffusion controlled.

4

Spin trapping and electron transfer

From the discussion of the Marcus theory above and equations (20) and (21), we see that the experimental data needed to judge the feasibility of ET steps involving spin traps and spin adducts are the redox potentials and A values of the ST"/ST and ST/ST'- couples, as well as those for hydroxylamine derivatives related to the operation of reactions (4)or (5). The electroactivity of the spin adducts themselves is also of interest since it must somehow be related to their lifetimes in a redox-active environment. Moreover, the excited-state redox potentials (of ST*/ST-- and ST'+/ST*) are also necessary for the understanding of photo-ET processes of spin traps. Table 1 lists anodic potentials of commonly employed spin traps, and Table

L. EBERSON

102

Table 1 Anodic peak potentials of common spin traps, determined by cyclic voltammetry in acetonitrile, unless otherwise stated.

E,,N vs SCE

ST Bu'NO, monomeric [l] (BU'NO)~ 2,4,6-(BUt)3C6HJ'J" [2] PhCH=N(O)Bu' [3] 4-PyCH=N(0)But' ~-PYOCH=N(O)BU* 4-MePyCH=N(0)Buti 4-NOzC6H4CH=N(O)BUt 4-MeOC6H,CH=N(O)Bu' CH2=N(0)Bu' DMPO [4] 3,3,5,5-Tetramethyl-1-pyrroline-l-oxide (TMPO) 1,l-(Bu')~C=CH~ [5]

Ref"

1.82, 1.80 1.45, 1.51, 1.40 1.40, 1.38 1.47, 1.53, 1.49' 1.93 1.37 2.32 1.87' 1.25'" 2.08, 1.7 1.63, 1.72, 1.68 1.78 1.60

b,d b.c,d

c.d b,c.f b b b

f f c. k b.c.2

I k

a The references follow the order of the potentials. McIntire et al., 1980. Sosonkin et al., 1982. dGronchi et al., 1983. 'In dichloromethane. 'Eberson et al., 1996a. KPy= pyridino. "PyO = 1oxidopyridino. ' 1-Methylpyridinio. Reversible potential. 'Eberson et al., 1994b. ' Eberson, 1994. J

Table 2 Cathodic peak potentials of common spin traps, determined by cyclic voltammetry in acetonitrile, unless otherwise stated. ST Bu'NO, monomeric [l] 2,4,6-(Bu')&HzNO [2] PhCH=N(O)Bu' [3] 4-PyCH=N(O)BU' 4-PyOCH=N(0)Butf 4-MePyCH=N(0)Bu' D W O [4] l,l-(Bu')zC=CHZ [5]

Ref."

E,JV vs SCE

b.c

-1.77, -1.76 - 1.25d -2.40 -1.85 -1.77 -0.88d -2.35 <-2.8

C

b b b b b h ~~

~

'The references follow the order of the potentials bMcIntire et al., 1980. 'Gronchi er al., 1983. Reversible potential. 'Py = pyridino. 'PyO = 1-oxidopyridino.8 1-Methylpyridinio. Eberson et al., 1994b.

2 the corresponding cathodic ones, in so far as they are known. Normally, the potentials given are peak potentials obtained by cyclic voltammetry, since the stability of the corresponding radical ion is not high enough to allow for the observation of the peak potential of the latter. However, it is not likely that the reversible potential will differ by more than 0.05-0.1 V from the peak potential. In Table 3 the redox potentials of excited states of spin traps are listed.

SPIN TRAPPING AND ELECTRON TRANSFER

Table 3 Excitation wavelengths A,,

ST Bu'NO, monomeric [l] 2,4,6-(But),C6H2NO [2] PhCH=N(O)Bu' [3] CH,=N(O)Bu' DMPO [4] l,l-(But)2C=CH2 [5]

103

and excited state potentials of spin traps."

E'(ST*/ST-)N

AexJnm 676 338 (746) 298 -250 242 185

Eo(STf/ST*)IV

0.12 2.4 (0.4) 1.8 3.0 2.7 3.6

0.03 -2.2 (-0.25) -2.2 -2.8 -3.4 -4.6

"Eberson QI al., 1994b.

Table 4 Reversible potentials for nitroxylhitrosonium (R2N-0/R2N=Of) couples, as determined by cyclic voltammetry in acetonitrile. Eo(R2N-0IR,N=O+)N vs SCE

Compound R2N-0 (BuL)2N-0 2,2,6,6-Tetramethylpiperidin-l-oxyl 4-Hydroxy-2,2,6,6-tetramethylpiperidin-l-oxyl 4-0~0-2,2,6,6-tetramethylpiperidin-l-oxyl Ph-PBN

0.57 0.63, 0.59 0.69, 0.64 0.80, 0.74 0.7

Ref.

" n,b 0.b 0.6 C

"Summerman and Deffner, 1975. Bobbitt and Flores, 1988. 'Bard et al., 1974.

These are of interest for judging mechanisms of photochemical spin trapping. In simple MO theory, an excited state has the electronic configuration of both a radical cation (half-filled HOMO) and a radical anion (half-filled LUMO) as a result of the excitation of an electron. Thus an excited state will exhibit powerful redox reactivity, the singlet state more so than the triplet state, acting as an oxidant towards donor molecules or as a reductant towards acceptor molecules, depending upon the experimental setup. The excited-state redox potentials are calculated from the ground-state potentials and the excitation energy AEo,oby equations (23) and (24) (Rehm and Weller, 1969).

= E"(ST+/ST)- AEo,o

(23)

E"(ST*/ST'-) = E"(ST/ST-) + AEo.0

(24)

E"(ST"IST*)

Table 4 shows redox potentials for nitroxyYnitrosonium couples, showing that nitroxyl radicals are relatively easy to oxidize.

L. EBERSON

104

KINETIC SCHEME FOR SPIN ADDUCT FORMATION

The detection by epr spectroscopy of a particular spin adduct is generally dependent on a complex kinetic scheme (Scheme l), which can be divided into three parts. Reactions giving intermediates leading to spin adducts

4

fractions of second-hours

Step leading to spin adduct formation

1

k = 106-1010 dm’ m o l - ’ ~ - ~

Reaction(s) leading to the decay of the spin adduct

1

seconds-days

Diamagnetic products Scheme 1 Time evolution and decay of spin adducts.

Thus the very act of spin trapping is only a single step among many in a complex interplay. Generally reactions of radicals X with spin traps are very rapid (Janzen and Haire, 1990), in the range of 104-10’0 dm3mol-’ s-’, which is necessary for spin trapping to be competitive with other possible pathways of radical decay. The alternative, inverted spin trapping between a spin trap radical cation and a nucleophile, should also be very fast, as for example indicated by the rate constants for reactions between aromatic radical cations and nucleophiles (Parker et al., 1989; Reitstoen and Parker, 1991; Workentin et al., 1994a,b; Wang er al., 1995). Practically no rates have been determined for the reaction between spin trap radical cations and nucleophiles, an exception being the PBN+-H,O reaction at 3.6 X lo4dm3mol-’ s-’ (Zubarev and Brede, 1994). The lifetime of a spin adduct is thus seen to be dependent on both its rate of formation and its decay, meaning that the half-life of a particular spin adduct under the conditions of its generation is a somewhat uncertain quantity, as frequently noticed in the literature. The “intrinsic” stability of a particular spin adduct should preferably be determined on isolated specimens, as for example reported for the trichloromethyl adduct of PBN, which was shown to survive almost unchanged in benzene solution after 90 days in the dark (Janzen et al., 1993).

SPIN TRAPPING AND ELECTRON TRANSFER

5

105

Evidence for the ST'+-nucleophile mechanism under thermal conditions

As already mentioned in Section 1,the radical ion-mediated mechanism, often to be denoted as inverted spin trapping in the following, was discussed and experimentally supported in several isolated cases in the period between 1975 and 1990, but was not subjected to systematic study. In addition, numerous studies were performed on systems which, in the light of later developments, must have involved inverted spin trapping, but which were interpreted differently. Two particularly interesting cases involve the formation of fluoro and acetoxy spin adducts. One of the dogmas of inorganic chemistry states that fluoride ion cannot possibly be oxidized to elemental fluorine in solution because of the very high redox potential of the F/l- couple. This excludes the possibility that the fluorine atom can be formed in any reaction in fluid solution. Treatment of PBN [3] with AgF2, a fluorinating reagent of high oxidizing power (Zweig et al., 1980) which presumably reacts by a radical cation-mediated mechanism (Burdon and Parsons, 1975),gave the bis-fluorinated nitroxide [6] in which not only a fluorine atom had formally been added to PBN, but also the a-H had been substituted by a fluorine atom as shown in reaction (25) (Janzen, 1971). With a milder reagent, F,N-0, which still has the structure of a respectable ET oxidant, the fluoro adduct [7] was formed in benzene and was slowly transformed into the difluoro compound. Since the fluorine atom cannot be an intermediate, the radical cation mechanism (26) and (27) is an attractive one for formation of both nitroxides. Ph-CH=N(O)Bu'

+ 2AgF2

+

+ 2AgF

Ph-CF,-N(O)Bu' [61

-

(25)

-e-! -H+. -e-

PBN+ + F-

-+

Ph--CH(F)-N(O)Bu' [71

[61

(27)

+F-

However, since it can be shown that fluoride ion also can participate in the addition-oxidation mechanism (p. 130), at least in media of low dielectric constant (Eberson, 1994), the latter possibility also cannot be excluded on the basis of existing information. Thus, the "fluorine dogma" makes fluoride ion a key species for mechanistic study (Eberson and Persson, 1997). Another mechanistically useful nucleophile is acetate ion and related carboxylates. Acetate ion is difficult to oxidize (Eberson, 1963) and reacts with radical cations in a bond-forming reaction (Eberson and Nyberg, 1976). The oxidation product, the acetoxyl radical, has properties which make trapping it very unlikely in that its decarboxylation rate constant is 1.3 X lo9 s-' (Hillborn

106

L. EBERSON

and Pincock, 1991; for other RCOO, see a compilation by Budac and Wan, 1992). Thus, if an acetoxy spin adduct is detected, it cannot be derived from the trapping of an acetoxyl radical. An early case of acetoxy spin adduct formation was discussed in the ozonolysis of dimethylacetylene in dichloromethane at -78°C (Pryor et al., 1982). The spin trap, PBN, was added at this temperature after the reaction had been performed, and then the solution was warmed slowly, whereupon CH,COO-PBN started to appear around -70°C. The acetoxyl radical was ruled out as an intermediate on kinetic grounds, and a version of the radical cation mechanism was proposed in which the ozone addition product [8] acted as an ET oxidant, as shown in reaction (28). Today, it would appear simpler to assume that acetate ion somehow had been formed during the reaction and reacted with PBN+. The key message was however clear: acetate ion should be a good mechanistic probe for inverted spin trapping.

OXIDATION OF PBN BY HEXACHLOROOSMATE(V) ION IN THE PRESENCE OF NUCLEOPHILES

It was the analogy between the mechanism of oxidative acetoxylation (anodic or metal ion-promoted) which provided the impetus for a more detailed study of spin trapping via the radical cation mechanism. The problems are actually identical; when a mixture of acetate ion and an organic compound (RH) in a suitable solvent is anodically oxidized or treated with a chemical oxidant, is acetate ion or RH the electron-donating species? And, in general, what is the species oxidized in any mixture of the composition RH/Nu-? Organic electrochemistry has given a firm basis for answering such questions (Eberson and Nyberg, 1976), and the answers are highly relevant for deciding about the outcome of oxidations of ST/Nu- mixtures. As a starting-point, E"(Nu'/Nu-) data are given in Table 5 for a number of systems of interest in connection with spin trapping. Data pertaining to water and a typical dipolar aprotic solvent, acetonitrile, are included, mainly to illustrate that negatively charged nucleophiles have a lower E" in a dipolar aprotic solvent. This is due to the decreased solvation ability of the latter solvent type towards anions, the effect being stronger the harder the anion is.

SPIN TRAPPING AND ELECTRON TRANSFER

107

Table 5 Standard potentials of redox couples Nu'/Nu- of interest in inverted spin trapping under oxidizing conditions." ~

NuF-

c1BrICNCNOSCN-

s0:-

N; NO; NO; HOCH3COOPhCOOCF,COO(N02)C Succinimide anions Benzotriazolate ion Benzotriazole Pyridine Triethyl phosphite

EON vs SCE in H 2 0 3.4 2.4 1.8 1.2 2.3 (2.3) 1.4 2.5 1.1 2.1 0.8 1.9 2.2 1.5 2.0 1.8 1.8-2.3

EON vs SCE in acetonitrile

2.7 1.9 1.5 1.o 2.0 (2.0) 1.3

0.8 0.7 1.3 1.6 0.9 1.4 1.7 1.3-1.8 1.1 2.0

2.2 2.1

"Eberson,1987; Carloni et al., 1996; Alberti et al., 1997.

On the other hand, in 1,1,1,3,3,3-hexafluoropropan-2-01, a solvent with even stronger ability to solvate anions than water and therefore of great usefulness for spin trapping studies (see below), E" values will presumably be higher than in water. The first application of the acetate ion probe used hexachloroosmate(V) as the oxidant (Eberson and Nilsson, 1993). The OsVC1;/Os"'C1~- system is a strong oxidant, its E" being 1.22V, and has been shown to oxidize organic compounds with redox potentials E"(RH+IRH) in the range between 1.3 and 2.1 V in an outer-sphere mechanism (Eberson and Nilsson, 1990). It was also found that 0s"CI; slowly reacted with carboxylates in an ET reaction, thus simulating the anodic Kolbe reaction (Eberson and Utley, 1983). When the reaction of Os"C1; and various RCOO- in dichloromethane was conducted in the presence of PBN to see whether the R spin adduct could be monitored, the ensuing reaction was very fast and only the acyloxyl adducts were detectable (Table 6). Evidently, OsvC1; reacts preferentially with PBN (E" = 1.5 V) in the presence of a carboxylate (for example, E" of acetate ion is 1.6 V). Other charged nucleophiles (trinitromethanide, tetramethylsuccinimide anion, benzotriazolate ion) behaved similarly, forming Nu-PBN

L. EBERSON

108

Table 6 Acyloxyl radical adducts formed in the oxidation of tetrabutylammonium carboxylates and PBN [3] by PbP0svCl6 in dichloromethane at 20°C."

Carboxylate source (Ac0)ZHAcOAcOH Bu'CH(CN)COO (Bu'COO),HPhCOOCF3COO-

Spin adduct from

aN/mT

aH/mT

1.36 1.33 1.36 0.81 1.34 1.36 1.35 0.81 1.33 0.81

0.16 0.14 0.17

CH3COO CHSCOO CH3COO

0.14 0.14 0.15

Bu'CH(CN)COO Bu'COO PhCOO

O*'

6'0

0.14

CF3CO0 O'b

~~

~

"Eberson and Nilsson (1990). bThe benzoyl t-butyl nitroxide [9], often detected in spin trapping experiments under oxidizing conditions.

whereas cyanide ion did not give any spin adduct, presumably because of the known instability of the cyano spin adduct (Eberson ef al., 1997). Neutral nucleophiles (benzotriazole, lH-1,2,3-triazole, 1H-tetrazole, triethyI phosphite) gave spin adducts, while 3,5-lutidine did not give any. 4-Nitro-PBN, a more oxidation-resistant spin trap (E" = 1.9 V), gave spin adducts upon oxidation by Os"C16, with trinitromethanide, benzotriazole and triethyl phosphite; with acetate ion, the nitro analogue of benzoyl N-t-butylnitroxide [9] was detected, most probably with the acetoxy adduct as the precursor in an EC-type mechanism (reaction (29)). As a nitroxyl radical, the acetoxy adduct should be easily oxidized (Table 4). Ph-CH(OAc)N(O)Bu'

-e-,

-n+.-e-

Ph-C(OAc),N(O)Bu'

~

Ph-C+(OAc)N(O)Bu'

--*

Ph-CO-N(0)Bu'

-

AcO-

(29)

[91

The competition between the Os"C1; reaction with a neutral compound (A) and a negatively charged one (A-), respectively, which is the experimental situation in some of the spin trapping reactions mentioned above, was analysed by the Marcus treatment for some model cases in dichloromethane or acetonitrile. These data are shown in Table 7, giving the details of the calculations in order to illustrate the use of equations (20) and (21) and the importance of the electrostatic factors, particularly in dichloromethane. The assumptions behind the calculations are given in the table heading and footnotes.

Table 7 Calculated rate constants of outer-sphere ET model reactions of OsC&/A and OsC&/A- type for substrates of three representative standard potentials, 1.5, 1.9 and 2.2 V."

Substrate

EON

AG"lkca1 mol-'

ES term in AG"'kca1 mol-'

AG"'kca1 mol-'

AGfkcal mol-'

AG:/kcal mol-le

AG*kcal mol-' f

log (kET/dm3 mol-' s-l)g

In CHzClz ( D = 8.9) A AA AA A-

1.5 1.5 1.9 1.9 2.2 2.2

6.9 6.9 16.1 16.1 23.1 23.1

- 12.2

-6.1 - 12.2 -6.1 - 12.2 -6.1

-5.3 0.8 3.9 10.0 10.9 17.0

3.9 6.7 8.3 12.2 12.9 17.6

0 6.1 0 6.1 0 6.1

3.9 12.8 8.3 18.3 12.9 23.7

8.1 1.6 5.1 -2.5 1S -6.4

6.9 6.9 16.1 16.1 23.1 23.1

-3.0 -1.5 -3.0 -1.5 -3.0 -1.5

3.9 5.4 13.1 14.6 20.1 21.6

8.4 9.2 14.5 15.7 20.3 21.7

0 1.5 0 1.5 0 1.5

8.4 10.7 14.5 17.2 20.3 23.2

4.8 3.1 0.3 -1.6 -3.9 -6.1

In CH&N (D = 36.2) A AA AA A-

1.5 1.5 1.9 1.9 2.2 2.2

'E"(OsQ-/OsC~-) was taken to be 1.20V, rI2 to be 6.1 8,and A to be 25 kcal mol-' (Eberson and Nilsson, 1993). Electrostatic term; second term of equation (21). 'Equation (21). dThe second, quadratic term of equation (20). 'Fist term of equation (20). fEquation (20). gEqual to 11 - AGt/(2.303RT).

L. EBERSON

110

To see how the data can be used to provide insights into the spin trapping process, PBN would correspond to A with E" = 1.5 V, and acetate ion to Awith E" = 1.5 V (Table 5 gives 1.6 V in acetonitrile, and 1.5 V is therefore somewhat too low, but then it is presumably adequate for dichloromethane). In dichloromethane, the OsVC1;-PBN reaction is estimated to be very fast, more than 6 powers of ten faster than the OsvC1;-acetate ion reaction, whereas in acetonitrile the absolute rates are still high but the ratio is only about 50. This difference resides only in the difference between electrostatic factors and illustrates the problems of understanding ET reactions in solvents of even lower dielectric constant such as benzene. The rate ratio between reactions of the two neutral compounds, PBN and triethyl phosphite (E" = 2.1 V), can also be estimated from Table 7. It should be between lo5 and lo6, indicating a very high preference for PBN oxidation.

OXIDATION OF PBN BY TRIS(CBROM0PHENYL)AMINIUM ION (TBPA') ([lo]'+) IN THE PRESENCE OF NUCLEOPHILES

TBPA+ is a commonly used ET oxidant. It also possesses electrophilic reactivity, reacting with bond formation to acetate, cyanide and chloride ion, all of which have high E" values (Table 5 ) , to give 2-substituted derivatives of [lo] (Eberson and Larsson, 1986,1987). Nucleophiles with lower E" (Br-, I-,

[lo] TBPA possibly trinitromethanide ion) react with TBPA" according to an ET mechanism. Thus the acetate probe is again of critical usefulness; it is probably unavoidable that it will react with TBPA" ( k = 6 X lo4dm3mol-' s-'), but it does not give the acetoxyl radical in this process. Its redox potential, E"(TBPA'+/TBPA) = 1.06V, makes the ET reaction with PBN relatively fast [calculated rate constant (equation (20)) for outer-sphere ET, lo3 dm3mol-' s-', using A = 20 kcal molp'], an important requirement in view of the rather short half-lives of acyloxyl spin adducts. Indeed, carboxylates and a host of other nucleophiles gave spin adducts when oxidized in dichloromethane by TBPA' together with PBN (Eberson, 1992). These results are shown in Table 8, with the nucleophiles listed in order of decreasing E"(Nu'/Nu-) (Table 5). Fluoride ion gave rise to both [6] and [7],

SPIN TRAPPING AND ELECTRON TRANSFER

111

Table 8 Spin adducts formed in the oxidation of PBN-Nu- by TBPA+ in dichloromethane, unless otherwise stated.a Nucleophile

FCNCN- in acetonitrile CNOCI Pyridine 3,s-Dimethylpyridine 2,6-Dimethylpyridine PhSO; NO; Succinimide anion Me,-succinimide anion Me,-succinimide anion in acetonitrile Triethyl phosphite CF3COOPhCOOAcO(AcO),H(ButC00)2HBrSCN(N02)SC(NO2)&- in acetonitrile N;

E"(Nu'INu-)N vs SCE 2.7 2.0 2.0 2.0 1.9 2.2 2.2 2.2 1.8 >1.6 >1.6 >1.6 2.1 1.4 0.9 1.6

1.5 1.3 1.7 1.7 0.8

Spin adduct from

F [71; F~ t61 -CN -CN -NCO =O l-Py+ 3,s-Dimethyl-l-Pyc 2,6-Dimethyl-l-Pyf =O =O N-Succinimido Me,-N-succinimido Me,-N-succinimido (EtO)2(OP=O PhCOO CHSCOO CH3COO t-BuCOO No adduct seen NCS(N02)3C (N02)sC N3

"Eberson (1992).

as found earlier for other oxidants [equations (25)-(27)]. The formation of spin adducts was not found to be related to E"(Nu7") in any obvious way, but they were obtained over the whole range between 3.4 and 1.1V. This indicates the nature of the mechanistic problem: at which E"(Nu'/Nu-) is there a switch between inverted and true spin trapping and what criteria should be used to distinguish between these situations? The Marcus theory can be of some help for a first sorting procedure, but experimental criteria are needed. A few nucleophiles either did not give any spin adduct with PBN or directly gave the benzoyl nitrone [9]. Bromide ion did not give any spin adduct, explicable by the very short lifetime of Br-PBN (Rehorek and Janzen, 1984) and trifluoroacetate, nitrate, phenylsulfinate and chloride ion produced [9]. This can either be explained by the rapid further oxidation of the spin adduct formed [similar to reaction (29); see Table 41 or a rapid solvolysis reaction of the latter (Scheme 2), forming [9] by reaction of the intermediate carbocation

112

xI

0-

Ph-CH-q But

-X

L. EBERSON

+

Ph-CH-N\

/0O '

But

*

OH I

Ph-CH-N\

0.

/

Bu'

ox

[9]

X =good taving group

Scheme 2

with adventitious water and further oxidation of the hydroxyl spin adduct. This type of substitution mechanism was demonstrated for spin adducts where the trapped radical corresponded to a good leaving group (Rehorek et al., 1984; Davies el al., 1992). DMPO is more difficult to oxidize than PBN by about 0.2 V (Table 1) and is therefore expected to engage in spin trapping via its radical cation with greater difficulty, as found for the OsC1;-4-N02-PBN reaction. Only acetate ion, tetramethylsuccinimide ion and triethyl phosphite gave the corresponding adducts upon oxidation with TBPA+ in dichloromethane in the presence of DMPO, whereas fluoride ion gave the hydroxyl adduct. The latter was probably formed from water available from the unavoidable hydration shell around fluoride ion in its tetrabutylammonium salt. 3,3,5J-Tetramethyl-l-pyrroline-l-oxide ([ll];TMPO) underwent inverted spin trapping but only with one nucleophile, triethyl phosphite. This is expected in view of the even lower redox reactivity of TMPO, E" = 1.8 V.

[ll]TMPO

As mentioned above, criteria for distinguishing inverted from true spin trapping are urgently needed. A promising approach for this purpose utilized the truly astounding capacity of 1,1,1,3,3,3-hexafluoropropan-2-o1(HFP) to suppress the reactivity of nucleophiles towards cationic species, especially radical cations. A review of the solvent properties of HFP (Eberson et al., 1996b) concludes that its hydrogen-bonding capacity to anions or negatively charged centres is probably the main cause of this effect. To illustrate the rate attenuations possible, Table 9 lists rate constants for the model reaction between TBPA" and nucleophiles in HFP and acetonitrile; rate attenuations sometimes amount to lo9 for the hard nucleophiles. In absolute rate terms,

SPIN TRAPPING AND ELECTRON TRANSFER

113

Table 9 Rate constants for the reaction between TBPA+ and Nu- (as Bu4N+ salts in the appropriate cases) in HFP at 20"C, compared to acetonitrile." ~

~~

log (k/dm3mol-' Nucleophile

HFP

c1-

-414 -3.5 -4.1 -3.1 -3.4 -2.7 >3 -2.8 -0.2

Br(AcO); (NOd3CPyridine 3J-Dimethylpyridine IBenzotriazolate ion Triethyl phosphite

s-l)

in

Acetonitrile

Difference

2.1 4.5 4.8 0.4

7.1 8.0 8.9 3.5

-

-

"Ebersonet at. (1996a).

Table 10 Inhibitory effect of HFP on the formation of spin adducts from the reaction of PBN, Nu- and TBPA' in dichloromethane." ~

Nucleophile (AcO),H-

c1-

3,SDimethylpyridine Triethyl phosphite ( N W C Benzotriazolate ion Tetramethylsuccinimide ion

-~

Spin adduct in neat HFP None None None (EtO),P+ None None None

-~

-

Percentage HFP above which no spin adduct is formed -

40 10 8.5

2

"Ebersoner at. (1996a).

among the species studied, only triethyl phosphite retains any reasonable nucleophilic reactivity in neat HFP. Iodide ion was not affected much, presumably because it undergoes ET to TBPA+ instead of nucleophilic interaction. 2,2,2-Trifluoroethanol has similar properties, although not as drastic as HFP, and has been used to influence the competition between nucleophilic and ET mechanisms (Workentin et al., 1994a,b). The oxidation of mixtures of PBN and Nu- by TBPA'+ in HFP gave no spin adducts from the commonly used nucleophiles (Table lo), except in the case of triethyl phosphite and related phosphorus compounds (Eberson et al., 1996a). Thus any PBN' formed must react so slowly with Nu- that the spin adduct concentration is too low to be detectable. "Titration" of the percentage HFP in dichloromethane which just barely allowed for the formation of the

L. EBERSON

114

appropriate spin adduct in some selected case showed that a relatively low percentage of E-lFP suffices to suppress formation of the spin adduct (Table 10). Thus the behaviour of the PBN-Nu--TBPA+ reaction in HFP and HFP-dichloromethane additionally supports the radical cation-nucleophile mechanism, although in a negative and not entirely conclusive way. 6 Properties of the PBN and DMPO radical cations

Recently, the radical cation of PBN has been characterized by matrix spectroscopy and its reactivity has been studied by fast spectroscopic methods (Zubarev and Brede, 1994), and found to conform to the behaviour deduced from the OsCC and TBPA+ studies. y-Radiolysis of PBN in a glassy matrix of isobutyl chloride or Freon-113 (CF2C1CFCl2)at 77 K produced an intensely green glass containing PBN+, the epr spectrum of which had an anisotropic nitrogen coupling constant All = 2.75 mT and gll= 2.0037. The mechanism of the radiolysis reaction is well established (Neta, 1976) and involves the formation of solvated electrons (e-), which add to the matrix species and produce chloride ion, and positive holes (h+)which eventually come to rest at the matrix component of lowest 1, (Symons, 1997), in this case PBN (see reactions (30) and (31)). e-

+ R-C1

+

R

+ C1-

h+ + PBN -P PBN+ PhCH+-N(O)Bu'

+ C1-

-+

PhCH(C1)-N(0)Bu'

(30) (31) (32)

[I21

Upon slow warming of the matrix, the colour disappeared and a new species with All = 3.24 mT and gll = 2.0038 appeared, assigned to the formation of the chloro spin adduct [12] (32); after melting of the matrix at 240K the characteristic solution epr spectrum of [12] was recorded. By y-radiolysis of the isomeric oxirane [13], which cannot sustain spin trapping, another way of direct matrix generation of PBN+ was available and thus made possible further confirmation of these results (Zubarev and Brede, 1995).

SPIN TRAPPING AND ELECTRON TRANSFER

115

Fast spectroscopy was also used to probe the reactivity of PBN+. The 266nm laser excitation of peroxydisulfate ion in aqueous solution at room temperature gives the powerful oxidant SOL, which oxidizes PBN in an exergonic reaction (by about 0.8eV, see Tables 1 and 5 ) with k = 3 X lo9 dm3mol-' s-'. The pseudo-first-order rate constant for the decay of PBN' by reaction with water to give HO-PBN was 2 X lo6s-', a relatively slow reaction (k = 3.6 X lo4dm3mol-' s-' at ambient temperature). Laser excitation of chloranil ([14], tetrachlorobenzoquinone) at 355 nm in acetonitrile-5% water gives its triplet state, 3[14]*, which is a strong oxidant (E0('[14]*/[14]'-) = 2.3 V) and will oxidize PBN to PBN+ with k = 8 X lo9 dm3mol-'s-'. In this medium the water reaction of PBN+ had k = 1.5 X lo5dm3mol-' s-l. There was also indication of a second decay pathway of PBN+ [equation (33)], involving attack upon a second molecule of PBN (k = 6 X lo8dm' mol-' s-'), and this was suggested to decompose rapidly into the benzoyl nitroxide [9] and the imine [15], both known products of PBN photolysis.

PEN + PEN'+

-

191 +

A study of the DMPO radical cation showed the same features; the radical cation is formed at 77 K in a CFC13 matrix and reacts with added traces of water to give HO-DMPO upon warming to 270K (Chandra and Symons, 1986). These studies show that PBN' and DMPO' are reactive radical cations possessing high electrophilic activity at the a-position. They have actually been classified as a-aminoxylcarbenium ions, for PBNf written as [16],

L. EBERSON

116

although this structure does not seem to have any of the expected appreciable delocalization of charge into the benzene ring. The similarity between the findings on PBN' and on DMPOf by matrix spectroscopy indicates that these radical cations will behave analogously in solution, the latter being the more reactive in view of its 0.2 V higher redox potential (Table 1). 7 Anodic spin trapping experiments The electrochemistry of RH-Nu- systems is well established (Eberson and Nyberg, 1976; Eberson et al., 1991; Childs et al., 1991). The radical cation mechanism has been shown to prevail for most situations where Nu- = F-, Cl-, RCOO-, OCN-, CN-, NO;, Py and triethyl phosphite, all of them nucleophiles that are difficult to oxidize (Table 5). The initial formation of Nu' is indicated for the redox-reactive SCN-, N;, I- and NO;, with Br- and (NO&C- occupying a somewhat indeterminate position. Table 11summarizes results of spin trapping experiments where PBN-Nuand other ST-Nu- systems have been oxidized anodically at platinum. Originally, all the reactions were suggested to proceed via Nu' radicals (Janzen et al., 1980; Walter et al., 1982), but the fact that PBN is oxidized at a lower potential than C1-, CNO- and CN- (Tables 1 and 5 ) clearly shows that the faster electrochemical process must be PBN PBN' at the potentials employed. On the other hand, azide ion is oxidized in a faster reaction than any of the spin traps used and thus azide radical is implicated as being trapped. The C1-4MePyPBN [17] system is a case where possibly C1' is involved in view of the high Epaof this spin trap. Tetrabutylboride ion, Bu4B-, is oxidized anodically with EPa= 0.35 V and thus combines low nucleophilicity with high redox reactivity. Electrolysis of its

-

Table 11 Anodic oxidation of ST-Bu4NNu in acetonitrile with tetraethylammonium perchlorate as supporting electrolyte."

Nu-

ST

Anode potentiaW vs SCE

c1-

PBN 4MePyBN PBN PBN PBN 4PyOBN 2Py0BNb 2SPBN'

0.95 0.90 0.90 0.85 0.80 0.80 0.80 0.80

c1CNO-

Spin adduct formed

Probable mechanism

Cl-PBN, [lo] C14MePyBN OCN-PBN NC-PBN N3-PBN N3-PBN N3-PBK N3-PBN

via PBN+ via C1' via PBN+ via PBN+ via N; via N3 via N; via N;

"Janzenet al. (1980);Walter et al. (1982). *Epa= 1.63 V. '2-Sulfonate of PBN, Epa= 1.34 V.

SPIN TRAPPING AND ELECTRON TRANSFER

117

H

[17] 4MePyBN

tetrabutylammonium salt together with PBN in cetonitrile at an a de potential of 0.4 V gave Bu-PBN under rigorous deoxygenation conditions and also the B u O adduct with oxygen present, showing that Bu' (the redox potential of Bu+/Bu' is estimated to be of the order of 1.9 V; Eberson, 1963) must be an intermediate in the decomposition of the oxidized boride species [reaction (34)] (Bancroft et al., 1979). Bu4B--e-

-*

Bu,B

+

PBN+ + Bu,B-

Bu3B + Bum- Bu-PBN

-*

B u ~ B+ BU-PBN

(34)

(35)

In principle, this reaction is a good model for the design of a proper spin trapping situation in an oxidative system (see Section 16). The radical to be trapped is formed from the initially reacting species in a secondary reaction, and the outcome of this reaction is not of a type that is likely to result from PBN+ in a single step (reaction (35)) even if there were a chance that PBN+ would be formed. The low anode potential additionally refutes the latter possibility. Somewhat surprisingly, no spin adduct was seen from the oxidation of Ph4B- (Epa= 0.92 V) under similar conditions, the anode potential being varied between 0.5 and 2.2 V. Since Ph-PBN could be independently formed in a thermal reaction and was stable under the anodic conditions used, and Ph' was judged to be electroinactive, it was concluded that P b B decomposed intramolecularly with direct formation of biphenyl. An interesting version of electrochemical oxidation is available in the photocatalytic oxidation of organic materials on semiconductor surfaces, for example on TiOa or CdS (for a review, see Fox, 1991). When light of a suitable wavelength is allowed to impinge on such a surface, electrons in the valence band are excited to the conduction band, and a potential difference, equal to the band gap, is set up between the two levels. The holes in the valence band will thus be capable of extracting electrons from an external substrate. TiOz, with a band gap of 3.2 eV, was successfully used for the photooxidation of acetate ion in acetic acid, a photochemical version of the Kolbe reaction (Kraeutler et af.,1978). The main products formed were methane and carbon dioxide, in addition to small amounts of ethane. The latter is the major product

L. EBERSON

118

in the anodic Kolbe reaction (for a review, see Eberson and Utley, 1983). In the presence of PBN, the methyl spin adduct was detectable, formed by oxidation of acetate ion at the illuminated TiOz surface. Again the experimental set-up, with the radical formed in a secondary reaction after initial ET, leaves little doubt that trapping of a radical has occurred, but now under conditions where both acetate ion and PBN are oxidizable. But why does not the concurrent PBN+-acetate ion reaction give CH3COO-PBN? This question can have several reasonable answers, one being the relative instability of CH3COO-PBN in comparison to CH3-PBN, making detection impossible on the time scale of the experiment. The second is that CH3COO-PBN is actually formed, as indicated by the shape of the published epr spectrum of CH3-PBN, but could only be defined with uncertainty by aN= 1.43 and aH< 0.3 mT owing to severely overlapping spectral lines. It would indeed be surprising if the acetoxyl adduct were not formed, and it may be that now available information on the properties of these spin adducts might help in the design of experiments to give more conclusive results on this point. Another study with TiOz as catalyst involved the photodecomposition of water in the presence of PBN or 4PyOBN ([MI; Jaeger and Bard, 1979). The results were complicated and difficult to interpret, but it was concluded that H O was formed in the photooxidation process and trapped as such. However, in view of the short half-life of HO-PBN established later ( T ~=, ~10 s in acetonitrile and 90s in water, see Section 15), the reported control experiments are not entirely convincing, CdS, with a band gap of 2.5 eV, has been used for the photooxidation of azide ion in the presence of PBN, resulting in a strong N3-PBN signal from proper trapping of N3(Amadelli et al., 1989).

H

[18] 4PyOBN 8 Photochemicalspin trapping experiments

Photochemical spin trapping experiments are the stock in trade, and the most difficult ones to judge with respect to mechanism because of their high complexity. The method became popular at a time when the effect of light upon molecules was believed to result mainly in homolysis of bonds, principally because of its ready use in combination with epr spectroscopy and

SPIN TRAPPING AND ELECTRON TRANSFER

119

its ability to sustain detectable steady-state concentrations of transient radicals. With the discovery of the widespread occurrence of photo-ET reactions (for two series of reviews, see Fox and Chanon, 1988; Mattay, 1989, 1990,1991,1992,1993), the situation has changed drastically. Excited states are usually strong redox reagents (see Table 3 for properties of spin traps in this respect) and will undergo ET with donor and/or acceptor species present, often resulting in the formation of radical ions which might decompose to give new radical species. Photochemical E T reactions can be classified in at least three categories (which can co-exist), namely (i) simple homolysis of bonds of neutral molecules to give radicals of low redox reactivity; (ii) excitation of a species D to produce an excited state D* which initiates a second-order ET reaction involving another component of acceptor type, A, with formation of the radical pair D f A - ; (iii) direct excitation of a charge transfer (CT) complex formed between two reaction components D and A to form the same radical pair D CA - . The first case is obviously an ideal situation if it can be realized, but this is seldom the case. The incursion or predominance of situations (ii) and/or (iii) in almost any system is possible, and precautions must be taken to avoid these complications. Much can be done by controlling the wavelength of the light source, but it is also possible to affect the chemistry in a predictable manner. The radical pair D'A- is one point where some corrective (or, if so desired, affirmative) action can be taken. If neither D" nor A - has a fast second-order pathway available for further reaction, back electron transfer to give D and A becomes the only reaction, as for example in the photoexcitation of the anthracene-tetracyanoethylene CT complex (Hilinski et al., 1984). Thus, no chemical consequence of the excitation process will be noticeable. Changing to an acceptor corresponding to a highly labile radical anion, such as in the anthracene-tetranitromethane CT complex, will almost completely eliminate back ET, since the lifetime of (NO2)&- is less than 3ps, and the radical cation, trinitromethanide ion and NO2 appear as a triad of reactive species [reaction (36)] (Kochi, 1988, 1990). The first reaction from the triad is the reaction between the radical cation and trinitromethanide ion. If in addition a protic acid is present, the trinitromethanide ion is eliminated by protonation and the slower reaction with NO2 will predominate (Eberson et al., 1993). Quinone acceptors behave in a similar way, using protonation of the radical anion to prevent back ET, as utilized, for example, in the oxidizing system 2,3-dichloro-5,6-dicyanoquinone(DDQ)-trifluoroacetic acid, with or without light, which is used for the generation of radical cations [reaction (37)] (Handoo and Gadru, 1986; Davies and Ng, 1995).

L. EBERSON

120

Q

Some examples of photochemical reactions which were deliberately set up to favour inverted spin trapping gave results which entirely supported the reasoning above (Eberson, 1994). A first approach involved the excitation of PBN with light of A 300 rim (Table 3), which would give the strong reductant PBN*. A weak outer-sphere ET oxidant, the tetrabutylammonium salt of 12-tungstocobalt(III)ate ( ( B U ~ N ) ~ C O ~ to ~ ~be Wdenoted ~ ~ O ~ Co"'W ~, (Eberson, 1983; Baciocchi et af., 1992, 1993, 1996; Bietti et af., 1996) with E"(CO"'W/CO'~W)about -0.1 V in acetonitrile) was present, together with a nucleophile. The expected working of this scheme is shown in reactions (38) and (39). It should be noted that the PBN+Co"W ion pair formed after ET must undergo predominantly back ET, since only the intermolecular PBN+Nu- reaction will lead to chemical change.

-

-

hv, A=300nrn

PBN

PBN* + Co"'W

PBN*

PBN+ + CoUW 1 NuNU-PBN

+

(39)

This scheme, set up with reactions in dichloromethane, gave spin adducts from several of the nucleophiles discussed above (F-, C1-, AcO-, CN-, tetramethylsuccinimide anion and triethyl phosphite), provided UV light was employed. With filtered light of A > 435 nm, no spin adducts were detected. This is expected, since PBN cannot then be excited. With water as the nucleophile, only benzoyl nitroxide [9] was seen, indicating that any HO-PBN disappears too rapidly to be detectable ( T ] , ~ 10 s in acetonitrile) and/or that its rate of formation from PBN+Co"W is too low (see above). The complication that nucleophilic addition-oxidation might compete was ruled out experimentally in dichloromethane, but detected for fluoride ion in chloroform, using dioxygen to oxidize the intermediate hydroxylamine anion. DMPO did not produce spin adducts in the Co"'W scheme, except with tetramethylsuccinimide ion. This was possibly due to the fact that the A,, of DMPO, at 242 nm, lies somewhat outside the lower limit of the wavelength region of the lamp used. A second scheme involved the use of a sensitizer, 2,4,6-tris(4methoxypheny1)pyrylium ion (E"(A*/A-) = 1.8V, A,, = 422 nm), to

-

SPIN TRAPPING AND ELECTRON TRANSFER

121

produce an excited state which would oxidize PBN in an ET reaction. Again, back-ET should be a significant competing reaction. Some of the nucleophiles were not chemically compatible with this sensitizer, but both tetramethylsuccinimide ion and 3,5dimethylpyridine gave spin adducts with PBN, using light of wavelength > 400 nm which could only excite the sensitizer. 9

Examples of problems in photo-initiated spin trapping

From the above, it is evident that every photochemical system must be carefully analysed in order to establish the nature of the process of spin adduct formation. Not all systems have the inbuilt diagnostic features of the fluoride or carboxylate nucleophiles, and it must therefore be accepted that mechanistic certainty will be difficult to attain. It also must be remembered that many studies in the past were designed without regard to the inverted spin trapping mechanism and are difficult to judge owing to lack of critical experiments to test this particular aspect.

FORMATION OF CHLORO SPIN ADDUCTS

One way of producing the chloro adduct of PBN [12] is to photolyse hexachloroethane with PBN present by UV light in a suitable solvent, such as acetonitrile (Rehorek et al., 1991). This leads to the homolytic cleavage of C-Cl bonds to give from each molecule of hexachloroethane two chlorine atoms [reaction (40)]. The light employed is filtered, the cutoff being 310 nm. This avoids to a large extent direct excitation of PBN (A,,, = 295 nm) which would introduce chloride ion via reaction (41) (the appropriate redox potential of GC16is not known, but is estimated to be similar to that of CC14, -0.8 V in acetonitrile; Ekstrom, 1988) and thus very rapidly produce Cl;- in the equilibrium of reaction (42). This species is not trapped by PBN.

-

h v , h=300nrn

PBN

c1- + C1'

-

Gc'6

PBN*

C&14+C1-+C1'

(41)

2 X I01"drn3rnol~1 sC'

c1Z'-

1 . 1 ~ 1 0 ~ ~

The kinetic scheme of reaction (40), followed by reaction (42), was used to investigate the influence of added chloride ion on the epr spectral intensity of Cl-PBN. This variable decreased by a factor of about 50 in going from

L. EBERSON

122

[Ph,PCl] = 0 to 50 mmol dm-3, indicating that [CY] is depleted by some mechanism, most probably by being tied up with chloride ion according to reaction (42). A quantitative analysis of the kinetics gave the rate constant of the spin trapping reaction as 1.2 X 10' dm3mol-' s-'. This study demonstrates that trapping of chlorine atoms from the photolysis of hexachloroethane is feasible, but it would be interesting to see what would happen with light capable of exciting PBN. Equally clear-cut cases of inverted spin trapping of chloride ion can be demonstrated during the UV photolysis of N-chlorosuccinimide [19]-C1 with PBN (Eberson et al., 1994a). The formation of Cl-PBN precedes that of the succinimidyl adduct ([19]-PBN, see below) in acetonitrile or dichloromethane, but is the only spin adduct seen in benzene where it is protected from the imidyl reaction products ([19]' is trapped by benzene). This scheme is outlined in reactions (43)-(45); the cleavage mode of [19]-Cr- to give chloride ion and [19]' was established by pulse radiolysis experiments (Lind er at., 1993). This reaction is fast, k = 8 X loss-', meaning that [19]-C1 is an acceptor which will induce chemical changes (see above). Its redox potential, Eo([19]-C1/[19]-CY-)= 0.1 V, makes it a good electron acceptor.

-

hv,A=300nrn

PBN

PBN*

PBN* + [19]-Cl+ PBN+ + [19]-Cr- + C1-

(43)

+ [19]'

PBN+ + C1- -+ Cl-PBN

(44) (45)

I

CI

[19]-C1 In agreement with this scheme, no chloro spin adduct was obtained when HFP was used as the solvent (Eberson et al., 1996a), as expected in view of the vast reactivity decrease of chloride ion in HFP (Table 9).

FORMATION OF CYAN0 AND THIOCYANO SPIN ADDUCTS

The electrochemical behaviour of PBN+-cyanide ion is identical to that found in the two cases of inverted spin trapping described above, namely that attack at PBN+ occurs via the softer carbon atom of CN-. This contrasts with an observation of the cyano adduct to PBN formed by irradiation of Mo(CN)i-

123

SPIN TRAPPING AND ELECTRON TRANSFER

-

in methanol-dichloromethane at -7O"C, where the cyano group becomes attached via nitrogen (Rehorek el al., 1979; see also Rehorek and Janzen, 1985). This might indicate differing selectivity in PBN+-CN- and PBN-CN reactant pairs, and needs further exploration. Similarly, the TBPA' oxidation of PBN in the presence of thiocyanate ion gives a thiocyanato adduct connected via the sulfur atom (Eberson, 1992). The redox potential of the SCN/SCN- couple in relation to the other reactants is such that kinetic arguments for an SCN-mediated mechanism would prevail, but this is not altogether sure. It is therefore of interest that the photolysis (light of A > 330 nm) of chloranil[14] and tetrabutylammonium thiocyanate in the presence of PBN in acetonitrile gives a spin adduct of a nitrogen-centred radical (Rehorek and Janzen, 1986). This was assigned to trapping of (SCNL-, but a more likely explanation is that trapping of the thiocyanate moiety occurs via nitrogen to give S=C=N-PBN. In such cases the explanation for the differing results might be the high redox potential of [14]* (2.3V) causing diffusion-controlled, indiscriminate ET oxidation of both SCN- and PBN and thus setting up conditions for both spin trapping mechanisms: the stability of S=C=N-PBN, formed by reaction of PBN' and SCN-, however, is such that it will predominate over NCS-PBN.

FORMATION OF THE TRINITROMETHYL SPIN ADDUCT

The formation of the trinitromethyl adduct of PBN by photolysis of PBN and tetranitromethane (Okhlobystina et al., 1975) is an unequivocal case of inverted spin trapping. These components give an orange-red CT complex in, for example, dichloromethane; when this solution is irradiated by light which only can excite the CT complex (A > 430 nm) the spin adduct (N02)3C-PBN is formed via reaction (46) (Eberson et al., 1994b). This adduct is highly persistent. When the solution is acidified by =2% trifluoroacetic acid, irradiation does not lead to spin adduct formation owing to protonation of trinitromethanide ion. C(NOz)d...PBN

+

(N02)3C-

+ NO2 + PBN+

(N02)3C-PBN

4

(46)

In HFP, where trinitromethanide is 3000 times less reactive than in acetonitrile, a weak signal of (NO&C-PBN is still obtained by the photolysis procedure (Eberson et al., 1996a). Evidently, some nucleophilic reactivity is retained by trinitromethanide ion in HFI?

FORMATION OF IMIDYL SPIN ADDUCTS

The trapping of the succinimidyl radical and its congeners is a classical problem (Lagercrantz and Forshult, 1969; Chalfont and Perkins, 1970;

L. EBERSON

124

Lagercrantz, 1971; Kaushal and Roberts, 1989). The photolysis of N haloimides ( I d , X = Br, C1) with MNP gives imidyl spin adducts, Im-MNP, which have been shown to originate from the excitation of MNP (A = 676 nm), suggesting an inverted spin trapping mechanism (Eberson et al., 1994a), combined with the nucleophilic displacement of chlorine by imidyl for Im-Cl cases (the cleavage mode of 1mCI'- is to give Im' and Cl-) and possibly also true trapping of Im' [reactions (47) and (48)]. To compound the mechanistic difficulties, a slow dark reaction giving Im-MNP' can also be monitored.

-

-, MNP*

hu

MNP

-

ImBr

hv

MNP

-

MNPf

ImCl

MNP*

+ Im- + Br'

MNPf

+ Im' + C1-

-

Im-MNP

+

-

Irn-MNP

(47)

Cl-MW (not seen)

In-

For PBN, photolysis with ImCl in dichloromethane or acetonitrile initially gives the chloro adduct (see above) which after a short time is replaced by the imidyl adduct, presumably via the same mechanism as given for MNP in reaction (48). Imidyl spin adducts were also formed from PBN-ImBr photolysis, analogously to reaction (47). 1,l-Di-t-butylethylene [5] has one unique property in comparison with other spin traps, in that it cannot be excited under normal photolysis conditions (A,, = 185 nm). Imidyl-[5]' are formed by UV photolysis of ImX-[5] solutions, and it is obvious that the mechanism must involve excitation of ImX to ImX* around 205 nm, followed by either homolytic cleavage of the excited state or oxidation of [5] (Table 3) [E"(ImX*/ImX-) is estimated to be very high, 6 V] to give [5]'+ and create conditions for inverted spin trapping. Conditions for favouring the observation of the homolytic cleavage of ImX should be ideal in HFP: the nucleophilicity of imide anions should be strongly suppressed (cf. Table 9) and, moreover, the pK, of HFP (9.3) ensures that the imide (pK, 9.5-11) exists in the protonated form. Thus both nucleophilic addition-oxidation and inverted spin trapping should be suppressed in HFI? Yet imidyl spin adducts from MNP and PBN can be obtained by photolysis by UV light, providing unambiguous cases of imidyl spin trapping (Eberson et al., 1996a).

TRAPPING OF AROYLOXYL RADICALS

As noted above, the rate of decarboxylation of the acetoxyl radical (k = 1.3 X lo9s-') is too high for spin trapping to be feasible. The rate of decarboxylation of the benzoyloxyl radical is -lo3 times slower,

125

SPIN TRAPPING AND ELECTRON TRANSFER

k = 2 X lo6 s-', and thus spin trapping would be competitive. The decarboxylation of 4-substituted ArCOO' is even slower, k = 104-105s-l (Budac and Wan, 1992). The generation of the benzoyloxyl radical relies on the thermal or photoinitiated decomposition [reaction (49)] of dibenzoyl peroxide (DBPO). An early study (Janzen et al., 1972) showed that the kinetics of the thermal reaction between DBPO and PBN in benzene to give PhCOO-PBN could be followed by monitoring [PhCOO-PBN] from 38°C and upwards. The reaction was first order in [DBPO] and zero order in [PBN], and the rate constants s-l at evaluated for the homolysis of the 0-0 bond in DBPO ( k = 3.7 X 38°C) agreed well with those of other studies at higher temperatures. Thus in benzene the homolytic decomposition mechanism of DBPO seems to prevail.

-

Aorhu

(PhC00)2

2PhCOO

(49)

Diacyl peroxides are, however, also electron transfer oxidants, which according to a theoretical analysis should possess standard potentials, Eo[(ArCOO),/RCOO RCOO-) of around 0.6 V in water, provided that the electron transfer process is of the dissociative type (50) (Eberson, 1982~). Such a value brings thermal ET steps involving DBPO within reach for redox-active organic molecules, as for example suggested by the so-called CIEEL mechanism of chemiluminescence (Schuster, 1982).

+

(PhC00)2 e -

4

PhCOO + PhCOO-

(50)

For a less reactive molecule like PBN, a Marcus calculation using a dm3mol s-l for the reorganization energy of 40 kcal mol-' gives k = reaction with DBPO in acetonitrile at 25"C, just to select a solvent which does not cause complications from the consideration of electrostatic terms. Clearly this is of a similar order of magnitude as the rate constants determined in benzene. In photochemical reactions, the role of DBPO will undoubtedly be that of an electron acceptor from an excited state species, as shown in reaction (51). Thus, inverted spin trapping will be feasible and an unambiguous interpretation of the appearance of PhCOO-ST will be difficult. In HFP the very strongly attenuated reactivity of benzoate ion should, however, make the homolysis mechanism predominate, as indicated by the appearance of both PhCOO-ST and Ph-ST (ST = PBN or DMPO) in the photolysis of DBPO and ST (Eberson et al., 1996a). (PhC00)z + RH*

4

PhCOO' + PhCOO-

+ RH'+

(51)

126

L. EBERSON

In general, the importance of the acceptor properties of all types of compounds containing an 0-0 bond should be emphasized. A likely function of a peroxidic compound (hydrogen peroxide, alkyl peroxides, acyl peroxides, peroxydisulfate, to mention a few commonly used ones) under photochemical conditions (UV light) should be that of an electron acceptor from an excited state. Moreover, the electron acceptor efficiency is high in view of the dissociative nature of the ET step.

10 Ionizing radiation and spin trapping

High-energy ionizing radiation, such as electron and y-ray beams, can be used for the generation and detection of radicals by spin trapping. The processes leading from these high-energy sources (keV to MeV) to chemistry in the usual energy range are complex, but can, at least for electron sources, be controlled by additives to produce either hydrated electrons, hydroxyl radicals or hydrogen atoms (Neta, 1976). Much work conducted in low-temperature matrices has shown that the primary chemical process induced by y-irradiation is formation of electrons (e-) and positive holes (h+),the latter eventually leading to the formation of radical cations of the cornponent(s) with the lowest ionization potential (Symons, 1997). This means that an added spin trap may be transformed into its radical cation by y-irradiation and thus create conditions for inverted spin trapping, as already described for PBN and DMPO above in experiments designed to study this aspect.

11 Spin trapping of radicals generated by ultrasound (sonolysis)

Ultrasound has chemical effects on liquid systems owing to the high temperatures (thousands of kelvins) and pressures (hundreds of atmospheres) produced during the collapse of acoustic cavitation bubbles. This creates microchambers where the vapour of species present can undergo pyrolytic reactions with formation of radicals. Thus, water on sonolysis produces H and H O , both of which can be trapped by DMPO or PBN or various water-soluble PBN derivatives (Makino et al., 1982a,b, 1990). The possibility that the spin adducts were formed by addition-oxidation andlor reduction of the spin trap by e& has been discussed. Experiments conducted in the presence of scavengers in combination with kinetic analysis supported the assumption that H and H O are formed directly; thus, for example, HO-DMPO could be gradually replaced by the formate adduct (-OzCDMPO) by adding increasing concentrations of sodium formate, until at

127

SPIN TRAPPING AND ELECTRON TRANSFER

[HCOO-] = 1 mmol dmp3 the latter adduct was the only one (see reactions (52)-(54)).

+ DMPOH O + HCOOHO

Ozc' + H2O

-+

- 0 Z C

+ DMPO

(52)

HO-DMPO

-+

(53)

-0,C-DMPO

(54)

Experiments in neat N,N-dimethylformamide, using 3,5-dibromo-4nitrosobenzenesulfonate [20] as the spin trap, avoided the difficulties of competing redox reactions since the species trapped, CH; and 'CH2N(CH3)CH0 [reactions (55)-(57)], cannot conceivably be generated except by homolysis of N,N-dimethylformamide (Misik et al., 1995). NO I

so3-

CHIN(CH3)CHO + CH; CH;

+ CH3N(CH;)CHO

CH; and 'CH2N(CH3)CH0+ [20]

-

-+

+ 'N(CH3)CHO

'CH2N(CH;)CHO

+ CH4

(55)

(56)

CH3-[20]' and '[20]-CH2N(CH3)CH0 (57)

12 Spin trapping in biochemical/biological systems

Spin trapping is an often-used technique in the study of possible radical production in biological systems (for reviews see Kalyanaraman, 1982; Mason, 1984; Mottley and Mason, 1989), particularly by the detection and monitoring of spin adducts of the hydroxyl and hydroperoxyl ('OOH) radicals in view of their relation to possible damage mechanisms. This is a large area of research which it is not possible to cover in a limited review, and the treatment will therefore be restricted to a discussion of the electron transfer properties of biochemical systems (for a review on the application of the Marcus theory to reactions between xenobiotics and redox proteins, see Eberson, 1985) and

L. EBERSON

128

Table U Potentials, E"', of redox proteins in water at pH 7.0 and 25°C." E"'N vs SCE

System (redox couple) ~~

~

~

~

~

Cytochrome c from horse (Fe3+/Fe2+) Haemoglobin (Fe3+/Fe2+) Myoglobin (Fe3+/Fe2+) Rubredoxin (Fe3+/Fe2+) Horseradish peroxidase (Fe3+/Fe2+) Horseradish peroxidase Compound I (Fe5+/Fe4+) High-potential protein from Chromatium vinosurn (Fe3+/Fe2+) Cytochrome P450 (Fe5+/Fe4+or Fe4+/Fe3+) Azurin (Cu2+/Cu+) Hastocyanin (Cu2+/Cu+) Ceruloplasmin (Cu2+/Cu+) Laccase (Cu2+/Cu+)

~

0.01 -0.07 -0.19 -0.30 -0.41 0.70 0.11 >0.56 0.14 0.13 0.15 0.18

"From a compilation in Eberson (1985).

their possible implications for the two redox spin trapping mechanisms under discussion here. In a later section, some specific problems connected with hydroxyl adducts will be discussed. A review on spin trapping artefacts in biological model systems has appeared (Tomasi and Iannone, 1993). Table 12 shows redox properties of some redox systems of biochemical nature. Generally, the redox potentials are modest, cytochrome P450 possibly being an exception. If cytochrome P450 functions as an electron transfer oxidant towards xenobiotic molecules, it is necessary to postulate a considerably higher potential (1.3-1.8 V) from considerations of the Marcus theory (Eberson, 1990). Otherwise, none of the systems listed in Table 12 would seem to be capable of oxidizing any of the common spin traps to their radical cations. One enzyme, lignin peroxidase, together with hydrogen peroxide, has been shown to oxidize organic substrates to epr spectrally detectable radical cations, as shown below in reaction (58) (Kersten ef al., 1990). The upper limit for a detectable radical cation is a respectable 1.34 V (1,4-dimethoxybenzene radical cation), indicating that the upper limit for radical cation formation might possibly touch spin traps of low E"(ST+IST),such as [2] and [3] (Table 1). 'Ikro other enzymes, horseradish peroxidase + hydrogen peroxide and laccase + oxygen, gave only the epr spectrum of 1,2,4,5-tetramethoxybenzene, the most redox reactive of the compounds studied. This compound is an easily oxidizable one, however, and not representative of commonly occurring organic substrates. On the other hand, the potentials of most redox proteins of Table 12 are well suited for an oxidative role in the addition-oxidation mechanism, being capable of oxidizing the hydroxylamine intermediate. A clear-cut example of this mechanism appears to be "trapping of cyano radical" by MNP [l] in solutions

SPIN TRAPPING AND ELECTRON TRANSFER

129

of cyanide ion-hydrogen peroxide-horseradish peroxidase (Moreno et al., 1988; Stolze et al., 1989), since one-electron oxidation of cyanide ion (E" = 2.3 V, see Table 5 ) by horseradish peroxidase should be excluded for kinetic reasons. An estimated rate constant for such an ET step is dm3 mol-' s-'.

-

lignin peroxidase

ArH

A r H + (epr-active in aqueous medium)

(58)

ArH = 1,4-dimethoxybenzene(1.34 V) 1,2,3,4-tetramethoxybenzene(1.25 V) hexamethoxybenzene (1.24V) 1,2,4,5-tetramethoxybenzene(0.81V)

13 Conclusions on the radical cation mechanism

As shown above, conditions for radical cation formation are easily established in spin trapping experiments, either by accident or design, and a scrutiny of the literature would no doubt turn up a large number of individual cases of suspected radical cation mechanisms. However, little of principal interest would be learned by such an exercise, and therefore this approach will not be followed. A scrutiny of the excellent compilation of spin adducts by Buettner (1987) shows that the general cases discussed above are the important ones. The formation of radical cations is a process of high-energy type, and needs reaction conditions capable of creating high-potential oxidaats. Thus the recognition of the radical cation mechanism is nearly always possible, even if it cannot always be distinguished from true spin trapping in a simple way.

14 Spin adduct formation via radical anions

It was already mentioned [reactions (8) and (9) and the associated text, p. 941 that the first situation in which a radical ion of a spin trap was suggested to be involved (Crozet et al., 1975) was the reaction between an alkyl iodide and a thiolate ion in the presence of TBN [2]. This compound is reduced reversibly at -1.25 V, and with E"(RS'IRS-) around 0.2 V reaction (8) is endergonic by 1.4 eV, not a favourable precondition for an E T reaction. Therefore, it is likely that some other mechanism is responsible for the observations made. The quite negative reduction potentials of spin traps (Table 2) make them less amenable to participation in the radical anion mechanism, as first established in the cathodic reduction of benzenediazonium salts at a controlled potential in the presence of PBN (Bard et al., 1974). In fact, the lower cathodic limit of the spin trapping method is set not by the nitrone but by the spin adduct formed.

L. EBERSON

130

A different electrochemical approach was applied to the cathodic reduction of sulfones in N,N-dimethylformamide (Djeghidjegh et al., 1988), for example t-butyl phenyl sulfone, which is reduced at a more negative potential (Epc= -2.5 V) than is PBN (-2.4 V). Thus, the electrolysis of a mixture of PBN and the sulfone would possibly proceed via both true and inverted spin trapping. If a mediator of lower redox potential, such as anthracene (-2.0 V), was added and the electrolysis carried out at this potential, it was claimed that only the sulfone was reduced by anthracene'- with formation of t-butyl radical and thus true spin trapping was observed. It is difficult to see how this can be reconciled with the Marcus theory, which predicts that anthracene'- should react preferentially with PBN. The ratio of ET to PBN over sulfone is calculated to be -20 from equations (20) and (21), if both reactions are assumed to have the same A of 20 kcal mol-'. The ready protonation of radical anions under conditions of proton availability causes other problems to appear, as for example shown by the stepwise cathodic reduction of PBN to the corresponding imine and amine [reactions (59) and (60)] during which the intermediate radicals [21] and [22] appear and become trapped by PBN (Simonet et al., 1990). Ph-CH=N(O)Bu' PhCH=N-Bu'

+ e- + H+

-

+ 2e- + H'

PhCH2NBu'

PI1

-

+ OH-

PhCH=N-Bu'

+ PhCHNHBu'

-

(59)

C e - . +Hi

[221

PhCH,NHBu' (60)

15 The nucleophilic addition-oxidation mechanism

This mechanism, involving the addition of a nucleophilic species to the nitroso or nitrone functionality [reactions (2) and (3)] with formation of a hydroxylamine, and oxidation of the latter to the nitroxyl radical is more difficult, if not sometimes impossible, to avoid. Hydroxylamines have low redox potentials, ElI2being in the range of -(0.4-0.5) V in aqueous medium for a series of alkyl- and arylhydroxylamines in their anionic form and 0.5-0.8 V in acetonitrile in their neutral form (Table 13). In dimethyl sulfoxide, the anions have EPain the range of -(0.7-1.1) V. This means that even a weak ET oxidant like dioxygen [E0(O2/0,'-)= -0.4 V in water] can oxidize these intermediates, two diagnostic examples being formation of F-PBN from fluoride ion, O2and PBN in chloroform (Eberson, 1994), and AcO-PBN from acetate ion, O2 and PBN (Forrester and Hepburn, 1971). The situation depicted in reaction (61) must indeed be very common, not least in biological systems containing proteins of the seemingly low redox reactivity shown by most redox proteins in Table 12. NU- + H+ + ST

-

ST(H)Nu El,z=O 5-0.8V

-

NU-ST

(61)

SPIN TRAPPING AND ELECTRON TRANSFER

131

Table 13 Anodic half-wave or peak potentials (vs SCE) of hydroxylamines RNHOH or the corresponding anions RNHO-.

Compound

Ep,IVb

Hydroxylamine N-Methylhydroxylamine N-Ethylhydroxylamine N-Propylhydroxylamine N-Isopropylhydroxylamine N-t-Butylhydroxylamine N-Cyclohexylhydroxylamine N-Phenylhy droxylamine N-(4-Bromophenyl)hydroxylamine N-Benzylhydroxylamine 2-Phenyl-2-hydroxylaminopropane 3-Methyl-3-hydroxylamino-2-butanol N,N-Dimethylhydroxylamine

-0.35 -0.48 -0.49 -0.49 -0.48 -0.47 -0.47 -0.48

N,N-Dibenzylhydroxylamine

-0.38

N-Hydroxypiperidine

E,,IV'

0.65

-0.75 -0.64

0.80 0.60 0.45 0.55

-0.52 -0.47 -0.51 0.50 - 1.10

"RNHO- oxidation in aqueous solution at pH 13 at an Hg anode (Iversen and Lund, 1969). bRNHO- oxidation in dimethyl sulfoxide-Et,NBF, at a PT anode (Bordwell and Liu, 1996); the published potentials were converted from the ferricinium/ferrocene reference to SCE by adding 0.51 V (Bordwell ef al., 1991). 'RNHOH oxidation in acetonitrile-NaC10, at a glassy carbon anode (Sayo ef al., 1973; Ozaki and Masui, 1978).

Outside the spin trapping field, reactions of nitrones with nucleophiles have been found to give initially products of addition, in the appropriate cases followed by elimination reactions or other steps leading to stable products (Breuer, 1989). Active nucleophiles include methanol, azide, thiocarboxylic acids, thiols, cyanide, carbanions, phosphorus ylids and organolithium or organomagnesium compounds. Trimethylsilyl cyanide reacts with nitrones to give a-cyano-0-trimethylsilyl products (Tsuge et al., 1980), which applied to PBN, for example, should give a mechanistically interesting precursor to NC-PBN. Few studies of reaction (61) have been performed with the goal of delineating the scope of the mechanism, and most spin trapping studies rule it out in a cursory way, if it is mentioned at all. It is therefore a matter of some urgency to define systems which are suitable for study of the additionelimination mechanism, particularly the characteristics of the initial equilibrium (reaction (2) or ( 3 ) ) and the rate constants for the oxidation of hydroxylamines. One recent study has shown that heteroaromatic bases of N-H type are prone to give N-nitroxyls from PBN or DMPO using weak oxidants such as dioxygen or chloranil (E" = -0.4 and 0.0 V, respectively), thus providing diagnostic cases of reaction (61) (Alberti et al., 1997). In the reaction of N-chlorobenzotriazole (BT-Cl) and PBN, the autocatalysed

L. EBERSON

132

formation of N-(1-benzotriazoly1)-PBN takes place with benzotriazole as the autocatalytic species (reactions (62)-(64)) (Carloni et af., 1996). The reaction was strongly inhibited in a solvent incapable of donating a hydrogen atom to BT,such as benzene. Benzotriazolate anion did not sustain the autocatalytic process, but addition of an equivalent amount of acid immediately started it. This indicates that the equilibrium of reaction (62) must be driven to the right by protonation in order to ensure a sufficiently high concentration of the hydroxylamine. Also it has been found that DMPO-benzotriazole in acetonitrile is oxidized relatively fast by Co"'W, the weak ET oxidant characteristics of which have already been mentioned (p. 120), with formation of N-benzotriazolyl-DMPO and Co"W (Alberti et al., 1997). PBN+BT-H + PBN(H)BT PBN(H)BT + BT-Cl + BT-PBN B T + RH (from solvent)

-

+ B T + HCl

BT-H + R

(62) (63)

(64)

The reaction exemplified by reactions (62)-(64) represents what may well be a general case of the addition-oxidation mechanism, the reaction between a spin trap and a weak, seemingly unreactive electron acceptor. If a catalyst HA is present as an impurity, an autocatalysed reaction is set up according to reactions (65)-(68). Such a mechanism may be the origin of many spin adduct sightings in thermal systems of the type spin trap-weak acceptor (X-Y), whose radical anion can cleave in a fast step, as exemplified by acceptors such as 3-chloroperoxybenzoic acid (Janzen et al., 1992b), N-haloimides (Lagercrantz, 1971; Kaushal and Roberts, 1989; Eberson et al., 1994a), N-chlorosulfonamides (Evans et af., 1985), and polyhalo compounds, such as trichloroacetonitrile (Sang et al., 1996; Eberson et af., 1997). ST+H-A + ST(H)A

(65)

ST(H)A + X-Y + A-ST + HX + Y

(66)

Y' + PBN

ST + HX

-

-

Y-PBN

(67)

ST(H)X; then a new cycle

(68)

TRAPPING OF THE HYDROXYL RADICAL

As already pointed out, determining the mechanism of formation of hydroxyl spin adducts in aqueous media under oxidizing conditions is a particularly urgent problem in view of its implications in biochemistry and biology. In

SPIN TRAPPING AND ELECTRON TRANSFER

133

itself, the notion of water and/or hydroxide being oxidized to hydroxyl radical is already a thermodynamically difficult proposition, since E"(HO/HO-) is as high as 1.7V (Table 5 ) and that of neutral water oxidation is significantly higher. Few oxidants are capable of effecting this reaction in aqueous systems. The class of Fenton reagents, Fe"(aq) or other Fe"(1igand) + H202,was once considered to be a source of free H O in aqueous medium and thus a good calibration benchmark for direct formation of hydroxyl spin adducts, but was recently shown to involve an iron complex, (ligand)Fe"OOH, which directly transfers the hydroxyl group to substrates present (Sawyer et al., 1993; Hage et al., 1995). The generation of H O by pulse radiolysis provides a way for investigating the kinetics of hydroxyl spin adduct formation. For PBN and some of its 4-substituted derivatives (ranging from 4-Me0 to 4-N02), rate constants in the range of (5-9) X lo9dm3mol-' s-l were determined (Greenstock and Wiebe, 1982). A study of the reaction of the water-soluble 2-, 3- and 4-PyBN[23] and H O showed that most of the hydroxyl radicals became attached to the heteroaromatic ring (Neta et al., 1980; Sridhar et al., 1986). Similar findings

were reported for PBN itself, where H O was shown to attack predominantly the phenyl ring with formation of cyclohexadienyl radicals (Zubarev et al., 1992).Thus PBN and PBNs should have low efficiencyfor the trapping of H O . The situation is different in DMPO, where no aromatic ring is present. Again generating H O by pulse radiolysis, it was possible to apply time-resolved epr studies to monitor the appearance of HO-DMPO and determine the rate constant, k = 2.8 X lo9dm3 mol-' sP1. The yield of trapping was high (94%) (Madden and Taniguchi, 1996).The rate of disappearance of hydroxyl adducts of PBN and its derivatives is fast and thus it is necessary to use a steady-state method for generation, usually photolysis in the presence of 1% H202. In aqueous phosphate buffer the half-life is approximately 90 s at pH 6 (Kotake and Janzen, 1991) and in acetonitrile 10 s (Janzen et al., 1992a). The effect of 4-substituents is small, Hammett's p being -0.6. The half-life of HO-DMPO is appreciably longer, -15 min at pH 7.0, when generated by the H202/hv method (Marriott et al., 1980). However, when HO-DMPO was generated by photolysis of DMPO together with aqueous peroxydisulfate, its half-life was too short to be measured (4 s) (Kirino et al., 1981). In a later study using the same method of generation, in which it was demonstrated that HO-DMPO actually can be formed also by nucleophilic substitution upon initially formed

L. EBERSON

134

-O,SO-DMPO, no indication of such a short lifetime was obtained (Davies et al., 1992). This difference illustrates the uncertain nature of spin adduct half-lives and their strong dependence on the exact reaction conditions. Thus there is little doubt that the hydroxyl radical, if generated by an unambiguous method such as pulse radiolysis, can be trapped by PBN or DMPO, even if the former has several deficiencies, among them low trapping efficiency and short half-life of HO-PBN. The problem in hydroxyl radical trapping thus rests with the possible competition from the nucleophilic addition-oxidation mechanism, as exemplified in reaction (69) for DMPO and Ox-Red as a general one-electron redox system, or the inverted spin trapping mechanism (70). The treatment to follow will mostly be limited to DMPO. DMPO + H 2 0 + HO-DMPO(H)

+ Ox

DMPO + Ox + Red + DMPO'

-+

-+

HO-DMPO

HO-DMPO

+ H+ + Red

+ H+

(69) (70)

The suitability of DMPO as a spin trap for hydroxyl and alkoxyl radicals was tested recently (Hanna et al., 1992) in response to a demonstration of easy nucleophilic addition of water to DMPO and oxidation to HO-DMPO by aqueous FeC13 (Makino et al., 1990). Oxidation of DMPO by Fe"' or Cu" in "0-enriched water showed that nucleophilic addition of water occurred at the nitrone carbon of DMPO and that this pathway was the major one leading to HO-DMPO. With H202 added, each of the metal ions promoted the formation of much stronger signals of HO-DMPO, and the proportion of nucleophilic addition decreased, most markedly for Fe'". This indicates either that nonlabelled H O is formed from H202and trapped as such, or that the stronger nucleophile H z 0 2 can undergo nucleophilic addition-oxidation as well, provided there is a rapid follow-up reaction for converting HOODMPO to HO-DMPO. In the presence of chelators, the nucleophilic mechanism was suppressed and thus it was concluded that the nucleophilic addition mechanism should not be of concern in biochemical systems. However, the proposed explanation for the suppression by chelating agents was dependent on the metal ion-promoted addition of water to DMPO, followed by oxidation of the hydroxylamine by the metal ion [reaction (7111. Fe"'

DMPO

+

H20

Fell'

DMPO(H)OH

HO-DMPO

(71)

The addition-oxidation mechanism does not require catalysis in the first step, and an alternative explanation of these results would be to assume that the chelator changes the standard potential of the metal ion and thus the oxidation rate. Table 14 summarizes some of the data required to test this assumption; unfortunately, standard potentials were not available for Fe"'/Fe" in all the

135

SPIN TRAPPING AND ELECTRON TRANSFER

Table 14 Formation of HO-DMPO by the oxidation of an aqueous solution of DMPO by metal redox couples."

Oxidizing system

E"IV vs SCE

Fe"' in water Fe"' in water, 2 mmol dm-' EDTA Fe"' in water, nitrilotriacetate Fe"'(CN)iCu" in water with C1~

~

"Hanna er trl. (1992). h " + " ,

______

Formation' of HO-DMPO

0.53 -0.12 0.09

0.17 0.30 ~

formed, "-", not formed.

buffer systems used. Yet one can see that the role of the chelatorlbuffer species might well be to change the standard potential of the redox-active metal and thus the oxidation rate of the second step of reaction (71). If so, the redox potential for oxidation of the hydroxylamine intermediate of DMPO should be somewhere between 0.2 and 0.3 V, some 0.3-0.6 V below the values of the hydroxylamine E,, values of Table 13. Thus the oxidation process is entirely feasible and in good agreement with the results for the DMPO-Nbenzotriazole-Co"'W system mentioned above (Alberti et al., 1997). The study above (Hanna et al., 1992) also addressed the problem of nucleophilic addition of alcohols to DMPO, using Fe"' as the oxidant in an aqueous-alcoholic solution (from 95 '30to 25 YOwater). Only primary alcohols engaged in this reaction, whereas 2-propanol or 2-methyl-2-propanol did not react even when the alcohol concentration was increased to 70%. This may depend on either decreased reactivity of secondary and tertiary alcohols, perhaps for steric reasons, or lower stability of the corresponding spin adducts. Later even more complexity was demonstrated (Makino et al., 1992) in the reaction between DMPO and Fe"' in water. The HO-DMPO formed was transformed into a hydroxamic acid [24] which is a tautomer of 2-hydroxyDMPO [25]; in a Fenton system transfer of a hydroxyl (cf. p. 133) from the ligand-Fe00H complex to either of these species leads to additional epr-active nitroxyls I261 and [27] in reaction (72). From the above it is clear that DMPO can undergo the addition-oxidation mechanism with water as the nucleophile, provided a suitable oxidant is present. With a primary alcohol competing, the 0-connected alkoxy spin adduct is formed in addition to HO-DMPO. O n the other hand, with a hydroxyl radical source a competing alcohol will undergo hydrogen abstraction by H O and form an a-hydroxyalkyl radical which forms a C-connected spin adduct. This criterion clearly can distinguish between the two mechanisms at least in model systems (for recent examples, see Reszka and Chignell, 1995; Janzen et nl., 1995; Thomas et al., 1996).

136

“ ” ‘ O H OH H3C

0. I

-

L. EBERSON

Ffl

H3c&L0H

H3C OH

A-

I

0’

16 Bona fide spin trappings: a recipe

The discussion so far has centred around cases where there is a risk of misinterpretation of spin trapping studies, and it is therefore fair to end with an appreciation of the method as used in the large number of unambiguously interpretable studies which have been performed. Common to most of these investigations is the use of a mechanistic and kinetic barrier against the radical cation and addition-oxidation mechanisms: the radical X to be trapped should be generated by the cleavage of a strong, thermally nondissociable bond in a more complex molecule X-Y, which is not an electron acceptor. It

Table 15 Examples of systems with a mechanistic bias towards proper spin trapping. Spin trap PBN DMPO, PBN DMPO, PBN DMPO DMPO PhNO DMPO POBN PhNO

Source of radical, X-Y Bu,Sn-Bu, hv Et,Hg-Et, hv PhCHzHg-CHzPh, hv HOCHZ-H, BP* HOMe,C-H, BP* PhMe2C-H, NiOz Ph-NHNH2, erythrocytes PhCH2CH2-NHNH2, microsomes

x

Ref.

Bu’ Et’ PhCH; HOCH; HOMezC PhMe,C

a

Ph’

PhCHZCH; 3,4-(Me0)zC6H3CH(OH)-CH(Me)Ph, Ph(Me,)CH Ligninase + HzOZ

a. b a,b b b c

d e

f

“Janzen and Blackburn, 1969. bJanzen and Liu, 1973. ‘Terabe and Konaka, 1972. dHill and Thornalley, 1982. “Rumyantseva et al., 1991. ’ H a m e l et al., 1986.

SPIN TRAPPING AND ELECTRON TRANSFER

137

is then vanishingly improbable that X-Y should donate X- to the spin trap radical cation or to the spin trap itself, these steps being necessary as a prelude to either of the undesirable mechanisms. Table 15 illustrates the principle for a number of spin trapping situations.

References Alberti, A., Carloni, P., Eberson, L., Greci, L. and Stipa, l? (1997). J. Chem. SOC.,Perkin Trans. 2, 887 Amadelli, R., Maldotti, A., Bartocci, C. and Carassiti, V. (1989). J. Phys. Chem. 93, 6448 Baciocchi, E., Crescenzi, M., Fasella, E. and Mattioli, M. (1992). J. Org. Chem. 57, 4684 Baciocchi, E., Bietti, M. and Mattioli, M. (1993). J. Org. Chem. 58, 7106. Baciocchi, E., Bietti, M. and Steenken, S. (1996). J. Chem. SOC., Perkin Trans. 2, 1261 Bancroft, E. E., Blount, H. N. and Janzen, E. G. (1979). J. Am. Chem. SOC.101, 3692 Bard, A. J., Gilbert, J. C. and Goodin, R. D. (1974). J. Am. Chem. SOC. 96,620 Bard, A. J., Ledwith, A. and Shine, H. J. (1976). Adv. Phys. Org. Chem. 13, 156 Baumann, H., Oertel, U., Timpe, H. J., Zubarev, V. E., Fok, N. V. and Mel’nikov, M. J. (1985). 2. Chem. 25, 182 Bietti, M., Baciocchi, E. and Engberts, J. B. F. N. (1996). J. Chem. Soc., Chem. Commun. 1307 Bobbitt, J. M. and Flores, C. L. (1988). Heterocycles 27, 509 Bordwell, E G. and Liu, W.-Z. (1996). J. Am. Chem SOC. 118, 8778 Bordwell, F. G., Cheng, J.-P., Ji, G.-Z., Satish, A. A. and Zhang, Z. (1991). J. Am. Chem. SOC. 113, 9790 Breuer, E. (1989). In Nitroneq Nitronates and Nitroxides (ed. S. Patai and Z . Rappoport). Wiley, Chichester, chaps. 2 and 3 Budac, D. and Wan, .’F (1992). J. Photochem. Photobiol. A : Chem. 58,135 Buettner, G. R. (1987). Free Radical Biol. Med. 3, 259 Burdon, J. and Parsons, I. W. (1975). Tetrahedron 31,2401 Carloni, P., Eberson, L., Greci, L., Sgarabotti, P. and Stipa, l? (1996). J. Chem. SOC., Perkin Trans. 2, 1297 Cerri, V., Frejaville, C., Vila, F., Allouche, A., Gronchi, G. and Tordo, I? (1989). J. Org. Chem. 54, 1447 Chalfont, G. R. and Perkins, M. J. (1970). J. Chem. Soc. B 401 Chandra, H. and Symons, M. C. R. (1986). J. Chem. Soc., Chem. Commun. 1301 Childs, W. V., Christensen, L., Klink, F. W. and Kolpin, C. F. (1991). In Organic Electrochemistry (ed. H. Lund and M. M. Baizer), 3rd edn. Dekker, New York, chap. 26 Crozet, M. P., Flesia, E., Surzur, J. M., Boyer, M. and Tordo, l? (1975). Tetrahedron Lett. 4563 Davies, A. G. and Ng, K.-M. (1995). Austr. J. Chem. 48, 167 Davies, M. J., Gilbert, B. C., Stell, J. K. and Whitwood, A. C. (1992). J. Chem. SOC., Perkin Trans. 2, 333 Djeghidjegh, N., El Badre, M. C., Simonet, J. and Mousset, G. (1988). J. Electroanal. Chem. 246,457 Eberson, L. (1963). Acta Chem. Scand. 17,2004

138

L. EBERSON

Eberson, L. (1982a). Acta Chem. Scand., Ser. B 36,533 Eberson, L. (1982b). Adv. Phys. Org. Chem. 18,79 Eberson, L. (1982~).Chem. Scr. 20, 29 Eberson, L. (1983). J. Am. Chem. SOC.105, 3192 Eberson, L. (1985). Adv. Free Radical Biol. Med. 1, 19 Eberson, L. (1987). Electron Transfer Reactions in Organic Chemistry. Springer-Verlag, Heidelberg Eberson, L. (1990). Acta Chem. Scand. 44,733 Eberson, L. (1992). J. Chem. SOC.,Perkin Trans. 2, 1807 Eberson, L. (1994). J. Chem. SOC.,Perkin Trans. 2, 171 Eberson, L. and Larsson, B. (1986). Acta Chem. Scand., Ser. B. 40,210 Eberson, L. and Larsson, B. (1987). Acta Chem. Scand., Ser. B. 41, 367 Eberson, L. and Nilsson, M. (1990). Acta Chem. Scand. 44,1062 Eberson, L. and Nilsson, M. (1993). Acta Chem. Scand. 47, 1129 Eberson, L. and Nyberg, K. (1976). Adv. Phys. Org. Chem. 12, 1 Eberson, L. and Persson, 0. (1997). J. Chem. SOC.Perkin Trans. 2, 893 Eberson, L. and Shaik, S. S. (1990). J. Am. Chem. SOC.112,4484 Eberson, L. and Utley, J. H. P. (1983). In Organic Electrochemistry (ed. M. M. Baizer and H. Lund), 2nd edn. Dekker, New York, chap. 14 Eberson, L., Utley, J. H. P. and Hammerich, 0.(1991). In Organic Electrochemistry (ed. H. Lund and M. M. Baizer), 3rd edn. Dekker, New York, chap. 25 Eberson, L., Gonzalez-Luque, R., Lorentzon, J., Merchdn, M. and Roos, B. 0. (1993). J. Am. Chem. SOC.115, 2898 Eberson, L., Lind, J. and Merenyi, G. (1994a). J. Chem. SOC., Perkin Trans 2, 1181 Eberson, L., Hartshorn, M. P., Radner, F. and Svensson, J. 0. (1994b). J. Chem. Soc., Perkin Trans 2, 1719 Eberson, L., Hartshorn, M. P. and Persson, 0. (1996a). J. Chem. SOC., Perkin Trans. 2, 141 Eberson, L., Hartshorn, M. P., Persson, 0. and Radner, F. (1996b). J. Chem. Soc., Chem. Commun. 2105 Eberson, L., MacCullough, J. J. and Persson, 0. (1997). J. Chem. Soc., Perkin Trans. 2, 133 Ekstrom, M. (1988). “Electron transfer in reductions of polyhalogenated alkanes.” Thesis, Lund University Evans, C. A. (1979). Aldrichirn. Acta 12(2), 23 Evans, J. C., Jackson, S. K. and Rowlands, C. C. (1985). Tetrahedron 41, 5191, 5195 Forrester, A. R. and Hepburn, S. €? (1971). J. Chem. SOC.(C) 701 Fox, M. A. (1991). In Organic Electrochemistry (ed. H. Lund and M. M. Baizer), 3rd edn. Dekker, New York, chap. 34, p. 1397 Fox, M. A. and Chanon, M., eds. (1988). Photoinduced Electron Transfer, part A-D. Elsevier, Amsterdam Greenstock, C. L. and Wiebe, R. H. (1982). Can. J. Chem. 60, 1560 Gronchi, G. and Tordo, P. (1993). Res. Chem. Intermed. 19, 733 Gronchi, G., Courbis, P., Tordo, P. Mousset, G. and Simonet, J. (1983). J. Phys. Chem. 83, 1343 Hage, J. I?, Llobet, A. and Sawyer, D. T. (1995). Bioorg. Med. Chem. 3, 1383 Hammel, K. E., Kalyanaraman, B. and Kirk, T. K. (1986). Proc. Natl. Acad. Sci. USA 83, 3708 Hammerich, 0. and Parker, V. D. (1984). Adv. Phys. Org. Chem. 20, 55 Handoo, K. L. and Gadru, K. (1986). Curr. Sci. 55,920 Hanna, P. M., Chamulitrat, W. and Mason, R. P. (1992). Arch. Biochem. Biophys. 296. 640

SPIN TRAPPING AND ELECTRON TRANSFER

139

Hilinski, E. F., Masnovi, J. M., Kochi, J. K. and Rentzepis, P. M. (1984). J. A m . Chem. SOC.106, 8071 Hill, H. A. 0. and Thornalley, l? J. (1982). Can. J. Chem. 60, 1528 Hillborn, J. W. and Pincock, J. A. (1991). J. A m . Chem. SOC.113, 2683 Iversen, P. E. and Lund, H . (1969). Anal. Chem. 41, 1322 Jaeger, C. D. and Bard, A. J. (1979). 1.Phys. Chem. 83,3146 Janzen, E . G. (1971). Accounts Chem. Res. 4, 31 Janzen, E. G. and Blackburn, B. J. (1969). J. A m . Chem. SOC.91, 4481 Janzen, E. G. and Coulter, G. A. (1984). J. A m . Chem. SOC.106, 1962 Janzen, E. G. and Haire (1990). In Advances in Free Radical Chemistry (ed. D. D. Tanner), vol. 1. JAI Press, London Janzen, E. G. and Liu, J. I. (1973). J. M a p . Res. 9, 510 Janzen, E. G., Evans, C. A. and Nishi, Y . (1972). J. Am. Chem. SOC.94, 8236 Janzen, E. G., Stronks, H. J., Nutter, Jr., D. E., Davis, E. R., Blount, H. N., Poyer, J. L. and McCay, P. B. (1980). Can. J. Chem. 58, 1596 Janzen, E. G., Hinton, R. D. and Kotake, Y. (1992a). Tetrahedron Lett. 33, 1257 Janzen, E. G., Lin, C.-R. and Hinton, R. D. (1992b). J. Org. Chem. 57, 1633 Janzen, E. G., Chen, G., Bray, T. M., Reinke, L. A., Poyer, J. L. and McCay, P B. (1993). J. Chem. SOC.,Perkin Trans. 2, 1983 Janzen, E. G., Zhang, Y. and Arimura, M. (1995). J. Org. Chem. 60,5434 Kalyanaraman, B. (1982). Rev. Biochem. Toxicol. 4, 73 Kaushal, P. and Roberts, B. I? (1989). J. Chem. SOC.,Perkin Trans. 2, 1559 Kersten, l? J., Kalyanaraman, B., Hammel, K. E., Reinhammar, B. and Kirk, T. K. (1990). Biochem. J. 268, 475 Kirino, Y., Ohkuma, T. and Kwan, T. (1981). Chem. Pharm Bull. 29, 29 Kochi, J. K. (1988). Angew. Chem., Int. Ed. Engl. 27, 1227 Kochi, J. K. (1990). Acta Chem. Scand. 44,409 Kotake, Y. and Janzen, E. G. (1991). J. Am. Chem. SOC.113, 9503 Kraeutler, B., Jaeger, C. D. and Bard, A. J. (1978). J. Am. Chem. SOC.100, 4903 Lagercrantz, C. (1971). J. Phys. Chem. 75, 3466 Lagercrantz, C. and Forshult, S. (1969). Acta Chem. Scand. 23, 708 Lind, J.. Jonsson, M., Eriksen, T., Merenyi, G. and Eberson, L. (1993). J. Phys. Chem. 97.1610 Lund, H., Daasbjerg, K., Lund, T.. Occhialini, D. and Pedersen, S. U. (1997). Acta Chem. Scand. 51, 135 Madden, K. F! and Taniguchi, H. (1996). J. Phys. Chem. 100,7511 Makino, K., Mossoba, M. M. and Riesz, l? (1982a). J. A m . Chem. SOC.104,3537 Makino, K., Mossoba, M. M. and Riesz, I? (1982b). J. Phys. Chem. 87,1369 Makino, K., Hagiwara, T., Hagi, A., Nishi, M. and Murakami, A. (1990). Biochem. Biophys. Res. Commun. 172, 1073 Makino, K.. Hagi, A., Ide, H., Murakami, A. and Nishi, M. (1992). Can. J. Chem. 70, 2818 Marcus, R. A. (1964). Anntc. Rev. Phys. Chem. 15, 155 Marcus, R. A. and Sutin, N. (1985). Biochim. Biophys. Acta 811, 265 Marriott, P. R., Perkins, M. J. and Griller, D. (1980). Can. J. Chem. 58, 803 Mason, R. F! (1984). In Spin Labeling in Phurmucology (ed. J. L. Holtzman). Academic Press, New York, p. 87 Mattay, J., ed. (1989). Photoinduced Electron Transfer I , Top Curr. Chem. 156 Mattay, J., ed. (1990). Photoinduced Electron Transfer II, Top Curr. Chem. 158 Mattay, J.. ed. (1991). Photoinduced Electron Transfer III, Top Curr. Chem. 159 Mattay, J.. ed. (1992). Photoinduced Electron Transfer N , Top Curr. Chem. 163 Mattay, J., ed. (1993). Photoinduced Electron Transfer V , Top Curr. Chem. 168

140

L. EBERSON

McIntire, G. L., Blount, H. N., Stronks, H. J., Shetty, R. V. and Janzen, E. G. (1980). J. Phys. Chem. 84,916 Misik, V., Kirschenbaum, L. J. and Riesz, P. (1995). J. Phys. Chem. 99,5970 Moreno, S. N. J., Stolze, K., Janzen, E. G. and Mason, R. P. (1988). Arch. Biochem. Biophys. 265, 267 Mottley, C. and Mason, R. l? (1989). In Biological Magnetic Resonance (ed. L. J. Berliner and J. Reuben), vol. 8. Plenum Press, New York, p. 489 Mottley, C., Connor, H. D. and Mason, R. P. (1986). Biochem. Biophys. Res. Comrnun. 141, 622 Neta, P. (1976). Adv. Phys. Org. Chem. 12,223 Neta, P., Steenken, S., Janzen, R. V. and Shetty, R. V. (1980). J. Phys. Chem. 84, 532 Okhlobystina, L. V., lkyrikov, V. A., Shapiro, B. I., Syrkin, Ya. K. and Fainzil’berg, A. A. (1975). Bull. Acad. Sci. USSR, Ser. Chem. (Engl. Transl.) 2323 Ozaki, S. and Masui, M. (1978). Chem. Pharm. Bull (Tokyo) 26,1364 Parker, V. D., Reitstoen, B. and Tilset, M. (1989). J. Phys. Org. Chem. 2, 580 Perkins, M. J. (1980). Adv. Phys. Org. Chem. 17, 1 Pryor, W. A., Govindan, C. K. and Church, D. E (1982). J. Am. Chem. SOC. 104, 7563 Rehm, D. and Weller, A. (1969). Ber. Bunsenges. Phys. Chem. 73,834 Rehorek, D. and Janzen, E. G. (1984). Z. Chem. 24,441 Rehorek, D. and Janzen, E. G. (1985). J. Prakt. Chem. 327,705 Rehorek, D. and Janzen, E. G. (1986). Inorg. Chim. Acta 118, L29 Rehorek, D., Salvetter, J., Hantschmann, A., Hennig, H., Stasicka, Z. and Chodkowska, A. (1979). Inorg. Chim. Acta 37, L471 Rehorek, D., Dubose, C. M. and Janzen, E. G. (1984). Inorg. Chim. Acta 83, L7 Rehorek, D., Janzen, E. G. and Kotake, Y. (1991). Can. J. Chem. 69,1131 Reitstoen, B. and Parker, V. D. (1991). J. Am. Chem. SOC. 113,6954 Reszka, K. and Chignell, C. F. (1995). Photochem. Photobiol. 61, 269 Reynolds, W. L. and Lumry, R. W. (1966). Mechanisms of Electron Transfer. Ronald Press, New York Rumyantseva, G. V., Kennedy, C. H. and Mason, R. P. (1991). J. Biol. Chem. 266, 21 422 Sang, H., Janzen, E. G. and Poyer, J. L. (1996). J. Chem. Soc., Perkin Trans. 2,1183 SavCant, J.-M. (1990). Adv. Phys. Org. Chem. 2 6 , l Sawyer, D. T., Kang, C., Llobet, A. and Redman, C. (1993). J. Am. Chem. SOC. 115, 5817 Sayo, H., Ozaki, S. and Masui, M. (1973). Chem. Pharm. Bull. (Tokyo) 21, 1988 Schuster, G. B. (1982). Adv. Phys. Org. Chem. 18, 187 Simonet, J., El Badre, M. C., Emir, B., Boujlel, K. and Kossai’, R. (1990). J. Electroanal. Chem. 279,205 Sosonkin, I. M., Belevskii, V. N., Strogov, G. N., Domarev, A. N. and Yarkov, S. l? (1982). J. Org. Chem. (USSR), Engl. Transl. 18, 1313 Sridhar, R., Beaumont, P. C. and Powers, E. L. (1986). J. Radioanalyt. Nucl. Chem. 101, 227 Stolze, K., Moreno, S. N. J. and Mason, R. P. (1989). J. Inorg. Biochem. 37,45 Siimmerman, W. and Deffner, U. (1975). Tetrahedron 31,593 Symons, M. C. R. (1997). Acta Chem. Scand. 51, 127 Terabe, S. and Konaka, R. (1972). J. Chem. SOC., Perkin Trans. 2,2163 Thomas, C. E., Ohlweiler, D. F., Carr, A. A., Nieduzak, T. R., Hay, D. A., Adams, G., Vaz, R. and Bernotas, R. C. (1996). J. Biol. Chem. 271,3097 Tomasi, A. and Iannone, A. (1993). Biological Magneric Resonance (ed. L. J. Berliner and J. Reubern), voi. 13. Plenum Press, New York, p. 353

SPIN TRAPPING AND ELECTRON TRANSFER

141

Tsuge, O., Urano, S. and Iwasaki, T. (1980). Bull. Chem. SOC.Jpn. 53,485 Walter, T. H., Bancroft, E. E., McIntire, G. L., Davis, E. R., Gierasch, L. M., Blount, H. N., Stronks, H. J. and Janzen, E. G. (1982). Can. J. Chem. 60, 1621 Wang, H., Zheng, G. and Parker, V. D. (1995). Acta Chem. Scand. 49,311 Workentin, M. S., Johnston, L. J., Wayner, D. D. M. and Parker, V D. (1994a). J. Am. Chem. SOC.116,8729 Workentin, M. S., Schepp, N. I?, Johnston, L. J. and Wayner, D. D. M. (1994b). J. Am. Chem. SOC.116, 1141 Yoshida, K. (1984) Electrooxidation in Organic Chemistry. Wiley, New York Zubarev, V. E. and Brede, 0. (1994). J. Chem. SOC., Perkin Trans. 2,1821 Zubarev, V. E. and Brede, 0. (1995). J. Chem. Soc., Perkin Trans. 2, 2183 Zubarev, V. E., Meinert, R. and Brede, 0. (1992). Radiat. Phys. Chem. 39,281 Zweig, A.. Fischer, R. G. and Lancaster, J. E. (1980). J. Org. Chem. 45, 3597

This Page Intentionally Left Blank

Secondary Deuterium Kinetic Isotope Effects and Transition State Structure OLLEMATSSON

Department of Organic Chemistry, Uppsala University, Uppsala, Sweden AND

KENNETH C. WESTAWAY

Department of Chemistry and Biochemistry, Laurentian University, Sudbury, Ontario, Canada

1 Introduction 2 Secondary a-deuterium KIEs in SN reactions The origin of secondary a-deuterium KIEs Using secondary a-deuterium KIEs to determine the symmetry of SN2 transition states The effect of a change in substituent on the secondary a-deuterium KIEs The Menshutkin reaction The effect of ion-pairing on the secondary a-deuterium KIEs The effect of a change in solvent on the secondary a-deuterium KIEs Secondary a-deuterium KIEs and the effect of ionic strength on transition state structure 3 Secondary P-deuterium KIEs Secondary @-deuterium KIEs in carbocation SN reactions Secondary @-deuterium KIEs and the case for negative ion hyperconjugation Secondary P-deuterium KIEs due to hyperconjugation in carbene and radical reactions 4 Secondary deuterium KIEs and tunnelling Large secondary deuterium KIEs in hydride transfer reactions Tunnelling in the hydron transfer step of p-elimination reactions The magnitude of the secondary hydrogen KIE as a criterion for tunnelling Predictions of tunnelling criteria based on model calculations The relationship between the magnitude of secondary deuterium and tritium KIEs and the rule of the geometric mean Temperature dependence of secondary tritium KIEs Structural effects on the secondary KIEs in elimination reactions Kinetic complexity as an alternative to tunnelling 5 Remote secondary deuterium KIEs

144 146 146 164 171 174 190 195 197 197 197 202 210 211 213 216 217 220 223 228 229 231 23 1

143 ADVANCES IN PHYSICAL ORGANIC CHEMISTRY VOLUME 31 0065-3160/98 $30.00

Copyrighr 0 1998 Academic Press AIl righis of reprodurnon in any form resewed

0. MATSSON AND K. C. WESTAWAY

144

6 New methods for the accurate determination of secondary deuterium KIEs Methods based on measuring optical activity Determining multiple KIEs using natural abundance nmr spectroscopy The chromatographic isotopic separation method The double labelling liquid scintillation technique 7 Conclusion Acknowledgements References

234 234 238 240 241 242 242 243

1 Introduction

This chapter is concerned with secondary deuterium kinetic isotope effects. It will not review the theory of kinetic isotope effects, which has been covered extensively in other publications (Bigeleisen and Wolfsberg, 1958; Buddenbaum and Shiner, 1977a,b;Melander, 1960a; Melander and Saunders, 1980a; Van Hook, 1970). Nor is it intended to be a comprehensive review of the literature. Rather, it attempts to illustrate some of the important recent advances in the interpretation and uses of these kinetic isotope effects to elucidate reaction mechanisms. Other excellent reviews of secondary deuterium KIEs (Halevi, 1963; Shiner, 1970a; Kirsch, 1977; Hogg, 1978; Cleland, 1987; McLennan, 1987; Westaway, 1987a) cover the early developments in this field. Secondary a-deuterium kinetic isotope effects (KIEs) have been widely used to determine the mechanism of SN reactions and to elucidate the structure of their transition states (Shiner, 1970a; Westaway, 1987a). Some of the significant studies illustrating these principles are presented in this section. A secondary deuterium kinetic isotope effect is observed when substitution of a deuterium atom(s) for a hydrogen atom(s) in the substrate changes the rate constant but the bond to the deuterium atom is neither broken nor formed in the transition state of the rate-determining step of the reaction. Several types of secondary hydrogen-deuterium (deuterium) KIEs are found. They are characterized by the position of the deuterium relative to the reaction centre. Thus, a secondary a-deuterium KIE is observed when an a-hydrogen(s) is replaced by deuterium [equations (1) and (2)], where L is either hydrogen or deuterium.

-

Y-

slow

RCL-X

RCL++X-

RCL-Y

(1)

SECONDARY D-KINETIC ISOTOPE EFFECTS

145

ZPE

E

Fig. 1 (a) A reaction in which AZPE,,,,,,,,, is greater than AZPE~t,,n,,t,on,t,t,) and (kH/kD)a > 1.O. (b) A reaction in which AZPE(,,,,,, is less than AZPE(t,n,i~,n,t,t,) and (kHlkD)=< 1.O. Reproduced, with permission, from Smith and Westaway (1982).

When the deuterium is at the P-carbon as in equation (3), a secondary P-deuterium KIE is found. -+

Y-

slow

RCLCH2-X

RCL&H:+X-

RCL-Y

(3)

Since the bond to the isotopic atom is not formed or broken in the transition state of the rate-determining step of the reaction, the difference between the rate constant for the reaction of the undeuterated and deuterated substrates is usually small. As a result, secondary deuterium KIEs are usually close to unity, i.e. the maximum secondary deuterium KIE is 1.25 per deuterium (Shiner, 1970a) and most of these KIEs are less than 1.10 (Westaway, 1987a). Therefore, careful kinetic measurements with an error of approximately 1% in each rate constant or specially designed competitive methods are required to determine them with an acceptable degree of accuracy. As with primary isotope effects, the origin of secondary isotope effects is considered to be mainly due to changes in force constants upon going from reactants to the transition state of the rate-determining step of the reaction. For the most part, secondary isotope effects depend on the change in zero-point energy (ZPE). Smaller force constants for the bonds to the isotopic nuclei in the transition state than in the reactant lead to an isotope effect greater than unity (Fig. la). When the force constants for the bonds to the isotope are greater in the transition state than in the reactant, on the other hand, an isotope effect of less than unity is observed (Fig. lb).

0. MATSSON AND K. C. WESTAWAY

146

2 Secondary a-deuterium KIEs in SN reactions THE ORIGIN OF SECONDARY a-DEUTERIUM KIEs

Secondary a-deuterium KIEs are determined when hydrogen is replaced by deuterium at the a or reacting carbon. Until recently, the generally accepted view based on the experimental work of Streitwieser et at. (1958), Bartell (1961) and Kaplan and Thornton (1967) was that the secondary a-deuterium kinetic isotope effects are primarily determined by the changes that occur in the C,-H(D) out-of-plane bending vibrations in going from the reactants to the transition state. Thus, solvolysis reactions proceeding via a carbocation are expected to give large normal isotope effects; for example, a value of (kHlkD), of approximately 1.15 per D atom is expected when the reactant is an alkyl chloride because the C,-H(D) out-of-plane bending vibrations are of much higher energy in the tetrahedral reactant than they are in the almost trigonal planar transition state (Fig. la). However, since the magnitude of (kHlkD), depends on the changes that occur in the C,-H(D) out-of-plane bending vibrations when the reactant is converted into the transition state, the KIE should be leaving group-dependent. This has, in fact, been observed (Hartshorn and Shiner, 1972; Westaway, 1987b) and is illustrated in Table 1. It is worth noting that Murr and Donnelly (1970a,b) have demonstrated that the secondary a-deuterium KIE is only approximately 75% of the theoretical maximum kinetic isotope effect when the ionization ( k , ) step of the reaction (Scheme 1) is fully rate determining, i.e. when the reaction occurs via a limiting SN 1 mechanism (Shiner, 1970b; Westaway, 1987~).

Scheme 1 Table 1 The maximum secondary a-deuterium KIEs expected for SN1 reactions with various leaving groups at 25°C." Leaving group Iodide Bromide Chloride Ammonia Fluoride Arenesulphonate 'Data taken from Westaway (1987b).

Maximum expected (kH/kD)per a-D at 25°C 1.09 1.13 1.15 1.19 1.22 1.22

SECONDARY D-KINETIC ISOTOPE EFFECTS

147

Smaller secondary a-deuterium kinetic isotope effects are observed for reactions proceeding via the SN2 mechanism. Until recently, these small KIEs have been attributed to steric interference by the leaving group and/or the incoming nucleophile with the C,-H(D) out-of-plane bending vibrations of the trigonal bipyramidal s N 2 transition state, [l],where Nu is the nucleophile and LG is the leaving group in the SN2reaction. This means that the change in the energy of the C,-H(D) out-of plane bending vibrations on going to the

i

L

I 6 $Nu---C---LG 6

1;

LL [I1

transition state is small. As a result, small inverse or small normal secondary a-deuterium KIEs are observed (Fig. lb). In fact, small or inverse isotope = 0.95-1.04, are generally observed for the s N 2 effects, (kH/kD)/a.deu,erium reactions of primary substrates (Humski et al., 1974). Recently, this view of secondary a-deuterium KIEs has had to be modified in the light of results obtained from several different theoretical calculations which showed that the C,-H(D) stretching vibration contribution to the isotope effect was much more important than previously thought. The first indication that the original description of secondary a-deuterium KIEs was incorrect was published by Williams (1984), who used the degenerate displacement of methylammonium ion by ammonia (equation (4)) to model the compression effects in enzymatic methyl transfer ( s N 2 ) reactions. +

NH3 + CH3- NH3-

["NH3---CH3---NH311'd N H 3 + CH3-NH3 +

Sr

(4)

Williams calculated molar Gibbs free energies for the reactants, the encounter complexes and the transition states for both the uncatalysed and enzyme-catalysed reactions using ab initio methods at the 4-31G level of SCF-MO theory. The secondary a-deuterium (kCH,/kCD3), the r2CI'3Cand "C/14C KIEs were also calculated for these reactions using equation ( 5 ) .

The secondary a-deuterium KIEs calculated for the uncatalysed reaction were in the range found experimentally for other SN2methyl transfers. The calculated KIE was also analysed in terms of the zero-point energy (ZPE), the molecular mass-moment of inertia (MMI) and the excitation (EXC) contributions to the total isotope effect. The inverse KIE was found to arise from an

0. MATSSON AND K. C. WESTAWAY

148

Table 2 Analysis of ZPE factor contributing to the calculated secondary a-kca,lkcD, for the conversion of the encounter complex to the transition state in the uncatalysed SN2reaction between ammonia and methylammonium ion at 298 K." Vibration KIE

C,-H stretch

C,-H deformation

0.786

1.339

C,-H

rock

2.585

N-H rock

Other

Total

0.296

1.047

0.843

"Datataken from Williams (1984).

appreciable inverse ZPE factor (0.843), which was partially counterbalanced by a normal EXC factor (1.093) and an MMI factor of 0.991. This gave a total km,/kCDa of 0.913. The ZPE factor of 0.843 for the kCH3/kCD3 was further dissected into contributions from the different vibrational modes (see Table 2). The degenerate rocking modes of the CH3and the leaving NH3groups are highly coupled and their frequencies are very isotopically sensitive. The contribution of the C,-H bending modes to the KIE is normal and is almost cancelled by the inverse contribution from the N-H rocking modes. The most interesting and surprising finding, however, was the large and inverse contribution from the C,-H(D) stretching vibrations to the KIE. This was in direct contrast to the traditional view that the magnitude of the secondary a-deuterium KIEs in sN2 reactions was determined by changes in the C,-H(D) out-of plane bending vibrations. This discovery that the C,-H(D) stretching vibrations contributed significantly to the magnitude of secondary a-deuterium KIEs has been supported by the results of several other theoretical investigations of sN2 reactions by Truhlar and co-workers (Zhao et at., 1991; Viggiano et al., 1991; Hu and Truhlar, 1995) by Wolfe and Kim (1991), by Boyd et al. (1993), by Barnes and Williams (1993) and by Poirier et af. (1994). For instance, calculations on the sN2 reactions between microhydrated chloride ion and methyl chloride ( 6 ) using canonical variational transition state theory with semiclassical transmission coefficients by Truhlar and coworkers (Zhao et al., 1991) also suggested that the stretching vibrations are a significant contributor to the secondary a-deuterium KIE. Cl-(H,O),

+ CH3-CI*

CHS--Cl+ Cl*-(H20),

n = 0-2

(6)

These workers found that the largest contributions to the isotope effect were associated with the high- and the low-energy vibrations, i.e. the high-energy C,-H(D) stretching vibrations and the low-energy torsional vibrations, and that smaller contributions were obtained from the medium-energy C,-H(D) out-of plane bending vibrations around 1400 cm-I (Table 3).

SECONDARY D-KINETIC ISOTOPE EFFECTS

149

Table 3 The contribution to the observed secondary a-deuterium KIEs for the sN2 reactions between microhydrated chloride ion and methyl chloride at 300 K." ~

(kHfkD)",b

(kH1kD)vib

Total Reactant

(kHfkD)vib

c1Cl - (HzO)

0.76 0.53

Cl-(H,O)*

0.51

(kH1kD)vib

from the high-energy vibrations

from the medium-energy vibrations

from the low-energy vibrations

0.71 0.71 0.71

1.26 1.25 1.23

0.85 0.60 0.59

"Data taken from Zhao et al. (1991).

Table 4 The temperature dependence of the secondary a-deuterium KIEs for the sN2

reaction between chloride ion and methyl chloride."

TemperatureIK 207 300 538

Experimental

Theory

0.81 2 0.03 0.81 5 0.03 0.89 ? 0.06

0.88 0.93 0.97

"Data taken from Viggiano et al. (1991).

In another study of the gas-phase SN2reaction between chloride ion and methyl bromide, Truhlar and co-workers (Viggiano et al., 1991) determined the temperature dependence of the secondary a-deuterium KIE experimentally and computationally. Their results (Table 4) show that the KIE becomes less inverse by approximately 10% when the temperature increases from 207 to 538 K. The theoretical calculations using the canonical variational transition state theory method indicate that (i) the inverse secondary a-deuterium KIEs are due to the low-frequency (<300 cm-I) vibrations that are present in the transition state but not in the reactant, and to the high-energy C,-H(D) stretching vibrations, and (ii) that the temperature dependence of the KIE is primarily due to the effect of temperature on the high-energy C,-H(D) stretching vibrations. Hu and Truhlar (1995) also found the same source and temperature dependence for the secondary a-deuterium KIE in a theoretical investigation of the secondary a-deuterium KIEs for three halide ion-methyl halide SN2 reactions (7). Y-

+ CH3-X

-

[Y---CH,---X]'-

-

CH3-Y

(Y- = C1, X = Br; Y- = C1, X = I; Y-

= Br,

+ XX

=

I)

(7)

0. MATSSON AND K. C. WESTAWAY

150

Table 5 The experimental and theoretical secondary a-deuterium KIEs and the components of the vibrational contribution to these KIEs for three SN2 reactions between halide ion and methyl halides at 300 K.” Total (kHlkD)a

Reaction

Expt. Theory

C1-

+ CH3Br

0.80‘ 0.8g

C1-

+ CH31

0.84

Br-

+ CH31

0.76

0.91 0.94 0.95 0.90 0.91 0.93 0.93 0.96

Level MP2/PDZ+ MP2/PTZ+ ME/PTZ++ MP2/PDZ+ MP2/PTZ+ MP2/PDZ+ MP2/PTZ+ MP4/PTZ+

(kii/kD)vib

( k ~ / k ~ ) v i b (kH/kD)vib

highenergyb vibrations

mediumenergy‘ vibrations

lowenergyd vibrations

0.74 0.74 0.74 0.75 0.74 0.75 0.75 0.75

1.18 1.21 1.22 1.14 1.17 1.17 1.19 1.22

0.84 0.84 0.84 0.84 0.84 0.83 0.83 0.83

“Data taken from Hu and Truhlar (1995). ’Contributions by the C,-H(D) stretching vibrations with llh > 2000 cm-’. ‘Contributions by vibrations with 11A > 300 cm-’ but (2000 cm-’. ‘Contributions by vibrations with llh < 300 cm-l. eGronert et al. (1991). fViggiano et al. (1992).

These calculations were carried out using extended basis set calculations with electron correlation for the reactants and transition states and the KIEs were calculated using the canonical unified statistical theory. The secondary a-deuterium KIEs and the vibrational components of these KIEs are presented in Table 5. The first observation is that the calculated KIEs are in reasonably good agreement with the experimental values for the two SN2 reactions where chloride ion is the nucleophile. Surprisingly,the experimental and theoretical values are not close when bromide ion is the nucleophile, although the theory does predict that the KIE is inverse. The second important observation is that an analysis of the data shows that the major contributor to the KIE is the inverse contribution from the high-energy C,-H(D) stretching vibrations. The second largest contributor to the KIE is the normal contribution from the C,-H(D) bending vibrations, which is almost cancelled by the contribution from the low-energy vibrations. It is interesting to note (vide infra) that the variation in the C,-H(D) bending contribution is greater than that in the contributions from either the high- or the low-energy vibrations. In fact, the contributions from both the high- and the low-energy vibrations are effectively identical in all three reactions. The third observation is that increasing the temperature makes these inverse secondary a-deuterium KIEs more inverse. The more inverse KIE at the higher temperature is observed because the increase in the contribution due to the high-energy

SECONDARY D-KINETIC ISOTOPE EFFECTS

151

1.2

0.8

0.6 200

I

I

400

600

I

800

I

lo00

TIK

Fig. 2 The temperature dependence of the vibrational contributions to the secondary a-deuterium KIE for the SN2reaction between chloride ion and methyl bromide by (a) the high energy C,-H(D) stretching vibrations, (b) the C,-H(D) bending vibrations and (c) the low-energy transition state vibrations. Modified, with permission, from Hu and Truhlar (1995).

C,-H(D) stretching vibrations is slightly larger than the decrease in the contribution due to the C,-H(D) bending vibrations. A typical relationship is shown for the chloride ion-methyl bromide reaction in Fig. 2. Wolfe and Kim (1991) also reported that the magnitude of a secondary a-deuterium KIE is primarily determined by the changes that occur in the C,-H(D) stretching vibrations when the reactant is converted into the transition state. Wolfe and Kim calculated the transition state structures and the secondary a-deuterium KIEs for a series of identity SN2 reactions of methyl substrates [reaction (S)] at various levels of theory ranging from 4-31G to MP416-31 + G*//6-31+ G*. The KIEs were partitioned into two contributions, those from the C,-H(D) stretching vibrations and those from the C@-H(D) bending vibrations.

r c I*L

(L = H, D; X = X*

=

J

CI, F, C=N, N=C, OCH3, OF, and HCIC)

0. MATSSON AND K. C. WESTAWAY

152

Table 6 The secondary a-deuterium KIEs for seven identity SN2 reactions at 298 K."

x = x* c1 F OF C=N

N=C OCH3

C=CH

(k" IkD) a

Stretching vibration contribution

Bending vibration contribution

0.94 0.98 0.96 0.84 0.95 0.93 0.81 0.88 0.78 0.81 0.69 0.71 0.79 0.85 0.68 0.71

0.71 0.75 0.75 0.74 0.78 0.81 0.91 0.87 0.70 0.70 0.72 0.71 0.56 0.49 0.68 0.68

1.33 1.32 1.29 1.14 1.23 1.15 0.90 1.02 1.12 1.18 0.96 1.oo 1.42 1.72 0.99 1.05

Total Level 6-31 + G* 4-31G MP2/6-31 + G* 6-31 + G* 4-31G MP216-31 + G* 6-31 + G* . 4-31G 6-31 + G* 4-31G 6-31 + G* 4-31G 6-31 + G* 4-31G 6-31 + G* 4-31G

"Data taken from Wolfe and Kim (1991).

The results in Table 6 support the contention that the contribution to the isotope effect arising from the stretching vibrations is larger than the contribution from the bending vibrations at all levels of calculation. For instance, the stretching contribution to the KIE for the reaction between chloride ion and methyl chloride at the 6-31 + G * level, is (1.0011.41) X 100% = 0.71 or 41% inverse, while the bending contribution is (1.3311.00) X 100% = 33% normal. As a result, the total KIE is inverse. In fact, the inverse stretching contribution to the KIE ranges from 10% to 104%, while the bending vibration contributions, which are both inverse and normal, only range from zero to 72% and in every case the stretching component of the KIE is larger than the bending component of the KIE. Another surprising result of these calculations was that they suggested the relationship between the magnitude of the secondary a-deuterium KIE and transition state structure that had been based on experimental results (Streitwieser et al., 1958; Bartell, 1961; Kaplan and Thornton, 1967) was incorrect. Wolfe and Kim plotted the calculated secondary a-deuterium KIE at various levels of theory versus a looseness parameter, L , for the transition state. The L parameter was defined as the sum of the percentage extension of the C-X and the C-X* bonds on going from the reactant (product) to

SECONDARY D-KINETIC ISOTOPE EFFECTS

A1

I-

153

0 1

2m A 2

.3 2 0 5. .4

-

30

1'

r

0.6

n 5

0 4

6007 =

0.934

I

6m7 I

I

40.0

I

70.0

1 I

100

L Fig. 3 The calculated secondary a-deuterium KIEs versus the looseness parameter L for seven identity sN2 reactions. Calculations: 4-31G, 0 6-31 +G*, A M E / 6-31 + G*. Data from Table 6. Modified, with permission, from Wolfe and Kim (1991).

the transition state [equations (9),

(lo)].

Y0c-x = %c-x. = 100[(Lf_x- Lc-x)/Lc-x]

(10)

The slope of the calculated secondary a-deuterium KIE versus L plot (Fig. 3) was negative with a correlation coefficient of 0.934. The negative slope of the (kHlkD)oversus L plot suggested that the largest kinetic isotope effect is associated with the tightest rather than the loosest SN2transition state, as the experimental results had suggested. This was a very important conclusion because, if it was correct, all of the secondary a-deuterium KIEs that had been used to determine transition state structure (Axelsson et al., 1990;Harris et al., 1979; Lee, 1995; Shiner, 1970a; Westaway, 1987a; Westaway and Ali, 1979; Westaway and Waszczylo, 1982; Yamataka and Ando, 1979) had been interpreted incorrectly. Finally, because (i) it was concluded that the secondary a-deuterium KIE was determined primarily by the C,-H(D) stretching vibrations and (ii) the lengths (strengths) of the C,-H bonds were effectively the same in all of the identity SN2 transition states, Wolfe and Kim concluded that the magnitude of the isotope effect was determined by the length of the C,-H bonds in the substrate. This was an unusual conclusion because it meant that all the secondary a-deuterium isotope effects for the SN2 reactions of a particular substrate should be the same; for example, it meant that the KIE should be

0. MATSSON AND K. C. WESTAWAY

154

Table 7 The secondary a-deuterium KIEs for SN2 reactions with different nucleophiles. Substrate

-

Nucleophile

CH3OCH20*NO2

HzO AcO1-

NO;!

CH30CH2&(CH3)2CJ15 F-

c11-

CH&H&HzCH*Cl

Naf -SC&

c6W-

(kHlk&

Reference

1.24 1.05 1.31

Craze el al. (1978)

0.99 1.13 1.18

Knier and Jencks (1980)

1.03 1.13

Westaway and Lai (1988)

independent of the nucleophile used in the SN2 reaction. This surprising, and testable, conclusion has been found to be incorrect because several workers have shown that the secondary a-deuterium KIEs for sN2 reactions are dependent on the nucleophile and are, therefore, not determined solely by the C,-H bonds in the substrate. Some examples illustrating this are given in Table 7. Wolfe and Kim’s view of the origin of secondary a-deuterium KIEs has been challenged by two different groups. Barnes and Williams (1993) calculated the transition state structures and the secondary a-deuterium KIEs for the identity sN2 reactions between chloride ion and several substituted methyl chlorides (reaction (11)). ci-

+ R~R~CH-CI*-CI-CHR~R~+ ci*-

(11)

(R’ and R2 = H, CH, or CH30)

The calculations were performed at the semiempirical level using AM1 parametrization. The results for the methyl chloride reaction (Table 8) supported Williams’ earlier findings for the methylammonium ion-ammonia reaction (p. 147) and the results by Wolfe and Kim in that the inverse secondary a-deuterium KIE arose from an increase in the C,-H stretching force constants which accompanied the change from sp3 hybridization at the a-carbon in the reactant to the sp2-like hybridization in the transition state. More important, however, were the observations that (i) the total KIE is dominated by the vibrational (ZPE) component of the KIE with which it correlates linearly, and (ii) that the inverse contribution from the C,-H(D) stretching vibrations is almost constant for all the reactions. This suggests that the contribution from the other vibrations, i.e. the rest in Table 8, determines the magnitude of the KIE. In fact, Barnes and Williams stated that the

SECONDARY D-KINETIC ISOTOPE EFFECTS

155

Table 8 The AM1 calculated semiclassical secondary a-deuterium KIEs, the stretching and other contribution to the KIEs and the C-C1 transition state bond lengths for the identity SN2 reactions between chloride ion and substituted methyl

chlorides.“

H H H Me

Me

H Me Me0 Me Me0

0.982 1.015 1.048 1.090 1.156

0.935 0.977 1.027 1.114 1.196

0.959 0.954 0.954 0.952 0.960

0.975 1.024 1.077 1.170 1.246

2.154 2.202 2.252 2.270 2.325

“Data taken from Barnes and Williams (1993).

contribution to the KIE by the other, i.e. the bending, vibrations correlate with the total KIE. Therefore, the magnitude of the total KIE for these substituted methyl chloride reactions is determined by the bending vibration contribution and the results support the conventional view that the variation in these isotope effects is governed by changes in the bending vibrations. An examination of the data in Table 8 shows that the contribution due to “the rest” of the vibrations, including the C,-H bending vibrations, increases and becomes more normal as more alkyl groups or electron-donating substituents are attached to the a-carbon. The KIE will, therefore, vary from inverse to normal depending on the exact balance between the ZPE contributions from the stretching and the rest of the vibrational modes. More important, however, is that the increases in the contribution by “the rest” of the vibrations are directly related to an increase in the C,-C1 bond length in the transition state. Although Barnes and Williams only indicate how the C,-C1 bond in the SN2transition state is changed when the structure of the substrate is altered, the results suggest that the transition state is looser when a larger secondary a-deuterium KIE is observed. This is exactly the trend suggested by the early experimental work and is in direct contrast to the conclusions drawn by Wolfe and Kim. Poirier, Wang and Westaway (1994) also investigated the relationship between transition state structure and the magnitude of the secondary a-deuterium KIE in a theoretical study of the SN2reactions between methyl and ethyl chlorides and fluorides with several different nucleophiles (reaction (12)).

(L = H,D; Nu

= F,

C1 NH2, OH, SH, SCH,; X = F, Cl)

0. MATSSON AND K. C. WESTAWAY

156

The transition state structures were calculated at the HF/6-31 + G* level and the BEBOVIB-IV program was modified to accept the ab initio geometries and force constants and to calculate the secondary a-deuterium KIEs. The KIEs were separated into translational (nans), rotational (Rot), vibrational (Vib) and tunnelling contributions using equation (13) and (kH/kD)Vib was further factored into a stretching and a bending contribution to the KIE using equation (14). (kH/kD)Oaer is the KIE due to the low-energy (torsional) vibrations below 1000 cm-'.

The results in Table 9 show that, although the changes that occur in both the C,-H(D) out-of-plane bending vibrations and the C,-H(D) stretching vibrations when the reactant is converted into the transition state, i.e. that both the (kH/k&tretching and (kH/kD)Bending contributions to the KIE are large, the major contributor to the magnitude of the secondary a-deuterium KIE is the change in the C,-H(D) stretching vibrations. For instance, the inverse (kH/kD)Stretching in the methyl fluoride reactions is, on average, 32% whereas ranges from 2% inverse to 22% normal. To this extent, the the (kH/kD)Bending conclusions of this study support those of Wolfe and Kim (1991) and of Barnes Table 9 The HF/6-31 + G* secondary a-deuterium KIEs and vibrational contributions to the KIEs for the SN2 reactions between methyl fluorides and chlorides with different nucleophiles at 25°C."

Methyl 5uoride reactions F OH NH2

c1

SH

0.846 0.871 0.904 1.002 1.010

0.677 0.686 0.692 0.665 0.663

0.978 1.024 1.054 1.157 1.215

3.6930 3.7845 3.9055 4.2611 4.3088

0.705 0.714 0.716 0.680 0.690 0.701

1.098 1.128 1.112 1.178 1.197 1.221

4.2611 4.3714 4.4599 4.7880 4.8844 4.8713

Methyl chloride reactions

F OH NHZ Cl SH SCH3

0.882 0.907 0.897 0.935 0.953 0.953

"Data taken from Pokier er al. (1994).

SECONDARY D-KINETIC ISOTOPE EFFECTS

157

1.020

0.960

0

0

0

kdkn

0.900 -

0

0

0

0.840

r = 0.983 o r=

0.780

I 0.95

I

1.00

I

I

1.05

1.10

0.985

I

I

1.15

1.20

I 1.25

(kdbfaendms

Fig. 4 The total secondary a-deuterium KIE versus the bending vibration contribution to the KIE for the SN2reactions of methyl fluorides and chlorides with different nucleophiles at 25°C. The open circles are for the methyl chloride reactions and the solid circles are for the methyl fluoride reactions. Data from Pokier et al. (1994), with

permission. and Williams (1993). However, the stretching vibration contribution to the KIEs for all the reactions with the same leaving group are virtually identical for both the methyl fluoride and the methyl chloride reactions. The stretching vibration contribution to the KIE varies only by 0.03, i.e. from 0.66 to 0.69, for the methyl fluoride reactions and the corresponding contribution in the methyl chloride reactions only varies by 0.04, i.e. from 0.68 to 0.72. The bending vibration contribution to the KIE, on the other hand, changes by 24%, i.e. from 0.978 to 1.215, in the methyl fluoride reactions and by 12%, or from 1.098 to 1.221, in the methyl chloride reactions. These conclusions are consistent with those published by Barnes and Williams (1993) in that, although the stretching contribution to the KIE is large, it does not determine the magnitude of the KIE or the changes in the secondary a-deuterium KIEs with a change in nucleophile or leaving group. The magnitude of the total secondary a-deuterium KIE was shown to be and directly related to (kH/kD)&?nding (Fig. 4) and both the total (kH/kD)a (kH/kD)Bend,ng were shown to correlate with RTS, the nucleophile-leaving group distance in the transition state, with very high correlation coefficients ranging from 0.957 to 1.000 (Figs 5 and 6, respectively). These figures clearly show that the magnitude of a secondary a-deuterium KIE is directly related to the nudeophile-leaving group distance in the SN2transition state as the out-of-plane bending vibration model suggested.

0. MATSSON AND K. C. WESTAWAY

158 1.020

0.960

kdkD

0.900

0.840

L

0

o

r = 1.000 r = 0.973

4.50

4.75

0.780 3.50

3.75

4.00

4.25

5.00

Rrs

Fig. 5 The total secondary a-deuterium KIE versus R , for the SN2 reactions of methyl fluorides and chlorides with different nucleophiles at 25°C. The open circles are for the methyl chloride reactions and the solid circles are for the methyl fluoride reactions. Data from Poirier el al. (1994), with permission.

"280

I 0

0 0 0

0

0 0

0

0

a 0

o 0

3.50

r = 0.907 r = 0.957

I

I

I

1

1

3.75

4.00

4.25

4.50

4.75

5.00

Rrs

Fig. 6 The bending vibration contribution to the secondary a-deuterium KIE versus RTs for the SN2reactions of methyl fluorides and chlorides with different nucleophiles at 25°C. The open circles are for the methyl chloride reactions and the solid circles are for the methyl fluoride reactions. Data from Poirier et al. (1994), with permission.

SECONDARY D-KINETIC ISOTOPE EFFECTS

159

Table 10 The HF/6-31 + G* secondary a-deuterium KIEs and vibrational contributions to the KIEs for the SN2reactions between ethyl fluorides and chlorides with different nucleophiles at 25°C."

Nucleophile

(kH/kDL

(k/b)Stretching

(k/kD)Bendmg

RTS

Ethyl fluoride reactions F OH NH2

c1

0.914 0.935 0.940 1.034

0.809 0.809 0.816 0.792

0.995 1.008 1.030 1.137

3.7729 3.8725 3.8725 4.3991

0.791 0.792 0.784 0.781

1.145 1.168 1.223 1.214

4.3991 4.5204 4.9750 5.0486

Ethyl chloride reactions F OH CI SH

0.960 0.976 1.017 1.018

"Data taken from Poirier ef al. (1994).

Identical relationships were found to hold for the sN2 reactions of ethyl fluorides and ethyl chlorides with various nucleophiles (Table 10; Figs 7-9). The results of this study and that of Barnes and Williams on different reactions using a different level of theory, clearly demonstrate that, although the magnitude of the secondary a-deuterium KIE is determined by the changes in both the stretching and bending vibrations that occur in the C,-H(D) bonds when the reactant is converted into the transition state in an sN2 reaction, the KIEs may be either inverse or normal. The observed value depends on whether the inverse stretching contribution to the KIE is larger or smaller than the bending contribution, which can be either inverse or normal. The data in Tables 9 and 10 show that the stretching contribution is smaller (less inverse) for more complex substrates (e.g. ethyl) than for methyl substrates. A more important conclusion is that the stretching vibration contribution to the KIE is constant and determined by the leaving group while the bending contribution and, therefore, the total KIE is related to the nucleophile-leaving group distance in the sN2 transition state. Therefore, the magnitude of these KIEs is determined by the looseness of the transition state as the out-of-plane bending vibration model had suggested. The most recent contribution to this controversy is Glad and Jensen's (1997) theoretical investigation at the MP2/6-31 ++ G(d,p) ab initio level of the secondary a-deuterium KIEs for the identity sN2 reactions of three methyl halides (reaction (15)). The KIEs were calculated for various AR$-x where AR&-x = and RSC_xand R",, represent the length of the C-X bond in the transition state and the reactant, respectively.

0. MATSSON AND K. C. WESTAWAY

160 1.040

00

1.000 0

k~lk 0.960 ~-

0

0

0.920 0

0.880 0.950

1,000

1.050

0

r = 0.994

o

r = 0.993

1.100 1.150 (kdkD)Ber*ling

1.200

1.250

Fig. 7 The total secondary a-deuterium KIE versus the bending vibration contribution to the KIE for the sN2 reactions of ethyl fluorides and chlorides with different nucleophiles at 25°C. The open circles are for the ethyl chloride reactions and the solid circles are for the ethyl fluoride reactions. Data from Poirier et al. (1994), with permission. 1.040 OQ)

1.000

0

kH/kD 0.960

-

0

0

0

0.920 -

0

o

0

0.880 3.50

I

I

I

3.75

4.00

4.25

r = 0.992 r = 0.995

I

I

4.50

4.75

I

5.00

RTS

Fig. 8 The total secondary a-deuterium KIE versus RTSfor the SN2reactions of ethyl fluorides and chlorides with different nucleophiles at 25°C. The open circles are for the ethyl chloride reactions and the solid circles are for the ethyl fluoride reactions. Data from Poirier et al. (1994), with permission.

SECONDARY D-KINETIC ISOTOPE EFFECTS

161

1.280

0

1.200

-

0

0

8

1.040

a r=

a

0.996

o r = 0.976

a a

0.960 3.75

I

I

4.00

4.25

I

I

4.50

I

4.75

4.50

5.25

RTS

Fig. 9 The bending vibration contribution to the secondary a-deuterium KIE versus RTSfor the SN2 reactions of ethyl fluorides and chlorides with different nucleophiles at 25°C. The open circles are for the ethyl chloride reactions and the solid circles are for the ethyl fluoride reactions. Data from Poirier et al. (1994), with permission.

L,C-x*

+ x-

-1

-

L tI x --c--.x*

Llt

L,c-x+x*-

(15)

(L = H or D; X = X* = F, C1, Br)

The results (Fig. 10) clearly show that the magnitude of the KIE is related to the looseness of the transition state. Thus, the results of this study are in agreement with the conclusions based on the theoretical calculations by Barnes and Williams (1993) and Poirier et al. (1994). Another interesting observation is that the slopes of the plots of KIE versus elongation of the C-X bond on going to the transition state are almost identical for the three identity reactions. Glad and Jensen also calculated at the MP2/6-31+ + G(d,p) level the secondary a-deuterium KIEs starting from the separated reactants like Poirier et al. and the ion-dipole complexes like Wolfe and Kim. They found the same trend in the KIEs with looseness of the transition state in both cases, although the magnitude of the KIEs depended on the starting point for the calculation. It is worth noting, however, that the differences between the separated reactant and ion-dipole KIEs are small and not all in the same direction (Table 11). The KIEs starting from the separated reactant have a large, and

0. MATSSON AND K. C. WESTAWAY

162

1.6 1.4

0.4

I 0.0

I

I

I

0.1

0.2

0.3

I

I

I

0.4

0.5

0.6

[email protected]

Fig. 10 The secondary cr-deuteriumKIE for the identity SN2reactions between halide ions and methyl halides versus the elongation of the C-X bond on going from the reactant to the transition state. Data from Glad and Jensen (1997), modified, with

permission. constant, normal rotational component that is smaller than the inverse vibrational component which, therefore, controls the KIE (Table 11).For the KIEs calculated from the ion-dipole complex, the rotational component is almost unity and the KIE is effectively determined by the inverse vibrational contribution to the KIE. The important observation, however, is that the same trend with looseness, i.e. with A a X , is observed regardless of the starting point for the calculation. Several other findings are in agreement with the conclusions drawn by Poirier et af. For instance, an analysis of the individual contributions to the KIEs indicated that the variation in the KIE with AR& is primarily due to changes in the vibrational components to the KIE and not due to changes in either the translational or the rotational contributions, which are identical for all three reactions (Table 11).Moreover, an analysis of the contributions by the individual vibrations to the KIE (Fig. 11) shows that the increase in the KIE with AR&x is primarily due to the increase that occurs in the two C,-H(D) out-of-plane bending vibrations and that neither the contribution due to the two C,-H(D) stretching vibrations nor that due to the X-C-X symmetric bending vibration changes significantly with AR&. The contribution due to the C,-H(D) in-plane bending vibration, on the other hand, decreases slightly as the transition state becomes looser. However, this decrease is small with respect to the increase due to the C,-H(D)

SECONDARY D-KINETIC ISOTOPE EFFECTS

163

Table 11 The translational, rotational and vibrational contributions to the secondary a-deuterium KIEs for the three halide ion-methyl halide identity SN2 reactions calculated from the separated reactants and from the ion-dipole complexes.0 Ion-dipole complex

Separated reactants Halide

F Br C1

(kHlkD) (kdkD) ( ~ H I ~ D(kHlkD) ) Total Trans. Rot. Vib. 0.937 0.957 0.989

1.229 1.239 1.223

1.045 1.022 1.036

(kdkD) ( k ~ l k ~(kHlkD) ) (~HI~D) Total Trans. Rot. Vib.

0.730 0.757 0.781

1.OOO 1.OOO 1.000

0.906 0.978 0.986

0.997 1.000 0.999

0.909 0.979 0.987

"Data taken from Glad and Jensen (1997). C,-H out-of-planeumberella

1.2

c,-H in-plane bend

......I

**-&*-a; )*-

,.#*@'

. h

0.8

.*'\

. .................. ..-

0'

..'-.. ...I .-...... I

r,L(;::

.-a

*.*

-*

c,-H out-of-planebend

.................................... i, ...................ll.....U.l.l..l.l..........~

C,-H strekh

0.6 0.0

0.1

0.2

0.3

0.4

0.5

AR&

Fig. 11 The contribution of individual vibrations to the secondary a-deuterium KIE for the SN2reaction between fluoride ion and methyl fluoride as a function of the elongation of the C-F bond on going from the reactant to the transition state. Data from Glad and Jensen (1997), modified, with permission.

out-of-plane bending vibrations and has little effect on the magnitude of the KIE. This again, confirms that the out-of-plane bending model is correct for interpreting the secondary a-deuterium KIEs found for SN2 reactions. Finally, Lee (1995) has used cross-interaction constants to model the transition states for several SN2reactions. Lee concluded that the magnitude of the secondary a-deuterium KIE increases as the SN2 transition state

0. MATSSON AND K. C. WESTAWAY

164

becomes looser. Therefore, the evidence, from both experimental and theoretical studies, is overwhelmingly in favour of the idea that the magnitude of the secondary a-deuterium KIE in s N 2 reactions is directly related to the looseness of the transition state and that the out-of-plane bending vibration model is correct.

USING SECONDARY a-DEUTERIUM KIEs TO DETERMINE THE SYMMETRY OF S N 2 TRANSITION STATES

Although it appears that the magnitude of the secondary a-deuterium KIE is determined by the nucleophile-leaving group distance in the s N 2 transition state, Westaway et al. (1997) have recently suggested that this is not always the case. The secondary a-deuterium and primary nitrogen (leaving group) KIEs (k14/k1') found for the s N 2 reactions between several p-substituted thiophenoxide ions and benzyldimethylphenylammonium ion at 0°C in DMF containing a high concentration of sodium nitrate [equation (16); Table 121 do not change with Z .

zo S-

-

+ ChHSCH2-A(CH&ChH5

Z

S-CHZChHS

+ (CH3)ZNChHS

(16)

n o explanations for the invariant secondary a-deuterium and nitrogen KIEs are possible a priori. One is that the transition state structure does not change when the substituent on the nucleophile ( Z ) is altered. This suggestion was discounted because (i) no one has ever observed a reaction where a change in substituent does not change transition state structure and (ii) it is unreasonable to conclude that a change in nucleophile, which changes the rate constant by a factor of 6.4, would not alter the energy and, therefore, the Table 12 The primary nitrogen k'4/k'Sand secondary a-deuterium KIEs for the SN2 reactions between several p-substituted sodium thiophenoxides and benzyldimethylphenylammonium nitrate in DMF at 0°C."

p-Substituent on the thiophenoxide ion CH30 H

c1

k'4/k15

(kH1kd.a

1.0162 -C 0.0007' 1.0166 2 0.0004 1.0166 ? 0.0005

1.221 ? 0.012' 1.215 ? 0.011 1.215 2 0.013 -~

"Data taken from Westaway ef al. (1997). bThe errors are the standard deviations of the mean. T h e error in each isotope effect = l//~,,[(Ak,,)~+ (kH/kD)'X (AkD)*I1'*where Ak,and AkD are the standard deviations for the rate constants for the undeuterated and deuterated substrates, respectively.

SECONDARY D-KINETIC ISOTOPE EFFECTS

165

Table 13 The secondary a-deuterium and primary nitrogen KIEs and Hammett p values for the SN2 reaction between sodium thiophenoxide and benzyldimethylphenylammonium nitrate at different ionic strengths in DMF at 0°C." Ionic strength

0.904 0.64

(kHlkD)a

1.215 2 0.011* 1.179 2 0.007

P

Correlation coefficient

1.62 2 0.01 -1.76? 0.19

1.000 0.994

k'4/k'5

1.0166 2 0.0004' 1.0200 2 0.0007

-

"Data taken from Westaway et al. (1997). 'The errors in the isotope effect = 1/kD[(AkH)* -I (kH/kDfZ X (AkD)2]1'2where AkH and AkD are the standard deviations for the rate constants for the undeuterated and deuterated substrates, respectively. T h e errors are the standard deviations of the mean of five different measurements.

structure of the transition state. The second, and more likely, possibility is that the change in nucleophile alters the structure of the transition state but that the change in transition state structure does not affect the nitrogen or the secondary a-deuterium KIEs. The identical nitrogen KIEs were interpreted in the usual fashion, i.e. that all three reactions have identical amounts of C,-N bond rupture in the transition state. This is reasonable because the magnitude of a nitrogen leaving group kinetic isotope effect is primarily determined by the change in the vibrational energy of the C,-N bond in going from the reactant to the transition state (the amount of C,-N bond rupture in the transition state') (Saunders, 1975). A comparison of the nitrogen (leaving group) and secondary a-deuterium KIEs and the Hammett p-values for the benzyldimethylphenylammonium ion-thiophenoxide ion reactions (16) at two different ionic strengths (Table 13) showed that the higher ionic strength transition states, i.e. where the KIEs are constant, are reactant-like with reasonably short C,-N bonds and very long S-C, bonds. This is because the nitrogen kinetic isotope effect of 1.0166 2 0.0004 is only approximately one-third of the theoretical maximum KIE of 1.044 for a nitrogen leaving group (Maccoll, 1974; Buddenbaum and Shiner, 1977b). This implies that C,-N bond rupture is not well advanced in the SN2 transition state'. The secondary a-deuterium KIEs found for the higher-ionic strength reactions, on the other hand, are the largest (-11% per a-D) that have been found for an SN2reaction of a quaternary ammonium ion. Thus, these KIEs indicate that the transition state is very loose (Poirier et al., 'Some of the vibrational energy lost when the C,-N bond breaks in the transition state will be partially replaced by the increased vibrational energy associated with the strengthening of the N-C(Pheny1) bond, i.e. by conjugation between the dimethylamino group and the benzene ring in the transition state. Although this will reduce the KIE's ability to detect a change in the extent of C,-N bond rupture in the transition state (Westaway and Ali, 1979), reactions with identical kinetic isotope effects must have identical amounts of C,-N bond rupture in the transition state.

166

0. MATSSON AND K. C. WESTAWAY

Fig. 12 The C,-H(D) out-of-plane bending vibrations for an unsymmetrical SN2 transition state. Modified, with permission, from Westaway et al. (1997).

1994). Since the C,-N bond is short and the transition states are very loose, the S-C, bond must be very long in these transition states. If one assumes (i) that the transition states for the reactions with the constant nitrogen and secondary a-deuterium KIEs are reactant-like with very long S-C, and fairly short C,-N bonds, (ii) that the structure of the transition state is altered by the change in substituent on the nucleophile, and (iii) that changing the substituent in the nucleophile does not alter the amount of C,-N bond rupture in the transition state (the nitrogen KIE) significantly (see p. 165), the change in substituent must affect the S-C, bond in the transition state markedly. However, since the transition state is very reactant-like, the changes in the length of the S-C, transition state bond occur too far from the a-carbon in the unsymmetrical transition state (Fig. 12) to affect the C,-(H)D out-of-plane bending vibrations (the magnitude of the secondary a-deuterium KIE). As a result, the magnitude of the secondary a-deuterium KIE in these reactions is only determined by what happens to the shorter C,-N bond when the substituent is changed. Since the C,-N bond does not change when the substituent in the nucleophile is altered, neither the nitrogen nor the secondary a-deuterium kinetic isotope effects change when the substituent on the nucleophile is altered in these sN2 reactions. These conclusions are interesting because they are consistent with the predictions of the “bond strength hypothesis” (Westaway, 1993), which suggests that “there will be a significant change in the weaker reacting bond but little or no change in the stronger reacting bond in an sN2 transition state when a substituent in the nucleophile, the substrate, or the leaving group is altered”. Since the carbon-sulfur bond is weaker than the carbon-nitrogen bond in these sN2 reactions (Westaway, 1993),the “bond strength hypothesis” predicts that adding an electron-withdrawing substituent to the nucleophile should not affect the C,-N bond significantly but should lead to a significant change in the length of the weaker S-C, bond. These are the exact changes suggested on the basis of the isotope effects. Finally, because the transition states for these reactions are reactant-like with very long S-C, and short C,-N bonds, the stronger (C,-N) bond is shorter than the weaker (S-C,) bond in these reactant-like (unsymmetrical) sN2 transition states.

SECONDARY D-KINETIC ISOTOPE EFFECTS

167

The same phenomenon, i.e. that the secondary a-deuterium KIE is determined by the changes in only the shorter reacting bond in the SN2 transition state rather than by the nucleophile-leaving group distance, has been found in a completely different reaction system. Matsson, Westaway and co-workers (Matsson et al., 1996) used k"/k'4 carbon incoming nucleophile, chlorine leaving group (Hill and Fry, 1962), and k1*/k14a-carbon (Fry, 1970) KIEs to model the transition states for a series of SN2 reactions between p-substituted benzyl chlorides and labelled cyanide ion (reaction (17)).

(* = "C or 14C)

The chlorine leaving group and the incoming group 1'C/'4C KIEs (Table 14) decrease when a more electron-withdrawing substituent is added to the benzene ring of the substrate. A more detailed analysis of these isotope effects (Matsson et al., 1996) showed that the change in the C,-C1 bond (the change in the chlorine KIE with respect to the maximum chlorine KIE) is approximately six times greater than the change in the N=C-C, bond (the change in the k"lkI4 incoming nucleophile KIE with respect to the maximum k"/k14) when the substituent on the benzene ring is altered. This means that adding a more electron-withdrawing substituent to the benzene ring of the substrate shortens the C,-CI bond significantly but has little or no effect on the C,-C=N bond. These effects are depicted in Fig. 13. The secondary a-deuterium KIEs for these reactions (Table 14) decrease slightly when a more electron-withdrawing substituent is on the substrate, but they are not very sensitive to the change in substituent (Westaway et al., 1997).

Table 14 The chlorine 35C1137C1leaving group, the incoming nucleophile l1C/I4C,the secondary a-deuterium, and the l2C/l4Ca-carbon KIEs for the SN2reactions between p-substituted benzyl chlorides and cyanide ion in 20% aqueous DMSO at 30°C."

CH3 H CI

1.0079 2 0.0004 1.0072 2 0.0003 1.0060 2 0.0002

1.0104 2 0.0001 1.0105 +- 0.002 1.0070 +- 0.001 -

1.008 2 0.003d 1.011 +- 0.001 1.002 +- 0.003 ~

1.090 1.102 1.106 ~

"Data taken from Westaway et al. (1997). *Measured in 20% aqueous dioxane at 30°C (Hill and Fry, 1962). "Measured in 20% aqueous dioxane at 40°C. No error limits were given for these isotope effects (Fry, 1970). 'The error in the isotope effect = l/kD[(AkH)' + (kH/kD)'X (AkD)2]'" where AkH and AkD are the standard deviations for the rate constants for the undeuterated and deuterated substrates, respectively.

0. MATSSON AND K. C. WESTAWAY

168

H

Fig, 13 The relative structures for the SN2 transition states for the reactions of p-substituted benzyl chlorides with cyanide ion. Reproduced, with permission, from Matsson et al. (1996).

Normally, the small change observed in the secondary a-deuterium KIE would lead one to conclude that the nucleophile-leaving group distance in the transition state does not change significantly when the substituent on the benzene ring is altered (Poirier et al., 1994). However, the chlorine KIEs indicate that the C,-Cl bond shortens significantly and that the N=C-C, bond shortens slightly when a more electron-withdrawing substituent is added to the benzene ring (Matsson et al., 1996). Thus, the transition state becomes tighter (the nucleophile-leaving group distance in these s N 2 transition states decreases) when a more electron-withdrawing substituent is on the benzene ring of the substrate. If the magnitude of the secondary a-deuterium KIE were determined by the nucleophile-leaving group distance in the transition state, the secondary a-deuterium KIEs should decrease as a more electronwithdrawing p-substituent is added to the substrate. The only way to rationalize the almost constant secondary a-deuterium KIEs found for these reactions was to assume that these benzyl chloride-cyanide ion SN2 transition states are unsymmetrical and that only the shorter reacting bond determines the magnitude of the isotope effect. A comparison of the incoming group 11C/14Cand the secondary a-deuterium KIEs in Table 14 shows that the two KIEs parallel each other exactly. Obviously, the factors that affect the secondary a-deuterium KIEs also affect the incoming nucleophile carbon

SECONDARY D-KINETIC ISOTOPE EFFECTS

169

KIEs. This suggests that these transition states have a short N=C-C, bond and that the C,-C1 bonds in these transition states are long, i.e. that the transition states are unsymmetrical and product-like as shown in Fig. 13. As a result, the changes that occur in the C,-Cl bond when thep-substituent on the benzene ring is altered do not affect the C,-(H)D out-of-plane bending vibrations significantly,and the magnitude of the secondary a-deuterium KIE is only determined by the length of the shorter N=C-C, transition state bond. The 12C/14Ca-carbon KIEs (Fry, 1970) for these sN2 reactions (Table 14) increase as a more electron-withdrawing group is added to the benzene ring of the substrate. Since the maximum a-carbon KIE is observed when the sN2 transition state is symmetrical, i.e. when the strength of the a-carbon bond with the nucleophile is equal to that with the leaving group (Sims et al., 1972; Gray er al., 1979), the transition state for the p-chlorobenzyl chloride reaction is the most symmetrical. The chlorine and the incoming nucleophile carbon KIEs indicate that the C,-C1 bond length in the transition state decreases significantly while the N=C-C, bond length decreases slightly when a more electron-withdrawing substituent is added to the substrate. If these transition states are product-like, with short N=C-C, and long C,-C1 bonds, shortening the C,-C1 bond by adding a more electron-withdrawing substituent would make the C,-Cl bond more equal in strength to the short N=C-C, transition state bond and a more symmetrical transition state and larger 12C/'4C a-carbon KIE would be observed.' Again, the changes that occur in transition state structure are consistent with the "bond strength hypothesis" (Westaway, 1993). As expected, the weaker C,-Cl bond changes significantly and there is little or no change in the stronger N=C-C, bond when the substituent in the substrate is altered. It is also worth noting that the weaker C,-C1 reacting bond is long and the stronger N=C-C, reacting bond is short in these unsymmetrical transition states. This is interesting because the weaker S-C, bond was longer than the stronger C,-N bond in the unsymmetrical sN2 transition states in the p-substituted thiophenoxide ion-benzyldimethylphenylammonium ion reactions (vide supra). Thus, the change in substituent has the same effect on transition state structure in these two reactions: (i) the weaker reacting bond in the SN2 transition state is changed significantly by the substituent while there is little or no change in the stronger reacting bond and (ii) the stronger reacting bond is shorter than the weaker reacting bond in both these unsymmetrical sN2 transition states, even though the change in substituent has been made at different positions, in the nucleophile and in the substrate, respectively, in these two reactions. Product-like transition states would be expected to have inverse k"/kL4incoming nucleophile carbon KIEs. The authors were unable to explain why all the k'1/k'4KIEs for these reactions were normal, i.e. between 1.0105 and 1.0070.

170

0. MATSSON AND K. C. WESTAWAY

Finally, Westaway et al. (1997) suggested that the change in the secondary deuterium KIE with a change in substituent could be used to determine whether an sN2 transition state was symmetrical or unsymmetrical. The magnitude of a secondary a-deuterium KIE effect for an sN2 reaction can either be determined (i) by the nucleophile-leaving group distance in a symmetrical transition state or (ii) by the length of the shorter, stronger, reacting bond of an unsymmetrical transition state. A comparison of the secondary a-deuterium KIEs for several SN2 reactions of benzyl substrates (Ando et al., 1984; Ashan et al., 1980; Koshy and Robertson, 1974; Lee et al., 1991; Shiner et af., 1970; Vitullo et af., 1980; Westaway et al., 1998; Westaway, 1987c; Westaway and Ali, 1979; Westaway and Koerner, unpublished results; Westaway and Waszczylo, 1982) given in Table 15 shows there are two different types of substituent effects on the KIE. In some SN2 reactions, the secondary a-deuterium KIE decreases markedly (by between 2.4% and 12%) when a more electron-withdrawing substituent is added to the substrate (Ashan et af.,1980;Koshy and Robertson, 1974; Shiner et al., 1970; Vitullo et al., 1980; Westaway and Waszczylo, 1982) or the leaving group (Westaway and Ali, 1979). In other sN2 reactions, the secondary a-deuterium KIE is virtually independent of substituent, the change in the isotope effect with substituent being less than 1% (the average change is 05%) even when the substituent is changed from strongly electron-donating to strongly electron-withdrawing.It was proposed that the SN2reactions where the secondary a-deuterium KIE varies with the substituent have reasonably tight, symmetrical transition states. In these transition states, the C,-(H)D out-ofplane bending vibrations are affected by both the nucleophile and the leaving group, and the magnitude of the secondary a-deuterium KIE is determined by the nucleophile-leaving group distance in the sN2 transition state. The SN2 reactions where the secondary a-deuterium KIE is independent of substituent have unsymmetrical transition states where the strongest reacting bond is short. In these reactions, the C,-(H)D out-of-plane bending vibrations that determine the magnitude of the KIE are only affected by the nucleophile in the shortest reacting bond. It is worth noting that in all the reactions with an unsymmetrical transition state found to date, the stronger reacting bond is short and the weaker reacting bond is long. Since the “bond strength hypothesis” predicts that little or no change will occur in the stronger (shortest) reacting bond in an s N 2transition state when a substituent on the nucleophile, the leaving group, or the substrate is altered (Westaway, 1993), the secondary a-deuterium KIE will be insensitive to a change in substituent. Although the relationship between the secondary a-deuterium KIEs and transition state structure is different for the two types of transition state and interpreting secondary a-deuterium KIEs is, therefore, more difficult, it appears that the change in the KIE with substituent should be a good indicator for determining whether an sN2 transition state is symmetrical or unsymmetrical.

SECONDARY D-KINETIC ISOTOPE EFFECTS

171

THE EFFECT OF A CHANGE IN SUBSTITUENT ON THE SECONDARY (Y-DEUTERIUMKIE

Westaway and co-workers have used secondary a-deuterium and nitrogen leaving group KIEs to learn how a change in substituent affects the structure of SN2transition states (Westaway and Ali, 1979; Westaway and Waszczylo, 1982). In a recent study to determine how a change in nucleophile affected the structure of the transition state, neither the nitrogen leaving group nor the secondary a-deuterium KIEs for the SN2 reactions between several p substituted thiophenoxide ions and benzyldimethylphenylammonium ion (equation (18)) at 0°C in DMF (Table 12, p. 164) change significantlywhen the substituent is altered.

The reaction mixtures contained a high concentration of sodium nitrate to keep the ionic strength constant so that accurate rate constants could be determined (Pham, 1993; Pham and Westaway, 1996). Because it is unreasonable to conclude that a change in nucleophile, which changes the rate constant by a factor of 6.4, would not alter the energy and structure of the transition state, it was concluded that the change in nucleophile changes the transition state without changing either isotope effect. If the magnitude of the nitrogen KIE increases with the percentage of C,-N bond rupture in the s N 2 transition state (Saunders, 1975), then all three reactions have identical amounts of C,-N bond rupture in the transition state. If the s N 2 transition state were unsymmetrical and the weaker (S-C,) reacting bond were very long and the stronger (C,-N) reacting bond was short, the magnitude of the secondary a-deuterium KIE would be determined by the length of only the shorter C,-N bond because the sulfur nucleophile would be too far away to affect the C,-H(D) out-of-plane bending vibrations that determine the magnitude of the KIE (Westaway et al., 1997). Identical secondary adeuterium KIEs would be found in these reactions because the nitrogen KIEs indicate that the stronger C,-N bond does not change with substituent. In fact, C,-N bond rupture is not well advanced in these transition states because (i) the nitrogen KIE is only approximately one-third of the theoretical maximum nitrogen leaving group KIE of 1.044 (Maccoll, 1974) and (ii) the nitrogen KIEs found when thiophenoxide ion was the nucleophile in these reactions (k14/k15= 1.0166 2 0.0004) is significantly smaller than the kI4/kI5 = 1.0200 t 0.0007 found for the same reaction in DMF at an ionic strength of 0.64 (Pham and Westaway, 1996). The (kH/kD), = 1.22 ? 0.01 found for this reaction, on the other hand, is the largest that has been found for an SN2 reaction of a quaternary ammonium ion. Therefore, the transition

Table 15 Some secondary a-deuterium KIEs for SN2reactions with symmetrical and unsymmetrical transition states.

(k"kda p-Substituent (Z) AKIE CH30

Substratelnucleophile

CH3

H

c1

1.096

1.056

1.046

Br

NOz

(YO)

1.039

8.7

Reference

s N 2 reactions with symmetrical transition states

1.126

Z-BzCl"/C&&iZ-BzCL/H,O Z-BzCVH,O in 10% aq. CH3CN Z-BzOBs/H20 in 90% aq. EtOH Z-BzCVHzO Z-BzBr/S20;* Z-BzBr/N; Z-BzBr/OHB z & M e z a Z/C6H5S-

1.032 1.008 1.061

1.028

1.0% 1.059 1.124 1.092 1.063 1.024 0.984

1.207

1.179

1.151

5.6

1.215 1.011

1.215 1.002

0.6 Westaway et al. (1997) 0.99 Westaway et al. (1997)

1.004 1.032 0.996

6.3 5.1 12.0 3.1 3.1 2.8 4.4

Westaway and Waszczylo (1982) Ashan et al. (1980) Ashan et a1 (1980) Shiner et al. (1970) Koshy and Robertson (1974) Vitullo et al. (1980) Vitullo et al. (1980) Vitullo et al. (1980) Westaway and Ali (1979)

sN2 reactions with unsymmetrical transition states 1.221

B&e2C6H5/Z-C&SZ-BzCYCN-

1.008

BzOS0,OZ /

1.096

1.102

0.6

Lee et al. (1991)

1.098

1.095

0.3

Lee et al. (1991)

CH,OGNHz

BzOS020Z/

wz

3

h

3

h

v

m

h

v

m

h

3 s

2 s v;

2

2

0.962

ul

f

2

0.984

w

2

0.981

i U

0.3

Lee et al. (1991) u

u

0.3

i

2

0.974

i U

Lee et af. (1991) QJ

u

d

0.971

I-

0.993

2

M e O S 0 2 e Z1 0 N H 2

0.990

02" 3

i U

z

U

QJ

3

2

Lee et al. (1991) u QJ

9

0.9

Lee et al. (1991) u

0

m

0.953

0.3

i

3

v,

.oZN

&

E t O S 0 2 0 Z/ O N H 2

M

E t O S O 2 0 Z/ C H 3 0 0 N H 2

O*N

Secondary a-tritium KIEsb 1.9

9 T-4

Ando etal. (1984)

1.026

0.7

Ando e t a l . (1984)

m

1.048

8

1.042

3

3

1.033

1.055

3

1.061

1.033

Z-@Me2 "Bz = P-z-C&CH~.

bSecondary a-tritium K E s are much larger and more sensitive to a change in transition state structure than secondary cu-deuterium KIEs

174

0. MATSSON AND K. C. WESTAWAY

state is very loose with a long S-N distance and, since the C,-N bond is short, the S-C, bond must be very long. The conclusion is warranted because the (kHlkD), found in this study is significantly larger than the (kHlkD), = 1.179 ? 0.007 found for the same reaction at an ionic strength of 0.64 (Westaway and Ali, 1979). Also, the Hammett p = -1.62? 0.01 in the = 1.22 is smaller that the Hammett p value of reaction where (kH/kD)u , 1.179. Since a larger -1.76 ? 0.19 found in the reaction for which ( k H / k D )= p value is observed when the change in charge on going from the reactants to the transition state is larger, i.e. when there is more bond formation between the nucleophile and the a-carbon in the transition state, the reaction with the larger (kHlkD)amust have the longer S-C, transition state bond. The most reasonable explanation of the constant nitrogen and secondary a-deuterium KIEs found in these reactions is that changing the substituent in the nucleophile does not affect the amount of C,-N bond rupture in the transition state but changes the length of the S-C, transition state bond significantly. Unfortunately, in this case, the nitrogen and secondary adeuterium KIEs for these sN2 reactions do not indicate how the change in substituent in the nucleophile affects the length of the S-C, transition state bond. It is worth noting that these changes in transition state structure are consistent with the predictions of the “bond strength hypothesis” (Westaway, 1993), which suggests that there will be a significant change in the weaker S-C, reacting bond but little or no change in the stronger C,-N reacting bond in the sN2 transition state when a substituent in the nucleophile is altered. Also, they support the idea that the shortest bond in these sN2 transition states is the strongest bond.

THE MENSHUTKIN REACTION

The discussion in the section on the origin of secondary a-deuterium KIEs demonstrated the power of combining theoretical calculations with experimental measurements for determining the origin and developing an understanding of the secondary a-deuterium KIEs in sN2 reactions. One SN2 reaction where a combination of experiment and theory has been used to elucidate the structure of the transition state is the Menshutkin reaction (19) (Abboud et al., 1993)

In the first experimental study where KIEs were determined for Menshutkin reactions, Leffek and MacLean (1965) measured the secondary a-

SECONDARY D-KINETIC ISOTOPE EFFECTS

175

Table 16 Secondary a-deuterium KIEs found in Menshutkin sN2 reactions at 50°C." Substrate CH31 CHJ CH31 CH31 CHJ CH31 CH31 CHJ CH31 CH3Br CH~OTS GHSI C2H5Br n-C3H7Br (CH3)ZCHI (CH3)2CHBr 4-CH30PhCH2Br C6H5CH2Br 4-N02PhCH2Br

Nucleophile (GH5)3N (C3H7)3N (C,H,),N Pyridine 2,6-Dimethylpyridine Pyridine Pyridine Pyridine Pyridine Pyridine Pyridine Pyridine Pyridine Pyridine Pyridine Pyridine (GH5)3N (GH5)3N (GH5)3N

Solvent

(kHlkD)per a-D

0.957 0.963 0.964 0.972 0.957 0.959' 0.972' 0.950' 0.962' 0.975' 0.994' 0.983' 0.983' 0.980' 1.004' 0.958' 1.014d 0.993d 0.98gd

"Data taken from Leffek and MacLean (1965). *Data taken from Leffek and Matheson (1972a). 'Data taken from Leffek and Matheson (1972b). "Measured at 25°C by VituUo et al. (1980).

deuterium KIEs in Table 16 for the Menshutkin reactions between several different amines and methyl iodide in benzene at 50°C.They found the KIEs were all inverse and effectively independent of the structure of the amine even when the reactivity of the amine was altered drastically. Later, Leffek and co-workers found (Table 16) that the inverse secondary a-deuterium KIEs for the Menshutkin reactions between methyl substrates and pyridine varied slightly with the solvent (Leffek and Matheson, 1972a) and also with the leaving group (Leffek and Matheson, 1972b). Somewhat smaller (less inverse) secondary a-deuterium KIEs (Table 16) were observed for the more highly substituted substrates such as ethyl and isopropyl (Leffek and Matheson, 1972b) and p-substituted benzyl bromides (Vitullo et al., 1980) which have KIEs near unity. The larger secondary a-deuterium KIEs that were observed for the more highly substituted substrates were expected because these reactions are thought to have looser sN2 transition states. While it is difficult to rule out an inductive contribution to the secondary a-deuterium KIEs in these Menshutkin sN2 reactions, the steric origin of these KIEs is clearly indicated by the inverse secondary a-deuterium KIEs found in an extensive study of the sN2 Menshutkin reactions between substituted

0. MATSSON AND K. C. WESTAWAY

176

Table 17 Secondary deuterium KIEs for the Menshutkin reactions of deuterated pyridines and alkyl iodides in nitrobenzene at 25°C." Substrate CH31 CH31 CHJ CH31 CH31 CH31 GH5I GH5I (CH3)XHI

Nucleophile

4-Methyl-d,-p yridine 3-Methyl-d3-pyridine 2-Methyl-d3-pyridine 2,6-Dimethyl-d6-pyridine 4-Deuterop yridine Perdeuteropyridine 2-Methyl-d3-pyridine

2,6-Dimethyl-d6-pyridine 2-Methyl-d3-pyridine

kHlkD 0.999 0.991 0.971 0.913 0.988 0.970 0.965' 0.933' 0.945'

"Datataken from Brown and McDonald (1966), Brown et al. (1966). bMeasuredat 75°C. 'Measured at 100°C.

pyridines and alkyl iodides in nitrobenzene (reaction (20); Table 17) by Brown and co-workers (Brown and McDonald, 1966; Brown et at., 1966). An inductive KIE in these reactions would be inverse because substituting a more electron-donating deuterium for hydrogen would increase the electron density (nucleophilicity) of the pyridine nitrogen and make the deuterated pyridine react faster. However, an inverse secondary steric KIE would also be expected because the deuterated pyridine with the shorter C-D bonds would have a less sterically crowded (lower energy) sN2 transition state and react faster.

Although the inverse KIEs for the sN2 reactions in Table 17 can be rationalized in terms of a steric and/or an inductive KIE, several facts suggest that these KIEs are primarily caused by the increase in steric crowding that occurs when the reactants are converted into the SN2transition state. The virtual absence of a KIE when the CD3 group is placed in the 3 or 4 position of the pyridine ring and the much larger (more inverse) KIE in the 2-methyl-d3-pyridine-methyl iodide reaction (entries 1-3, Table 17) is consistent with a steric but not an inductive KIE. If inductive effects were predominant, the KIEs should increase in a regular manner as the deuterium is moved closer to the nitrogen. The much larger KIE found when the methyl group is in the ortho-position in the 2-methylpyridine-methyl iodide reaction is consistent with a secondary steric KIE, however, because the KIE does not

SECONDARY D-KINETIC ISOTOPE EFFECTS

177

change significantly until the methyl group is near the reaction centre and causes considerable steric crowding in the sN2 transition state. The KIE in the 2,6-dimethylpyridine-methyl iodide reaction is more than twice the KIE in the 2-methylpyridine-methyl iodide reaction. This is also consistent with a steric origin for the KIE because the 2,6-dimethylpyridine transition state must be much more sterically crowded than the 2-methylpyridine transition state. If the increase had been due to an inductive effect, the increase in the KIE in the 2,6-dimethylpyridine reaction should have been approximately twice the KIE for the 2-methylpyridine reaction, i.e. approximately 0.94 rather than the 0.91 that was observed. The steric rather than the inductive origin of the secondary deuterium KIE is also suggested because kH/kD = 0.994 per deuterium found in the perdeuteropyridine-methyl iodide reaction is smaller (less inverse) than the kH/kD = 0.988 per deuterium found for the 4-deuteropyridine reaction. A secondary inductive KIE should be more inverse when a deuterium is substituted for a hydrogen nearer the reaction centre, i.e. at the meta- or orthorather than at the para-position of the pyridine ring. Thus, if the KIE were inductive in origin, the KIE in the perdeuteropyridine reaction should be more inverse than that observed for the 4-deuteropyridine reaction. If the observed KIE were the result of a steric KIE, on the other hand, a less inverse KIE per deuterium could be found in the perdeuteropyridine reaction, i.e. a less inverse KIE per deuterium would be expected if there were little or no increase in steric hindrance around the C-H(D) bonds as the substrate was converted into the SN2 transition state. Since the KIE per D for the perdeuteropyridine reaction is less than 1%,the transition state must not be sterically crowded and the KIE must be steric in origin. Finally, the secondary deuterium KIEs observed in the reactions between 2-methyl-d3-pyridineand methyl-, ethyl- and isopropyl iodides (entries 3, 7 and 9, Table 17) are not consistent with an inductive KIE. If an inductive KIE were important in these reactions, one would expect the same KIE for all three reactions because the deuteriums would increase the nucleophilicity of the pyridine by the same amount in each reaction. The different KIEs for these three reactions are consistent with a steric KIE because the most inverse KIE is observed in the isopropyl iodide reaction, which would be expected to have the most crowded transition state, and the least inverse KIE is found in the methyl iodide reaction, where the transition state is the least crowded. Kaplan and Thornton (1967) also concluded that the secondary deuterium KIEs found in these sN2 reactions were not inductive in origin. Although the large inverse secondary deuterium KIE (kHlkD= 0.883 2 0.008) found in the C6H5N(CL& + CH3-OTs

-

CH3-6(CL3)2C6H5 OTs

(L = H, D)

178

0. MATSSON AND K. C. WESTAWAY

Menshutkin reaction between N,N-dimethyl-d6-anilineand methyl tosylate in nitrobenzene at 51.3"C(21) is consistent with an inductive KIE resulting from the increased electron density on the nitrogen (the nucleophilicity), a steric explanation was preferred for three reasons. First, the much smaller (less inverse) kHlkD= 0.952 in the corresponding reaction between dimethyl-d6-phenylphosphineand methyl tosylate is consistent with the lower steric crowding in the transition state of the phosphine reaction, i.e. the CD3 groups on the larger phosphorus atom are farther away from the methyl group of the substrate in the sN2 transition state and would not increase the steric crowding in the transition state to the same extent they do in the Menshutkin reaction. Secondly, a vibrational analysis of the reactants and products of the Menshutkin reaction showed that the C-H and C-D stretching frequencies were the same, whereas the bending force constants for these bonds were significantly different, in the reactants and products. This suggested that the KIE resulted from changes that occurred in the bending vibrations in going to the product-like transition state. This is indicative of a steric KIE because the C-H(D) out-of-plane bending vibrations are affected markedly by increases in steric crowding. The final, and strongest piece of evidence supporting the steric origin of the KIE, was that the secondary deuterium KIE calculated using the force constants found in the vibrational analysis of the reactants and products (effectively a steric explanation of the KIE) was identical to the experimental KIE. Further evidence against the inductive origin of the secondary deuterium KIEs in the Menshutkin reaction was provided by Leffek and Matheson (1971). These workers measured the temperature dependence of the secondary deuterium KIE for the sN2 reaction between N,N-dimethyl-d6-aniline and methyl tosylate in nitrobenzene. This reaction was chosen because Kaplan and Thornton had concluded that the incoming nucleophile secondary deuterium KIE in this reaction was mainly determined by steric effects on the C-H(D) out-of-plane bending vibrations in the transition state. Leffek and Matheson found that this reaction was enthalpy controlled, i.e. (AH: - AH;) was equal to -134 cal mol-I whereas T(A% - A S ) was only -45 cal mol-' at 300K, and concluded that the KIE was steric in origin. In fact, Leffek and Matheson concluded that the deuterium KIEs in all nonsolvolytic sN2 reactions were determined by changes in steric crowding in going from the reactants to the transition state. The results described above force one to conclude that secondary &-deuteriumKIEs in Menshutkin reactions are the result of changes in steric crowding that occur around the C,-H(D) bonds when the substrate is converted into the sN2 transition state. This conclusion is supported by the results from several recent theoretical investigations which demonstrate that the magnitude of these KIEs are related to the change in the C,-H(D) out-of-plane bending vibrations (steric crowding) when the reactants are converted into the SN2transition state (vide supra). Finally, it is worth noting

SECONDARY D-KINETIC ISOTOPE EFFECTS

179

Table 18 The secondary a-deuterium KIEs for the Menshutkin reaction between 3,s-disubstitutedpyridines and methyl iodide in 2-nitropropaneat 25°C."

3-x

5-Y

(kH/kD)u

CH3 CH3 H

CH3

0.908

H H

0.851 0.850

C1

H

0.835

C1

0.810

c1 "Data taken Gom Hams et al. (1981).

that the inverse values found for these KIEs suggest that the transition states for these Menshutkin reactions are reasonably tight, with N-C, bond formation greater than C,-LG bond rupture. In a more recent study, Harris and co-workers (Harris ef al., 1981) found that the secondary a-deuterium KIEs (Table 18) were larger (more inverse) when a poorer nucleophile was used in the SN2 reactions (22) between 3,5-disubstituted pyridines and methyl iodide in 2-nitropropane at 25°C.

c

ON+CH3-I

Y

-

d0

CH3-+

1-

(22)

Y

The important observation is that all of the isotope effects are large and inverse. Therefore, the transition states in these reactions must be very crowded, i.e. the C,-H(D) out-of-plane bending vibrations in the transition state must be high energy (Poirier et af., 1994). As a result, these workers concluded that nitrogen-a-carbon bond formation is more advanced than a-carbon-iodine bond rupture in the transition state. It is interesting, however, that in spite of the small secondary a-deuterium KIEs, these authors concluded that the N-C, bond formation is only approximately 30% complete in the transition state. Other types of KIEs have been measured in an attempt to determine the structure of the transition states of Menshutkin reactions. For example, Bourns and Hayes (quoted in Smith and Westaway, 1982) and Kurz and co-workers (Kurz et al., 1986a,b) found very small incoming nucleophile nitrogen KIEs in Menshutkin reactions (Table 19). These very small isotope effects, which are only slightly larger than the error in the KIEs, are not affected significantly by a change in the leaving group, the solvent, or even the substrate. It is important to note, however, that these very small incoming nucleophile nitrogen KIEs suggest that the transition state is early with only a small amount of N-C,

0. MATSSON AND K. C. WESTAWAY

180

Table 19 The incoming nucleophile nitrogen KIEs found for Menshutkin reactions with various amines in several solvents. Nucleophile 4-Methylpyridine Pyridine

3-Acetylpyridine 2,6-Dimethylpyridine Quinuclidine

N,N-Dimethyl-4-toluidine

N,N-Dimethyl-4-bromoaniline N,N-Dimet hyl-4-methoxyaniline Et3N Me3N

Substrate

Solvent

CH3OMs CH30Tf CH30Ms CH30Ts CH3OTf CH~OTS CH30Tf CH30Ts CH30Tf CH30Tf CH30Tf CH3Cl CH3Br CH31 CH3Th+ CH3Th+ CH30Ms CH30Ms CH30Ts CH30Tf CH30Ms CH~OMS CH~OMS CH30Ms

HzO MeCN H20 H20 H20 H20 H*0 H20 H20 MeCN DCE HzO H20 H20 H20 MeCN H20 H20 H20 MeCN H20 MeCN MeCN (wet) MeCN , Me2C=0 BZOSO2C6H5 Me2C=0 BZOS02C6H4Cl Me,C=O 3-BrBzOS0,C6H4Cl Me2C=0 B Z O S O ~ C ~ H ~ C I Me,C=O BzOSO~C~H,CI Me2C=0 CH31 Benzene CH3CH2I Benzene CH3CH2Br Benzene CH3CH2Br Benzene

kl4/k15a 0.9969 0.9937 0.9971 0.9978 0.9965 0.9978 0.9965 0.9978 0.9965 0.9946 0.9942 0.9960 0.9972 0.9950 0.9976 0.9964 0.9972 0.9978 0.9977 0.9941 1.om2 0.9962 1.0024' 1.0027' 1.0028' 1.0028' 1.0027' 1.0020" 1.0038' 1.0019' 1.0009d 1.0015' 0.9991" 0.9994

"Measured at 25°C. Data taken from Kurz et al. (1986a,b). bMeasured at 50°C. Data taken from Kurz et al. (1986a.b). 'Measured at 35°C. p-Substituents in substrate. Data taken from Ando et al. (1985). dMeasured at 8°C. Data taken from Smith and Westaway (1982). Weasured at 25°C. Data taken from Smith and Westaway (1982). fMeasured at 6°C. Data taken from Smith and Westaway (1982).

bond formation. In fact, BEBOVIB-IV calculations (Sims and Lewis, 1985), which assume total constant bonding at the a-carbon in the transition state, suggest that the nitrogen-a-carbon bond order ,in these transition states is between 0.2 and 0.3. This is in excellent agreement with the results from

SECONDARY D-KINETIC ISOTOPE EFFECTS

181

Harris's study, which also concluded that nitrogen-a-carbon bond order in these transition states is between 0.2 and 0.3. In another study, Paneth and O'Leary (1991) measured the incoming nucleophile nitrogen and the secondary a-deuterium KIEs for the Menshutkin reaction between N,N-dimethyl-p-toluidine and methyl iodide in methanol at 25°C. They found a very small incoming nucleophile nitrogen KIE of 1.0019 -+ 0.0001 in good agreement with the KIEs presented in Table 19. The secondary a-deuterium KIE for this reaction was 0.83 2 0.04, in good agreement with the KIEs reported by Harris et al. (1981). A large a-carbon12karbon-14 KIE = 1.12 has been found for this reaction (Buist and Bender, 1958). On the basis of the a-carbon and nitrogen KIEs and AM1 calculations, Paneth and O'Leary concluded that the transition state was symmetrical or slightly late with nitrogen-a-carbon bond formation more advanced than a-carbon-iodide bond rupture. However, they noted that the secondary a-deuterium KIE suggested a late transition state. Matsson and co-workers (Axelsson et al., 1987) measured the carbon1lkarbon-14 a-carbon KIEs for the Menshutkin reaction between N,Ndimethyl-p-toluidine and labelled methyl iodide in methanol at 30°C. The "C-labelled methyl iodide required for this study was prepared in three steps [reaction (23)] from the "C atoms produced in a cyclotron (LAngstrom et al., 1987).

The 14C-labelledmethyl iodide used to measure k"/kl4 was commercially available. The very large value of k'1/k'4 = 1.202 2 0.008 is almost twice the magnitude of k12/k14of 1.12 2 0.01 measured for the same reaction at 48.5"C (Buist and Bender, 1958). The value of the k"/kl4 can be estimated from the observed k"ik14 using the equation In k"/k14/ln kl2lkI4= 1.6 (Axelsson et al. 1991). This gives a k"/k14 = 1.199, which is in excellent agreement with the observed k"/kL4= 1.0202 and it is safe to conclude that this Menshutkin reaction has a very large a-carbon KIE and a reasonably symmetrical transition state (Saunders, 1975; Sims et al., 1972). The BEBOVIB-IV calculations (Sims and Lewis, 1985) that were used to model the transition state for this reaction (Axelsson er al., 1991) suggested that the transition state is early with a nitrogen-a-carbon bond order of approximately 0.3 and an a-carbon-iodide bond order of approximately 0.7. This is in good agreement with the conclusions reached by Harris and co-workers (1981). Large (near the theoretical maximum) a-carbon KIEs for the Menshutkin reactions between 3,5-disubstituted pyridines and methyl iodide [reaction (22); Table 201 have also been reported by Yamataka and co-workers (Ando et al., 1987). Although the a-carbon KIEs increase slightly as more electronwithdrawing substituents are added to the nucleophile, they are all large and

0. MATSSON AND K. C. WESTAWAY

182

Table 20 The a-carbon-12lcarbon-13 KIEs for the Menshutkin reaction between 3,5-disubstituted pyridines and methyl iodide in 2-nitropropane at 25°C." 3-x

5-Y

(k'2/k'3),

CH3 H H H

1.063? 0.004 1.062 2 0.002 1.066 ? 0.005 1.074 ? 0.002 1.076 2 0.016

c1

"Data taken from Ando et al. (1987).

effectively constant for a wide range of nucleophiles. For example, the KIE only changes by 0.013 when the nucleophile is changed from 3,5-dimethylpyridine to 3,5-dichloropyridine and the rate constant decreases by 340 times. This suggests that the transition state for the Menshutkin reaction is not very susceptible to changes in the structure of the reactants. One way to rationalize the secondary a-deuterium measured by Harris et al. (Table 18, p. 179) with the a-carbon KIEs found by Ando et al. for these reactions is to assume that all the transition states are symmetrical with short N-C, and C,-I bonds. Ando et al. (1984) measured both the a-carbon k1*/kI4and the secondary a-hydrogen-tritium KIEs for the Menshutkin reactions between m- and p-substituted N,N-dimethylanilines and substituted benzyl benzenesulfonates in acetone at 35°C [reaction (24)]. They found large (near the theoretical maximum) carbon-12/carbon-14and small secondary a-tritium KIEs for these reactions (Table 21). These a-carbon KIEs agree, in general, with the large a-carbon KIEs reported by other workers (vide supra). It is important to note, however, that these carbon KIEs go through a maximum when the leaving group is changed. In fact, this was the first experimental evidence that the a-carbon KIEs in sN2 reactions pass through a maximum as the theoretical calculations suggest (Saunders, 1975; Sims et al., 1972).

The secondary a-tritium KIEs in these reactions, on the other hand, are small and normal. Benzyl substrates have looser sN2 transition states than methyl substrates (vide infra) and thus, the benzyl substrate reactions would be expected to have slightly larger (normal) KIEs rather than the inverse KIEs

SECONDARY D-KINETIC ISOTOPE EFFECTS

183

Table 21 The a-carbon-12/carbon-14and secondary a-hydrogen-tritium KIEs for the SN2 reactions between Y-substituted N,N-dimethylanilines and Z-substituted benzyl X-substituted benzenesulfonates in acetone at 35°C."

Z Y

X

p-CH30 p-CH30 p-CH30 P-CHS P-CHs P-CH3 P-CH3 P-CH3 H H H m-CH, p-Br m- N02

= m-Br

Z=H

kl2/kl46

(kHWb

klZjkl4b

(kHWb

p-c1 H P-CH3

1.130

1.033

1.142

1.061

m-NO,

1.151 1.148 1.137 1.141 1.141

1.140 1.148 1.119 1.149 1.162 1.156 1.147 1.158 1.143 1.135

1.056 1.055 1.043 1.033 1.035

1.139 1.127

1.048

p-c1 H P-CH3 P-CH~O m-N02 p-c1 H p-c1 p-c1

m-NO2

1.129 1.117

1.041 1.026 1.030 1.031

1.033

1.042

"Data taken from Ando et al. (1984). 'The errors in the / ~ ' * / kare ' ~ between 20.003 and 50.005 while those for the ( / C ~ / / C ~ range )~ from 20.008 to 50.012.

found for the Menshutkin reactions of methyl substrates. Thus, these tritium KIEs are in general agreement with the inverse secondary a-deuterium KIEs that have been reported for other Menshutkin reactions and suggest a reasonably tight transition state. It is worth noting that the tritium KIEs for the m-bromobenzyl benzenesulfonate reactions are smaller than those for the benzyl benzenesulfonates. This is expected because the transition state is invariably tighter, with a shorter nucleophile-leaving group distance, when a more electron-withdrawing substituent is on the phenyl ring on the a-carbon (Westaway and Waszczylo, 1982; Westaway et al., 1997). Finally, although the transition states in these Menshutkin reactions appear to be slightly looser than those found for the methyl substrates, the KIEs are consistent with a fairly symmetrical transition state with significant nitrogen-a-carbon bond formation and less a-carbon-oxygen bond rupture. The newest type of KIE that has been used to characterize the transition state of the Menshutkin reaction is a secondary incoming nucleophile deuterium KIE. These KIEs are measured by using primary amines labelled with deuterium at the nitrogen of the nucleophile [reaction (25)]. In fact, Lee and collaborators (Lee et al., 1991) measured both the secondary a-deuterium and the secondary incoming nucleophile deuterium KIEs for four different Menshutkin reactions (Table 22). The secondary

0. MATSSON AND K. C. WESTAWAY

184

(L = H, D)

a-deuterium KIEs for the benzyl benzenesulfonate reactions are large and normal. This indicates that these reactions have a loose transition state with long nucleophile-a-carbon and/or a-carbon-leaving group bonds. The methyl and ethyl substrate reactions, on the other hand, have inverse secondary a-deuterium isotope effects like those found for the Menshutkin reactions of Table 22 The secondary a-deuterium and secondary incoming nucleophile deuterium KIEs found for the SN2 reactions between p-substituted anilines and benzylamines with benzyl, methyl and ethyl p-substituted benzenesulfonates in acetonitrile at 30°C." Substituent on the nucleophile

p-Substituent on the leaving group

(kHlkD)a

(kHlkD)Nuci ~

Benzyl p-substituted benzenesulfonates with p-substituted anilines m-NO, p-CH30

m-NO, p-CH30

1.089 2 0.005 1.096 2 0.009 1.095 ? 0.010 1.102 ? 0.010

CH3 CH3

NO, NO,

0.973b 0.955 0.951 0.898

Benzyl p-substituted benzenesulfonates with p-substituted benzylamines m-NO, p-CH3O

m-NO, p-CH30

-

CH3 CH3

NO2 NO,

0.966' 0.952 0.953 0.940

Methyl p-substituted benzenesulfonates with p-substituted anilines m-NO, p-CH30

m-NO, p-CH3O

0.971 5 0.009 0.990 ? 0.008 0.974 t 0.007 0.993 5 0.007

CH3 CH3

NO, NO,

0.963 2 0.009 0.978 2 0.008' 0.968 t 0.009' 0.984 2 0.007'

Ethyl p-substituted benzenesulfonates with p-substituted anilines m-NO, p-CH30

m-NO, p-CH3O

0.963 ? 0.009 0.978 ? 0.008 0.968 ? 0.009 0.984 2 0.007

CH3 CH3

NO, NO,

0.851b" 0.862'*' 0.8586.' 0.869'"

~~

"Datataken from Lee et al. (1991). bThe authors did not give error limits for these isotope effects They imply that the error is less than 1%. 'At 65°C.

SECONDARY D-KINETIC ISOTOPE EFFECTS

185

other methyl substrates (see Table 16,17; p. 175,176). This indicates that these reactions have tight transition states with short nucleophile-a-carbon and/or a-carbon-leaving group bonds. The secondary incoming nucleophile deuterium KIEs are all inverse. This is because the energy of both the N-H(D) bending and stretching vibrations in the transition state increases as the steric crowding increases (the nitrogen-a-carbon bond forms). Therefore, when nitrogen-a-carbon bond formation is more complete in the transition state and steric crowding around the N-H(D) bonds is greater, the KIE is more inverse. Thus, these new isotope effects are useful because they indicate the degree of nitrogen-acarbon bond formation in the transition state. Because the incoming nucleophile secondary deuterium KIEs are less inverse for the benzyl substrates, the authors concluded that the transition states for the Menshutkin reactions of the benzyl substrates were early (reactant-like) with nitrogen-a-carbon bond formation lagging behind acarbon-oxygen bond rupture. The transition states for the Menshutkin reactions of methyl and ethyl substrates, on the other hand, are tight (product-like) with nitrogen-a-carbon bond formation greater than acarbon+xygen bond rupture. Finally, it is worth noting that different substituent effects have been found for the Menshutkin reactions of the benzyl and the methyl and ethyl substrates. For the benzyl substrates, changing to a better nucleophile, i.e. changing the substituent on the nucleophile from the rn-nitro to a p-methoxy substituent, leads to a later, more productlike transition state with a more inverse secondary incoming nucleophile deuterium KIE. However, the same change in nucleophile in the reactions with the methyl and ethyl substrates leads to an earlier transition state and a smaller (less inverse) secondary incoming nucleophile deuterium KIE. There is no explanation as yet for the different substituent effects found in these reactions. Recently, Paneth and co-workers published an extensive experimental (Szylhabel-Godala et al., 1996) and theoretical (Gawlita et al., 1996) study of the Menshutkin reaction. The experimental study involved measuring the incoming nucleophile nitrogen and the secondary a-deuterium KIEs for the Menshutkin reactions between three p-substituted N,N-dimethylanilines and methyl iodide in ethanol at 25°C. The nitrogen KIEs in Table 23 are all small, in agreement with those measured by Kurz and co-workers (see Table 19, p. 180), but decrease when a more electron-withdrawing substituent is on the nucleophile. Since the magnitude of an incoming nucleophile KIE decreases as the amount of N-C, bond formation increases in the transition state (Matsson et al., 1996), N-C, bond formation is greater when the more electron-withdrawing acetyl group is on the nucleophile. The secondary cY-deuteriumKIEs for these reactions, on the other hand, become smaller (less inverse) when a more electron-withdrawing group is added to the nucleophile. This suggests that the transition state is looser (the nucleophile-leaving group

0. MATSSON AND K. C. WESTAWAY

186

Table 23 The secondary cY-deuteriumand incoming nucleophile nitrogen KIEs found for the SN2reactions between p-substituted N,N-dimethylanilinesand methyl iodide in ethanol at 25°C.” ~~

~

p-Substituent on the nucleophile 1.0019 ? 0.0004’ 1.0036 ? 0.0003 1.0032 ? 0.0002 0.9989 ? 0.0003

0.973’ 0.927 ? 0.004 0.968 t 0.004 1.143 ? 0.003

“Data taken from Szylhabel-Godala et al. (1966). *Measured in methanol at 25°C.

distance is greater) when a more electron-withdrawing substituent is on the nucleophile. Therefore, the 4-acetyl-N,N-dimethylaniline-methyliodide reaction should have the shortest N-C, bond and the loosest transition state. The change in the incoming group nitrogen KIEs with substituent is small. In fact, it is less than 10% of the theoretical maximum nitrogen KIE of 1.04(Maccoll, 1974). This suggests that the N-C, bond only changes slightly with substituent. The secondary a-deuterium KIEs, on the other hand, change markedly with substituent. This suggests that there is a significant loosening of the SN2transition state with substituent. The authors indicated that there was a problem because the Hammond effect of the substituent on the nitrogen KIEs, which is opposite to the very small trend in nitrogen KIE reported by Kurz and El Nasr (1982), suggested that the transition state was more product-like in the reaction with the most electron-withdrawing substituent on the nucleophile. This was not consistent with the larger secondary adeuterium KIEs (the looser transition state) found when there was a more electron-withdrawing substituent in the nucleophile. Although it is not apparent why a more product-like transition state needs to be tighter and show a smaller secondary a-deuterium KIE, the authors proposed that a solvent molecule that affected the magnitude of the secondary a-deuterium KIEs but not the nitrogen KIEs might be present in the transition state. However, they presented evidence against this conclusion, reporting that there is no solvent KIE in these reactions. An alternative explanation of these results is based on Westaway’s “bond strength hypothesis”, i.e. that the weaker transition state bond will change markedly and that the stronger reacting bond will not change significantly when a substituent in the reaction is altered. If this hypothesis is applied to these reactions, one would expect a significant change in the weaker C,-I bond and only a small change in the stronger N-C, bond when the substituent in the nucleophile is changed. Then, the larger secondary a-deuterium KIEs found when a more electronwithdrawing substituent is used simply means that the stronger N-C, bond

SECONDARY D-KINETIC ISOTOPE EFFECTS

187

shortens slightly and the weaker C,-I transition state bond lengthens markedly when a more electron-withdrawing substituent is used. In the theoretical portion of their study of the Menshutkin reaction, Paneth and co-workers (Gawlita et al., 1996) calculated the transition state structure and the nitrogen incoming nucleophile, the secondary a-deuterium and the a-carbon-12karbon-13 KIEs for the Menshutkin reactions between methyl iodide and the 4-methyl-, the unsubstituted and the 4-acetyl-N,Ndimethylanilines using several different semiempirical hamiltonians and continuum solvent models. An AM1 (Dewar et al., 1985) hamiltonian with the SM2 (Cramer and Truhlar, 1991), the SCRF2 (Karelson et al., 1989) and the COSMO (Klamt and Schuurmann, 1993) solvation model, a PM3 hamiltonian (Stewart, 1989) with the SCRM. and the COSMO solvation model and a MNDO hamiltonian (Dewar and Thiel, 1977) with the eigenvector follow optimizer and the COSMO solvation model were used to calculate the geometry and the force fields for the reactants and transition states and the KIEs were calculated at 298K by Paneth’s ISOEFF Version 6 program. The geometries and KIEs for the reaction between the p-methyl N,Ndimethylaniline and methyl iodide were the same in methanol and in ethanol when the AMl/COSMO method was used, so all of the other KIEs (Table 24) were calculated using the dielectric constant for methanol. It is also worth noting that the KIEs calculated from the separated reactants and the encounter complex were identical. This undoubtedly was observed because the encounter complexes were only 0.5 kcal mol-’ more stable than the separated reactants. The PM3 hamiltonian gave the best estimates of the a-carbon KIEs (Table 24), although all three methods were capable of calculating these KIEs. However, only the A M K O S M O method predicted the correct trend and magnitude of the nitrogen incoming group KIEs and only the AMllSCRF method predicted secondary a-deuterium KIEs that were near the observed values and this method did not predict the observed crossover from the inverse to the large, normal, secondary a-deuterium KIE found in the 4-acetyl-N,N-dimethylanilinereaction. Thus, none of the methods was able to predict the observed KIEs. The structures of the transition states calculated by each method were different and, moreover, the transition state structures predicted by each method were not altered by the substituent. All three hamiltonians predicted the relative activation energies, although the best results were obtained with the PM3 hamiltonian and the SM2 model. Also, all the hamiltonians were able to predict reasonable models for the reactants. As a result, the authors attributed the failure to predict the observed KIEs to the method’s inability to calculate the structure of the SN2 transition state. Important differences in the transition state calculated by the different methods are that (i) the N-C, transition state bond is longer when the PM3 rather than the AM1 hamiltonian is used; (ii) the COSMO solvent method predicts that the N-C, bond is approximately 0.2 A longer than the SCRF

0. MATSSON AND K. C. WESTAWAY

188

Table 24 The calculated and experimental incoming nucleophile nitrogen, a-carbon and secondary a-deuterium KIEs for the SN2 Menshutkin reactions between p-substituted N,N-dimethylanilines and methyl iodide at 298 K.”

4-Methyl-NJV-dimethylaniline AM1 SCRF AM1 COSMO PM3 SCRF PM3 COSMO Experimental Experimental

0.9947 1.0008 1.0015 1.0026 1.0036 ? 0.0003’ 1.0019 ? O.OOOld

1.0563 1.0552 1.0704 1.0661 1.0655 t 0.0024‘ 1.060 ? 0.023‘

0.891 0.800 1.108 1.073 0.927 t 0.004’ 0.83 2

NJV-dimethylaniline AM1 SCRF AM1 COSMO PM3 SCRF PM3 COSMO Experimental

0.9955 1.0002 1.0016 1.0034 1.0032 & 0.0002’

1.0552 1.0555 1.0705 1.0697 1.0615t 0.002’

0.892 0.818 1.107 1.124 0.968 t 0.004’

1.052 1.055 1.058 1.075 1.075 1.064

0.800 0.942 0.768 1.135 1.084 0.644 1.143 & 0.003’

4-Acetyl-NJV-dimethylnnilme AM1 SM2 0.9995 AM1 SCRF 0.9972 AM1 COSMO 0.9994 PM3 SCRF 1.0018 PM3 COSMO 1.0019 MNDO COSMO 1.om Experimental 0.9989 5 0.0003*

-

“Data taken from Gawlita et al. (1996). bFrom Szylhabel-Godala et al. (1996) in ethanol. ‘Estimated from Axelsson et al. (1987). dFrom Paneth and O’Leary (1991) in methanol. ‘Estimated from Buist and Bender (1958).

model; and (iii) all three methods predict that the C,-I transition state bond order is approximately 0.25, whereas the N-C, transition state bond is strongly dependent on the hamiltonian. The N-C, transition state bond order is only 0.1 when the PM3 hamiltonian is used but varies from 0.3 to 0.5 when the AM1 or the MNDO hamiltonians are used. Thus, these latter hamiltonians suggest that the transition states are product-like with a total N-C, plus C,-I bond order of approximately 0.7. However, since these methods seem incapable of calculating the structure of the SN2 transition states, this description of the transition state and the substituent effects on the Menshutkin reaction is at best speculation. This paper is an important contribution, however, because it demonstrates that this approach to determining transition state structure is at best, questionable.

SECONDARY D-KINETIC ISOTOPE EFFECTS

189

Table 25 The carbon-ll/carbon-14 KIEs for the SN2reactions between several amine nucleophiles and labelled methyl iodide in acetonitrile or dimethoxyethane at 30°C and 15"C, respectively." Nucleophile

Solvent

Temperature

k"lkL4

PKa

2P-Lutidine 2,6-Lutidine Quinuclidine (CH,CH,),N

Acetonitrile Acetonitrile DME DME

30.00 30.00 15.00 15.00

1.189 2 0.012 1.220 2 0.009 1.220 2 0.005 1.221 2 0.006

6.72 6.77 10.95 10.65

"Data taken from Persson et al. (1995).

Finally, Persson et al. (1995) measured the 11C/14CKIEs for the SN2 reactions between several amine nucleophiles and labelled methyl iodide in dimethoxyethane or acetonitrile at 15°C and 30"C, respectively, to determine how sterically hindered nucleophiles affects the transition state of a Menshutkin reaction. The results in Table 25 show that all the k1'lk14-valuesfor these reactions are large. In fact, they are all near the theoretical maximum value for these KIEs. Secondly, the KIE for the reaction with the more sterically hindered amine, 2,6-lutidine, is larger than that for the less sterically hindered 2,4-lutidine. It is important to note that 2,6-lutidine and 2P-lutidine have almost the same pK,, so there is little or no electronic effect in these reactions. Le Noble and Miller (1979) found a larger chlorine leaving group KIE (k35/k37= 1.0038 2 0.0003) for the 2,6-lutidine-methyl chloride reaction than for the corresponding pyridine reaction (k35/k37 = 1.00355 ? 0.00008) in bromobenzene at 100°C. Thus, it appears that carbon-chlorine bond rupture is more advanced in the reaction with the more sterically hindered nucleophile, although the difference could be due to the fact that 2,6-lutidine is also a better nucleophile than pyridine. Since the a-carbon KIE in both the 2,4- and the 2,6-lutidine reactions are near the theoretical maximum value, both transition states must be reasonably symmetrical. Therefore, the KIEs suggest that the transition state for the reaction with the more sterically hindered nucleophile is looser with a longer a-carbon-chlorine bond. Finally, BEBOVIB-IV calculations also suggested that a looser transition state should be found when a more sterically hindered nucleophile was used in this reaction and the authors concluded that the transition states for these reactions were early but that the reaction with the more sterically hindered nucleophile and the larger a-carbon k"/k14 was looser. It is worth noting that the early transition states suggested by this study are consistent with the small nitrogen incoming nucleophile KIEs measured by Kurz and coworkers and by Bourns and Hayes (see Table 19, p. 180). Unfortunately, the same trend in k"lk14 is not observed in the triethylamine/quinuclidine reactions with methyl iodide. Here, identical 1*C/14CKIEs are found for both nucleophiles. It is possible that the identical

0. MATSSON AND K. C. WESTAWAY

190

isotope effects are due to the cancellation of two effects, a steric effect and an electronic effect, because triethylamine is both a stronger base and a more sterically hindered nucleophile than quinuclidine. It is interesting that increasing the steric hindrance in the triethylamine/quinuclidine-methyl chloride reactions has a different effect on the chlorine KIEs than in the 2,6-lutidine/pyridine-methyl chloride reactions. Swain and Hershey (1972) found a smaller chlorine KIE in the reaction with the more sterically hindered nucleophile, triethylamine (P5/P7= 1.00640 2 0.00009), than for the quinuclidine-methyl chloride (P5/P7 = 1.0071 2 0.0001) reaction. Again, it is not yet understood why different steric effects were found in the two reaction series. THE EFFECT OF ION-PAIRING ON THE! SECONDARY a-DEUTERIUM KIEs

Westaway and co-workers (Westaway and Lai, 1988; Lai and Westaway, 1989; Fang and Westaway, 1991) found that the secondary a-deuterium KIE in the sN2 reaction between butyl chloride and thiophenoxide ion (reaction (26)) was concentration dependent. CH3CH2CH2CH2-Cl+ C6H5S-(Na+)-CH3CH2CH2CH2-SC&

+ (Na+)Cl(26)

For instance, (kH/kD), was 1.085 2 0.011 when the sodium thiophenoxide concentration was 0.0086 mol dm-3 and 1.129 ? 0.010 when the concentration of the sodium thiophenoxide was reduced to 0.0040 mol dm-j in DMF at 20°C. Conductivity and UV studies of sodium thiophenoxide solutions in DMF, in DMSO, in methanol and in diglyme, and reactions done in the presence of the macrocyclic poiyether 15-crown-5(Westaway and Lai, 1988), showed that the change in the secondary a-deuterium KIE was due to a change in the form of the reacting nucleophile from a solvent-separated ion-pair complex at the higher concentration (see (27)) to a free ion at the lower concentration. xM+ + X-SC&

[Mf(solvent);SC,H5],

(27)

In fact, this behaviour shown in Table 26 appears to be typical of organic solvents because the same phenomenon was observed in DMSO, DMF, methanol and diglyme (Westaway and Lai, 1988; Lai and Westaway, 1989; Fang and Westaway, 1991). These secondary a-deuterium KIEs clearly indicate that the transition state is different when the form of the reacting nucleophile is changed in this sN2 reaction. In fact, the KIE becomes larger in every solvent when the nucleophile is converted from the contact ion-pair, into the solvent-separated ion-pair complex, or from the solvent-separated ion pair into the free ion. This suggests that the sN2 transition state is looser when the nucleophile is the free ion, i.e. has a longer S-C, and/or C,-C1 bond. Unfortunately, these KIEs do not indicate which bond is longer in the free ion transition state.

SECONDARY D-KINETIC ISOTOPE EFFECTS

191

Table 26 The secondary a-deuterium KIEs for the SN2 reactions between butyl chloride and thiophenoxide ion when the nucleophile is a contact ion-pair, a solvent-separated ion-pair complex and a free ion at 20°C." ~~

(kH/kD)ol

Form of the reacting nucleophile C6H5S-Li [M+ (~olvent);SC6H5], C6H& +

DMF

DMSO

MeOH

Diglyme 1,070' 1.131'

-

-

-

1.084

1.034 1.116

1.100 1.132

1.127

-

"Data taken from Westaway and Lai (1988). 'These KIEs were measured using lithium thiophenoxide as the nucleophile. The contact ion-pair was converted into the solvent-separated ion-pair complex by adding between 1.0% and 7.5% water to dry diglyme (Fang and Westaway, 1991).

In an effort to learn exactly how a change in the form of the reacting nucleophile from a free ion to an ion-pair affects the structure of the sN2 transition state, Westaway and co-workers determined both the secondary a-deuterium (Lai and Westaway, 1989) and nitrogen leaving group (Jiang, 1996) KIEs for the sN2 reactions between benzyldimethylphenylammonium nitrate and sodium p-substituted thiophenoxides in methanol at 20°C (reaction (28)).

The secondary a-deuterium KIEs shown in Table 27 decrease for both the ion-pair and the free ion reactions, when a more electron-withdrawing p-substituent is on the nucleophile. Because the magnitude of secondary a-deuterium KIEs is directly related to the S-N distance in the SN2transition state (Poirier et al., 1994), adding a more electron-withdrawing substituent to the nucleophile leads to a transition state with a shorter S-N distance. The primary nitrogen KIEs for the free ion and the ion-pair reactions, on the other hand, increase very slightly; the C,-N transition state bond length increases slightly when a more electron-withdrawing substituent is added to the nucleophile. Because the C,-N transition state bond length increases slightly while the S-N distance shortens, the S-C, transition state bond must shorten significantly when a more electron-withdrawing substituent is on the nucleophile in both the free ion and the solvent-separated ion-pair complex reactions. This unbalanced change from an early to a later transition state is illustrated in Fig. 14 using the free ion as the nucleophile.

0. MATSSON AND K. C. WESTAWAY

192

Table 27 The secondary a-deuterium and primary nitrogen KIEs for the free ion and ion-pair SN2 reactions between benzyldimethylphenylammonium nitrate and p substituted thiophenoxide ions in methanol at 20°C."

p-Substituent

(kHlkD)a

The nucleophile is the free thiophenoxide ion CH30 1.271 5 0.013' H 1.222 2 0.013 1.1215 0.014 c1

k'41k15

1.0162 5 0.0005' 1.0166 2 0.0008 1.0169 2 0.0005

The nucleophile is a solvent-separated ion-pair complex CH30 H

c1

1.216 2 0.012' 1.207 2 0.008 1.150t 0.009

1.0161 2 0.0005' 1.0162 2 0.0010 1.0166 2 0.0003

"Data taken from Jiang (1996). *The error in the isotope effect is l/kD[hkH)* + ( k ~ / k ~X)(' h k ~ ) ' ] where ' ~ , AkHand AkD are the standard deviations for the rate constants for the reactions of the undeuterated and deuterated substrates, respectively. 'Standard deviation of the average kinetic isotope effect.

The later transition states found when a more electron-withdrawing substituent is attached to the nucleophile may be found because a poorer nucleophile would have to be closer to the a-carbon to distort the C,-N+ bond and cause the reaction to occur. The greater change in the S-C, bond with substituent can be understood in terms of the bond strength hypothesis (Westaway, 1993). The S-C, bond is weaker than the C,-N+ bond and the bond strength hypothesis predicts that the greatest change will occur in the weaker S-C, bond and that there will be little or no change in the stronger C,-N bond when the p-substituent on the nucleophile changes. A second observation is that the substituent effect is greater in the free ion reactions than in the ion-pair reactions, (kH/kD),changing by 15% in the free ion reactions but by only 7% in the ion-pair reactions. The corresponding change in kI4/kl5is 0.0007 in the free ion reactions but only 0.0005 in the ion-pair reactions. A possible explanation for this is that the change in charge on the nucleophilic sulfur atom with substituent is greater in the free ions than in the ion-pairs. A CND0/2 calculation (Lai and Westaway, 1989) showed that the negative charge on the sulfur of the free ion decreases by 0.0208 electrons when the p-substituent is changed from methoxy to chloro. The corresponding decrease is only 0.0171 electrons when the substituent is changed in the ion-pair. Thus, the change in the amount of negative charge on the nucleophilic sulfur atom with substituent is larger for the free ion than for the ion-pair. Although the difference in the change in the negative charge seems small, only 0.0037, the substituent effect on the negative charge on the sulfur

SECONDARY D-KINETIC ISOTOPE EFFECTS

193

(kH/kD),= 1.27 k’4/k’s = 1.0162

(kH/kD), = 1.22 kl4/kI5 = 1.0166

(kH/ko).= 1.12 k’4/k’5 = 1.0169

Fig. 14 The relative transition state structures for the SN2 reactions between benzyldimethylphenylammonium ion and free p-substituted thiophenoxide ions in methanol at 20°C. Reproduced, with permission, from Westaway (1996).

atom is 22% greater for the free ion than for the ion-pair. Therefore, it is not surprising that the substituent effect is greater in the free ion reactions. Finally, the changes in transition state structure found in this study can be used to test the theories for predicting the substituent effects on the transition state structures. Thornton’s reacting bond rule (Thornton, 1967), the More O’Ferrall-Jencks energy surface method (More O’Ferrall, 1970; Jencks, 1972) and the Pross-Shaik method (Pross and Shaik, 1981) all fail to predict the change in transition state structure that was found in this study. Only the Bell, Evans and Polanyi principle (Dewar and Dougherty, 1975), which predicts an earlier transition state when a better nucleophile is used, and the bond strength hypothesis (Westaway, 1993), which predicts that there will be a significant change in the weaker S-C, reacting bond and little or no change in the stronger C,-N reacting bond when the p-substituent in the nucleophile is changed, are consistent with the experimental results. The secondary a-deuterium and primary nitrogen leaving group KIEs for the free ion and ion-pair reactions in Table 27 show how ion-pairing affects the structure of the transition state for the SN2 reactions between benzyldimethylphenylammonium nitrate and sodium p-substituted thiophenoxides in methanol at 20°C. The nitrogen KIEs for each free ion and the ion-pair reaction are identical within the experimental error of the method. Therefore, C,-N bond rupture must be equally advanced in the transition states of the free ion and the ion-pair reactions (Saunders, 1975). Although the incoming sulfur nucleophile KIEs have not been measured, one can deduce how the S-C, bond changes o n ion-pairing by combining the information provided by

194

0. MATSSON AND K. C. WESTAWAY

the secondary a-deuterium and the nitrogen KIEs. The magnitude of the secondary a-deuterium KIE is determined by the S-N distance in the s N 2 transition state (Barnes and Williams, 1993; Poirier et al., 1994) and since the nitrogen isotope effects show that the C,-N transition state bond is not altered by ion-pairing, the change in the secondary a-deuterium KIE caused by ion-pairing must be due to a change in the S-C, transition state bond. Therefore, the secondary a-deuterium KIEs indicate that the free ion S-C, bond (a) is significantly longer than the ion-pair S-C, bond when the nucleophile is p-methoxythiophenoxide ion; (b) is longer than the ion-pair S-C, bond when the nucleophile is p-methylthiophenoxide ion; (c) is slightly longer than the ion-pair S-C, bond when the nucleophile is thiophenoxide ion; but (d) is shorter than the ion-pair S-C, bond in the p-chlorothiophenoxide ion reaction. Finally, the Hammett p values found by changing thep-substituent on the nucleophile for the free ion and ion-pair s N 2 reactions were identical; the p values were -0.85 t 0.14 and -0.84 2 0.11 for the free ion and ion-pair reactions, respectively. This suggests that the change in charge on the nucleophilic sulfur atom on going to the transition state is identical for the free ion and ion-pair reactions and, therefore, the S-C, transition state bond is not altered significantly when the nucleophile changes from a free ion to an ion-pair. Although this conclusion is at odds with that based on the secondary a-deuterium KIEs, the identical p values for the free ion and ion-pair reactions may be observed because, on average, the S-C, bond for the free ion reaction is identical to the S-C, bond in the ion-pair reactions, i.e. the S-C, bond in the free ion reaction is longer when the p-substituent is MeO, Me or H but is shorter when the p-substituent is C1. One explanation for the longer S-C, bond in the free ion transition state is that the sodium ion reduces the electron density on the sulfur atom, making it a poorer nucleophile. This would lead to a more product-like transition state (vide supra). Unfortunately, this is not true for the p-chlorothiophenoxide ion reaction, which has a shorter S-C, bond in the free ion transition state. Two other observations are obvious when one considers how ion-pairing affects the structure of the SN2transition state. First, the major change in bonding occurs in the weaker S-C, bond and there is little or no change in the stronger C,-N reacting bond as the bond strength hypothesis predicts. The second observation is that the change in transition state structure on ion-pairing is greatest in the reaction with the best nucleophile and becomes smaller as a more electron-withdrawing substituent is added to the nucleophile. This suggests that the difference between the electron density on the sulfur atom of a free ion and of an ion-pair nucleophile is smaller when a more electronwithdrawing substituent is on the nucleophile. This probably occurs because the strength of the ionic bond between the solvent-separated sodium ion and the sulfur anion of the p-substituted thiophenoxide ion decreases significantly with the negative charge (the electron-withdrawing ability of the substituent) on the nucleophile. If this were the case, the difference between the free ion

SECONDARY D-KINETIC ISOTOPE EFFECTS

195

and the ion-pair secondary a-deuterium KIEs should decrease when a more electron-withdrawing group is attached to the nucleophile. This trend is found when the nucleophile is the p-methoxythiophenoxide ion, the p-methylthiophenoxide ion and the thiophenoxide ion. The effect of ion-pairing on the transition state for the p-chlorothiophenoxide ion reaction does not fit this trend, however.

THE EFFECT OF A CHANGE IN SOLVENT ON THE SECONDARY CY-DEUTERIUM KlEs

The secondary a-deuterium and primary nitrogen leaving group KIEs in Table 28 were determined for the ion-pair sN2 reactions between sodium thiophenoxide and benzyldimethylphenylammonium nitrate [reaction (28)] in DMF at 0°C (Westaway and Ali, 1979) and in methanol at 20°C (Jiang, 1996) to learn how a change in solvent affects the structure of the sN2 transition state. Regrettably, the KIEs were measured at different temperatures. Applying an average temperature-dependent decrease of 0.008 per 20°C to the secondary a-deuterium KIE of 1.179 found at 0°C in DMF (Shiner et al., 1969; Koshy and Robertson, 1974) suggests that this KIE would be approximately 1.17 at 20°C. The temperature dependence of a nitrogen isotope effect would appear to be small (Smith and Westaway, 1982); a change of 20°C would change the isotope effect by less than 0.1%. Therefore, the nitrogen KIE for the reaction in DMF would be 21.019 at 20°C.

Table 28 The secondary a-deuterium and primary nitrogen KIEs for the ion-pair SN2 reactions between sodium thiophenoxide and benzyldimethylphenylammonium nitrate in DMF at 0°C and in methanol at 20°C."

Temp. Solvent Methanol DMF DMF

("C)

(kHlkDL

20 0 20

1.215 ? 0.012' 1.179 ? 0.010 1.17'

k'4/k15 1.01622 O.OOIOd 1.0200 2 0.0007 a1.019'

Hammett p valueb 0.84 5 0.11" 1.70 ? 0.05

"Data taken from Jiang (1996) and from Westaway and Ah (1979). 'The Hammett p value was obtained by changing the p-substituent on the nucleophile. T h e error in the isotope effect is I/kD[AkH)*+ (k,,/kl,)*X (AkD)2]1'2,where AkH and AkD are the standard deviations for the rate constants for the reactions of the undeuterated and deuterated substrates, respectively. "Standard deviation of the average KIE. 'The standard error of coefficient of the p value. 'These KIEs are estimated at 20°C.

0.MATSSON AND K. C.WESTAWAY

196

In methanol

(kHlkD),= 1.215 = 1.0162

k14/k'5

Fig. 15 The relative transition state structures for the ion-pair SN2reactions between sodium thiophenoxide and benzyldimethylphenylammonium nitrate in DMF and in methanol. Reproduced, with permission, from Westaway (1996).

The larger secondary a-deuterium KIE of 1.215 in methanol indicates that the S-N transition state distance is greater in methanol than it is in DME The primary nitrogen KIE, on the other hand, is smaller in methanol than in DMF, indicating that the C,-N transition state bond is considerably shorter in methanol than in DME Because the transition state in methanol has a longer S-N distance but a shorter C,-N bond, the S-C, bond must be much longer in methanol than in DME Therefore, an earlier transition state with a much longer S-C, and a shorter C,-N bond is found in methanol (Fig. 15). The Hammett p values in Table 28, found by changing the p-substituent on the nucleophile in DMF and methanol, support this conclusion. The larger p value in DMF indicates that the change in charge on the nucleophilic sulfur atom is greater on going from the reactant to the transition state in DMF. Therefore, the S-C, transition state bond is shorter in DMF than in methanol. The earlier transition state in methanol can be rationalized as follows. The SN2 transition state for this reaction will be solvated primarily at the sulfur atom because access of solvent molecules to the partial positive charges on the a-carbon and on the nitrogen atom is sterically hindered. The second assumption is that the structure of the transition state will depend on its stability in that solvent, i.e. that the transition state that is the most stable in each solvent will be found. An early transition state would be more stable than a product-like transition state in methanol because solvation of the sulfur atom by hydrogen-bonding would lower the energy of an early transition state where there is a greater negative charge on the sulfur atom; the solvent stabilizes an earlier transition state more than a late transition state. As a result, a transition state with a longer and weaker S-C, transition state bond

SECONDARY D-KINETIC ISOTOPE EFFECTS

197

would be expected in methanol, and in DMF, a late, less ionic (dipolar) transition state which would be more strongly solvated by DME SECONDARY a-DEUTERIUM KIEs AND THE EFFECT OF IONIC STRENGTH ON TRANSITION STATE STRUCTURE

Finally, we refer again to the secondary a-deuterium and primary nitrogen KIEs in Table 29 for the ion-pair S,2 reactions between sodium thiophenoxide and benzyldimethylphenylammonium nitrate which were measured at two different ionic strengths in DMF at 0°C (Pham and Westaway, 1996). The larger secondary a-deuterium and the smaller nitrogen KIE found in the higher ionic strength reaction indicate that the S-N distance in the transition state is greater and that the C,-N bond is shorter in the reaction at a high ionic strength. This means that the S-C, bond is longer in the transition state for the high ionic strength reaction. The earlier, more ionic, transition state is probably found at the high ionic strength because the more ionic transition state will be more stable (more highly solvated) in the more ionic solvent. The important conclusion, however, is that use of inert salts to increase the ionic strength in reactions so that accurate rate constants can be measured changes the structure of the transition state markedly. 3 Secondary p-deuterium KIEs SECONDARY P-DEUTERIUM KIEs IN CARBOCATION SNREACTIONS

Secondary P-deuterium KIEs are observed when the hydrogen(s) on the ,)~ P-carbon are replaced by deuterium(s). These isotope effects ( l ~ ~ / l C ,are Table 29 The secondary a-deuterium and primary nitrogen KIEs and the relative transition state structures for the ion-pair SN2 reactions between sodium thiophenoxide and benzyldimethylphenylammonium nitrate in DMF at different ionic strengths at 0°C.“ Ionic strength

(kHlkD)*

kl4/kl4

0.640

1.179 2 O.OIOb 1.0200 2 0.0007‘

0.904

1.215 2 O.O1lb 1.0166 2 0.0004‘

Relative transition state structure

ss-

ss+

S---C_______ N

66+ S________ C____ N

“Data taken from Pham and Westaway (1996). ’The error in the isotope effect is 1/kD[Akb,)’ + (kf,/kD)* X (AkD)2]1’2,where AkH and Ak, are the standard deviations for the rate constants for the reactions of the undeuterated and deuterated substrates, respectively. ‘The standard deviation of the average KIE.

198

0. MATSSON AND K. C. WESTAWAY

greater than unity for nucleophilic substitution reactions. For example, the isotope effect per CD3 group increases from approximately 1.03 for ethyl compounds, which undoubtedly react by an sN2 mechanism, to approximately 1.34 for a t-butyl compound which reacts by a limiting s N 1 mechanism (Evans and Lo, 1966; Frisone and Thornton, 1964; Hakke el al., 1965). In fact, the magnitude of the isotope effect increases as the amount of positive charge (carbocation character) on the a-carbon in the transition state [2] is increased.

A wealth of experimental evidence (Shiner, 1970b) indicates that the secondary p-deuterium KIEs in carbocation reactions are primarily a result of hyperconjugation (Melander and Saunders, 1980b; Meot-Ner (Mautner), 1987; Shiner, 1970c; Shiner and Humphrey, 1963; Sunko and Hehre, 1983). Hyperconjugation was introduced by Mulliken to describe a conjugative type of interaction between methyl groups and other groups containing multiple bonds (Mulliken, 1933, 1935, 1939; Mulliken et al., 1941) and Gold (1983) defines hyperconjugation as the interaction of a a-bond with a .Ir-network.This type of interaction provides a mechanism for electron release from a methyl group to an electron-deficient centre such as a carbocation. For example, the p-deuterium KIE of 2.4 (kH/kDper p-CD, = 1.34) found for the solvolysis of t-butyl-d, chloride in several solvents (Hakke et al., 1965; Frisone and Thornton, 1964) shows that the C,-H bonds are weaker (the force constants are smaller) in the transition state than in the reactant state. This suggests that the t-butyl carbocation, which is closely related to the transition state of the rate-determining ( k , ) step of the solvolytic reaction, can be described by resonance structures of the type [3].

Corresponding to this valence bond view is a molecular orbital picture. The three a-orbitals of a CH3group are regarded as a basis from which three group orbitals may be constructed. One of the possible combinations of the a-orbitals has the same local symmetry as the vacant p-orbital on the cationic centre, and hence may overlap with it. Therefore, a withdrawal of electrons from the methyl group can take place. The orbital from which electron density

SECONDARY D-KINETIC ISOTOPE EFFECTS

199

is removed is essentially the C,-H bond and, therefore, an increase in the C,-H bond length and a decrease in the stretching force constant are expected. These changes in the C,-H bonding lead to the observed normal p-deuterium KIE. The shift in electron density also causes diminished repulsive interaction between pairs of out-of-plane hydrogens, causing a corresponding increase in the H-C,-H bond angles. These effects on geometry have been confirmed by quantum-mechanical calculations. Other studies by Shiner and co-workers (Shiner and Humphrey, 1963; Shiner and Jewett, 1965) have demonstrated that the magnitude of secondary p-deuterium KIEs is related to the dihedral angle between the C,-H(D) orbital and the developing p-orbital on the a-carbon. The study in which Shiner and Humphrey (1963) measured the secondary p-deuterium KIEs for the hydrolysis of 11-methyl-1l-chloro-9,10-dihydro-9,lO-ethanoanthracene [4] and its 1 2 , 1 2 4 [5] and 9,10-d2[6] analogues in 60% aqueous ethanol at 45°C is particularly elegant.

[41

151

A value of kHlkD= 1.07 per p-D was observed when the deuteriums were on the bridge at C-12 [5] and the dihedral angle between the p-orbital of the carbocation and the C,-H bonds was approximately 30" and hyperconjugation could occur. When [6] was used, the dihedral angle was 90" and there was no overlap between the empty p-orbital of the carbocation and the C,-H bonds; i.e. no hyperconjugation could occur and only a small inverse, inductive KIE, kHlkD= 0.99, was observed. This study and other studies by Shiner and co-workers (Shiner, 1970b) have established that the maximum secondary p-deuterium KIE in any system is observed when the dihedral angle is either 0" or 180", i.e. where the overlap between the C,-H and the p-orbital on the

200

0. MATSSON AND K. C. WESTAWAY

a-carbon is a maximum. For instance, the KIE per p-D decreased from 1.30 when the dihedral angle between the p-orbital of the carbocation and C,-H bond was Oo, to 1.07 when the dihedral angle was 30°,to 1.01when the dihedral angle was 60", to 0.99 when the dihedral angle was 90". In fact, Sunko et al. (1977) extended this idea by suggesting that the empirical equation (29), which related the magnitude of the hyperconjugative component of the secondary p-deuterium KIE to the dihedral angle, 8, between the developingp-orbital of the carbocation and the C,-H bond, could be used to determine transition state structure in carbocation reactions.

Although the geometric relationship suggested by Shiner and by Sunko and their co-workers clearly demonstrates that hyperconjugation is the major contributor to the secondary p-deuterium KIE in carbocation reactions, Williams (1985) has suggested that there is a significant inductive component to these KIEs. Williams used ab initio MO methods to calculate the geometries of the substrates and the isopropyl carbocation formed in a gas-phase heterolysis (30) of series of isopropyl derivatives at the RHF/4-31G level. (CH3)ZCH-X F=+ (CH3)zCH++ X(X = H, F, OH, OH;, N: at 25°C)

(30)

These structures were then used to generate the force fields and calculate the secondary p-deuterium-d, equilibrium isotope effects (EIEs) for the formation of the isopropyl carbocation (Table 30). Because the transition states for formation of the carbocation will be close to the structure of the carbocation, these KIEs should be excellent approximations of the maximum secondary P-deuterium KIEs expected for the limiting SN1 solvolytic reaction. An examination of the EIEs in Table 30 shows that the trend in the EIEs cannot be explained by hyperconjugation alone. If these KIEs were determined by hyperconjugation alone, the p-deuterium EIE would be expected to decrease as the substrate acquired more cationic character at the &-carbon,i.e. as the C,-C, bond shortened. However, this trend was not observed. As the leaving group improves along the series X = H, OH, F, OH:, the a-carbon of the substrate becomes more planar, the C,-C, bond shortens and the C,-H bond in the substrate becomes shorter, i.e. the amount of cationic character at the a-carbon increases and the secondary P-deuterium EIE increases. This trend in the EIE is obviously not consistent with a model where hyperconjugation is the sole contributor to the EIE. The calculated EIEs were factored into two components; a hyperconjugative (conformationally dependent contribution related to cos2 8) and an inductive (conformationally independent) contribution. The contributions of the hyperconjugative and inductive EIEs to the total EIE given in the last

P 4 Z

rn

3

0

Table 30 Structural features for the substrates in the gas-phsae heterolysis of several (CH&CH-X P-deuterium-d6EIEs and the hyperconjugative and inductive contributions to the EIE."

X H OH F OH;

N;

Planarity at C,

C-C, bond length

0.09 0.12 0.26 0.56 0.99

(4

1.530 1.523 1.512 1.502 1.456

C,-H bond length

(A>

1.0840 1.0832 1.0818 1.0815 1.0854

Calculated Hyperconjugative (kH/kD)lP-d6 contribution EIE to the EIE 1.544 1.570 1.678 1.776 1.119

2.68 2.70 2.53 2.22 1.12

v,

at 25"C, the calculated

9

B rn

Inductive contribution to the EIE

Percentage inductive EIE

0.58 0.58 0.66 0.80 1.00

43 43 34 20 0

rn

; g v, 4

"Data taken from Williams (1985).

N

9

202

0. MATSSON AND

K. C. WESTAWAY

three columns of Table 30 show that the inductive effects are smaller than the hyperconjugative effects in all cases, as Shiner and Sunko and their collaborators had concluded. However, the inductive contribution to the EIE is significant for most of the reactions; it accounts for a maximum of 43% of the EIE when the poorer leaving groups H and OH are used and still accounts for 20% of the EIE when a good leaving group, OH:, is used. It is interesting to note that the inductive contribution to the EIE is reduced to zero when the very good leaving group N,' is used. Thus, although hyperconjugation is the major contributor to the EIEs, the inductive contribution is significant for all but the best leaving groups, i.e. the hyperconjugative contribution increases and the inductive contribution decreases with increasing leaving group ability. Finally, as is the case for the secondary a-deuterium KIEs, the P-deuterium KIE is assumed to vary in magnitude from near unity for a reactant-like transition state to a maximal value for a transition state resembling the carbocation formed in an SN1reaction. The experimentally determined KIE may, therefore, be used as a measure of transition state structure provided that the maximum value of the KIE, i.e. the EIE for the formation of the carbocation, is known. SECONDARY p-DEUTERIUM KIEs AND THE CASE FOR NEGATIVE ION HYPERCONJUGATION

Negative ion hyperconjugation, which has also been called anionic or negative hyperconjugation, was proposed by Roberts et al. (1950) to account for the substituent effects caused by the trifluoromethyl group. The valence bond picture of negative ion hyperconjugation is provided by no-bond resonance structures like the one shown for the fluoroethyl anion in [7]. Fluorine negative ion hyperconjugation has been thoroughly discussed by Holtz (1971).

Negative ion hyperconjugation has also been proposed for methyl groups bonded to a carbanion centre. For instance, Hehre and co-workers (DeFrees et al., 1977,1979a) suggested that, depending on the actual need, the methyl group can either supply electron density to an adjacent electron-deficient centre (carbocation hyperconjugation) or accept it from an electron-rich centre (negative ion hyperconjugation). In negative ion hyperconjugation, valence bond structures like those in [7], but with a negative charge on the P-hydrogens, are used to explain how a methyl group stabilizes a carbanionic

SECONDARY D-KINETIC ISOTOPE EFFECTS

203

centre. This means that negative ion hyperconjugation should be dependent on geometry in the same way as hyperconjugation is in carbocation reactions (vide supra), i.e. negative ion hyperconjugation should be maximum when the dihedral angle between the C,-H bond and the lone-pair of electrons on the carbanion is either 180" or 0". This relationship has been examined extensively in theoretical studies. In one such study, DeFrees ef al. (1979b) investigated the dependence of the p-deuterium KIE on conformation for the hypothetical formation of an ethyl carbocation, anion and radical using MO theory at the 4-31G level. For all three species, the hyperconjugative component of the KIE was directly related to cos2 t9 (equation (29)), where 0 is the dihedral angle between the C,-H bond and the vacant, the singly occupied, or the doubly occupied orbital of the cationic, radical or anionic centre, respectively. Additional support for negative ion hyperconjugation has been provided by an important theoretical study of P-substituted ethyl carbocations and ethyl anions by Hoffmann et al. (1972). These workers found significant barriers to internal rotation in both p-carbocations and carbanions, as would be expected if both negative ion and positive ion hyperconjugation were important. Finally, since the methyl group acceptor orbital of appropriate symmetry is antibonding, a hyperconjugative interaction with an electron-rich centre would reduce the force constants for the C,-H bond. On the basis of this theoretical argument, a normal secondary P-deuterium KIE is anticipated for a reaction where negative charge is accumulating on an a-carbon adjacent to a deuterated methyl group. Although the secondary p-deuterium KIEs in reactions proceeding via carbocation intermediates have been investigated extensively (vide supra), comparatively few @deuterium KIEs have been published for reactions involving carbanion, radical, or carbene intermediates. The secondary pdeuterium KIEs in carbanion reactions where negative ion hyperconjugation can occur are more complex (see, for example, Apeloig, 1981;Friedman ef al., 1985; Schleyer and Kos, 1983; Streitwieser et af.,1981). Some of the secondary P-deuterium KIEs found for the formation of anions or anion ion-pairs that have been attributed either wholly or partly to hyperconjugative stabilization of the anionic centre are discussed in the following paragraphs. Hehre and co-workers (DeFrees et al., 1977, 1979a) have published both experimental and theoretical evidence in support of negative ion (anionic) hyperconjugation. These workers determined the free energies for the gas-phase hydron3 transfer equilibria (31), (32) and (33) by pulsed ion cyclotron resonance spectroscopy (Wolf et al., 1976). These equilibria, which involve the gas-phase formation of a methylamino, a methoxy and a thiomethoxy anion, all lie to the right, i.e. the formation of the isotopically light anion is favoured. These results were rationalized in terms of the MO

'A hydron (symbol L) is a proton, a deuteron or a triton (IUF'AC Commission, 1988).

0. MATSSON AND K. C. WESTAWAY

204

Table 31 The calculated C-H bond lengths for some neutral molecules and their conjugate bases at the 4-31G level."

Molecule

C-H bond length (A) Conjugate base C-H bond length (A)

CH3NH2 CH30H CH3SH _____

1.084 1.080 1.076 ~

F

~

CH3m CH30CH3S-

1.112 1.122 1.088

F

"Data taken from DeFrees et al. (1979b).

description of negative ion hyperconjugation. Because hyperconjugation puts electron density into the antibonding C,-H orbitals and weakens the C,-H bonds in the anions, the deuterium is concentrated in the stronger (shorter) C,-H bonds of the neutral species and the EIE is >LO. Ab initio MO calculations at the 4-31G level (DeFrees et al., 1979b) support this qualitative explanation based on the presence of negative ion hyperconjugation. The calculated C,-H bond lengths for the methylamino, the methoxy and the thiomethoxy anions in Table 31 were, indeed, significantly longer than those in the corresponding amine, alcohol and thiol. CD3NH+ CH3NH2

CD3NH2 + CH3NH

AGO = -0.37

? 0.08 kcal molF'

(31) CD30- + CH30H

CD30H + CH30-

AGO = -0.50

? 0.10 kcal mol-'

(32) CD3S- + CH3SH

+

CD3SH CH3S-

AGO = -0.30

? 0.08 kcal mol-'

(33)

In another study, Streitwieser and Van Sickle (1962) measured the secondary P-deuterium KIEs for the formation of carbanions from hydrocarbons with lithium cyclohexylamide in cyclohexylamine at 49.9"C. The rate constants needed for determining these KIEs for the formation of the carbanion (reaction (34)) were obtained by analysing the deuterium in the ethylbenzene recovered from the reaction at various times by mass spectrometry.

(L = H, D)

The normal secondary P-deuterium KIE of 1.11? 0.03 found for the formation of the benzyl carbanion from ethylbenzene labelled with three deuteriums at the P-carbon was thought to be indicative of a transition state with substantial carbanionic character. However, the KIE was rationalized in

SECONDARY D-KINETIC ISOTOPE EFFECTS

205

terms of an inductive effect (Halevi, 1963; Melander and Saunders, 1980c) rather than a hyperconjugative effect. The inductive KIE was normal because deuterium is electron-donating relative to hydrogen and the deuterated transition state is higher in energy than the undeuterated transition state. The authors also referred to an argument presented by Bartell (1960) that the secondary P-deuterium KIE is normal because diminished non-bonding repulsions between the leaving group and a P-hydrogen in the transition state lower a transition state C,-H bending force constant. The KIE arises because the C,-D bending vibrations for the shorter C,-D bonds, are not reduced as much when the substrate is converted into the transition state. More O’Ferrall and Slae (1970) reported a secondary P-deuterium KIE of 1.01 per deuterium ( ( k H / k D )=, 1.03 at 25°C) for the rate-determining formation of the carbanion formed in the first step of the Elcb p-elimination reaction of 9-fluorenylmethanol (reaction (35)).

The rate constants for the KIEs were measured using UV spectroscopy in separate kinetic runs using the undeuterated and deuterated substrates. Although this normal secondary P-deuterium KIE could be due to hyperconjugation, the authors, like Streitwieser and Van Sickle (1962), preferred to attribute it to an inductive effect. Kresge et al. (1974) determined that the secondary P-deuterium EIE for the ionization of 2-nitropropane (reaction 36)) in aqueous solution at 25°C was KHIKD = 1.23 2 0.03 (1.035 per P-D). (CL~)~CH--NO~ + B-

-

(CL,)~C--NO~ + BH

(36)

(L = H, D)

The secondary P-deuterium KIEs observed for the reaction of the same substrate with hydroxide ion and with tris(hydroxymethy1)methylamine in aqueous solution at 25°C were small, i.e. k H / k D= 1.09 -+ 0.01 and 1.10 2 0.01, respectively. While Kresge argued that the EIE was primarily due to hyperconjugation, the secondary P-deuterium KIEs were attributed partly to hyperconjugation and partly to a polar (inductive) effect. The rate constants for the evaluation of both the EIE and the KIEs were determined in separate kinetic runs by following the increase in the absorbance due to the nitronate ion by UV spectroscopy. Matsson (1985) determined the secondary p-deuterium KIE for the basecatalysed stereospecific 1,3-prototropic rearrangement (37) of 1-methylindene

0. MATSSON AND K. C. WESTAWAY

206

YP

B H CL3

CL3

(L = H or D)

using several structurally related rigid tertiary amines [8]-[ll] as the bases : B in toluene and in DMSO at 20°C. The reaction proceeds via an ion-pair mechanism with the formation of the carbanion-ammonium ion-pair being rate determining. The KIEs were measured by the isotopic quasi-racemate method (Bergson et af., 1977) described in Section 6. The secondary @-deuteriumKIEs presented in Table 32 were inversely related to primary deuterium KIEs in both solvents. In DMSO, where the reaction is faster, the primary KIEs are larger and the secondary KIEs are slightly smaller than those in toluene. The larger primary KIE suggests that there is an earlier, more symmetric, transition state in the DMSO reactions (Melander, 1960b; Westheimer, 1961). Matsson concluded that the KIE would obviously be affected by both the inductive effect of the @-deuteriumsand by negative ion hyperconjugation and concluded that these secondary deuterium KIEs were partially due to the inductive effect and partially due to negative ion hyperconjugation. Saunders (1997) used quantum-chemical methods in the most recent attempt to establish the mechanistic importance of negative ion hyperconjugation. Earlier, Saunders (1976) had proposed that one could distinguish between an E2 and an irreversible Elcb(irr) mechanism (Lowry and Richardson, 1987) by measuring the leaving group KIE for the reaction. The

SECONDARY D-KINETIC ISOTOPE EFFECTS

207

Table 32 Primary and secondary /3-deuterium KIEs for the rearrangement of 1-methylindene to 3-methylindene using the rigid amines [8]-[ll] as catalysts in toluene and DMSO at 20°C." Toluene Base

PI 191 I 101 [I11

Primary k ~ f k ~

DMSO

Secondary (kCH3&D3)P

6.06 t 0.06 5.03 t 0.06 5.75 2 0.06

1.096 -C 0.002 1.103t 0.001 1.1002 0.001

-

-

Primary

Secondary

kHlkD

(kCH,lkCD,),

7.83 t 0.1 7.29 5 0.07 7.76 i 0.07 7.62 t 0.07

1.082 2 0.002 1.088 t 0.002 1.084 5 0.002 1.085 2 0.002

"Datataken from Matsson (1985).

observation of a leaving group KIE would indicate that the reaction proceeded via an E2 mechanism, whereas no leaving group KIE would be observed for an Elcb(irr) reaction because the bond to the leaving group is intact in the transition state of the rate-determining hydron transfer reaction. However, Thibblin and Ahlberg (1977) pointed out that a leaving group KIE could be observed for an Elcb reaction if negative ion hyperconjugation in the carbanion formed in the hydron abstraction reaction weakened the bond to the leaving group in the transition state. Saunders (1997) investigated this possibility by calculating the structure of three fluorine-substituted ethyl carbanions, [12]-[14], at the MP2/631 + G*//MP2/6-31+ G* level. CF~H-CH~

CF,-CH,

CF,-C FH

Three criteria were used to detect the presence of negative ion hyperconjugation; (i) a comparison of the lengths of the C,-F bond anti to the unshared pair of electrons in the carbanion and in its conjugate acid, AR,,; (ii) the relative charges on the F anti to the unshared pair of electrons in the carbanion and its conjugate acid, AQF; and (iii) the I8F/l9FEIE, K18/K19,for the anti F on ionization of the conjugate acid. In order to allow for inductive or other nonconformationally dependent effects, the same three criteria were also applied to the fluorine gauche to the unshared pair of electrons. All three criteria (Table 33) indicated that the anti C,-F bond in the carbanion was significantly weaker than the gauche C,-F bond; (i) the length of the anti C,-F bond increases much more than the length of the gauche C,-F bond upon ionization; (ii) the negative charge on the anti C,-F bond becomes much larger than that on the gauche C,-F bond; and (iii) the fluorine EIE for the anti fluorine is much larger than for the gauche fluorine. This provides strong

0. MATSSON AND K. C. WESTAWAY

208

Table 33 The change in the C,-F bond length, the charge on the fluorine atoms and the fluorine EIE due to negative ion hyperconjugation for three fluorinated carbanions." Carbanion Quantityb A&F(,Ii)

(4

ARC-F(gauche)(A> AQF(pnri< AQF(gauche)e

K "lK *&I,) K 181Kfgwche)

CFZH-CH;

CF3-CH;

CF3-CFH-

1.4120' 0.0086 -0.5068' -0.0254 1.0246' 1.0004

0.1536 0.0327 -0.1292 -0.0403 1.0152 1.0040

0.0823 0.0261d -0.0840 -0.0397d 1.0093 1.0028d ~~

~~~

~~

"Data taken from Saunders (1997). bValue for the carbanion minus the value for the conjugate acid. The anti fluorine in this carbanion is almost a free fluorine ion, i.e. the carbanion is effectively an F-CFH=CH2 ion-dipole complex. "Average value for the two nonequivalent gauche fluorines. Watural population analysis charges.

support for negative ion hyperconjugation to a p-fluorine in a carbanion. It is important to note that the calculations showed that hyperconjugation is also important in the transition state for the formation of the carbanion. Finally, it is satisfying that the amount of hyperconjugation suggested by all three criteria decreased from [12] to [14], i.e. with increasing inductive stabilization of the carbanion by fluorine atoms. All the possible primary and secondary KIEs for the rate-determining hydron transfer step of the Elcb(irr) reaction between hydroxide ion and l,l,l-trifluoroethane were also calculated (Table 34). The large secondary P-deuterium KIE of 1.146 suggested that there was significant negative charge on the p-carbon in the transition state and the large primary deuterium KIE of 4.7076 is suggestive of a reasonably symmetrical transition state (Melander, 1960b;Westheimer, 1961). Therefore, the author concluded that the transition state for the rate-determining hydron transfer step is reasonably symmetrical with the hydron slightly more than 50% transferred to base. However, the important observation is that small, but significant, fluorine KIEs were found for the Elcb(irr) mechanism, i.e. k''/kI9 = 1.0058 for the anti fluorine and 1.0022 for the gauche fluorine. The estimated maximum KIE is 1.032 for cleavage of a C-F bond (Matsson et al., 1993) and the largest k1'/kI9 observed experimentally is 1.027 (Matsson et al., 1993; Persson et al., 1996). Therefore, it seems that Thibblin and Ahlberg were correct and that small primary leaving group KIEs due to negative ion hyperconjugation can be found in Elcb(irr) reactions. Most of the reported secondary p-deuterium KIEs attributed to negative ion hyperconjugation are for anion or carbanion ion-pair-forming hydron

SECONDARY D-KINETIC ISOTOPE EFFECTS

209

Table 34 Calculated KIEs at various positions for the proton transfer transition state for the Elcb(irr) reaction between hydroxide ion and l,l,l-trifluoroethane at 20°C." k'81k'9 anti

gauche

k1'/k19

1.0058

1.Go22

(kCH3IkCD3)P secondary

kHlkD primary

1.1462

4.7076

(k1Z/k'3), (k'Z/k13)p

1.0199

1.0158

"Data taken from Saunders (1997).

transfer reactions. However, some anion-forming decarboxylations have been examined by Kluger and Brandl (1986a,b). For the decarboxylation reaction of 2,2-dimethylbenzoylaceticacid (reaction (38)), Kluger and Brandl (1986a) found a secondary p-deuterium KIE of 1.052 (= 1.0085 per D) in 0.1 mol dm-3 hydrochloric acid at 48.5"C and 1.111(= 1.0177 per D) for the reaction of the acid anion (conjugate base of the substrate) at pH 10 and 56.8"C.

(L = H, D)

The larger KIE is thus observed for the decarboxylation of the conjugate base which is the slower reacting substrate. This is consistent with the idea that the highest energy transition state is more product-like and, therefore, has the greatest need for hyperconjugative stabilization. The corresponding KIEs for the reactions of the substrates with only one deuterated methyl group were 1.027 (= 1.0089 per D) in 1 mol dm-' HCl and 1.07 (= 1.023 per D) at pH 10. This shows that the isotope effects are cumulative, i.e. that the rule of the geometric mean (Melander and Saunders, 1980d) holds. Kluger and Brandl (1986b) also studied the decarboxylation and basecatalysed elimination reactions of lactylthiamin, the adduct of pyruvate and thiamin (Scheme 2). These reactions are nonenzymic models for reactions of the intermediates formed during the reaction catalysed by the enzyme pyruvate decarboxylase. The secondary P-deuterium KIE for the decarboxylation was found to be 1.09 at pH 3.8 in 0.5 mol dm-3 sodium acetate at 25°C. In the less polar medium, 38% ethanolic aqueous sodium acetate, chosen to mimic the nonpolar reactive site in the enzyme, the reaction is significantly faster but the KIE was, within experimental error, identical to the KIE found in water. This clearly demonstrates that the stabilization of the transition state by hyperconjugation is unaffected by the change in solvent. A larger secondary P-deuterium KIE of 1.12 was found for the elimination reaction at pH 11 in water buffered with sodium carbonate. On the basis of these KIEs, it was concluded that the electron-rich centre developing in these reactions is stabilized by negative ion hyperconjugation.

0. MATSSON AND

210

/

C H ~ y T T H c H 2 c H 2 0 H

N\

K. C. WESTAWAY

CH:NYs

R

COOH

L,CCCOO

I

\

+ /%YS H

(L = H, D)

Scheme 2

SECONDARY P-DEUTERIUM KIEs DUE TO HYPERCONJUGATION IN CARBENE AND RADICAL REACTIONS

One report of a secondary P-deuterium KIE for a carbene insertion reaction has appeared recently. Pascal and Mischke (1991) found that the P-deuterium KIE for the insertion of dichlorocarbene into the benzylic C-H bond of cumene (reaction (39)) was (k,lk,& = 1.250 and 1.22 when the KIE was based on GC-MS analyses and 'H nmr, respectively.

This corresponds to an isotope effect of approximately 3.5% per deuterium. In comparison, the secondary P-deuterium KIEs in s N 1 reactions are all normal and range from 5% to 15% per deuterium. Because the normal KIEs in SN1 reactions result from the weakening of the C,-L bond by a hyperconjugative interaction with the incipient carbocation in the transition state, the authors concluded that hyperconjugative interactions are present also in the transition state for the insertion reaction. The normal secondary p-deuterium KIE observed for the insertion reaction is consistent with the dipolar three-centre transition state structure [15] proposed by Seyferth et a!. (1970a,b) because the partial positive charge on the a-carbon is stabilized by hyperconjugation.

SECONDARY D-KINETIC ISOTOPE EFFECTS

6-

c12q- -

211

---p

Only a few secondary p-deuterium KIEs have been measured for homolytic reactions and the mechanistic utility of these, invariably small, KIEs is limited. The first secondary @deuterium KIE for a radical reaction was published by = 1.052 (= 1.017 Seltzer and Hamilton in 1966. They found a small (kHlkD)p per D) for the thermal decomposition of the azo-bis(1-phenylethane) (reaction (40)) at 105°C in ethylbenzene and concluded that it was due to hyperconjugation C,H,-CH-N=N-

I

CL

CH-C6HS

-*

Nl + 2 C6HS-CH

I

I

CJ-3

(40)

CL3

Koenig and Wolf (1967) reported a similarly small value of 1.195 (=1.02 per D) for the formation of t-butyl radical from the perester of pivalic acid (reaction (41)). 40 (CLdF, 0-OH

-

(CL3)3C'

+

c02

(41)

This KIE was also attributed to hyperconjugation. The authors suggested that smaller secondary deuterium KIEs were found in radical reactions because hyperconjugation was less important in radicals than in carbocations. Finally, Holm and 0gaard Madsen (1992) determined the secondary P-deuterium KIEs for the addition of P-deuterated Grignard reagents to unsaturated ketones. The authors attributed the small, <5%, KIEs they found in these radical reactions to opposing steric and hyperconjugative effects. 4

Secondary deuterium KIEs and tunnelling

Quantum-mechanical tunnelling has been recognized as a possible contributor to the rate of a chemical reaction for many years. For instance, the theory of tunnelling for proton transfer reactions was developed by Bell (1959) in his famous book The Proton in Chemistry. Later, Bell (1980a) published a more thorough treatment of tunnelling in his book The Tunnel Effect in Chemistry.

212

0. MATSSON AND K. C.WESTAWAY

According to the theory, three factors promote tunnelling: (i) low mass of the particle being transferred; this makes tunnelling more likely for hydrogen than for the heavier isotopes deuterium and tritium; (ii) low temperature: tunnelling becomes more important as the temperature decreases and vanishes at the high-temperature limit; and (iii) a narrow potential energy bamer for the chemical reaction; the probability of tunnelling increases as the energy barrier becomes narrower. nnnelling has sometimes been regarded as a mysterious phenomenon by chemists. It is worth stressing, therefore, that tunnelling has the same firm foundation in quantum mechanics as zero-point energy, which is the most important component of a KIE; both these phenomena are a consequence of Heisenberg’s uncertainty principle. Because of their dependence on mass, KIEs have been used in two ways to detect tunnelling. One is that primary deuterium KIEs are larger than predicted on the basis of zero-point energy alone when tunnelling makes a significant contribution to the KIE. For example, primary deuterium KIEs larger than 25 have been reported (Lewis and Funderburk, 1967;Wilson ef al., 1973) for proton transfer reactions where tunnelling is important. The second method of detecting tunnelling relies on the fact that the primary hydrogen KIE shows an anomalous temperature dependence when significant tunnelling takes place. In the absence of tunnelling, the temperature dependence of the rate constant should follow the Arrhenius equation (42) Ink = 1nA - EJRT

(42)

where k is the rate constant, E, is the activation energy, A is the Arrhenius pre-exponential factor and T is the absolute temperature. In the presence of tunnelling, the rate constant at low temperature will be greater (the E,/R term will be smaller) than predicted by the equation and the Arrhenius plot (Fig. 16) will be curved. When this occurs, extrapolation to 1/T = 0 yields a value of In A which is smaller (In Annnel)than that predicted by the linear equation (In Aclass).Because of the mass dependence of tunnelling, AH (for protium) will be underestimated more than AD (for deuterium) and AT (for tritium). As a result, the isotopic ratios of the pre-exponential factors, AH/AD and AHIAT, will be less than unity. These ratios are expected to be close to unity in the absence of tunnelling. Of the two methods, the temperature dependence of the KIE has generally been regarded as the more reliable criterion for tunnelling. Because unexpectedly large primary deuterium KIEs are observed in reactions where tunnelling is important, and unexpectedly large secondary deuterium KIEs have been observed in some hydron transfers in elimination and enzyme-catalysed hydride transfer reactions, Saunders (vide infra) wondered whether very large secondary deuterium KIEs were also indicative of tunnelling.

SECONDARY D-KINETIC ISOTOPE EFFECTS

213

I

1/T

Fig. 16 The curved Arrhenius plots found when tunnelling is important. Reproduced, with permission, from Bell (1980a).

LARGE SECONDARY DEUTERIUM KIEs IN HYDRIDE TRANSFER REACTIONS

Isotope effects have been used to determine whether the hydride transfer from the enzyme cofactor nicotinamide-adenine dinucleotide (NADH) (reaction (43))-takes place as a hydride ion transfer in a single kinetic step or in a multistep reaction via an uncoupled electron and hydrogen transfer. H CONHp

I

R

In one of these studies, Kurz and Frieden (1980) observed the first unexpectedly large secondary a-deuterium KIE. They found that the secondary a-deuterium KIE for the nonenzymatic hydride ion reduction of 4-cyano-2,6-dinitrobenzenesulfonateby NADH (reaction (44)) was 1.156 2 0.018 and 1.1454 ? 0.0093 using direct and competitive kinetic methods, respectively. The corresponding equilibrium isotope effects (EIEs) were found to be 1.013 3 0.020 and 1.0347 t 0.0087, respectively. Thus, the secondary deuterium KIE was much larger than the EIE. The magnitude of a secondary a-deuterium KIE is normally attributed to the rehybridization of the a-carbon that takes place when the reactant is transformed into the

0. MATSSON AND K. C. WESTAWAY

214

transition state. Therefore, the expectation is that the secondary KIE should increase monotonically as the transition state changes from reactant-like to product-like. If this model is correct, the EIE is a good estimate of the maximum secondary a-deuterium KIE for the reaction, i.e. the KIE for a very product-like transition state.

+

NADH

-

@

+

NAD

(a)

oz$NozCN Kurz and Frieden (1980) offered two different explanations for the fact that the secondary a-deuterium KIE is much larger than the EIE. One explanation, attributed to Kresge, is based on the idea that the transition state has a feature that is not present in the reactant or product. If the transition state is early and the a-hydrogen has not yet moved significantly into the plane of the ring, a hyperconjugative interaction in which the a-hydrogen stabilizes the developing positive charge on the dihydropyridine nitrogen could occur. Such an interaction is possible for the transition state but not for the reactants or products. A consequence of this hyperconjugation is that the C,-H(D) bond is weakened in going to the transition state. This means that a normal (>1.0) secondary deuterium KIE will be observed. The second explanation was suggested by Kurz and Frieden. Because the masses of the leaving group (the hydride) and the isotopes generating the secondary a-deuterium KIE are comparable, it may be possible that bending motion of the a-hydrogen (deuterium) is part of the reaction coordinate motion. In these circumstances, the large observed KIE is not purely secondary but is the product of a secondary and a primary KIE. It is worth noting that finding a secondary a-deuterium KIE larger than the EIE is not unique. In fact, it has been found in several other reactions. For instance, Cleland and co-workers (Cook et al., 1980,1981; Cook and Cleland, 1981a,b) found unexpectedly large secondary a-deuterium KIEs in some enzymatic reactions; for example, a secondary a-deuterium KIE of 1.22 for the reduction of acetone catalysed by yeast alcohol dehydrogenase and a KIE of 1.34 for the reduction of cyclohexanone catalysed by horse-liver dehydrogenase. The consequences on the magnitude of the secondary a-deuterium KIE of coupling the motion of the nontransferring a-hydrogen into the reaction coordinate motion, as suggested by Kurz and Frieden, was investigated in some model calculations by Huskey and Schowen (1983). They used two different models to calculate the secondary isotope effects for the hydride transfer reaction (45).

SECONDARY D-KINETIC ISOTOPE EFFECTS

215

The first model was very simple. It is symmetrical with respect to the reactants and products and it violates the so called “cut-off” rules of Wolfsberg and Stern which state that atoms positioned two bonds away from the isotopic site should be included in the calculations (Wolfsberg and Stem, 1964; Stern and Wolfsberg, 1966). This model was chosen so that a large number of calculations could be made rapidly. The second, more realistic, model using the reactants shown in reaction (45) was used to calculate the KIEs for certain selected transition state structures suggested by the simpler model. Huskey and Schowen displayed their results as a function of (i) the imaginary frequency at the reaction barrier, which is related to the curvature at the top of the reaction barrier and (ii) the amplitude of the motion of the nontransferring a-hydrogen relative to the transferring hydrogen in the reaction coordinate motion. When the imaginary frequency at the reaction barrier was small and, therefore, little tunnelling would occur, the secondary a-deuterium KIE was smaller than the EIE and the calculated primary KIE was normal, i.e. close to values observed experimentally, when there was little motion of the a-hydrogen in the reaction coordinate (point A in Table 35). When the motion of the nontransferring a-hydrogen (HI) was large relative to that for the transferring hydrogen (H2) and there was little tunnelling, the secondary deuterium KIE increased, but at the same time the primary KIE decreased to well below the experimental values (point B, Table 35). Hence, when there is little tunnelling, the secondary a-deuterium KIE can only increase at the cost of the primary deuterium KIE. When tunnelling was increased by making the reaction coordinate frequency larger and the motion of the secondary (nontransferring) hydrogen in the reaction coordinate vibrational mode was small, large primary KIEs and normal secondary KIEs Table 35 Calculated secondary a-and primary deuterium KIEs for the model hydride transfer reaction (4.9“

Point A B C D

Imaginary frequencylcm-’ <200

lo00 >lo00

Relative amplitude of H2/HI motion in reaction coordinate vibration

(kdkD)u

0.06 1.7 0.04 0.6

1.027 1.078 1.027 1.252

“Datataken from Huskey and Schowen (1983).

Primary k ~ l k ~

5.99 1.40 13 4.7

216

0. MATSSON AND K. C. WESTAWAY

(less than the EIE) were obtained (point C). However, when tunnelling and a large amplitude of the nontransferring a-hydrogen in the reaction coordinate were combined, the observed phenomena were reproduced, i.e. the secondary a-deuterium KIE was larger than the EIE and the magnitude of the primary KIE was normal (point D). It is worth noting that these calculations reproduced another aspect of the experimental results found by Kurz and Frieden and by Cook et al., namely that increasing the mass of the transferring hydrogen by deuteration reduces the secondary a-deuterium KIE. For example, Cook et al. (1981) reported that the secondary a-deuterium KIE decreased from 1.22 to 1.07 when deuterium rather than hydrogen was transferred in the formate dehydrogenase-catalysed reaction between DPN-4-d and formate ion. This point has been explored further by Saunders (vide infra). Increase of the transferring mass from 1amu for hydrogen to 16 amu, thereby simulating reactions such as solvolysis and acyl transfer, made the anomalous phenomena disappear. Therefore, an increase in the reduced mass of the reaction coordinate motion, whether caused by deuteration or transferring a heavier group, reduces the importance of tunnelling and its observable consequences. TUNNELLING IN THE HYDRON TRANSFER STEP OF P-ELIMINATION REACTIONS

In a series of papers, Saunders (1984,1985) has systematically investigated the effect of tunnelling on the magnitude of the secondary a-deuterium KIEs for the hydron transfer step in p-elimination reactions. Some of this work has already been reviewed (Saunders, 1992). An important aspect of this investigation was the fruitful combination of experiment and theory. Experimental data were not just reproduced by the model calculations, but predictions based on the model calculations were made, tested and verified by experiment. These studies were also important because, in calculations of this kind, one must choose several parameters arbitrarily. In spite of this apparent problem, the predictions based on Saunders’ calculations were confirmed experimentally and, moreover, the set of arbitrarily chosen parameters reproduced not just one but several sets of data. Therefore, one can have considerable trust in both the results and the method. While these calculations have identitied reaction systems where tunnelling is significant, the most important contribution of the work is that it developed new criteria for determining when tunnelling occurs in chemical reactions. This work has demonstrated that (i) the magnitude of the secondary deuterium KIE, (ii) the temperature dependence of the secondary deuterium KIE, (iii) the relationship between the secondary tritium and deuterium KIEs and (iv) the isotope effects on isotope effects, i.e. the failure of the rule of the geometric mean (Melander and Saunders, 1980d) for the KIEs found when several positions are labelled, can all be used to detect the presence of tunnelling.

SECONDARY D-KINETIC ISOTOPE EFFECTS

217

THE MAGNITUDE OF THE SECONDARY HYDROGEN KIE AS A CRITERION FOR TUNNELLING

In these papers, Saunders (Subramanian and Saunders, 1981; Saunders, 1984) reported a secondary tritium KIE of k H / k D = 1.31 for the E2 reaction between ethoxide ion and 2-phenylethyltrimethylammonium ion (46) in ethanol at 40°C. C6HS-CHL-CH2-fiMe3

+ EtO-

-

C6H5-CL=CH2 + NMe,

+ EtOH

(46)

Since the reaction is not reversible, the EIE could not be measured. However, the secondary deuterium EIE could be estimated using the fractionation factors published by Hartshorn and Shiner (1972). This approach predicted would be equal to 1.115 at 45°C. This that the secondary EIE, (KHIKD)sec, corresponds to a (KHIKT)sec = 1.170 in the absence of tunnelling. Because the secondary tritium KIE is much larger than the EIE, it seems likely that tunnelling is important in this reaction. Saunders (1984) carried out BEBOVIB calculations (Sims and Lewis, 1985) for the model reaction in (47) in an attempt to determine whether the large secondary tritium KIE found for the reaction in equation (46) was due to tunnelling. CCHDCH2-CI

+ OH-

-

CCD=CHZ + C1- + H20

(47)

Five different models were examined in these calculations with (i) different curvatures of the reaction barrier (the amount of tunnelling) and (ii) different coupling between the bending vibrations of the nontransferring p-hydrogen and the C,-H stretching vibration of the transferring @hydrogen in the reaction coordinate vibrational mode. lhnnelling was accounted for by the truncated Bell equation (48) (Bell, 1980b)

Ql=

0.5 u, sin (0.5 u,)

where Qt is the tunnel correction and the tunnelling frequency ut = 1%1, the absolute value of the imaginary reaction coordinate frequency. Only two of the five models (Fig. 17) were able to reproduce the experimental results; the secondary tritium KIE only exceeds the EIE when (i) strong coupling between the bending vibrations of the nontransferring P-hydrogen and the stretching vibrations of the transferring P-hydrogen characterizes the reaction coordinate vibration and (ii) the curvature of the barrier is large and tunnelling is important. Interestingly, for models 3 to 5 in Fig. 17, there is no correlation between the magnitude of the secondary tritium KIE and the extent of rehybridization of the P-carbon as measured by the

0.MATSSON AND K. C. WESTAWAY

218

‘ i / I

I\

I

I

1.20 - - -

k~I k~ 1.10

- -

-

I

I

I

I

I

I

0.1

0.3

0.5

0.7

0.9

4.0

Fig. 17 Plot of the calculated secondary deuterium KIE versus the extent of 0-H bond formation for the model elimination reaction at 45°C. Models 1and 2 have different imaginary frequencies and no coupling of the C,-D bending vibrational motion with the C,-H stretching motion in the transition state. Models 3 , 4 and 5 have increasing extents of coupling between the C,-D bending and C,-H stretching motion in the transition state. Reproduced, with permission, from Saunders (1997).

0 - H transition state bond order, nfi-. Instead, the secondary @tritium KIE reflects the extent to which motion of the nontransferring P-hydrogen is involved in the motion along the reaction coordinate. More experimental data where the secondary KIE was larger than the EIE were subsequently published by Subramanian and Saunders (1984). The 2-arylethyl system was employed in these studies because other relevant data, such as the primary deuterium KIE, were available for this reaction. Special techniques were developed to determine the primary and the secondary tritium KIEs for this system. Three isotopically distinct elimination reactions (49-51) are possible for a 2-arylethyl derivative tracer labelled with tritium at the 2-position. Ar-CH,CH,-X

+ -OR

Ar-CHTCH2-X

+ -OR

Ar-CHTCH,-X

+ -OR

-* H!

Ar-CH=CH2

+ ROH + X-

(49)

Ar-CH=CH,

+ ROT + X-

(50)

Ar-CT=CH2

+ ROH + X-

(51)

k&

SECONDARY D-KINETIC ISOTOPE EFFECTS

219

The primary tritium KIE (G/@)and the secondary tritium KIE (G/k&) were determined in two different experiments carried out in the following way. If R, is the radioactivity of the original reactant and RRoHis the radioactivity of ROH (ROT) determined at low (4%) extents of reaction, the primary tritium KIE is given by equation (52). If the radioactivity of the styrene isolated at low extents of reaction, Rs, is also measured, the secondary tritium isotope effect can be calculated from (53).

(k3G)= f @ d R R O H )

(52)

( k g / k a = f(Ro/Rs)

(53)

The secondary deuterium KIEs obtained by converting the secondary tritium KIEs found for the E2 reactions of several different 2-arylethyl substrates into secondary deuterium KIEs with the Swain-Schaad equation (Swain et al., 1958) are in Table 36. As discussed above, one would expect the secondary deuterium isotope effect to reflect the extent to which rehybridization of the p-carbon from sp3 of the reactant to sp2 in the product has taken place in the transition state. According to this reasoning, the secondary tritium EIE should be a good estimate of the maximum secondary tritium KIE. Because these reactions were not reversible, the EIE could not be measured. However, one can estimate the EIE (the maximum expected secondary KIE) using Hartshorn and Shiner’s (1972) fractionation factors. The predicted EIE (KH/KD)values were 1.117 at 40°C and 1.113 at 50°C. Seven of the reactions Table 36 Secondary p-tritium KIEs for the E2 elimination reactions of PhCHTCH2-X at 50°C.” Leaving group NMe3 NMe3 me3

NMe3 NMe, NMe3 SMe2 SMe2 SMe, SMe, OTsBrBr-

Base/solvent

kglk;

EtO-EtOH EtO-IEtOH OH-/7O% aq. Me,SO OH-/60% aq. Me,SO OH-/50% aq. Me,SO EtO-/EtOH EtO-EtOH OH-/70% aq. Me2S0 OH-/60% aq. Me,SO OH-/50% aq. Me2S0 Bu‘O-/Bu‘OH EtO-IEtOH Bu‘O-/Bu‘OH

1.259 2 0.010 1.311-+ 0.014 1.235 2 0.016 1.250 2 0.023 1.243 2 0.031 1.284 2 0.030 1.157 t 0.022 1.119 2 0.023 1.144 2 0.026 1.134 2 0.026 1.239 2 0.023 1.110 2 0.024 1.071 ? 0.026

“Data taken from Subramanian and Saunders (1984). bCalculatedfrom k H l k D = ( ~ H / / c T ) ~ ’ ~ . ~ . ‘At 40°C. dSubstrate is p-CIC&CHTCH26Me3

kglkg 1.17’ 1.21“ 1.16 1.17 1.16 1.w 1.11 1.08 1.10 1.09 1.16 1.08 1.05

220

0. MATSSON AND K. C. WESTAWAY

in Table 36 have secondary deuterium KIEs that are significantly larger than the EIE. Since the primary KIEs for these seven reactions range from 3.0 to 8.0 (Brown et al., 1981; Saunders and Edison, 1960; Kaldor and Saunders, 1979) and are too large to be compatible with very product-like E2 transition states, i.e. the Melander-Westheimer curve indicates that reactions where the p-hydrogen is almost completely transferred to base in the transition state have small primary deuterium KIEs (Melander 1960b;Westheimer, 1961), one would expect the secondary p-deuterium KIEs for these reactions to be significantly smaller than the EIE. The authors, therefore, concluded that rehybridization cannot be the sole cause of these large secondary p-deuterium KIEs. On the basis of the model calculations by Saunders (vide supra), the large secondary KIEs were attributed to tunnelling. The tunnel correction enters because the bending vibrations of the nontransferring hydrogen are coupled to the C,-H(T) stretching vibrations of the transferring hydrogen. In other words, the extent to which motion of the nontransferring hydrogen contributes to the motion along the reaction coordinate determines the degree to which tunnelling causes an enlarged secondary KIE. In fact, since the calculations by Saunders (1984) suggest that the secondary KIE is largest for the reactions where the hydrogen is between 30% and 50% transferred to the base in the transition state (Fig. 17), the magnitude of the secondary p-deuterium KIE is clearly not related to the amount of rehybridization at the p-carbon in the transition state. An examination of the kwlkz KIEs in Table 36 shows that the nature of the leaving group rather than a change in solvent or base is primarily responsible for the variation in the KIE. For instance, changing the base and solvent from ethoxide ion in ethanol to hydroxide ion in 50% aqueous DMSO when the leaving group is trimethylamine only changes kglkz from 1.26 to 1.24. Similar small changes from 1.16 to 1.13 are found when the leaving group is dimethyl sulfide. The variation in the KIE due to leaving group was thought to be related to the amount of carbanion character on the p-carbon of the transition state, i.e. the more Elcb-like the transition state, the greater the secondary p-deuterium KIE and therefore the greater the amount of tunnellinglphydrogen motion in the reaction coordinate.

PREDICTIONS OF TUNNELLING CRITERIA BASED ON MODEL CALCULATIONS

Saunders (1985) extended his investigation of the effect of tunnelling on the magnitude of secondary deuterium KIEs in a theoretical study of the E2 reaction between hydroxide ion and a model substrate (reaction (54)). C-CH&HZ-Cl+

OH-

C-CH=CHz

+

+ C1- + HZO

(54)

The KIE calculations were carried out on the model reaction using the BEBOVIIB approach (Sims and Lewis, 1985) and assuming that the changes

SECONDARY D-KINETIC ISOTOPE EFFECTS

221

in all the reacting bonds were synchronous, i.e. that the total bond order to a given atom was conserved [16]. The order of the bond between the transferring /3-proton

Hob[I61

and the base in the transition state, n&-, was varied so as to create a series of transition states ranging from reactant-like to product-like. Off-diagonal force constants were chosen that yielded reaction coordinate motions in the transition state that converted reactants smoothly to products. The parameters used in that procedure were chosen to vary the coupling between the stretching vibration for the transferring hydron and the bending vibrations of the nontransferring hydron. The so called curvature parameter, D, was also varied. Increasing D , i.e. increasing the curvature of the potential barrier at the transition state, increases the value of the imaginary frequency and increases the amount of tunnelling. The tunnel corrections were calculated from the first term of the Bell equation (48). This is the simplest, and most commonly used, method for estimating the tunnel correction to an isotope effect, although it is debatable whether such a simple model yields meaningful estimates. The four models investigated in these BEBOVIB calculations had similar contributions of hydron transfer to the reaction coordinate motion, but the amount of coupling with the bending vibrations of the nontransferring hydron was increased from totally uncoupled (model 1) to highly coupled in model 4. In models 3 and 4, the curvature parameter, D,was increased to increase the amount of tunnelling. The KIEs calculated for these model transition states (Table 37) were very interesting. The first observation was that all the models yielded reasonable values for the primary deuterium KIE, i.e. between 1.00 and 8.76. The second observation was that the secondary deuterium KIE, in the absence of tunnelling, increases monotonically towards the EIE of KHIKD= 1.124 as nA-o increases. This is, in fact, the semiclassical trend expected for secondary deuterium KIEs. The same pattern is observed when tunnelling is included (model 1) but there is no coupling between the stretching and bending vibrations of the transferring and nontransferring hydrons. However, for the models with such coupling (models 2, 3 and 4), the tunnelling contribution to the KIE becomes very large and is the dominant factor in determining the magnitude of the secondary deuterium KIE when

0. MATSSON AND K. C. WESTAWAY

222

Table 37 The calculated secondary deuterium and tritium JSIEs for the E2 reaction between OH- and CCH2CH2-Cl at 45°C."

Model 1

2

3

4

n1:o-H

(kHHlkHD)scl'

(kslk:)

(kElk&b

0.1 0.3 0.5 0.7 0.9 0.1 0.3 0.5 0.7 0.9 0.1 0.3 0.5 0.7 0.9 0.1 0.3 0.5 0.7 0.9

1.014 1.025 1.048 1.082 1.113

1.016 1.027 1.053 1.086 1.113

1.022 1.033 1.067 1.112 1.158

1.025 1.037 1.074 1.118 1.158

1.027 1.032 1.058 1.088 1.103

1.032 1.069 1.088 1.093 1.103

1.042 1.042 1.077 1.120 1.140

1.051 1.098 1.123 1.128 1.410

1.037 1.031 4.020 1.107 1.110

1.055 1.184 6.665 1.142 1.111

1.057 1.038 1.056 1.148 1.151

1.084 1.266 1.197 1.202 1.153

1.038 1.027 1.052 1.110 1.114

1.064 1.270 1.267 1.162 1.115

1.056 1.032 1.067 1.152 1.158

1.097 1.392 1.385 1.232 1.160

(kEIk3

"Data taken from Saunders (1985). bSemiclassical(no tunnelling correction used).

= 0.3-0.7. For such transition states, the secondary deuterium KIE exceeds the EIE. These results are qualitatively similar to those in Fig. 17. It is interesting that the tunnelling contribution to the KIE is small for very reactant-like and very product-like transition states. Finally, the secondary tritium KIEs show the same trends as the secondary deuterium KIEs in Table 37. However, McLennan (1979) calculated larger than normal primary deuterium KIEs, of magnitudes similar to those found in reactions where tunnelling was important, for hydron transfer reactions with loose transition states, i.e. where the total bond order to the transferring hydrogen is not conserved in going to the transition state. This hypothesis was tested by putting nA-o = nLH= 0.2 in Saunders' BEBOVIB models. The resulting secondary deuterium KIE was increased, but not to the extent that it exceeded the EIE. Therefore, it appears that secondary deuterium KIEs larger than the EIE are only observed when the stretching and bending vibrations for the transferring and nontransferring hydrogens are coupled and tunnelling is important.

SECONDARY D-KINETIC ISOTOPE EFFECTS

223

The temperature dependence of the KIEs and the isotopic ratios of the pre-exponential factors AHIAD and AHIAT were calculated for each of the five transition states for the model reaction (54) since these ratios have been taken as reliable evidence for tunnelling in primary KIEs. The AHIAD and AHIAT values for the primary KIEs were close to unity for the models with very little tunnelling, but were less than 1.0 for the models with extensive tunnelling. For example, for model 4 where n&-o = 0.3, AHIAD = 0.033 and AHIAT = 0.019. Thus, the calculations show that these isotopic ratios of pre-exponential factors for primary hydrogen KIEs are excellent indicators of tunnelling. The corresponding AHIAD and AHIAT for the secondary KIEs were 0.284 and 0.190, respectively. This suggests that observing a small AHIAD or a small AHIAT for a secondary KIE is also a criterion for tunnelling. Another criterion that has been used for detecting the presence of tunnelling is that the Swain-Schaad relation (55)

which is based on only the semiclassical KIE and does not consider tunnelling (Swain et al., 1958), fails to predict the experimental kH/kT from the known primary kHlkD or vice versa. Although this criterion for tunnelling has not proved very useful, Saunders calculated the primary and secondary deuterium and the secondary tritium KIEs for the four models of reaction (54) in order to determine whether the Swain-Schaad relation held for deuterium and tritium KIEs when tunnelling occurred. The Swain-Schaad relationship failed to predict the correct primary kHlkTfrom the known primary kHlkDor vice versa when tunnelling was important, but (55) gave the experimental secondary KIE even when tunnelling was important. This suggests that the failure of the Swain-Schaad equation for secondary KIEs will not be of diagnostic value for tunnelling.

THE RELATIONSHIP BETWEEN THE MAGNITUDE OF SECONDARY DEUTERIUM AND TRITIUM KIEs AND THE RULE OF THE GEOMETRIC MEAN

Earlier calculations by Huskey and Schowen (1983) (p. 215) had shown that the secondary deuterium KIE decreased when the transferring atom in a hydride transfer reaction was changed from hydrogen to deuterium. Saunders (1985) also used calculations on his model E2 reaction (54) to determine whether the secondary deuterium KIE in a proton transfer reaction decreased with an increase in the mass of the transferring hydron. In fact, Saunders observed the same phenomenon for his hypothetical E2 elimination reaction. The secondary KIE calculated for models 3 and 4 (with coupling and a large curvature to the reaction barrier) with the tunnelling correction decreased when the transferring hydron was changed to a heavier isotope. For instance,

0. MATSSON AND K. C. WESTAWAY

224

the secondary KIE in model 4 decreased from 1.267 (Table 37) to 1.161 to 1.135 (Table 38) when the transferring hydron was changed from a proton to a deuteron to a triton in the model reaction and nfI-o was 0.5. Therefore, it appears that the mass of the transferring hydron has a significant effect on the magnitude of secondary deuterium KIEs when tunnelling is important. However, as previously noted, the effect is only observed when the mass of the transferring atom is similar to the mass of the nontransferring hydron. Saunders also used calculations on his model reaction (54) to determine the relationship between the secondary hydrogen-deuterium (secondary hydrogen-tritium) KIEs and secondary deuterium-tritium KIEs for doubly labelled substrates and to investigate how tunnelling affects this relationship. To obtain secondary KIEs that could be checked experimentally, Saunders calculated the secondary k,Hlk,D and e l k : KIEs for the hypothetical E2 reaction of a pair of doubly labelled substrates [17] and [18]. C-CD2CHz-Cl

C-CDTCH2-Cl

If this mixture were used as the substrate in an experiment, one could measure the primary k$k: and the secondary kElk; KIEs in the same way that Subramanian and Saunders (1981) determined the primary k$k? and the secondary kglkT, KIEs using equations (52) and (53). Saunders calculated the primary and secondary hydrogen-deuterium and hydrogen-tritium KIEs for his model reaction (i) assuming that the KIEs are semiclassical and there is no tunnelling and (ii) by including the correction for tunnelling in equation (48). The semiclassical primary and secondary hydrogen-deuterium and hydrogen-tritium KIEs can be estimated from the semiclassical primary and secondary deuterium-tritium KIEs using the relationships (56) and (57).

kHlkD= (kD/kT)2.26

(56)

kHlkT= (kD/kT)3.26

(57)

It is important to note that these equations are based on the Swain-Schaad relationship, which assumes that there is no tunnelling in any of the isotopic reactions (the KIEs are semiclassical) and that the relationship between the KIEs is determined only by the masses of the hydrogen, deuterium and tritium atoms. The secondary k,Hlk,D, and Pdk; KIEs calculated both with and without the tunnelling correction are virtually identical when there is little or no coupling (models 1 and 2, Table 38), i.e. equations (56) and (57) hold. However, the values calculated using the tunnelling correction are much larger than the semiclassical KIEs predicted by (56) and (57) when tunnelling and

SECONDARY D-KINETIC ISOTOPE EFFECTS

225

coupling are important (models 3 and 4, Table 38) and the difference between the semiclassicaland the tunnelling-corrected KIEs increases with the amount of coupling and tunnelling. For the primary KIEs, the semiclassical KIEs are smaller than the KIEs calculated using the tunnelling correction when tunnelling is important (models 1 4 ,Table 38). Obviously, the Swain-Schaad type relationships fail for both the primary and the secondary KIEs when tunnelling is important. These results indicate that a more reliable method for detecting tunnelling is to use doubly labelled substrates. Because tunnelling decreases as the mass of the transferring atom increases, the relative contribution of tunnelling to a deuterium-tritium KIE is substantially less than that to either a hydrogendeuterium or a hydrogen-tritium KIE. Therefore, comparison of a deuteriumtritium KIE with a hydrogen-deuterium or a hydrogen-tritium KIE via (56) or (57) is a much better way of detecting tunnelling than comparing a hydrogen-deuterium and a hydrogen-tritium KIE. The results of these calculations have implications on the applicability of the rule of the geometric mean, which indicates that the KIE for a doubly labelled species should be the product of the KIEs for the corresponding singly labelled substrates. For instance, the KIE for the doubly labelled [17] should be the product of the secondary deuterium KIE, kElkE, associated with the nontransferring hydrogen and the primary deuterium KIE, kElkg, produced by the transferring hydrogen (equation 58)).

This is only true when

However, the calculations show that (59) does not apply when tunnelling is significant. Therefore, the rule of the geometric mean is not expected to hold if isotopic substitution in one position affects the tunnelling contribution to a KIE at another position. Saunders and co-workers (Amin er al., 1990) used E2 elimination reactions in the p-substituted 2-phenylethyl system to test the new criteria for tunnelling suggested by the above calculations. The actual substrates and baseholvent bromide, systems they used were: (2-phenylethyl-2-t)-trimethylammonium [19], with sodium ethoxide in ethanol, 2-phenylethyl-2-r bromide, [20], with potassium t-butoxide in t-butyl alcohol and 2-(p-chlorophenyl)ethyl-2-? tosylate, [21], with potassium t-butoxide in t-butyl alcohol. When equation (57) was applied to the experimental secondary (kglkT,)KIEs in Table 39, the calculated kE/kLKIEs were 1.106 2 0.033 and 1.092 ? 0.026 for [19] and [XI,

Table 38 Calculated primary and secondary deuterium and tritium KLEs for the E2 reaction between OH- and CCH2CH2-Cl at 45°C."

1.014 1.021 1.029 1.034 1.038

1.567 2.022 2.023 1.567 1.304 1.529 1.898 1.884 1.642 1.279

1.582 2.295 2.276 1.784 1.306 1.548 2.134 2.099 1.692 1.281

1.014 1.014 1.024 1.040 1.041

1.020 1.043 1.052 1.051 1.042

1.540 1.829 1.817 1.615 1.275

1.014 1.013 1.022 1.041 1.043

1.023 1.052 1.062 1.057 1.044

1.539 1.804 1.7% 1.607 1.273

1.588 2.200 2.141 1.697 1.280 1.606 2.245 2.171 1.703 1.279

0.1 0.3 0.5 0.7 0.9 0.1 0.3 0.5 0.7 0.9

1.013 1.030 1.056 1.013 1.123 1.021 1.037 1.064 1.092 1.113

1.014 1.031 1.059 1.014 1.123 1.024 1.052 1.078 1.096 1.113

1.012 1.032 1.060 1.012 1.127

1.012 1.033 1.063 1.012 1.127

1.006 1.010 1.019 1.006 1.044

1.007 1.011 1.021 1.007 1.045

1.017 1.038 1.066 1.094 1.117

1.019 1.046 1.074 1.096 1.117

1.011 1.013 1.022 1.032 1.038

3

0.1 0.3 0.5 0.7 0.9

1.029 1.043 1.070 1.110 1.119

1.039 1.104 1.134 1.134 1.120

1.024 1.046 1.075 1.111 1.123

1.030 1.083 1.117 1.129 1.124

4

0.1 0.3 0.5 0.7 0.9

1.030 1.042 1.069 1.114 1.123

1.044 1.129 1.161 1.149 1.124

1.025 1.047 1.076 1.115 1.127

1.035 1.099 1.135 1.142 1.128

1

2

"Data taken from Saunders (1985). bSemiclassical(no tunnelling correction used).

227

SECONDARY D-KINETIC ISOTOPE EFFECTS

respectively. Both these KIEs are much smaller than the directly observed secondary tritium KIEs of 1.2042 and 1.1907, respectively.

Table 39 The secondary P-hydrogedtritium and deuteriudtritium KIEs for the E2 reactions of p-YC&,CLTCH2X at 50°C." YK

Base/solvent

L

elk:

WNMe3 H/NMe3 CYOTs CYOTs

EtO-IEtOH EtO-IEtOH Bu'O -/Bu'OH Bu'O-ISu'OH

H D H D

1.2042 ? 0.0149 1.0314 2 0.0099 1.1907 2 0.0122 1.0274 ? 0.0077

"Data taken from Amin

et al.

(1990).

Qualitatively the results are explained in the following way. Although the transferring deuterium atom does not introduce a primary isotope effect due to zero-point energy differences into kglkT,, there is less tunnelling when deuterium is transferred than when hydrogen is transferred. Therefore, the tunnel correction to the secondary kB/k', is small relative to that for kglk;. Thus, the experimental results are in agreement with the results of the model calculations. Deviations from equation (57) have also been used to demonstrate that tunnelling is important in the enzyme-catalysed oxidation of benzyl alcohol to benzaldehyde by NAD' and yeast alcohol dehydrogenase (YADH) (reaction (60)) (Cha er at., 1989; Klinman, 1991).

~

L

&'

~ + O

I

R

H

Y*DH

.

L +

H

+ H+

(60)

0. MATSSON AND K. C. WESTAWAY

228

Table 40 The primary and secondary deuterium-tritium and hydrogen-tritium KIEs for the oxidation of benzyl alcohol to benzaldehyde with NADf and yeast alcohol dehydrogenase at 25°C."

KIE

Primary

Secondary

kDlkT

1.73 t 0.02 5.91 t 0.20 7.13 t 0.07

1.03 t 0.006 1.11 t 0.02 1.35 t 0.015

(kHlkT)wb

kHlkT

"Data taken from Cha et al. (1989). bThis semiclassical KIE was calculated from the experimental k D k T using equation (57) and assuming no tunnelling.

Table 41 Isotopic ratios of the Arrhenius parameters for the secondary p-tritium KIEs in the E2 reactions of p-YC6H4CHTCH2X! EaT - E a H

Y/X

Baselsolvent

(kcal mol-')

H/NMe3 WBr ClIOTs

EtO-EtOH Bu'O-IBu'OH Bu'O-/Bu'OH

0.326 t 0.022 0.200 t 0.028 0.340 t 0.021

A H ~A T

0.705 t 0.024 0.927 t 0.040 0.704 t 0.023

"Data taken from Amin et al. (1990).

Both the primary and secondary hydrogen-tritium KIEs found experimentally for this reaction (Table 40) are significantly larger than the semiclassical hydrogen-tritium KIEs calculated from the experimental deuterium-tritium KIEs using equation (57). These results are important because they confirm the results, based on the theoretical calculations, that the failure of equation (57) to give the kHlkTfrom the experimental kD/kT is positive evidence for tunnelling in a hydrogen transfer reaction.

TEMPERATURE DEPENDENCE OF SECONDARY TRITIUM KIEs

Saunders and co-workers (Amin et al., 1990) determined the secondary tritium KIEs for the E2 reactions of [19] with sodium ethoxide in ethanol, [20] with potassium t-butoxide in t-butyl alcohol and [21] with potassium t-butoxide in t-butyl alcohol over a temperature range of 40°C. The Arrhenius parameters were found for each isotopic reactant and the AHIAT ratios were calculated (Table 41). The AHIAT ratios for the reactions of [19] and [21] are both less than unity, confirming that, in agreement with the model calculations, tunnelling is important in these reactions. The AHIAT ratio for the reaction of

SECONDARY D-KINETIC ISOTOPE EFFECTS

229

[20] is close to unity and indicates that tunnelling does not occur even though the secondary tritium KIE = 1.27 was larger than the maximum predicted KIE (the EIE estimated from Hartshorn and Shiner's fractionation factors) of 1.17. This suggests that observing a secondary tritium KIE larger than the EIE is not sufficient grounds for concluding that tunnelling is important in a reaction.

STRUCTURAL EFFECTS ON SECONDARY KIEs IN ELIMINATION REACTIONS

Very little is actually known about the characteristics of a chemical reaction that promote tunnelling. It has been assumed that steric hindrance between the reacting partners in the transition state increases the amount of tunnelling by narrowing the potential energy barrier for the reaction. This argument was invoked to explain some large primary deuterium KIEs found for the proton transfer reactions between nitroalkanes and sterically hindered lutidines (Lewis and Funderburk, 1967). However, steric hindrance does not always lead to increased tunnelling. In a recent paper, Lin and Saunders (1994) reported experiments aimed at identifying the structural factors that promote tunnelling in E2 elimination reactions. In that investigation, the alkyl group of the doubly labelled 2-arylethyl substrates, [22]-[24], were varied while the leaving group was kept constant.

L is the hydrogen or deuterium atom that is not transferred in the elimination reaction and T is tritium that is present in tracer quantities. These substrates were chosen so that the reactions would have transition states ranging from very ElcB-like for [22], to central or intermediate for [23], to El-like for [24]. For practical reasons, the baseholvent system could not be kept constant as was originally intended. EtO-/EtOH was used in the reaction with substrates [22] and [23] whereas Bu'O-/Bu'OH was used with substrate [24]. Although the secondary tritium KIE (when L = H) for the reaction of [22] was

0. MATSSON AND K. C. WESTAWAY

230

Table 42 The secondary hydrogen-tritium and deuterium-tritium KIEs in the E2 reactions of several arylethyltrimethylammonium bromides at 50°C." Substrate

Baselsolvent

(k: lka>sc"

kElk&

P I

EtO-EtOH EtO-/EtOH Bu'O-/Bu'OH

1.032 2 0.003 1.031 2 0.003 1.029 2 0.003

1.266 2 0.013 1.238 ? 0.004 1.224 2 0.005

~ 3 1 ~ 4 1

"Data taken from Lin and Saunders (1994). bThesemiclassical KIE was calculated from the experimental secondary kglk; using equation (57) and assuming no tunnelling.

Table 43 Isotopic ratios of the Arrhenius parameters for the secondary tritium KIEs in the E2 reactions of several arylethyltrimethylamonium bromides." Substrate

Baselsolvent

P I

EtO-EtOH EtO-EtOH Bu'O-/Bu'OH

~ 3 1 P41

(&T

- EaH)

0.478 -+ 0.028 0.264 2 0.021 0.203 2 0.015

A

H

4

0.602 2 0.026 0.821 2 0.027 0.898 2 0.020

"Data taken from Lin and Saunders (1994).

significantly larger than those for the reactions of [23] and [24], all the secondary tritium KIEs shown in Table 42 are much larger than the maximum KIE (the EIE of 1.17 estimated from Hartshorn and Shiner's fractionation factors). This clearly suggests that tunnelIing occurs in all three reactions. The isotopic ratios of the Arrhenius pre-exponential parameters for these reactions (Table 43) are significantly different. The ratio is smallest (0.602) for the reaction of [22] then increases significantly to 0.821 for the reaction of [23] and to 0.898 for the reaction of [24]. However, all of the isotopic ratios are less than unity. Therefore, this second criterion for tunnelling indicates there is a sizable tunnelling contribution in all three reactions. Finally, tunnelling does not affect the magnitude of the experimental k$kz significantly. Therefore, converting the experimental (in the absence of tunnelling) secondary hydrogen-tritium KIE, kE/kT,, into the secondary hydrogen-tritium KIE, (kElk;),, using equation (61) gives the semiclassical (sc) KIE in the absence of tunnelling.

If tunnelling is important in the reaction, the experimental k z l k z will be larger than the (kElk;),, calculated using equation (61). The (kElkT,),, calculated for all three reactions by substituting the experimental kE/kgs into equation (61) are all significantly smaller than the

SECONDARY D-KINETIC ISOTOPE EFFECTS

231

experimental kElkT, values (column 4, Table 42). This indicates that tunnelling is important in all three reactions. Although all three criteria suggest that there is significant tunnelling in all three reactions, all three criteria suggest that tunnelling is most important in the reaction of [22]. Therefore, it has been concluded that the greater the carbanion character in the transition state, the greater the probability that tunnelling will occur. Finally, it is important to realize that the application of several criteria is advisable if reliable estimates of tunnelling are to be obtained. Just one criterion, e.g. the magnitude of the secondary KIE, may be misleading since a large KIE may be the result of a small tunnelling contribution and a large zero-point energy contribution.

KINETIC COMPLEXITY AS AN ALTERNATIVE TO TUNNELLING

Several workers have pointed out that observations that are taken as evidence of tunnelling may be due to other factors. For example, this problem can arise (i) when there is a kinetically complex situation where several steps in a reaction are partially rate limiting (see, e.g. Klinman, 1991) or (ii) where there is a competition between two or more reactions for a common intermediate (Thibblin, 1988;Thibblin and Ahlberg, 1989). It is important to note, however, that the results discussed by Saunders and co-workers are not impaired by any such kinetic complexity and are only due to a substantial tunnelling contribution to the KIE.

5

Remote secondary deuterium KlEs

Recently, Brown and co-workers (Nagorski et al., 1994; Slebocka-Tilk et al., 1995) have found large remote secondary deuterium KIEs in their extensive investigation of the electrophilic addition of bromine to 7-norbornylidene-7’norbornane under a variety of conditions.

A large inverse secondary deuterium KIE of 0.64 was observed in acetic acid at 25°C when the perdeutero (&,) compound was the deuterated substrate. This large inverse deuterium KIE was attributed to the KIE for the rate-determining formation of the bromonium ion (62). Although a portion of this KIE is undoubtedly due to the inductive effect (deuterium is more electron-donating than hydrogen and the deuterated bromonium ion would

232

0. MATSSON AND K. C. WESTAWAY

Fig. 18 The bromonium ion formed when bromine reacts with 7-norbornylidene-7’norbornane. Reproduced, with permission, from Nagorski et al. (1994).

be more stable than the undeuterated bromonium ion), Brown and coworkers concluded that the inductive KIE would be small, i.e. less inverse than 0.71 (between 1% and 2% per D) for all 20 deuteri~ms.~ Since the observed secondary deuterium KIE is much more inverse than 0.71 and since there are no P-hydrogens that could stabilize the bromonium ion by hyperconjugation (the hydrogens p to the positive centre are orthogonal to the orbitals of the bromonium ion), the greatest contributor to the observed KIE is a secondary steric KIE (a C-D bond being shorter than a C-H bond) associated with the four remote endo-hydrogens (Fig. 18). This explanation seems reasonable because MMX calculations of the bromonium ion indicate that the C-2, C-2’, C-3, and C-3’ endo-hydrogens are separated by only 2.11 8,or by substantially less than their van der Waals radii of 2 X 1.2 8, or 2.4 8,. If one assumes that all of the inverse KIE is due to these four remote hydrogens, the KIE per deuterium would be (0.64)1’4,or 0.89. It is worth noting that remote secondary steric deuterium KIEs have been suggested by a theoretical study of the addition reaction between cyclohexene and bromine (Brown, 1996, personal communication). The results suggested that approximately 67% of the observed KIE found when the substrate was cyclohexene-dlo arose from changes in the C-H bonds to the two vinyl hydrogens, bonds that are not present in the 7-norbornylidene-7‘-norbornanesystem. The remaining 33% of the observed isotope effect was attributed to the remote hydrogens on carbons 4 and 5 of the cyclohexane ring. The addition of bromide ion (Table 44) greatly retards the addition reaction in both methanol and acetic acid by making the kl step more reversible

4111e present authors (O.M. and K.W.) believe that the inductive KIE would be very much less than 0.71 because only the deuteriums close to the reaction centre would have a significant inductive contribution to the KIE.In fact, a secondary inductive KIE might be as small as a few per cent, i.e., kHlkD 0.96. i=

SECONDARY D-KINETIC ISOTOPE EFFECTS

233

Table 44 The rate constants and secondary deuterium KIEs for the addition of bromine to 7-norbornylidene-7’-norbornanein acetic acid and in methanol at 25°C.” [LiBrIi mol dm-3

0.00 0.01 0.02 0.03 0.04

k(HOAc)l

kHJkD

dm3 mol-’ s-’

(HOAc)

37.6 2.23 0.90 0.32 0.30

0.64 0.61 0.57 0.55 0.55

k(MeOH 1 dm3 mol-’s-’

k

H

k

(MeOH)

-

-

17.7 6.24 3.13 1.85

0.53 0.53 0.52 0.59

“Data taken from Slebocka-Tilk er al. (1995).

(Scheme 3); a large inverse secondary deuterium KIE is observed both in the presence, and in the absence, of added bromide ion. However, in acetic acid, the KIE becomes even more inverse as the concentration of bromide ion increases, i.e. it decreases from kHlkD= 0.64 to a minimum of 0.55 as the bromide ion concentration increases from zero to 0.04 mol dmP3(Table 44). In methanol, on the other hand, the inverse secondary deuterium KIE of 0.56 _t 0.04 is effectively independent of bromide ion concentration.

Scheme 3

In the mechanism preferred by the authors, the observed KIE is the product of the EIE for the reversible formation of the bromonium ion and the KIE for the rate-determining formation of the P-bromocarbocation (Scheme 3). Because the steric crowding of the C-2, C-2’,C-3 and C-3’ endo-hydrogens in the bromonium ion would be relieved in going to the P-bromocarbocation intermediate, one would expect the secondary deuterium KIE for the k2 step of the reaction to be normal, i.e. >1.00. If this is the case, the EIE for the formation of the bromonium ion must be significantly more inverse than the KIE for the k , step of the reaction, i.e. the KIE for the formation of the

234

0. MATSSON AND K. C. WESTAWAY

bromonium ion. This suggestion seems reasonable because the steric crowding in the bromonium ion would undoubtedly be greater than in the transition state for its formation. The more inverse KIE = 0.56, therefore, is effectively the EIE for the formation of the bromonium ion. 6 New methods for the accurate determination of secondary deuterium KIEs

Secondary isotope effects are small. In fact, most of the secondary deuterium KIEs that have been reported are less than 20% and many of them are only a few per cent. In spite of the small size, the same techniques that are used for other kinetic measurements are usually satisfactory for measuring these KIEs. Both competitive methods where both isotopic compounds are present in the same reaction mixture (Westaway and Ali, 1979) and absolute rate measurements, i.e. the separate determination of the rate constant for the single isotopic species (Fang and Westaway, 1991), are employed (Parkin, 1991). Most competitive methods (Melander and Saunders, 1980e) utilize isotope ratio measurements based on mass spectrometry (Shine et al., 1984) or radioactivity measurements by liquid scintillation (Ando et al., 1984;Axelsson et al., 1991). However, some special methods, which are particularly useful for the accurate determination of secondary KIEs, have been developed. These newer methods, which are based on polarimetry, nmr spectroscopy, chromatographic isotopic separation and liquid scintillation, respectively, are described in this section. The accurate measurement of small heavy-atom KIEs is discussed in a recent review by Paneth (1992).

METHODS BASED ON MEASURING OPTICAL ACTIVITY

The isotopic quasi-racemate method (IQRM)

This method is based on the polarimetric measurement of the optical activity induced by the KIE in a reaction mixture containing an isotopic quasiracemate, i.e. an approximately 50/50 mixture of the (+)-H and (-)-D substrate or vice versa, as one of the reactants. Variants of the method were independently reported by Bergson et al. (1977), Nadvi and Robinson (1978) and Tencer and Stein (1978). Later the method was successfully applied, particularly by Matsson and co-workers (Matsson, 1985;HussCnius et al., 1989; HussCnius and Matsson, 1990) to determine both primary and secondary KIEs in proton transfer reactions, and by Sinnott and co-workers (Bennet et al., 1985; Ashwell et al., 1992; Zhang et al., 1994) to determine both primary and secondary as well as heavy-atom KIEs for reactions of carbohydrate derivatives. The isotopic quasi-racemate or differential polarimetric method is a kinetic

235

SECONDARY D-KINETIC ISOTOPE EFFECTS

method which permits the simultaneous determination of the rate constant for both isotopic species and the rate constant ratio (the KIE) in one kinetic experiment. This has the advantage of eliminating any interexperimental errors. It is particularly useful for measuring very small KIEs since it is based on the direct measurement of the difference in an observable quantity for the two reacting isotopic species. The method is illustrated by reaction (63), where AH and AD are two isotopically substituted substrates with opposite optical rotation, mixed so that the initial optical rotation is close to zero.

Nonchiral products (-)-AD

When a reaction which transforms the reactants into nonchiral products is started, AH and AD, owing to the KIE, are consumed at different rates and an optical rotation is induced in the mixture. For a reaction that follows first-order or pseudo-first-order kinetics with the rate constants kH and kD,the time dependence of the optical rotation, a,is described by a two-exponential function (64). a = al exp( -kHt)

+ az exp( -kDt)

(64)

In the simplest case, where (+)-AH and (-)AD are isotopically pure, al = [aIH[AHloand a2 = [ c x ] ~ [ A D where ] ~ a is the specific rotation of the AH and AD isotopomers, respectively, and [AHIoand [ADIoare the concentrations of the substrates in g ml-’ at time t = 0. When the substrate is neither isotopically nor enantiomerically pure, corrections must be made in calculating al and u2 (Bergson et al., 1977). It is important to note that the pre-exponential factors, ul and a2, which contain the information about the starting conditions, can be determined with high accuracy. The extreme, a, (the maximum or minimum value of the optical rotation in the optical rotation versus time plot) and the corresponding reaction time, t,, are functions of the rate constant ratio (S = kH/kD) (65) and the difference between the rate constants (66), respectively. a, = a,[(a,/az)S]s’(l-s)- a*[(a&z)

S]1’(1 -s)

(65)

Hence, it is possible to calculate the KIE and the individual rate constants in one experiment. A typical optical rotation versus time plot is shown in Fig. 19.

236

0. MATSSON AND K. C. WESTAWAY

-4

-3 Optical rotation/ degrees

-

-

-2

7

-1

-

0 0

1

1

I

2

4

6

I

a

10

Timelh

Fig. 19 The optical rotation (degrees) versus time (h) for a DABCO-catalysed rearrangement of l-methyl-hitroindene in o-dichlorobenzene at 20°C.

An alternative to evaluating the KIE and the rate constants from the above equations is to apply nonlinear least-squares fitting to the complete kinetic set of a and t values. This latter procedure has the advantage that errors in the reaction model, e.g. an incorrect mechanism, or extraneous data points are more easily discovered. This method was applied by Bergson et al. (1977) and Matsson (1985) in the determination of both the primary deuterium and secondary a-deuterium KIEs in the l-methylindene rearrangement to 3methylindene (reaction (67)). For example, a secondary P-deuterium KIE of 1.103 ? 0.001 was determined very accurately in toluene at 20°C using this method (Bergson ef al., 1977).

In a variation of this method, Tencer and Stein (1978), mixed the isotopic quasi-racemate to near, but not exactly, zero rotation so that at a certain time, t,, the observed optical rotation of the reaction mixture was zero. The equations for this type of kinetic experiment enable one to calculate the difference between the individual isotopic rate constants from tz and the ratio of rate constants (the KIE) from t, and t, provided that the ratio of the initial rotations for the two isotopic substrates is known. Usually it is preferable to

SECONDARY D-KINETIC ISOTOPE EFFECTS

237

use the t, value to calculate the KIE since the extreme value in the optical rotation versus time plot is very well defined and can be measured very accurately. Isotopically engendered chirality

Another polarimetric method for the accurate determination of KIEs bears a strong resemblance to the isotopic quasi-racemate method, described above. In this method, Bach and co-workers (1991) utilized what they called isotopically engendered chirality to determine the primary deuterium KIE for an elimination reaction. In theory, the method can be used for any reaction where a substrate with a plane of symmetry yields, under normal conditions, a racemic mixture. For instance, if the plane of symmetry in the unlabelled

starting material [25] is removed by the stereospecific substitution of a deuterium atom to give [26] with two stereogenic centres, the elimination reaction yields a pair of isotopic quasi-enantiomers ("nominal enantiomers") (Scheme 4) and the KIE causes an enantiomeric excess (ee) of one of the isotopic enantiomers in the product.

Scheme 4

The KIE is calculated from equation (68), where eei and eef are the initial and final optical purities and aiand a, are the optical rotation of the starting material and the final rotation of the alkene, respectively. kH - eei - eef - [a]:' - [a]:5 k, eei + eef [a]:' + [a]?

It is worth noting that the KIE on the optical rotation of the product was neglected in this work. Bach et at. (1991) used this method to measure the

238

0. MATSSON AND K. C. WESTAWAY

primary deuterium KIE (kHlkD= 1.39 2 0.02) for the NaWDMSO basepromoted suprafacial elimination reaction forming the alkene (E)-cyclooctene [27] at 25 "C (reaction (69)).

-DX Base

I

kD

[27]( - )-R-l

[27] ( + )-S-1-3dl

It is worth noting that the KIEs determined by this polarimetric method are in excellent agreement with those found by the classical mass spectrometric method. Also, as the authors pointed out, the method can be extended to the measurement of secondary deuterium and even heavy-atom KIEs.

DETERMINING MULTIPLE KIEs USING NATURAL ABUNDANCE nmr SPECTROSCOPY

Deuterium nmr spectroscopy has been utilized for the last decade to determine large (primary deuterium) KIEs in reactions with isotopes present at the natural abundance level (Pascal el al., 1984,1986; Zhang, 1988). A great advantage of this approach is that labelled materials do not have to be synthesized. Neither is there any need for selective degradation procedures, which are often necessary to produce the molecules of low mass, e.g. COz, acceptable for isotope ratio mass spectrometry. Moreover, the KIEs for several positions can be determined from one sample. However, until quite recently the relatively low precision of the nmr integrations that are used for the quantitative assessment of the amount of deuterium at specific molecular sites has limited the applicability of this technique for determining small (secondary deuterium) KIEs. Singleton and Thomas (1995) suggested and demonstrated that naturalabundance nmr spectroscopy could also be used to measure small KIEs if the

SECONDARY D-KINETIC ISOTOPE EFFECTS

239

isotopic enrichment becomes significant during the reaction. The way to obtain a sample with sufficient isotopic enrichment when the KIE is small and natural abundance samples are used is to isolate the unreacted starting material at very high degrees of conversion. For instance, a 25% enrichment of the heavy isotope is observed in the unreacted substrate recovered after 99% reaction if the KIE is 1.05. Singleton and Thomas tested this method on the Diels-Alder reaction of isoprene with maleic anhydride (reaction (70)).

In this experiment unreacted isoprene was recovered from the reaction mixture at 98.9% of completion and the amounts of deuterium and I3Cat the various positions were compared to those in the starting material using nmr. The KIEs for the various atoms were then calculated from these data using equation (71), where f denotes the fraction of reaction, Rfis the ratio of the isotopes in the unreacted starting material and R, is the corresponding ratio in the original starting material.

The methyl group was used as the internal standard, i.e. they assumed that no change of its isotopic composition takes place during the reaction. The 1.00 (assumed)

' 2

l.OOO(3)

0.968(5)

Fig. 20 The deuterium and carbon-13 KIEs calculated for the Diels-Alder reaction between isoprene and maleic anhydride using the isotopic enrichment in the unreacted isoprene recovered from a reaction taken to 98.9% of completion. The numbers in parentheses represent the error in the KIE. Reproduced, with permission, from Singleton and Thomas (1995).

240

0. MATSSON AND K. C. WESTAWAY

resulting KIEs are displayed in Fig. 20. The excellent precision of these KIE measurements, even when the KIEs are very small, illustrates the impressive ability of this method, especially when one considers that all of these KIEs were determined from one experiment. Although the technique is impressive and gives excellent results even for small KIEs, there is a major limitation. One problem with this method is that the reaction must be run on a large scale and that one must be able to separate quantitatively a small amount of reactant from a large quantity of product. This is necessary so that one can determine f accurately and have enough sample to obtain a good nmr spectrum. The Diels-Alder reaction, for example, was run using 13 moles of the substrate! Another restraint is that the reaction must be clean, i.e. give 100% of the product that is being analysed. This is required because a side reaction would have an isotope effect and would change the isotopic ratio in the compound used in the nmr analysis. THE CHROMATOGRAPHIC ISOTOPIC SEPARATION METHOD

Holm and co-workers have been able to determine very small secondary &-deuterium (Holm, 1994a,b; 1996; Holm and Crossland, 1996), secondary p-deuterium (Holm and 0gaard Madsen, 1992), as well as primary I3C(Holm, 1993, 1994a,b; Holm and Crossland, 1996) KIEs by separating the isotopic compounds by capillary column gas chromatography. Some small secondary p-deuterium KIEs that have been measured for the reaction between labelled Grignard reagents and ketones by this technique are shown in Table 45. The baseline separation of deuterated and undeuterated compounds that is required for calculating the KIEs is possible on 50-100 m capillary columns provided that the number of isotopically substituted atoms in the molecule is 3 or more. However, the actual length of the capillary column required for the baseline separation of the isotopomers is determined by the deuterium Table 45 Secondary P-deuterium KIEs in the free radical reactions between Grignard reagents and ketones at 25°C." SubstratelGrignard reagent 2-Octanone Benzophenone 4,CDimethyl-1-phenyl1-penten-3-one 1,3-Diphenyl-2-propene1-one

(k€I/kdp CD3CH2MgBr

(CD,),CHMgBr

(CD,),CMgBr

0.985 t 0.003

1.014 ? 0.003 1.034 t 0.003

0.940 t 0.005 1.050 t 0.003 0.987 -+ 0.003

1.016 t 0.003 0.974 t 0.007

1.050 t 0.010

1.001 2 0.008

0.997 2 0.007

'Data taken from Holm and Bgaard Madsen (1992).

-

SECONDARY D-KINETIC ISOTOPE EFFECTS

241

Retention time

Fig. 21 Separation of the products from the reaction of CH,CD,MgBr and CD,CD,MgBr with benzophenone on a 100m X 0.2 mm X 0.33 pm HP-5 capillary column at 160 "C.

content in relation to the molecular mass. A sample separation is shown in Fig. 21. The secondary deuterium KIEs can be calculated from equation (71) using the ratio of the deuteratedhndeuterated Grignard reagents at the beginning of the reaction and the product ratios obtained at various extents of reaction from the gas chromatographic analysis.

THE DOUBLE LABELLING LIQUID SCINTILLATION TECHNIQUE

Remote double labelling techniques have been used successfully in the determination of enzyme KIEs (Kiick, 1991). A variant of this technique was applied to a nonenzymatic reaction by Matsson and co-workers (Axelsson el al., 1990). They determined the primary carbon and secondary deuterium KIEs for the SN2reaction between methyl iodide and hydroxide ion in 50% dioxane-water at 25°C. The a-carbon KIE was determined by the 'lC method (Axelsson et al. 1987,1991). In this method, a mixture of substrate molecules labelled with "C (tllZ= 20.4 min) and I4C is used. The reactants and products

242

0.MATSSON AND K. C. WESTAWAY

in samples removed from the reaction mixture at several times throughout the reaction were separated by liquid chromatography and collected in a scintillation cocktail. Then, the total radioactivity in each reactant and product fraction was measured. After decay of the short-lived "C radionuclide was complete, the amount of the long-lived 14Cradionuclide in the same samples was determined. Finally, the carbon KIE was calculated from the radioactivity data. One advantage of this method is that the heavy-atom (carbon) KIE measured in these experiments is large owing to the large mass difference in the carbon isotopes. The KIEs were then determined by the same method but using a substrate mixture where either the 'lC- or the 14C-labelled methyl iodide was doubly labelled with deuterium, i.e. the substrate was either "CH3 I and 14CD3for 11CD31and 14CH31.Assuming that the carbon and deuterium KIEs are multiplicative, i.e. that the rule of the geometric mean holds, the secondary deuterium KIEs could be calculated from these isotope effects and the previously determined carbon KIEs. The secondary a-deuterium KIEs obtained in this way were 0.881 2 0.012 and 0.896 2 0.011 when the doublelabelled substrates were "CD3 I and 14CD3I, respectively.

7 Conclusion

This chapter has attempted to demonstrate how secondary deuterium and tritium KIEs can be used to elucidate the mechanisms of reactions and determine the structure of their transition states. In particular, the advantages of using both theoretical calculations and experimental data to solve these problems has been emphasized. Unfortunately, several important topics where the combination of theoretical calculations and experimental work has been very useful in extending our understanding of KIEs could not be discussed. In particular, the extensive studies on the Diels-Alder and the Cope rearrangement by Houk and co-workers (Beno et al., 1996; Houk et al., 1992; Storer et al., 1994) are noteworthy.

Acknowledgements The authors gratefully acknowledge the financial support provided by the Swedish Natural Science Research Council (to O.M.) and the Natural Science and Engineering Research Council of Canada (to K.W.). Olle Matsson is indebted to David Tanner for allowing him to complete a portion of this chapter at the Denmark Technical University. Finally, the authors dedicate this chapter to Goran Bergson, Art Bourns and Peter Smith, who interested us in physical organic chemistry and taught us how to use KIEs to solve interesting and important problems.

SECONDARY D-KINETIC ISOTOPE EFFECTS

243

References Abboud, J.-L., Notario, R. Bertran, J. and Sola, M. (1993). In Progress in Physical Organic Chemistry (ed. R. W. Taft) vol. 19. Wiley, New York, p. 1 Amin, M., Price, R. C. and Saunders, W. H., Jr. (1990). J. Am. Chem. SOC.112,4467 Ando, T., Tanabe, H. and Yamataka, H. (1984). J. Am. Chem. SOC. 106,2084 Ando, T., Yamataka, H. and Wada, E. (1985). Isr. J. Chem. SOC.26,354 Ando, T., Kimura, T. and Yamataka, H. (1987). In Nucleophilicity (ed. J. M. Harris and S. P. McManus). American Chemical Society, Washington, DC, p. 108 Apeloig, Y. (1981). J. Chem. SOC., Chem. Commun. 396 Ashan, M., Robertson, R. E., Blandamer, M. J. and Scott, J. M. W. (1980). Can. J. Chem. 58,2142 Ashwell, M., Guo, X.and Sinnott, M. L. (1992). J. Am. Chem. SOC. 114, 10158 Axelsson, B. S., LBngstrom, B. and Matsson, 0. (1987). J. Am. Chem. SOC. 109, 7233 Axelsson, B. S., Matsson, 0. and Ungstrom, B. (1990). J. Am. Chem. SOC. 112,6661 Axelsson, B. S., Matsson, 0. and LBngstrom, B. (1991). J. Phys. Org. Chem. 4, 77 Bach, R. D., Knight, J. W. and Braden, M. L. (1991). J. Am. Chem. SOC. 1l3,4712 Barnes, J. A. and Williams, I. H. (1993). J. Chem. SOC., Chem. Commun. 1286 Bartell, L. S. (1960). Tetrahedron Lett. (6), 13 Bartell, L. S. (1961). J. Am. Chem. SOC.83, 3567 Bell, R. P. (1959). The Proton in Chemistry. Cornell University Press, Ithaca, NY Bell, R. P. (1980a). The Tunnel Effect in Chemistry, Chapman and Hall, London Bell, R. I? (1980b). The Tunnel Effect in Chemistry. Chapman and Hall, London, PP. 60-63 Bennet, A. J., Sinnott, M. L. and Sulochana Wijesundera, W. S. (1985). J. Chem. SOC., Perkin Trans. 2, 1233 Beno, B. R., Houk, K. N. and Singleton, D. A. (1996). J. Am. Chem. SOC.118,9984 Berg, H., Chanon, M., Gallo, R. and Rajemann, M. (1995). . I . Org. Chem. 60,1975. Bergson, G., Matsson, 0. and Sjoberg, S. (1977). Chem. Scr. 11,25 Bigeleisen, J. and Wolfsberg, M. (1958). Adv. Chem. Phys. 1, 15 Boyd, R. J., Kim, C.-K., Shi, Z., Weinberg, N. and Wolfe, S. (1993). J. Am. Chem. SOC. 115, 10147 Brown, H. C. and McDonald, G. J. (1966). J. Am. Chem. SOC.88,2514 Brown, H. C., Azzaro, M. E., Koelling, J. G. and McDonald, G. J. (1966).J. Am. Chem. SOC. 88,2520 Brown, K. C., Romano, F. J. and Saunders, W. H. Jr. (1981). J. Org. Chem., 46,4242. Buddenbaum, W. E. and Shiner, V. J., Jr. (1977a). In Isotope Effects on Enzymecatalyzed Reactions (ed. W. W. Cleland, M. H. O’Leary and D. B. Northrup). University Park Press, Baltimore, pp. 1-33. Buddenbaum, W. E. and Shiner, V. J., Jr. (1977b). In Isotope Effects on Enzymecatalyzed Reactions (ed. W. W. Cleland, M. H. O’Leary and D. B. Northrup). University Park Press, Baltimore, p. 18 Buist, G. J. and Bender, M. L. (1958). J. Am. Chem. SOC.80,4308 Cha, Y., Murray, C. J. and Klinman, J. P. (1989). Science 243, 1325 Cleland, W. W. (1987) Secondary isotope effects on enzymatic reactions. In Isotopes in Organic Chemistry, vol. 7 (ed. E. Buncel and C. C. Lee). Elsevier, Amsterdam, pp. 61-114. Cook, F! F. and Cleland, W. W. (1981a). Biochemistry 20, 1797 Cook, P. F. and Cleland, W. W. (1981b). Biochemistry 20, 1805 Cook, P. F., Blanchard, J. S. and Cleland, W. W. (1980). Biochemistry 19,4853 Cook, P. F., Oppenheimer, N. J. and Cleland, W. W. (1981) Biochemistry 20,1817

244

0.MATSSON AND K. C. WESTAWAY

Cramer, C. J. and 'Ikuhlar, D. G. (1991). J. Am. Chem. SOC.113,8305 Craze, G.-A., Kirby, A. J. and Osborne, R. (1978). J. Chem. SOC.Perkin Trans. 2,357 DeFrees, D. J., Bartmess, J. E., Kim, J. K, McIver, R. T., Jr. and Hehre, W. J. (1977). J. Am. Chem. SOC.99,6451 DeFrees, D. J., Taagepera, M., Levi, B. A., Pollack, S. K., Summerhays, K. D., Taft, R. W., Wolfsberg, M. and Hehre, W. J. (1979a). J. Am. Chem. SOC.101,5532 DeFrees, D. J., Hehre, W. J. and Sunko, D. E. (1979b). J. Am. Chem. SOC. 101, 2323 Dewar, M. J. and Dougherty, R. C. (1975). The PMO Theory of Organic Chemistry. Plenum, New York, pp. 219,220 Dewar, M. J. S. and Thiel, W. (1977). J. Am. Chem. SOC.99,4907 Dewar, M. J. S., Zoebisch, E. G., Healy, E. F., and Stewart, J. J. F? (1985).J. Am. Chem. SOC.107,3902. Evans, J. C. and Lo, G. Y.4. (1966). J. Am. Chem. SOC.88,2118 Fang, Y-R. and Westaway, K. C. (1991). Can J. Chem. 69,1017 Friedman, D. S., Franc], M. M. and Allen, L. L. (1985). Tetrahedron 41,499 Frisone, G. J. and Thornton, E. R. (1964). J. Am. Chem. SOC.86,1900 Fry, A. (1970) In Isotope Effects in Organic Reactions (ed. C. J. Collins, Jr. and N. S. Bowman), A.C.S. Monograph 167. Van Nostrand Reinhold, New York, p. 380 Gawlita, E., Szylhabel-Godala, A. and Paneth, F! (1996). J. Phys. Org. Chem. 9, 41 Glad, S. S. and Jensen, F. (1997). J. Am. Chem. SOC.119, 227 Gold, V., ed. (1983). Pure A&. Chem. 55, 1281 Gray, C. H., Coward, J. K., Schowen, B. K. and Schowen, R. L. (1979). J. Am. Chem. SOC.101,4351 Gronert, S., DePuy, C. H. and Bierbaum, V. M. (1991). J. Am. Chem. SOC.113,4009 Hakke, L., Queen, A. and Robertson, R. E. (1965). J. Am. Chem. SOC.87,161 Halevi, E. A. (1963). Progz Phys. Org. Chem. 1, 109 Harris, J. M., Shafer, S. G., Moffatt, J. R and Becker, A. R. (1979). J. Am. Chem. SOC. 101,3295 Harris, J. M., Paley, M. S. and Prasthofer, T. W. (1981). J. Am. Chem. SOC. 103, 5915 Hartshorn, S. R. and Shiner, V. J., Jr. (1972). J. Am. Chem. SOC.94, 9002 Hill, J. W. and Fry, A. (1962). J. Am. Chem. SOC.84,2763 Hoffmann, R., Radom, L., Pople, J. A., Schleyer, F! vR., Hehre, W. J. and Salem, L. (1972). J. Am. Chem. SOC.94,6221 Hogg, J. L. (1978). In Transition States of Biochemical Processes (ed. R. D. Gandour and R. L. Schowen). Plenum Press, New York, pp. 201-224. Holm, T. (1993). J. Am. Chem. SOC.115,916 Holm, T. (1994a). Acra Chem. Scand. 48, 362 Holm, T. (1994b). J. Am. Chem. SOC.116,8803 Holm, T. (1996). J. Organomet. Chem. 506,37 Holm, T. and Crossland, I. (1996). Acta Chem. Scand. 50,90 Holm, T. and 0gaard Madsen, J. (1992). Acta Chem. Scand. 46,985 Holtz, D. (1971). Prog. Phys. Org. Chem. 8, 1 Houk, K. N., Gustafson, S. M. and Black, K. A. (1992). J. Am. Chem. SOC. 114, 8565 Hu, W. F? and Tmhlar. D. G. (1995). J. Am. Chem SOC.117,10726 Humski, H., Sendijarevic, V.' and'Shiner, V. J., Jr. (1974). J. Am. Chem. SOC. 96, 6187 Huskey, W. P. and Schowen, R. L. (1983). J. Am. Chem. SOC. 105,5704 HussCnius, A. and Matsson, 0. (1990). Acra Chem. Scand. 44, 845 HussBnius, A., Matsson, 0. and Bergson, G. (1989). J. Chem. SOC., Perkin Trans. 2, 851

SECONDARY D-KINETIC ISOTOPE EFFECTS

245

IUPAC Commission on Physical Organic Chemistry (1988). Pure Appl. Chem. 60, 1115 Jiang, W. (1996). MSc. Dissertation, Laurentian University, Sudbury, Ont., Canada Jencks, W. I? (1972). Chem. Rev. 72,705 Kaldor, S. B. and Saunders, W. H. Jr. (1979). J. Am. Chem. SOC. 101,7594. Kaplan, E. and Thornton, E. R. (1967). J. Am. Chem. SOC.89,6644 Karelson, M. M., Tamm, T., Katritzky, A. R., Cato, S. J. and Zerner, M. C. (1989). Tetrahedron Comput. Methods 2, 295 Kiick, D. M. (1991). In Enzyme Mechanism from Isotope Effects (ed. I? F. Cook). CRC Press, Boca Raton, Fla., pp. 313-330. Kirsch, J. F. (1977). In Isotope Effects on Enzyme-catalyzed Reactions (ed. W. W. Cleland, M. H. O’Leary and D. B. Northrop). University Park Press, Baltimore, pp. 100- 122. Klamt, A. and Schuurmann, G. (1993). J. Chem. SOC., Perkin Trans. 2,799 Klinman, J. (1991). In Enzyme Mechanism from Isotope Effects (ed. F’. F. Cook). CRC Press, Boca Raton, Fla., pp. 127-151. Kluger, R. and Brandl, M. (1986a). J. Org. Chem. 51,3964 Kluger, R. and Brandl, M. (1986b). J. Am. Chem. SOC.108,7828 Knier, B. L. and Jencks, W. I? (1980). J. Am. Chem. SOC. 102,6789 Koenig, T. and Wolf, R. (1967). J. Am. Chem. SOC.89,2948 Koshy, K. M. and Robertson, R. E. (1974). J. Am. Chem. SOC.96, 914 Kresge, A. J., Drake D. A. and Chiang, Y. (1974). Can. J. Chem. 52,1889 Kurz, J. L. and El-Nasr, M. M. S. (1982). J. Am. Chem. SOC.104,5823 Kurz, L. C. and Frieden, C. (1980). J. Am. Chem. SOC. 102,4198 Kurz, J. L., Daniels, M. W., Cook, K. S. and Nasr, M. M. (1986a). J. Phys. Chem. 90, 5357 Kurz, J. L, Pantano, J. E., Wright, D. R. and Nasr, M. M. (1986b). J. Phys. Chem. 90, 5360 Lai, 2.G. and Westaway, K. C. (1989). Can. J. Chem. 67,21 LAngstrom, B., Antoni, G., Gullberg, P., Halldin, C., Malmborg, P., Nagren, K., Rimland, A. and Svard, H. (1987). 1 Nucl. Med. 28,1037 Lee, I. (1995). Chem. SOC. Rev. 223 Lee, I., Koh, H. J., Lee, B.-S., Sohn, D. S. and Lee, B. C. (1991). J. Chem. Soc., Perkin Trans. 2 1741 Leffek, K. T. and MacLean, J. W. (1965). Can. J. Chem. 43,40 Leffek, K. T. and Matheson, A. F. (1971). Can. J. Chem. 49,439 Leffek, K. T. and Matheson, A. F. (1972a). Can. J. Chem. 50, 982 Leffek, K. T. and Matheson, A. F. (1972b). Can. J. Chem. 50,986 Le Noble, W. J. and Miller, A. R. (1979). J. Org. Chem. 44,889 Lewis, E. S. and Funderburk, L. H. (1967). J. Am. Chem. SOC.89,2322 Lin, S. and Saunders, W. H., Jr. (1994). J. Am. Chem. SOC.116,6107 Lowry, T. H. and Richardson, K. S. (1987) Mechanism and Theory in Organic Chemistry. Harper and Row, New York, pp. 588-600. Maccoll, A. (1974). Annu. Rep. A: The Chemical Society, (London) 71,77 Matsson, 0.(1985). J. Chem. SOC.,Perkin Trans. 2221 Matsson, O., Persson, J., Axelsson, B. S.and Ldngstrom, B. (1993). J. Am. Chem. SOC. 115,5288 Matsson, O., Persson, J., Axelsson, B. S., Lfingstrom, B., Fang, Y.-R. and Westaway, K. C. (1996). J. Am. Chem. SOC.118,6350 McLennan, D. J. (1979). Aust. J. Chem. 32,1883 McLennan, D. J. (1987). Model calculations of secondary isotope effects. In Isotopes in

246

0.MATSSON AND K. C. WESTAWAY

Organic Chemistry, vol. 7 (ed. E. Buncel and C. C. Lee). Elsevier, Amsterdam, pp. 393480. Melander, L. (1960a). In Isotope Effects on Reaction Rates. Ronald Press, New York, pp. 7-40. Melander, L. (1960b). In Isotope Effects on Reaction Rates. Ronald Press, New York, pp. 24-32. Melander, L. and Saunders, W. H., Jr. (1980a). Reaction Rates of Isotopic Molecules. Wiley-Interscience,New York Melander, L. and Saunders, W. H., Jr. (1980b). In Reaction Rates of Isotopic Molecules. Wiley-Interscience,New York, pp. 4-28 Melander, L. and Saunders, W. H., Jr. (1980~).In Reaction Rates of Isotopic Molecules Wiley-Interscience,New York, pp. 197-199 Melander, L. and Saunders, W. H., Jr. (1980d). In Reaction Rates of Isotopic Molecules. Wiley-Interscience,New York, p. 209 Melander, L. and Saunders, W. H., Jr. (1980e). In Reaction Rates of Isotopic Molecules. Wiley-Interscience,New York, pp. 119-125 Meot-Ner (Mautner), M. (1987). J. Am. Chem. SOC.109,7947 More O’Ferrall, R. A. (1970). J. Chem. SOC.B 274 More O’Ferrall, R. A. and Slae, S. (1970). J. Chem. SOC.B 260 Mulliken, R. S. (1933). J. Chem. Phys. 1,492 Mulliken, R. S. (1935). J. Chem. Phys. 3,520 Mulliken, R. S. (1939). J. Chem. Phys. 7,339 Mulliken, R. S., Rieke, C. A. and Brown, W. G. (1941). J. Am. Chem. SOC.63,41. Murr, B. L. and Donnelly, M. F. (1970a). J. Am. Chem. SOC.92,6686 Murr, B. L. and Donnelly, M. E (1970b). J. Am. Chem. SOC.92,6688 Nadvi, N. S. and Robinson, M. J. T. (1978). Abstracts from the Fourth IUPAC Conference on Physical Organic Chemistry, York, UK, p. 141 Nagorski, R. W., Slebocka-Tik, H. and Brown, R. S. (1994). J. Am. Chem. SOC. 116, 419 Paneth, I? (1992). In Isotopes in Organic Chemistry, vol. 8 (ed. E. Buncel and W. H. Saunders, Jr.), Elsevier, Amsterdam, Ch. 2. Paneth, F! and O’Leary, M. H. (1991). J. Am. Chem. SOC.113,1691 Parkin, D. W. (1991). In Enzyme Mechanism from Isotope Effects (ed. F? F. Cook). CRC Press, Boca Raton, ma., Ch. 10, pp. 269-290. Pascal, R. A. and Mischke, S. (1991). J. Org. Chem. 56,6954 Pascal, Jr., R. A., Baum, M. W., Wagner, C. K.and Rodgers, L. R. (1984). J. Am. Chem. SOC. 106,5377 Pascal, Jr., R. A., Baum, M. W., Wagner, C. K., Rodgers, L. R. and Huang, D.-S. (1986). J. Am. Chem. SOC.108,6477 Persson, J., Berg, U. and Matsson, 0. (1995). J. Org. Chem. 60,5037 Persson, J., Axelsson, S. and Matsson, 0. (1996). J. Am. Chem. SOC.118, 20 Pham, T. V. (1993). MSc. Dissertation, Laurentian University, Sudbury, Ont., Canada Pham, T.V, and Westaway, K. C. (1996). Can. J. Chem. 74,2528 Poirier, R. A., Wang, Y. and Westaway, K. C. (1994). J. Am. Chem. SOC.116,2526 Pross, A. and Shaik, S. S. (1981). J. Am. Chem. SOC.103,3702 Roberts, J. D., Webb, R. L. and McElhill, E. A. (1950). J. Am. Chem. SOC.72,408 Saunders, W. H., Jr. (1975). Chem. Scripta. 8,27 Saunders, W. H., Jr. (1976). Acc. Chem. Res. 8,19 Saunders, W. H., Jr. (1984). J. Am. Chem. SOC.106,2223 Saunders, W. H., Jr. (1985). J. Am. Chem. SOC.107,164 Saunders, W. H., Jr., (1992). Croat. Chem. Acta 65, 505

SECONDARY D-KINETIC ISOTOPE EFFECTS

247

Saunders, W. H., Jr. (1997). J. Org. Chem. 62, 244 Saunders, W. H. Jr and Edison, D. H. (1960). J. Am. Chem. SOC. 82, 138. Schleyer, l? v R. and Kos, A. J. (1983). Tetrahedron 39, 1141 Seltzer, S. and Hamilton, E. (1966). J. Am. Chem. SOC.88,3775 Seyferth, D., Burlitch, J. M., Yamamoto, K., Washburne, S. S. and Attridge, C. J. (1970a). J. Org. Chem. 35, 1989 Seyferth, D., Mai, V. A. and Gordon, M. E. (1970b). J. Org. Chem. 35, 1993 Shine, H. J., Park, K. H., Brownawell, M. L. and San Filippo, J., Jr. (1984). J. Am. Chem. SOC. 106,7077 Shiner, V. J., Jr. (1970a) In Isotope Effects in Chemical Reactions (ed. C. J. Collins, Jr. and N. S. Bowman), A.C.S. Monograph 167. Van Nostrand Reinhold, New York, pp. 90-159. Shiner, V. J., Jr. (1970b). In Isotope Effects in Chemical Reactions (ed. C. J. Collins, Jr. and N. S. Bowman), A.C.S. Monograph 167. Van Nostrand Reinhold, New York, p. 137. Shiner, V. J., Jr. (1970~).In Isotope Effects in Chemical Reactions (ed. C. J. Collins, Jr. and N. S. Bowman), A.C.S. Monograph 167. Van Nostrand Reinhold, New York, p. 138. Shiner, V. J., Jr. and Humphrey, J. S., Jr. (1963). J. Am. Chem. SOC. 85,2416 Shiner, V. J., Jr. and Jewett, J. G. (1965). J. Am. Chem. SOC.87, 1382 Shiner, V. J., Jr., Dowd, W., Fisher, R. D., Hartshorn, S. R., Kessik, M. A., Milakofsky, L. and Rapp, M. W. (1969). J. Am. Chem. SOC.91,4838 Shiner, V. J., Jr., Rapp, M. W. and Pinnick, H. R., Jr. (1970). J. Am. Chem. SOC. 92, 232 Sims, L. B. and Lewis, D. E. (1985). In Isotopes in Organic Chemistry, vol. 6 (ed. E. Buncel and C. C. Lee). Elsevier, Amsterdam, pp. 161-257 Sims, L. B., Fry, A., Netherton, L. T., Wilson, J. C., Reppond, K. D. and Crook, S. W. (1972). J. Am. Chem. SOC. 94, 1364 Singleton, D. A. and Thomas, A. A. (1995). J. Am. Chem. SOC.117,9357 Slebocka-Tilk, H., Motallebi, S., Nagorski, R. W., "hmer, P., Brown, R. S. and McDonald, R. (1995). J. Am. Chem. SOC.117, 8769 Smith, l? J. and Westaway, K. C. (1982). In The Chemistry of the Functional Groups, Supplement F: The Chemistry of Amino,Nitroso and Nitro Compounds and Their Derivatives (ed. S. Patai). Wiley, London, p. 1277 Stewart, J. J. P. (1989). J. Comput. Chem. 10,221 Storer, J. W., Raimondi, L. and Houk, K. N. (1994). J. Am. Chem. SOC. 116,9675 Streitwieser, A., Jr. and Van Sickle, D. E. (1962). J. Am. Chem. SOC. 84, 254 Streitwieser, A. Jr., Jagow, R. H., Fahey, R. C. and Suzuki, S. (1958). J. Am. Chem. SOC. 80,2326 Streitwieser, A., Jr., Berke, C. M., Schriver, G. W., Grier, D. and Collins, J. B. (1981). Tetrahedron, Suppl. I . 37, 345 Stern, M. J. and Wolfsberg, M. (1966). J. Chem. Phys. 45, 4105 Subramanian, Rm. and Saunders, W. H., Jr. (1981). J. Phys. Chem. 85,1099 Subramanian, Rm. and Saunders, W. H., Jr. (1984). J. Am. Chem. SOC. 106,7887 Sunko, D. E. and Hehre, W. J. (1983). In Prog. Phys. Org. Chem. 14,205 Sunko, D. E., Szele, I. and Hehre, W. J. (1977). J. Am. Chem. SOC.99,5000 Swain, C. G. and Hershey, N. D. (1972). J. Am. Chem. SOC.94,1901 Swain, C. G., Stivers, E. C., Reuwer, J. F., Jr. and Schaad, L. J. (1958). J. Am. Chem. SOC. 80,5885 Szylhabel-Godala, A., Madhavan, S., Rudzinski, J., O'Leary, M. H. and Paneth, P. (1996). J. Phys. Org. Chem. 9, 35 Tencer, M. and Stein, A. R. (1978). Can. J. Chem. 56, 2994

248

0. MATSSON AND K. C. WESTAWAY

Thibblin, A. (1988). J. Phys. Org. Chem. 1, 161 Thibblin, A. and Ahlberg, P. (1977). J. Am. Chem. SOC.99,7926 Thibblin, A. and Ahlberg, I? (1989). Chem. SOC.Rev. 18,209 Thornton, E. R. (1967). J. Am. Chem. SOC.89,2915 Van Hook, W. A. (1970). In Isotope Effects in Chemical Reactions (ed. C. J. Collins, Jr. and N. S . Bowman), A.C.S. Monograph 167. Van Nostrand Reinhold, New York, pp. 1-12. Viggiano, A. A., Paschkewitz, J. S., Morris, R. A., Paulson, J. F., Gonzalez-Lafont, A. and Truhlar, D. G. (1991). J. Am. Chem. SOC.113,9404 Viggiano, A. A., Morris, R. A., Paschkewitz, J. S. and Paulson, J. F. (1992). J. Am. Chem. SOC.114,10477 Vitullo, V. P., Grabowski, J. and Sridharan, S. (1980). J. Am. Chem. SOC.102,6463 Westaway, K. C. (1987a). In Isotopes in Organic Chemistry, vol. 7 (ed. E. Buncel and C. C. Lee). Elsevier, Amsterdam, pp. 275-392. Westaway, K. C. (1987b). In Isotopes in Organic Chemistry, vol. 7 (ed. E. Buncel and C. C. Lee). Amsterdam, Elsevier, p. 311 Westaway, K. C. (1987~).In Isotopes in Organic Chemistry, vol. 7 (ed. E. Buncel and C. C. Lee), Elsevier, Amsterdam, p. 312 Westaway, K. C. (1993). Can. J. Chem. 71,2084 Westaway, K. C. (1996). In The Chemistry of the Functional Groups, Supplement F: The Chemistry of Amino, Nitroso, and Nitro and Related Groups (ed. S. Patai). Wiley-Interscience, New York. Westaway, K. C. and Ali, S. F. (1979). Can. J. Chem. 57,1354 Westaway, K. C. and Lai, Z. G. (1988). Can. J Chem. 66,1263 Westaway, K. C. and Waszczylo, Z. (1982). Can. J. Chern. 60,2500 Westaway, K. C., Pham, T. V and Fang, Y.-R. (1997). J. Am. Chem. SOC.119,3670. Westaway, K. C., Fang, Y.-R., Persson, J. and Matsson, 0. (1998). J. Am. Chem. SOC. 120, 3340. Westheimer, F. H. (1961). Chem Rev. 61, 265 Williams, I. H. (1984). J. Am. Chem. SOC.106, 7206 Williams, I. H. (1985). J. Chem. SOC.,Chem. Commun. 510 Wilson, H., Caldwell, J. D. and Lewis, E. S. (1973). J. Org. Chem. 38,564 Wolf, J. F., Devlin, 111, J. L., Taft, R. W., Wolfsberg, M. and Hehre, W. J. (1976). J. Am. Chem. SOC.98,287 Wolfe, S. and Kim, C.-K. (1991). J. Am. Chem. SOC.113, 8056 Wolfsberg, M. and Stern, M. J. (1964). Pure Appl. Chem. 8,225 Yamataka, H. and Ando, T. (1979). J. Am. Chem. SOC.101,266 Zhang, B.-L. (1988). Magn. Res. 26, 955 Zhang, Y., Bommuswamy, J. and Sinnott, M. L. (1994). J. Am. Chem. SOC.116,7557 Zhao, X . G., 'hcker, S. C. and Truhlar, D. G. (1991). J. Am. Chem. Soc., 113,826

Catalytic Antibodies G. MICHAEL BLACKBURN,* ANITA DATTA,HAZEL DENHAM AND PAUL WENTWORTH JR Krebs Institute, Department of Chemistry, University of Shefield, UK

Glossary 1 Introduction Antibodies and their biological role The quest for a new class of biocatalyst First examples of catalytic antibodies Stages in the production of catalytic antibodies 2 Approaches to hapten design Transition state analogues Bait and switch Entropy traps Desolvation Augmentation of chemical functionality 3 Spontaneous features of antibody catalysis Spontaneous covalent catalysis Spontaneous metal ion catalysis 4 Performance analysis of catalytic antibodies 5 A case study: NPN43C9 - an antibody anilidase Antibody production Mechanistic analysis Site-directed mutagenesis and computer modelling 6 Rescheduling regio- and stereo-chemistry of chemical reactions Diels-Alder cycloadditions Disfavoured regio- and stereo-selectivity Cationic cyclizations 7 Difficult processes Diastereoisomeric resolution Acetal and glycoside cleavage Phosphate ester cleavage Amide hydrolysis 8 Reactive immunization 9 Medical potential of abzymes Detoxification by catalytic antibodies Prodrug activation by catalytic antibodies Cell viability as an abzyme screen 10 Industrial potential of abzymes 11 Conclusions Appendix References

250 253 253 255 256 257 261 261 264 268 273 274 276 277 277 278 281 281 282 284 285 286 288 289 292 292 294 296 299 301 304 304 304 308 309 311 313 385

249 ADVANCES IN PHYSICAL ORGANIC CHEMISTRY VOLUME 31 0065-3160/98$30.00

Copyrighr 0 1998 Academic Press

All rights of reprodudon in any form reserved

250

G. BLACKBURN ETAL

Glossary

Abzyme An alternative name for a catalytic antibody (derived from Antibody-enzyme). Affinity labelling A method of identifying peptides located in the antigen binding site. The antibody is treated with a hapten which binds to the binding site and to proximal amino acid residues. Upon hydrolysis of the antibody, peptide fragments bound to the hapten are separated and identified. Antibodies Proteins of the immunoglobulin superfamily, carrying anfigenbinding sites that bind noncovalently to the corresponding epifope.They are produced by B lymphocytes (B cells) and are secreted from plasma cells in response to antigen stimulation. Antigen A molecule, usually peptide, protein or polysaccharide, that elicits an immune response when introduced into the tissues of an animal. B cells (also known as B lymphocytes) Derived from the bone marrow, where they differentiate into antibody-forming plasma cells and B memory cells, these cells are mediators of humoral immunity in response to antigens. Bait and switch A strategy whereby the charge-charge complementarity between antibody and hapten is exploited. By immunizing with haptens containing charges directed at key points of the reaction transition state, complementary charged residues are induced in the active site which are then used in catalysis of the substrate. BSA Bovine serum albumin, derived from cattle serum and used as a carrier protein. Carrier protein Macromolecule to which a hapten is conjugated, thereby enabling the hapten to stimulate the immune response. catELISA Similar to an ELZSA, except that the assay detects catalysis as opposed to simple binding between hapten and antibody. The substrate for a reaction is bound to the surface of the microtitre plate, and putative catalytic antibodies are applied. Any product molecules formed are then detected by the addition of anti-product antibodies, usually in the form of a polyclonal mixture raised in rabbits. The ELISA is then completed in the usual way, with an anti-rabbit “second antibody” conjugated to an enzyme, and the formation of coloured product upon addition of the substrate for this enzyme. The intensity of this colour is then indicative of the amount of product formed, and thus catalytic antibodies are selected directly. Conjugate In immunological terms this usually refers to the product obtained from the covalent coupling of a protein (e.g. a carrier protein) with a hapten, with a label such as fluorescein or with an enzyme. Conjugation The process of covalently bonding (multiple) copies of a hapten to a carrier protein, usually by means of a linker to distance the hapten from the surface of the carrier protein by a chain of about six atoms. ELISA (Enzyme-linked immunosorbent assay) An immunoassay in which antibody or antigen is detected. To detect antibody, antigen is first adsorbed

CATALYTIC ANTIBODIES

251

onto the surface of microtitre plates, after which the test sample is applied. Any unbound (non-antigen-specific) material is washed away, and remaining antibody-antigen complexes are detected by an antiimmunoglobulin conjugated to an enzyme. When the substrate for this enzyme is applied, a coloured product is formed which can be measured spectrophotometrically. The intensity of the coloured product is proportional to the concentration of antibody bound. Enhancement ratio, ER Quantified as kat/kUnat, is used to express the catalytic power of a biocatalyst. It is a comparison between the catalysed reaction occurring at its optimal rate and the background rate. Entropic trap A strategy aimed at improving the efficiency of catalytic antibodies, via the incorporation of a molecular constraint into the transition state analogue that gives the hapten a higher energy conformation than that of the reaction product. Epitope The region of an antigen to which antibody binds specifically. This is also known as the antigenic determinant. Fab’ The fragment obtained by pepsin digestion of immunogl~bulin~, followed by reduction of the interchain disulfide bond between the two heavy chains at the hinge region. The resulting fragment is similar to a Fab fragment in that it can bind with antigen univalently, but it has the extra hinge region of the heavy chain. Fab The fragment obtained by papain hydrolysis of immunoglobulins. The fragment has a molecular weight of -45 kDa and consists of one light chain linked to the N-terminal half of its corresponding heavy chain. A Fab contains one antigen binding site (as opposed to bivalent antibodies), and can combine with antigen as a univalent antibody. Hapten Substance that can interact with antibody but cannot elicit an immune response unless it is conjugated to a carrier protein before its introduction into the tissues of an animal. Haptens are mostly small molecules of less than 1kDa. For the generation of a catalytic antibody, a TSA (4.v.) is attached to a spacer molecule to give a hapten of which multiple copies can be linked to a carrier protein (qv.). Hybridoma Cell produced by the fusion of antibody-producing plasma cells with myelomakarcinoma cells. The resultant hybrids have then the capacity to produce antibody (as determined by the properties of the plasma cells), and can be grown in continuous culture indefinitely owing to the immortality of the myeloma fusion partner. This technique enabled the first continuous supply of monoclonal antibodies to be produced. IgG The major immunoglobulin in human serum. There are four subclasses of IgG; IgG1, IgG2, IgG3 and IgG4, but this number varies in different species. All are able to cross the placenta, and the first three subclasses flx complement by the classical pathway. The molecular mass of human IgG is 150 kDa and the normal serum concentration in man is 16 mg I&*. Immunoglobulin Member of a family of proteins containing heavy and light

252

G. BLACKBURN ETAL

chains joined together by interchain disulfide bonds. The members are divided into classes and subclasses,with most mammals having five classes (IgM, IgG, IgA, IgD and IgE). k,, The rate constant for the formation of product from a particular substrate. k,,, is obtained by dividing the Michaelis-Menten parameter, V,,,, by the total enzyme concentration. In real terms, the constant is a measure of how rapidly an enzyme can operate once its active site is occupied. KLH Keyhole limpet haemocyanin, used for its excellent antigenic properties. It is used as a currier protein in order to bestow immunogenicity in small haptens. K , The Michaelis-Menten constant, which is defined as the substrate concentration at which the biocatalyst is working at half its maximum rate (Vmax).In practice, K , gives a measure of the binding affinity between the substrate and biocatalyst; the smaller the value, the tighter the binding in the complex. Library A collection of antibodies, usually Fab or SCFVfragments, in the range of lo6 to 10'' and displayed on the surface of bacteriophage whose DNA gene contains a DNA sequence capable of expression as the antibody protein. Thus, identification of a single member of the library by selection can be used to generate multiple copies of the phage and sizeable amounts of the antibody protein. Monoclonal antibody, mAb Describes an antibody derived from a single clone of cells or a clonally obtained cell line. Its common use denotes an antibody secreted by a hybridoma cell line. Monoclonal antibodies are used very widely in the study of antigens, and as diagnostics. Polyclonal antibodies Antibodies derived from a mixture of cells, hence containing various populations of antibodies with different amino acid sequences. They are of limited use in that they will not all bind to the same epitopes following immunization with a haptenlcarrierprotein conjugate. They are also difficult to purify and characterize, but have been used with success in the catELZSA system. Positive clones A phrase usually used to describe those hybridoma clones which bind reasonably to their respective hapten in an enzyme-linked immunosorbent assay, thereby eliminating non-specific antibodies raised to different epitopes of the haptedcarrier conjugate. Residues General term for the unit of a polymer, that is the portion of a sugar, amino acid or nucleotide that is added as part of the polymer chain during polymerization. Single-chain antibody (SCFV) Comprises a VL linked to a VH chain via a polypeptide linker. It is thus a univalent functioning antibody containing both of the variable regions of the parent antibody. Site-directed mutagenesis Induced change in the nucleotide sequence of DNA aimed at particular nucleotide residues, usually in order to test their function.

CATALYTIC ANTIBODIES

253

Somatic hypermutation Mutations occurring in the variable region genes of the light and heavy chains during the formation of memory B cells. Those B cells whose affinity is increased by such mutations are positively selected by interaction with antigen, and this leads to an increase in the average affinity of the antibodies produced. Specificity constant Defined as kcat/Km.It is a pseudo-second-order rate constant which, in theory, would be the actual rate constant if formation of the enzyme-substrate complex were the rate-determining step. TSA (Transition state analogue) Frequently a stable analogue of an unstable, high-energy reaction intermediate that is close to related energy barriers in a multi-step reaction.

1 Introduction

This review addresses most of the important advances that have occurred in the field of catalytic antibodies since the first reports a decade ago (Pollack et al., 1986; Tramontano et al., 1986). One of the most stimulating features of this subject is that it is not confined to a single scientific discipline. Therefore, although this article looks at catalytic antibodies and their activities from a physical organic chemistry viewpoint, it seeks to provide a self-contained review requiring only a rudimentary biochemical knowledge of antibody structure, function and production. Adequate details of these matters have been supplied, including a glossary of many of the immunological terms employed written in general chemical language; these are included to stimulate rather than discourage the reader. The survey does not seek to be fully comprehensive, but rather focuses on the more significant parts of a subject which, in a little over ten years, has achieved much more than most pundits expected from this scientific prodigy in its infancy. However, a fairly complete survey of the literature is presented in the form of an Appendix, which tabulates over 120 examples of reactions catalysed, the haptens employed, and the kinetic data reported.

ANTIBODIES AND THEIR BIOLOGICAL ROLE

The immune response provides one of the most important biological defence mechanisms for higher organisms. It depends on the rapid generation of structurally novel proteins that can identify and bind tightly to foreign substances of potential harm to the parent organism. This family of proteins are the immunoglobulins. In their simplest form, they are made up of two pairs of polypeptide chains of different length and interconnected by disulfide bridges. The two light and two identical heavy chains contain repeated

254

G. BLACKBURN ETAL

Fig. 1 Schematic structure of the peptide components of an IgG immunoglobulin showing the two light (L) and two heavy (H) polypeptide chains, the disulfide bridges connecting them (-S-S-); the four variable regions of the light (V,) and heavy (V,) chains, and the 8 “constant” regions of the light (C,) and heavy (CHl,C d , Cd) chains (shaded rectangle). Hypervariable regions that provide antigen recognition and binding are located within six polypeptide loops, three in the VL and three in the VH sections (shaded circle, top left). These can be excised by proteolytic cleavage to give a fragment antibody, Fab (shaded lobe, top right).

homologous sequences of about 110 amino acids which fold individually into similar structural domains, essentially a bilayer of antiparallel P-pleated sheets. This leads to an IgG immunoglobulin molecule whose core structure is formed from 12 similar structural domains: 8 from the two heavy chains and 4 from the two light chains (Fig. 1) (Burton, 1990). By contrast, the N-terminal regions of antibody light and heavy chains vary greatly in the sequence and number of their constituent amino acids and thereby provide binding regions of enormous diversity, approaching 10” in number for higher mammals. The remarkable property of the immune system is its ability to respond to single or multiple alien species by rapid diversification of the sequences of these hypervariable regions through mutation, gene splicing, and RNA splicing. This generates a vast number of different antibodies which are selectively amplified in favour of those with the strongest affinity for the alien species.

CATALYTIC ANTIBODIES

255

THE QUEST FOR A NEW CLASS OF BIOCATALYST

In the mid-1940s Linus Pauling clearly stated the theory that enzymes work by their complementarity to the transition state for the reaction to be catalysed (Pauling, 1948). This concept was, with hindsight, a logical extension of the then relatively new transition state theory that had been developed to explain chemical catalysis (Evans and Polanyi, 1935; Eyring, 1935). Its fundamentals support the proposition that the rate of a reaction is related to the difference in Gibbs free energy (AG") between the ground state of reactant(s) and the transition state for the given reaction. For catalysis to occur, either the energy of the transition state has to be lowered (transition state stabilization) or the energy of the substrate has to be elevated (substrate destabilization). Pauling applied this to enzyme catalysis by stating that an enzyme preferentially binds to and hence stabilizes the transition state for a reaction over ground state of substrate(s) (Fig. 2). This has become a classical dogma in enzymology and is widely used to explain the way in which such biocatalysts are able to enhance specific processes with rate accelerations of up to lo'? over background (Albery and Knowles, 1976,1977; Albery, 1993 for a recent review). Pauling apparently did not bring ideas about antibodies into his concept of enzyme catalysis, though there is a tantalizing photograph in the volume of Pauling's Silliman lectures at Yale in 1947 which shows on a single blackboard cartoon both an energy profile diagram for the lowering of a transition state energy profile and also reference to an immunoglobulin (Pauling, 1947). And so it fell to Bill Jencks in his unsurpassed 1969 work on catalysis (Jencks, 1969)

Profile for

L

+ progress of reaction

-

*

Fig. 2 Catalysis is achieved by lowering the free energy of activation for a process, i.e. a catalyst must bind more strongly to the transition state (TSI) of the reaction than to either reactants or products. Thus: AAG* AAGca,:sand AAG,,,:,

*

256

G. BLACKBURN ETAL

to bring together the opportunity for synthesis of an enzyme using antibodies positively engineered in the immune system: “One way to do this [i.e. synthesize an enzyme] is to prepare an antibody to a haptenic group which resembles the transition state of a given reaction.”’ The practical achievement of this goal was held up for 18 years, primarily because of the great difficulty in isolation and purification of single-species proteins from the immune repertoire. During that time, many attempts to elicit catalysis by inhomogeneous (i.e. polyclonal) mixtures of antibodies were made and failed (e.g. Raso and Stollar, 1975; Summers, 1983). The problem was resolved in 1976 by Kohler and Milstein’s development of hybridoma technology, which has made it possible today both to screen rapidly the “complete” immune repertoire and to produce in vitro relatively large amounts of one specific monoclonal antibody species (Kohler and Milstein, 1975; Kohler et al., 1976). While transition states have been discussed in terms of their free energies, there have been relatively few attempts to describe their structure at atomic resolution for most catalysed reactions. Transition states are high-energy species, often involving incompletely formed bonds, and this makes their specification very difficult. In some cases these transient species have been studied using laser femtochemistry (Zewail and Bernstein, 1988), and predictions of some of their geometries have been made using molecular orbital calculations (Houk et af., 1995). Intermediates along the reaction coordinate are also often of very short lifetime, though some of their structures have been studied under stabilizing conditions while their existence and general nature can often be established using spectroscopic techniques or trapping experiments (March, 1992b). The Hammond postulate predicts that if a high-energy intermediate occurs along a reaction pathway, it will resemble the transition state nearest to it in energy (Hammond, 1955). Conversely, if the transition state is flanked by two such intermediates, the one of higher energy will provide a closer approximation to the transition state structure. This assumption provides a strong basis for the use of mimics of unstable reaction intermediates as transition state analogues (Bartlett and Lamden, 1986; Alberg et al., 1992).

FIRST EXAMPLES OF CATALYTIC ANTIBODIES

In 1986, Richard Lerner and Peter Schultz independently reported antibody catalysis of the hydrolysis of aryl esters and of carbonates, respectively (Pollack et al., 1986; Tramontano et al., 1986). Reactions of this type are well

’Jencks apparently was not aware of Pauling’s idea when he made this statement (Jencks, 1997, personal communication).

CATALYTIC ANTIBODIES

[ l ] Reactant

257

[2] Tetrahedral Intermediate

Products

Fig. 3 The hydrolysis of an aryl ester [l] (X = CH2) or a carbonate [l] (X = 0) proceeds through a tetrahedral intermediate [2] which is a close model of the transition state for the reaction. It differs substantially in geometry and charge from both reactants and products.

known to involve the formation and breakdown of an unstable tetrahedral intermediate, and so this can be deemed to be closely related to the transition state (TS’) of the reaction (Fig. 3). Transition states of this tetrahedral nature have now been mimicked effectively by a range of stable analogues, including phosphonic acids, phosphonate esters, a-difluoroketones, and hydroxymethylene functional groups (Jacobs, 1991). Lerner’s group elicited antibodies to a tetrahedral anionic phosphonate hapten [3] (Appendix entry 2.9)’ whilst Schultz’s group isolated a protein with high affinity for p-nitrophenyl cholyl phosphate [4] (Fig. 4) (Appendix entry 3.2).

STAGES IN THE! PRODUCTION OF CATALYTIC ANTIBODIES

It is appropriate at this stage in the review to consider the stages in production of a catalytic antibody and to put in focus the relative roles of chemistry, immunology, biochemistry, and molecular biology. Nothing less than the full integration of these cognate sciences is essential for the fullest realization of the most difficult objectives in the field of catalytic antibodies. In broad terms, the top section of the flow diagram for abzyme production (Fig. 5 ) involves chemistry, the right-hand side is immunology, the bottom sector is biochemistry, and molecular biology completes the core of the scheme. Chemistry

At the outset, chemistry dominates the selection of the process to be investigated (see Scheme 1later). The chosen reaction should meet most if not all of the following criteria: ’It might be helpful to the reader to indicate that the pyridine-2,6-dicarboxylate moiety in [3] was intended for an additional purpose, not used or needed for the activity described in the present scheme (Fig. 4).

G. BLACKBURN ETAL

258

Abzyme identity

Conditions

K,

PI

k,, [51

6D4"

pH 8; 25°C

1.9 p~

0.027 s-l

Ki

"31

0.16 p~

[61

[41

Abzyme identity

Conditions

Km [61

kca, 161

Ki [4l

MOPC167

pH 7; 30°C

208 p~

0.007 s-'

5 PM

Fig. 4 Lerner's group used phosphonate [3] as the hapten to raise an antibody which was capable of hydrolysing the ester [S] shown alongside it. Schultz found that naturally occurring antibodies using phosphate [4] as their antigen could hydrolyse the correspondingp-nitrophenylcholine carbonate [6]. (Those parts of haptens [3] and [4] required for antibody recognition have been emphasized with bold bonds.)

(a) (b) (c) (d) (e)

have a slow but measurable spontaneous rate under ambient conditions; be well analysed in mechanistic terms; be as simple as possible in number of reaction steps; be easy to monitor; lead to the design of a synthetically accessible TSA of adequate stability.

As we shall see later, most catalytic antibodies achieve rate accelerations in the range lo3 to lo6. It follows that for a very slow reaction, e.g. the alkaline hydrolysis of a phosphate diester with koH lo-'' M - ~s-' direct observation of the reaction is going to be experimentally problematic. Given that concentrations of catalytic antibodies employed are usually in the 1-10 p~ range, it has proved far more realistic to target the hydrolysis of an aliphatic ester, with koH 0.1 M-' s-' under ambient conditions. The need for a good understanding of the mechanism of the reaction is well illustrated by the case of amide hydrolysis. Many early enterprises sought to employ transition state analogues (TSAs) that were based on a stable anionic

-

-

2 59

CATALYTIC ANTIBODIES

ENTER

Hybridoma SCREEN for binding

Fig. 5 Stages in the production of a catalytic antibody.

tetrahedral intermediate, as had been successful for ester hydrolysis, and indeed identified catalytic antibodies capable of ester hydrolysis but not of amide cleavage! However, there is good evidence that, for aliphatic amides, breakdown of the tetrahedral intermediate (TI) is the rate-determining step and protonation of the leaving nitrogen is very important, and this must be built into TSA design. The importance of minimizing the number of covalent steps in the process to be catalysed is rather obvious. Single-step and double-step processes dominate the abzyme scene. However, there is substantial evidence that some acyl transfer reactions involve covalent antibody intermediates and so must proceed by up to four covalent steps. Nonetheless, such antibodies were not elicited by intentional design but rather discovered as a consequence of efficient screening for reactivity (Section 5). Direct monitoring of the catalysed reaction has most usually been carried out in real time by light absorption or fluorescent emission analysis and some initial progress has been made with light emission detection. The low quantity of abzyme usually available at the screening stage puts a premium on the sensitivity of such methods. However, some work has been carried out of necessity using indirect analysis, e.g. by hplc or nmr. Finally, this area of research might well have supported a JouraaE of Unsuccessful Abzyrnes. It is common experience in the field that some three out of four enterprises fail, and for no apparent reason. It is therefore imperative that chemical synthesis of a TSA should not be the ratedetermining step of an abzyme project. The average performance target is to achieve hapten synthesis within a year: one or two examples have employed

260

G. BLACKBURN ETAL

TSAs that could be found in a chemical catalogue, the most synthetically demanding cases have perforce employed multistep routes of considerable sophistication (e.g. Appendix entry 13.2). And, lastly, the TSA has to survive in vivo for at least 2 days to elicit the necessary antigenic response. Immunology

The interface of chemistry and immunology requires conjugation of multiple copies of the TSA to a carrier protein for production of antibodies by standard monoclonal technology (Kohler and Milstein, 1976). One such conjugate is used for mouse immunization and a second one for ELISA screening purposes. The carrier proteins selected for this purpose are bovine serum albumin (BSA), RMM 67 000, keyhole limpet haemocyanin, RMM 4 X lo6, and chicken ovalbumin, RMM 32 000. All of these are basic proteins of high immunogenicity and with multiple surface lysine residues that are widely used as sites for covalent attachment of hapten. Successful antibody production can take some 3 months and should deliver from 20 to 200 monoclonal antibody lines for screening, preferably of IgG isotype. Screening in early work sought to identify high affinity of the antibody for the TSA, using a process known as ELISA. This search can now be performed more quantitatively by BIAcore analysis, based on surface plasmon resonance methodology (LofAs and Johnsson, 1990). A subsequent development is the catELISA assay (Tawfik et al., 1993), which searches for product formation and hence the identification of abzymes that can generate product. Methods of this nature are adequate for screening sets of hybridomas but not for selection from much larger libraries of antibodies. So, most recently, selection methods employing suicide substrates (Section 7) (Janda et al., 1997) or DNA amplification methodology (Fenniri et al., 1995) have been brought into the repertoire of techniques for the direct identification of antibodies that can turn over their substrate. However, the tedious screening of hybridomas remains the mainstay of abzyme identification. Biochemistry

A family of 100 hybridoma antibodies can typically provide 20 tight binders and these need to be assayed for catalysis. At this stage in the production of an abzyme, the benefit of a sensitive, direct screen for product formation comes into its own. Following identification of a successful catalyst, the antibody is usually recloned to ensure purity and stabilization of the clone, then protein is produced in larger amount (-10 mg) and used for determination of the kinetics and mechanism of the catalysed process by classical biochemistry. Digestion of such protein with trypsin or papain provides fragment antibodies, Fabs, that contain only the attenuated upper limbs of the intact IgG (Fig. 1). It is these components that have been crystallized, in some

CATALYTIC ANTIBODIES

261

cases with the substrate analogue, product, or TSA bound in the combining site, and their structures have been determined by X-ray diffraction. Molecular biology

Only a few abzymes have reached the stage where mutagenesis is being employed to search for improved performance (Miller et al., 1997). Likewise, Hilvert is the first to have reached the stage of redesign of the hapten to attempt the production of antibodies with enhanced performance (Kast et al., 1996). So, the circle of production has now been completed for at least one example, and chemistry can start again with a revised synthetic target.

2 Approaches to hapten design

One can now recognize a variety of strategies in addition to the earliest ones deployed for hapten design. Some of these were presented originally as discrete solutions of the problem of abzyme generation, but it is now recognized that they need not be mutually exclusive either in design or in application. Indeed, more recent work often brings two or more of them together interactively. They can be classified broadly into five categories for the purposes of analysis of their principal design elements. The sequence of presentation of these here is in part related to the chronology of their appearance on the abzyme scene: 1. Transition state analogues 2. Bait and switch 3. Entropy traps 4. Desolvation 5. Functionality augmentation.

TRANSITION STATE ANALOGUES

As has clearly been shown by the majority of all published work on catalytic antibodies, the original guided methodology, i.e. the design of stable transition state analogues (TSAs) for use as haptens to induce the generation of catalytic antibodies, has served as the bedrock of abzyme research. Most work has been directed at hydrolytic reactions of acyl species, perhaps because of the broad knowledge of the nature of reaction mechanisms for such reactions and the wide experience of deploying phosphoryl species as stable mimics of unstable tetrahedral intermediates. More than 80 examples of hydrolytic antibodies have been reported, including the 47 examples of acyl group transfer to water listed below (Sections 1-5 of the Appendix).

G. BLACKBURN ETAL

262

Most such acyl transfer reactions involve stepwise addition of the nucleophile followed by expulsion of the leaving group with a transient, high-energy, tetrahedral intermediate (TI) separating these processes. The faster such reactions generally involve good leaving groups and the addition of the nucleophile is the rate-determining step. This broad conclusion from much detailed kinetic analysis has been endorsed by computation for the hydrolysis of methyl acetate (Teraishi et al., 1994). This places the energy for product formation from an anionic TI- some 7.6 kcal mol-’ lower than for its reversion to reactants. So, for the generation of antibodies for the hydrolysis of aryl esters, alkyl esters, carbonates and activated anilides, the design of hapten has focused on facilitating nucleophilic attack, and with considerable success. The tetrahedral intermediates used for this purpose initially deployed phosphorus(V) systems, relying on the strong polarization of the P=O bond (arguably more accurately represented as P+-O-). The range has included many of the possible species containing an ionized P-OH group (Scheme 1). One particularly good feature of such systems is that the P-0- bond is intermediate in length (1.521 A) between the C-0- bond calculated for a TI(0.2-0.3A shorter) and for the C..-O breaking bond in the transition state (some 0.6 8, longer) (Teraishi et al., 1992). Other tetrahedral systems used have included sulfonamides (Shen, 1995) and sulfones (Benedetti et al., 1996), secondary alcohols (Shokat el al., 1990), and a-fluoroketone hydrates (Kitazume et al., 1994). It is clear that phosphorus-based transition states have had the greatest success, as shown by the many entries in Sections 1-5 of the Appendix. This may be a direct result of their anionic or partial anionic character, a feature not generally available for the other species illustrated in Scheme 1, though a-difluorosulfonamides might reasonably also share this feature as a result of their enhanced acidity.

Phosphate diester

Phosphorothioate diester

Phosphinic acid

Sulfonamide

Phosphonate monoester

Sulfone Scheme 1

Phosphonamidate

Ketone hydrate

CATALYTIC ANTIBODIES

263

Fig. 6 Binding site details for antibody 4867 complexed with hapten p-nitrophenyl 4-carboxybutanephosphonate (Patten et al., 1996). N.B.: Amino acid residues in antibodies are identified by their presence in the light (L) or heavy (H) chains with a number denoting their sequence position from the N-terminus of the chain.

Not surprisingly, most of the catalytic antibody binding sites examined in structural detail have been found to contain a basic residue that provides a coulombic interaction with these TSAs, for which the prototype is the natural antibody McPC603 to phosphorylcholine, where the phosphate anion is stabilized by coulombic interaction with ArgH5’ (Padlan el af., 1985). In particular, X-ray structures analysed by Fujii (Fujii et af., 1995) have shown that the protonated HisH27din catalytic antibodies 6D9,4B5,8Dll and 9C10 (Appendix entry 1.8) is capable of forming a hydrogen bond to the oxyanion in the transition state for ester hydrolysis. In similar vein, Knossow has identified HisH35located proximate to the oxyanion of p-nitrophenyl methylphosphonate in the crystalline binary complex of antibody CNJ206 and TSA, a system designed to hydrolyse p-nitrophenyl acetate (cf. Appendix entry 2.7) (Charbonnier et al., 1995). A third example is seen in Schultz’s structure of antibody 48G7, which hydrolyses methylp-nitrophenyl carbonate (Appendix entry 3.1~).The hapten p-nitrophenyl 4-carboxybutanephosphonate is proximate to ArgL96and also forms hydrogen bonds to HisH35,5 r H 3 3and 5 r L 9 4(Fig. 6) (Patten et al., 1996). Clearly, the oxyanion hole is now as significant a feature of the binding site of such acyl transfer abzymes as it is already for esterases and peptidases and not without good reason. Knossow has analysed the structures of three esterase-like catalytic antibodies, each elicited in response to the same phosphonate TSA hapten (Charbonnier ef al., 1997). Catalysis for all three is accounted for by transition state stabilization and in each case there is an

264

G. BLACKBURN ETAL

oxyanion hole involving a tyrosine residue. This strongly suggests that evolution of immunoglobulins for binding to a single TSA hapten followed by selection from a large hybridoma repertoire by screening for catalysis leads to antibodies with structural convergence. Furthermore, the juxtaposition of X-ray structures of the unliganded esterase mAb D2.3 and its complexes with a substrate analogue and with one of the products provide a complete description of the reaction pathway. D2.3 acts at high pH by attack of hydroxide on the substrate with preferential stabilization of the oxyanion TIintermediate, involving one tyrosine and one arginine residue. Water readily diffuses to the reaction centre through a canal that is buried in the protein structure (Gigant et al., 1997). Such a clear picture of catalysis now opens the way for site-directed mutagenesis to improve the performance of this antibody.

BAIT AND SWITCH

Charge-charge complementarity is an important feature involved in the specific and tight binding of antibodies to their respective antigens. It is the amino acid sequence and conformation of the hypervariable (or complementarity-determining regions, CDRs) in the antibody combining site that determine the interactions between antigen and antibody. This has been exploited in a strategy dubbed “bait and switch” for the induction of antibody catalysts which perform p-elimination reactions (Shokat et al., 1989; Thorn et al., 1995), acyl-transfer processes (Janda et af.,1990b, 1991c; Suga et al., 1994a; Li and Janda, 1995), cis-trans alkene isomerizations (Jackson and Schultz, 1991) and dehydration reactions (Uno and Schultz, 1992). The bait and switch methodology deploys a hapten to act as a “bait”. This bait is a modified substrate that incorporates ionic functions intended to represent the coulombic distribution expected in the transition state. It is thereby designed to induce complementary, oppositely charged residues in the combining site of antibodies produced by the response of the immune system to this hapten. The catalytic ability of these antibodies is then sought by a subsequent “switch” to the real substrate and screening for product formation, as described above. The nature of the combining site of an antibody responding to charged haptens was fist elucidated by Grossberg and Pressman (1960), who used a cationic hapten containing a p-azophenyltrimethylammonium ion to elicit antibodies with a combining site carboxyl group, essential for substrate binding (as shown by diazoacetamide treatment). The first example of “bait and switch” for catalytic antibodies was provided by Shokat (Shokat et af., 1989), whose antibody 43D4-3D12 raised to hapten [7] was able to catalyse the p-elimination of [8] to give the trans-enone [9] with a rate acceleration of 8.8 X lo4 over background (Fig. 7; Appendix entry 8.1).

265

CATALMI C ANTI BODIES

PI

Abzyme identity

Conditions

Km

43D4-3D12

pH 6; 37°C

182 p~

kcat

[81

0.003 s-’

Ki

PI

0.29 p~

Fig. 7 Using the “bait and switch” principle, hapten [7] elicited an antibody, 43D4-3D12, which catalysed the p-elimination of [8] to a trans-ene-one [9]. The carboxyl function in [7] is necessary for its attachment to the carrier protein.

Subsequent analysis has identified a carboxylate residue, G ~ as uthe ~ catalytic function induced by the cationic charge in [7] (Shokat ef al., 1994). A similar “bait and switch” approach has been exploited for acyl-transfer reactions (Janda et al., 1990b, 1991~).The design of hapten [lo] incorporates both a transition state mimic and the cationic pyridinium moiety, designed to induce the presence of a potential general acid/base or nucleophilic amino acid residue in the combining site, able to assist in catalysis of the hydrolysis of substrate [ l l ] (Appendix entry 2.6). Some 30% of all of the monoclonal antibodies obtained using hapten [lo] were catalytic, and so the work was expanded to survey three other antigens based on the original TSA design (Janda et al., 1991~). The carboxylate anion in [12] was designed to induce a cationic combining site residue, whilst the quaternary ammonium species [ 131 combines tetrahedral mimicry and positive charge in the same locus. Finally, the hydroxyl group in [14] was designed to explore the effects of a neutral antigen (Fig. 8). Three important conclusions arose from this work. (i) A charged functionality is crucial for catalysis. (ii) Catalytic antibodies are produced from targeting different regions of the binding site with positive and negative haptens (though more were obtained in the case of the cationic hapten used originally). (iii) The combination of charge plus mimicry of the transition state is required to induce hydrolytic esterases.

Esterolytic antibodies have also been produced by Suga using a different “bait and switch” strategy (Appendix entry 2.1) (Suga et al., 1994a). A 1,2aminoalcohol function was designed for generating not only esterases but also amidases. Of three haptens synthesized, [15], [16] and [17], two contained

~

G. BLACKBURN ETAL

266

NHCO(CH2)3GOOR

NHCO(CH2)3COOR

NHCO(CH2)&OOR

NHCO(CH2)&OOR

[I31

[ 141

R = succinimidyl

0 NHCO(CH2)3COOH

[111

Fig. 8 The original hapten [lo] demonstrated the utility of the “bait and switch” strategy in the generation of antibodies to hydrolyse the ester substrate [ll].T h e e haptens, [12]-[14], were designed to examine further the effectiveness of point charges in amino acid induction. Both charged haptens, [12] and [13], produced antibodies that catalysed the hydrolysis of [ll],whereas the neutral hapten, [14],generated antibodies which bound the substrate unproductively.

ammonium cations and one a protonated amine, in order to elicit an anionic combining site for covalent catalysis. The outcome was interpreted as suggesting that haptens containing an NMe: group were too demanding sterically, so that the induced anionic amino acid residues in the antibody binding pocket were too distant to provide nucleophilic attack at the carbonyl carbon of substrate [MI. An alternative explanation may be that coulombic interactions lacking any hydrogen-bonding capability will not be sufficiently short range for the purpose intended. The use of secondary hydroxyl groups in the haptens [15] and [16] was designed to mimic the tetrahedral geometry of the transition state (as in Janda’s work), while the third hapten [17] replaced the neutral OH with an anionic phosphate group, designed to elicit a cationic combining site residue to stabilize the transition state oxyanion. However, this function in [17] may have proved too large to induce a catalytic residue close enough to the developing oxyanion, since weaker catalysis was observed relative to haptens [15] and [16] (kcatlkuncat = 2.4 X lo3, 3.3 X lo3, and -1 X lo3 for [15], [16], and [17] respectively) (Fig. 9). To achieve catalysis employing both acid and basic functions, an alternative zwitterionic hapten was proposed in which the anionic phosphoryl core is incorporated alongside the cationic ammonium moiety (cf. [171) (Suga et al.,

CATALYTIC ANT I8OD I ES

267

Fig. 9 Three haptens, [15]-[17], containing a 1,2-aminoalcohol functionality were investigated as alternatives for esterase and amidase induction. Of antibodies raised against hapten [15], 50% were shown to catalyse the hydrolysis of ester [18], thereby establishing the necessity for a compact haptenic structure. Hapten [19] along with [16] was employed in a heterologous immunization programme to elicit both a general and acidlbase function in the antibody binding site.

1994b). The difficulty in synthesizing such a target hapten can be overcome by stimulating the immune system first with the cationic and then with the anionic point charges using the two structurally related haptens [16] and [19], respectively. Such a sequential strategy has been dubbed “heterologous immunization” (Fig. 9) and the results of this strategy were compared with those from the individual use of haptens 1161 and [19] in a “homologous immunization” routine. Of 48 clones produced as a result of the homologous protocols, 7 were found to be catalytic, giving rate enhancements up to 3 X lo3. By contrast, 19 of the 50 clones obtained using the heterologous strategy displayed catalysis, the best being up to 2 orders of magnitude better. A final example of the bait and switch strategy (Thorn et al., 1995) focuses on the base-promoted decomposition of substituted benzisoxazole [20] to give cyanophenol [21] (Appendix entry 8.4). A cationic hapten [22] was used to mimic the transition state geometry of all reacting bonds. It was anticipated that if the benzimidazole hapten [22] induced the presence of a carboxylate in the binding site, it would be ideally positioned to make a hydrogen bond to the N-3 proton of the substrate. The resultant abzymes would thus have general base capability for abstracting the H-3 in the substrate (Fig. 10). %o monoclonals, 34E4 and 35F10, were found to catalyse the reaction with a rate acceleration greater than lo’, while the presence of a carboxylate-

G. BLACKBURN ETAL

268

0.

H CN

0[201

[211

Fig. 10 The use of a cationic hapten [22] mimics the transition state of the base-promoted decomposition of substituted benzisoxazole [20] to cyanophenol [21] and also acts as a “bait” to induce the presence of an anion in the combining site that may act as a general base.

containing binding site residue was confirmed by pH-rate profiles and covalent modification by a carbodiimide, which reduced catalysis by 84%. The bait and switch tactic clearly illustrates that antibodies are capable of a coulombic response that is potentially orthogonal to the use of transition state analogues in engendering catalysis. By variations in the hapten employed, it is possible to fashion antibody combining sites that contain individual residues to deliver intricate mechanisms of catalysis.

ENTROPY TRAPS

Rotational entropy

An important component of enzyme catalysis is the control of translational and rotational entropy in the transition state (Page and Jencks, 1971). This is well exemplified for unimolecular processes by the enzyme chorismate mutase, which catalyses the isomerization of chorismic acid [23] into prephenic acid [24]. This reaction proceeds through a cyclic transition state having a pseudo-diaxial conformation [25] (Addadi et al., 1983). With this analysis, Bartlett designed and synthesized a transition state analogue [26] which proved to be a powerful inhibitor for the enzyme (Bartlett and Johnson, 1985). X-ray structures of mutases from Escherichia coli (Lee et al., 1995), Bacillus subtilis (Chook et al., 1993, 1994) and Saccharomyces cerevisiae (Xue and Lipscomb, 1995) complexed to [26] show completely different protein architectures although the bacterial enzymes have similar values of k,,/k,,,, (3 X lo6) and of Ki for [26]. It appears that these enzymes exert their catalysis through a combination of conformational control and enthalpic lowering. Supporting this, Hillier has carried out a hybrid quantummechanical/molecular mechanics calculation on the B. subtilis complex with substrate [23]. He concluded that interactions between protein and substrate are maximal close to the transition state [25] and lead to a lowering of the energy barrier greater than is needed to produce the observed rate acceleration (Davidson and Hillier, 1994).

CATALYTIC ANTIBODIES

269

copI

c02-

OH

OH Chorismate [23]

Prephenate [24]

hc0*-

-Opt\

H

-.

-02c

OH

OR

Transition state [25]

TS Analogue [26]

Schultz employed TSA [26] as a hapten to generate antibodies to catalyse this same isomerization reaction [23]-[24] (Jackson et al., 1988). His kinetic analysis of one purified antibody revealed that it increases the entropy of activation of the reaction by 12 cal mol- K-' (Table 1, Antibody 11F1-2E11, Appendix entry 13.2b), and gives a rate enhancement of lo4.He suggested that this TSA induces a complementary combining site in the abzyme that constrains the reactants into the correct conformation for the [3,3]-sigmatropic reaction and designated this strategy as an "entropic trap". Table 1 Kinetic and thermodynamic parameters for the spontaneous, enzymecatalysed and antibody-catalysed conversion of chorismic acid [23] into prephenic acid [24].

Catalyst

AS*/ Relative AG*/ A p l calmol-' rate kcal mol-'kcal mol-' K-' K, [23]

Spontaneousa 1 Chorismate 3 X lo6 Mutaseh Antibody 1F7" 250 11Fl-2Elld 10000

24.2 15.9

20.5 15.9

-12.9 0

21.3 18.7

15.0 18.3

-22 51 p M -1.2 260 p~

45pM

"At 25°C. *E. coli enzyme at 25°C. 'pH 7.5; 14°C. dpH 7.0; 10°C.

k,,, [23]

1.35s-I

Ki[26] 75pM

0.072min-' 60011~ 0.27 min-' 9.0 p~

270

G. BLACKBURN FTAL

Hilvert’s group used the same hapten [26] with a different spacer to generate an antibody catalyst which has very different thermodynamic parameters. It has a high entropy of activation but an enthalpy lower than that of the wild-type enzyme (Table 1, Antibody 1F7, Appendix entry 13.2a) (Hilvert et al., 1988;Hilvert and Nared, 1988).Wilson has determined an X-ray crystal structure for the Fab‘ fragment of this antibody in a binary complex with its TSA (Haynes et at., 1994) which shows that amino acid residues in the active site of the antibody catalyst faithfully complement the components of the conformationally ordered transition state analogue (Fig. 11) while a trapped water molecule is probably responsible for the adverse entropy of activation. Thus it appears that antibodies have emulated enzymes in finding contrasting solutions to the same catalytic problem. Further examples of catalytic antibodies that are presumed to control rotational entropy are AZ-28, which catalyses an oxy-Cope [3.3]-sigmatropic rearrangement (Appendix entry 13.1) (Braisted and Schultz, 1994; Ulrich et at., 1996) and 2E4, which catalyses a peptide bond isomerization (Appendix entry 13.3) (Gibbs et al., 1992b; Liotta et al., 1995). Perhaps the area for the greatest opportunity for abzymes to achieve control of rotational entropy is in the area of cationic cyclization reactions (Li et al., 1997). The achievements of the Lerner group in this area (Appendix entries 15.1-15.4) will be discussed later in this article (Section 6).

Translational entropy The classic example of a reaction that demands control of translational entropy is surely the Diels-Alder cycloaddition. It is accelerated by high pressure and by solutions 8 M in LiCl (Blokzijl and Engberts, 1994; Ciobanu and Matsumoto, 1997; Dell, 1997) and proceeds through an entropically disfavoured, highly ordered transition state, showing large activation entropies in the range of -30 to -40 cal mol-’ K-’ (Sauer, 1966). While it is one of the most important and versatile transformations available to organic chemists, there is no unequivocal example of a biological counterpart. Hence, attempts to generate antibodies which could catalyse this reaction were seen as an important target. The major task in producing a “Diels-Alderase” antibody lies in the choice of a suitable haptenic structure, because the transition state for the reaction resembles product more closely than reactants (Fig. 12). The reaction product itself is an inappropriate hapten because it is likely to result in severe product inhibition of the catalyst, thereby preventing turnover. Tetrachlorothiophene dioxide (TCTD) [27] reacts with N-ethylmaleimide [28] to give an unstable, tricyclic intermediate [29] that spontaneously extrudes SO2 to give a dihydrophthalimide as the bicyclic adduct [30] (Raasch, 1980). This led to the design of hapten as a bridged dichloro-tricycloazadecene derivative [31] which closely mimics the high-energy intermediate [29] whilst

H

/Arg5'

)A1928

N

H

Glu58 Fig. 11 Schematic diagrams of X-ray crystal structures show the hydrogen-bonding (dashed lines) and electrostatic interactions between the transition state analogue [26] (in grey) with relevant side chains of (a) antibody 1F7 (Haynes et al., 1994) and (b) the active site of the E. coli enzyme (Lee et al., 1995).

272

G. BLACKBURN ETAL

Fig. 12 The Diels-Alder cycloaddition of TCI’D[27] and [28] proceeds through an unstable intermediate [29] which spontaneously extrudes SO2 to give the dihydrophthalimide adduct [30]. Hapten [31] was designed as a stable mimic of [29] that would be sufficiently different from product [30] to avoid product inhibition of the antibody catalyst.

being sufficiently different from the product [30] to avoid the possibility of end-product inhibition (Hilvert et al., 1989). Several antibodies raised to the hapten [31] accelerated the Diels-Alder cycloaddition between [27] and [28]. The most efficient of these, 1E9, performs multiple turnovers, showing that product inhibition has been largely avoided. Comparison of k,, with the second-order rate constant for the uncatalysed reaction (kuncat= 0.04 M-’ min-’, 25°C) gives an effective molarity: EM, of 110 M (Appendix entry 17.1) (Hilvert et al., 1989). This value is several orders of magnitude larger than any attainable concentration of substrates in aqueous solution, and therefore the antibody binding site confers a significant entropic advantage over the bimolecular Diels-Alder reaction. A number of further examples of Diels-Alder catalytic antibodies have been described (Appendix entries 17.2-17.5) and they must needs benefit from the same entropic advantage over spontaneous reactions, albeit without Hilvert’s ingenious approach to avoiding product inhibition. Their success in achieving control of regio- and stereo-chemistry will be discussed later (Section 6). Of greater long-term significance is the control of translational entropy for antibody-catalysed synthetic purposes. Benkovic’s description of an antibody ligase capable of joining an activated amino acid (e.g. [32]) to a second amino acid to give a dipeptide and to a dipeptide (e.g. [33]) to give a tripeptide with only low product inhibition is particularly significant (Scheme 2) (Appendix entry 18.4) (Smithrud et al., 1997). Antibody 16G3 can achieve 92% conversion of substrates for tripeptide formation and 70% for tetrapeptide synthesis within an assay time of 20 min. A concentration of 20 PM antibody can produce a 1 . 8 m ~solution of a dipeptide in 2h. The very good regio-control of the catalysed process is shown by the 80: 1ratio of formation 3The EM is equivalent to the concentration of substrate that would be needed in the uncatalysed reaction to achieve the same rate as achieved by the antibody ternary complex (Kirby, 1980).

CATALYTIC ANTI BODIES

0

273

NH~.HcI Ind = 3-indolyl

Scheme 2

of the programmed peptiL,: [34] compared to t..e unprogrammed product [35], whereas the uncatalysed reaction gives a 1: 1 ratio.

DESOLVATION

The Kemp decarboxylation of 6-nitro-3-carboxybenzisoxazole[36] is a classic example of rate acceleration by desolvation. Moving from water to a less polar environment can effect a 107-foldrate acceleration, which has been ascribed to a combination of (i) substrate destabilization by loss of hydrogen-bonding to solvent and (ii) transition state stabilization in a dipolar aprotic solvent (Kemp et al., 1975). Both Hilvert and Kirby have sought to generate abzymes for this process (Appendix entry 9.1) (Lewis et al., 1991; Sergeeva et al., 1996). Hilvert generated several antibodies using TSA [37] and the best, 25E10, gave a rate acceleration of 23 200 for decarboxylation of [36], comparable to rate accelerations found in other mixed solvent systems but much less than for hexamethylphosphoric triamide ( X lo8). In particular, it is of some concern that the K,,, for this antibody is as high as 25 mM, which reflects the tenuous relationship between the hapten design and the substrate/transition state structure. Unfortunately, apparently better-designed TSAs, e.g. [38] (Sergeeva ef al., 1996), fared worse in outcome, probably through the absence of a counter cation in the binding site. This may offer an opportunity for protein engineering to induce the presence of an N,N,N-trimethyllysine residue in the active site to provide a non-hydrogen-bonding salt pair. Selenoxide syn-eliminations are another reaction type favoured by less polar solvents (Reich, 1979). The planar 5-membered, pericyclic transition state for syn-elimination of [39] was mimicked by the racemic proline-based cis-hapten [39] to give 28 monoclonal antibodies (Appendix entry 8.5) (Zhou et al., 1997). Abzyme SZ-cis-42FV converted substrate [40] exclusively into

G. BLACKBURN ETAL

274

t361

trans-anethole [41] with an enhancement ratio (ER) of 62 (R = Me, X = NOz) and with a low K , of 33 PM. Abzyme SZ-cis-39C11 gave a good acceleration, k,,, 0.036 min-', k&K, 2400 M - ~min-' (substrate [40],R = H = X) comparable to the rate in 1,2-dichloroethane solution. Unexpectedly, the catalytic benefit appears to be mainly enthalpic both for the antibody and for the solvent switch, as shown by the data in Table 2.

AUGMENTATION OF CHEMICAL FUNCTIONALITY

Several antibodies have been modified to incorporate natural or synthetic groups to aid catalysis (Pollack et al., 1988). Pollack and Schultz reported the first example of a semi-synthetic abzyme through the introduction of an

CATALYTIC ANTIBODIES

275

Table 2 Parameters at 25°C for the syn-elimination of selenoxide [39] (R = X = H) in water, DCM, and catalysed by antibody SZ-cis-39C11. Catalyst

AG*

AH/

AS*/

kcal mo1-l cal mol-' K-'

26.3 t 0.15 26.3 ? 0.15 +0.014 ? 0.47 -7.8 ? 4.1 SZ-~is-39C11 22.2 2 1.2 19.7 ? 1.2 DCM 21.8 ? 0.5 20.3 5 0.5 -4.8 5 1.7

kc,t/Kru

M - ~min-'

Water

2400

(kobsV min-' ER

kcat

1.6 x 10-5 3.5 X lo-* 2200 4.4 X lo-' 2750

"25°C.

U

Fig. 13 A semi-synthetic abzyme. Selective derivatization of lysine-52 in the heavy chain of MOPC315 creates a thiol, then bonded to an imidazole, which gives an abzyme capable of improved hydrolysis of coumarin ester [42] with k,,, = 0.052 min-'.

imidazole residue into the catalytic site by selective modification of the thiol-containing antibody MOPC315 (Pollack and Schultz, 1989). This yielded a chemical mutant capable of hydrolysing coumarin ester [42] with k,, 0.052 min-' at pH 7.0,24"C. Incorporation of the nucleophilic group alone was previously shown to accelerate hydrolysis of the ester by a factor of lo4 over background controls (Pollack ef al., 1988). The process of modification is shown in Fig. 13. Lys-52 is first derivatized with 4-thiobutanal and then a catalytic imidazole is bonded through a disulfide bridge into the active site. This can now act as a general basehucleophile in the hydrolysis of [42], as was verified first by the pH-rate profile and then by complete deactivation of the antibody by diethyl pyrocarbonate (an imidazole-specific inactivating reagent). The first success in sequence-specific peptide cleavage by an antibody was claimed by Iverson (Iverson and Lerner, 1989). He used hapten [43] containing an inert Co"'(trien) complexed to the secondary amino acid of a

G. BLACKBURN ETAL

276

Fig. 14 A metal complex [43] used as hapten to raise antibodies capable of incorporating metal co-factors to facilitate the cleavage of [44] at the position

indicated

(1).

tetrapeptide. This approach was planned in the expectation of eliciting monoclonal antibodies with a binding site that could simultaneously accommodate a substrate molecule and a kinetically labile complex such as Zn"(trien) or Fe"'(trien), designed to provide catalysis. Much early work by Buckingham and Sargeson had shown that such cobalt complexes are catalytic for amide hydrolysis via polarization of the carbonyl group, through nucleophilic attack of metal-bound hydroxide, or by a combination of both processes (Sutton and Buckingham, 1987; Hendry and Sargeson, 1990). Of 13 peptidolytic monoclonals, 287F11 was selected for further analysis. At pH 6.5, cleavage of substrate [44]was observed with a variety of metal complexes. The Zn"(trien) complex was the most efficient, with 400 turnovers s-' (Fig. 14). per antibody combining site and a turnover number of 6 X While this approach is undoubtedly ingenious, there are some doubts about its actual performance. The site of cleavage of peptide [44] is not between the N-terminal phenylalanine and glycine, as expected from the design of the hapten, but rather between glycine and the internal phenylalanine. Moreover, attempts to repeat this work have not been overly successful. A major achievement in augmenting the chemical potential of antibodies has been in the area of redox processes. Many examples now exist of stereoselective reductions, particularly recruiting sodium cyanoborohydride (Appendix Section 22). A growing number of oxidation reactions can now be catalysed by abzymes, with augmentation from oxidants such as hydrogen peroxide and sodium periodate (Appendix Section 21).

3 Spontaneous features of antibody catalysis

While the presentation thus far has emphasized the programmed relationship of hapten design and consequent antibody catalytic activity, there is a growing number of examples where the detailed examination of catalysis reveals mechanistic features that were not evidently design features of the system at the outset. Such discoveries are clearly a strength rather than a weakness of

CATALYTIC ANTI B 0 DIES

277

the abzyme field, and two of these outturns are described in the following sections.

SPONTANEOUS COVALENT CATALYSIS

The nucleophilic activity of serine in the hydrolysis of esters and amides by many enzymes is one of the classic features of covalent catalysis by enzymes. So it was perhaps inevitable that an antibody capable of catalysing the hydrolysis of a phenyl ester should emerge having the same property. Scanlan has provided just that example with evidence from kinetic and X-ray structural analysis to establish that the hydrolysis of phenyl (R)-N-formylnorleucine [45] proceeds via an acyl antibody intermediate with abzyme 17E8 (Appendix entry 2.3) (Zhou et al., 1994).The antibody reaction has a bell-shaped pH-rate profile corresponding to ionizable groups of pK, 9.1 and 10.0. On the basis of X-ray analysis, the latter appears to be LysHY7, while a candidate for the former is TyrH'". This system is deemed to activate SerHg9as part of a catalytic diad with HisH35(Scheme 3 [46]). In addition to the kinetic and structural evidence

H

3 H

[461

Scheme 3

for this claim, a trapping experiment with hydroxylamine generated a mixture of amino acid and amino hydroxamic acid products from substrate [45] in the presence of antibody. In a similar vein, antibody NPN43C9 appears to employ a catalytic histidine, HisLg1,as a nucleophilic catalyst in the hydrolysis of a p-nitrophenyl phenylacetate ester, as discussed in detail below (Section 5; Appendix entry 2.8) (Gibbs et al., 1992a; Chen et al., 1993).

SPONTANEOUS METAL ION CATALYSIS

Janda and Lerner sought to establish that a metal ion or coordination complex need not be included within the hapten used for the induction of abzymes so that they can (i) bind a metallo-complex and thereby (ii) provide a suitable

G. BLACKBURN ETAL

278

environment for catalysis (Wade et al., 1993). The pyridine ester [48] was screened as a substrate for 23 antibodies raised against [47] as hapten. Antibody 84A3 proved to be capable of hydrolysis of [48] only in the presence of zinc, with a rate enhancement of 12 860 over the spontaneous rate and 1230 over that seen in the presence only of zinc. Other metals, CdZf,Co2+,Ni2+, were without activity. The affinity of 84A3 for the substrate was high at 3.5 p ~ , whereas the affinity for zinc in the presence of substrate was only 840 p ~This . is far weaker than any affinity of real use for the incorporation of metal ion activity into the catalytic antibody repertoire (plasma [Zn’+] is 17.2 p ~ ] . However, the resources of mutagenesis can readily be targeted on this problem with expectations of success. Given the great importance of the metalloproteinases, it seems inevitable that further work will be directed at this key area either by designed or opportunistic incorporation of metal ions into the catalytic apparatus of abzymes.

NHCO(CH2)3C02R

0

NHR’

4 Performance analysis of catalytic antibodies

In the first years of abzyme research, a majority of examples was concerned with acyl group transfer reactions. Many of these endeavours have been based on mimicry of the high-energy, tetrahedral intermediate that lies along such reaction pathways (Section 2) and which, though not truly a “transition state analogue”, provides a realistic target for production of a stable TSA. Most, though not all, were themselves based on four-coordinate phosphoryl centres. In 1991, Jacobs analysed 18 examples of antibody catalysis of acyl-transfer reactions as a test of the Pauling concept, i.e. delivering catalysis by TSS stabilization. The range of examples included the hydrolysis of aryl carbonates and of both aryl and alkyl esters. In some cases more than one reaction was catalysed by the same antibody, in others the same reaction was catalysed by different antibodies. Much earlier, Wolfenden (Westerick and Wolfenden, 1972) and Thompson (1973), established a criterion for enzyme inhibitors working as TSAs. They proposed that such activity should be reflected by a linear relationship between the inhibition constant for the enzyme Kiand its inverse second-

CATALYTIC ANTIBODIES

279

Fig. 15 A thermodynamic cycle linked to transition state theory gives an equation relating the enhancement ratio for a biocatalysed process to the ratio of equilibrium constants for the complex between the biocatalyst and (i) substrate and (ii) the transition state for the reaction. These two values can be estimated as K,, and Ki for the TSA, respectively.

order rate constant, K,/k,,,, for pairs of inhibitors and substrates that differ in structure only at the TSA/substrate locus. That has been well validated, inter alia, for phosphonate inhibitors of thermolysin (Bartlett and Marlowe, 1983) and pepsin (Bartlett and Giangiordano, 1996). In order to apply such a criterion t o a range of catalytic antibodies, Jacobs assumed firstly that the spontaneous hydrolysis reaction proceeds via the same TS* as that for the antibody-mediated reaction and secondly that all corrective factors due to medium effects are constant. By treating the hydrolysis reactions as pseudofirst-order processes, one can derive a simple relationship with approximations of KTsand K s to provide a mathematical statement in terms of Ki, K,, k,,, and k,,,,, (Fig. 15) (Wolfenden, 1969; Jencks, 1975; Benkovic et al., 1988; Jacobs, 1991). A log-log plot using K,, K,, k,,, and k,,,,, data from the 18 separate cases of antibody catalysis exhibited a linear, albeit scattered, correlation over four orders of magnitude and with a gradient of 0.86 (Fig. 16).4 Considering the assumptions made, this value is sufficiently close to unity to suggest that the antibodies do stabilize the transition state for their respective reactions. However, even the highest k,,,lk,,,,, value of lo6 in this series (Tramontano et al., 1988) barely compares with enhancement ratios seen for weaker enzyme catalysts (Lienhard, 1973).

'It may also be worth mentioning here that many early estimates of Kd for the affinity of the antibody to their TSA were upper limits, being based o n inhibition kinetics using concentrations of antibody that were significantly higher than the true K , being determined.

280

G. BLACKBURN ETAL

5

(K,/K~) 4-

Gradient = 0.86 r2 = 0.8

1

0

Fig. 16 Jacobs’ correlation between the enhancement ratio (kcat/kunmt)and the relative affinity for the TSA with respect to the substrate (Km/Ki)(Jacobs, 1991). The slope is an unweighted linear regression analysis.

The fact that many values of K J K , fall below the curve (Fig. 16) suggested that interactions between the antibody and the substrate are largely passive in terms of potential catalytic benefit. This conclusion exposes a serious limitation in the design of haptens, were that to be restricted solely to the transition state concept. It is well known that enzymes utilize a range of devices to achieve catalysis as well as dynamic interactions to guide substrate towards the transition state, which is then selectively stabilized. However, as has been illustrated above, the original concept of transition state stabilization has been augmented by a range of further approaches in the generation of catalytic antibodies and with considerable success. A second use of this type of analysis has been presented by Stewart and Benkovic (1995). They showed that the observed rate accelerations for some 60 antibody-catalysed processes can be predicted from the ratio of equilibrium binding constants to the catalytic antibodies for the reaction substrate, K,, and for the TSA used to raise the antibody, Ki. In particular, this approach supports a rationalization of product selectivity shown by many antibody catalysts for disfavoured reactions (Section 6) and predictions of the extent of rate accelerations that may be ultimately achieved by abzymes. They also used the analysis to highlight some differences between mechanism of catalysis by enzymes and abzymes (Stewart and Benkovic, 1995). It is interesting to note that the data plotted (Fig. 17) show a high degree of scatter with a correlation coefficient for the linear fit of only 0.6 and with a slope of 0.46, very different from the “theoretical slope” of unity. Perhaps of greatest significance are the

CATALYTIC ANTIBODIES

281

P Reaction Catalysed

6-

Ester Hydrolysis Ether Cleavage Claisen Decarboxylation Elimination Miscellaneous

Ester Synthesis

Amide Synthesis

0-

,' y = 0.462~ + 1.801 r = 0.597

DieldAkler

Fig. 17 The Stewart-Benkovic plot of rate enhancement vs relative binding of substrate and TSA for 60 abzyme-catalysed reactions (Stewart and Benkovic, 1995). The theoretical unit slope (---) diverges from the linear regression slope (-) for these data (for which the equation is shown).

many positive deviations from the general pattern. These appear to show that antibody catalysis can achieve rather more than is predicted from catalysis through transition state stabilization alone. 5 A case study: NPN43C9 - an antibody anilidase

At this point, we can integrate much of what has been discussed above in a single case study. Antibody NPN43C9 was reported in 1988 as the first example of catalysis of hydrolysis of an amide bond, in fact of an active anilide. Its structure and mode of action have been well studied (Janda et al., 1988b), which makes it an appropriate example for this purpose.

ANTIBODY PRODUCTION

Hapten design

Amide hydrolysis at alkaline pH involves a tetrahedral anionic intermediate, which was mimicked by the transition state analogue [49], an N-aryl arylphosphonamidate, appropriately related to substrate anilide 1501 (Fig. 18) (Appendix entry 2.8).

G. BLACKBURN ETAL

282

Abzyme identity

Conditions

K, [50]

NPN43C9

pH 9; 37°C

370 ~

L M

kcat

[SO]

0.05 min-'

Kd

[491

0.8 n M

Fig. 18 Antibody NPN43C9, raised against the phosphonamidate hapten [49], was capable of accelerating the hydrolysis of the anilide [50]. Whilst hapten [49] satisfies the stereoelectronic requirements for the TI- for amide hydrolysis, the resulting immune response may be dominated by the nitrophenyl and benzylic ring systems. Thus, antibodies generated will necessarily be anilidases and not amidases. NPN43C9 is, none the less, an important and interesting antibody in terms of the nitrogen leaving group in the reaction it catalyses and also because of the modelling and sequencing work carried out on it (vide infra). Bacterial expression The total cDNA construct of NPN43C9 was expressed efficiently in E. coli cells and protein purified, and its catalytic properties were assessed in both the monoclonal antibody and the single-chain antibody (scFv) 7A4-11212 for the hydrolysis of p-nitroanilide [50] and the related p-chlorophenyl ester [Sla] (Fig. 19). Virtually identical k,,, and K , values were obtained for both 7A4-1/212 and NPN43C9. This activity was decreased in both cases by the addition of the inhibitor m-nitroanilide [52], which gave Ki = 800 p~ and 400 p~ for the NPN43C9 and 7A4-1/212, respectively.

MECHANISTIC ANALYSIS

Kinetic analysis NPN43C9 was shown to give a rate acceleration for hydrolysis of [50] of approximately 1.5 X lo5, and its values of K , and V,,, were approximately the same as those for its Fab fragment, whose RNA sequence was subsequently used in cloning and expression of Fabs in a bacteriophage A system (Huse et al., 1989). Such an enterprise is capable of giving a greatly

CATALYTIC ANTIBODIES

283

1511 a X=CI b X=COCH,

cX=CHO d X=CH,

1521

eX=H f X=NO,

R = NHC(O)(CH,),CO,H

Fig. 19 Ester [sla] was used to investigate the comparative catalytic efficiency of the scFv 7A4-11212 and the parent mAb NPN43C9. This activity was inhibited by

m-nitroanilide [52]. expanded number of potential catalysts. It prompted a further study in which the coding sequences of the variable heavy (V,) and variable light chain (V,) fragments were used in the assembly of a single-chain antibody (Gibbs et al., 1991). The phosphonamidate [49] used to elicit 43C9 was designed to encourage general acid-base catalysis via oxyanion stabilization and protonation of the amide nitrogen in the tetrahedral transition state. However, results of pH-rate profiles in both D 2 0 and H 2 0 indicated that the mechanism involved an anionic transition state, probably progressing from the TI- (Benkovic et al., 1990,1991). The behaviour of the Michaelis-Menten parameters, k,,,lK,,, and k,,, as a function of pH shows that catalytic activity increases with increasing pH to a maximum with an apparent pK, of 9.0. Furthermore, the analysis helps to explain the deviation by almost lo3 of the value of k,,,lk,,,,, above that predicted on the basis of K,IK, (Section 4). Benkovic has postulated that this deviation may be a consequence of chemical catalytic processes (e.g. general acid-base or nucleophilic catalysis) being involved in the binding site for 43c9. The occurrence of a kinetic isotope effect in the pH-dependent region but its absence in the plateau region has been interpreted as suggesting the existence of two chemically distinct processes. The k,,, value at p H > 9 correlates with the rate-limiting formation of an acyl-antibody intermediate, whilst at low pH there is hydroxide-mediated hydrolysis of this intermediate. Moreover, I8O incorporation experiments showed that very little I8O exchange occurs in the NPN43C9-catalysed reaction relative to the uncatalysed one, which is consistent with acyl-intermediate formation preventing exchange (Janda et al., 1991a). The existence of a covalent acyl-antibody intermediate was further supported by analysis of the effects of a range of p-substituents on phenyl ester hydrolysis (Gibbs et al., 1992a). The antibody was found to catalyse hydrolysis of less reactive substrates [51a-e] within a rate factor of 10 of that for the p-nitroester substrate [51f], indicating that breakdown of the intermediate is the rate-determining step.

284

G. BLACKBURN E T A L

Substrate variations

Analysis of the substituent effects on NPN43C9 catalysis was achieved using a Hammett a - p correlation. A large p value of +2.3 was seen for the antibody-catalysed reaction. Such a large dependency on the leaving group is characteristic of nucleophilic attack by a neutral nitrogen nucleophile such as imidazole. By contrast, hydrolysis via general base catalysis would result in little charge build-up on the phenol oxygen and a low p value of 0.5-0.7 would be expected. Nucleophilic attack by, for example, hydroxide would lead to greater charge build-up in the TS' and a higher p value of -1.0-1.2. That a histidine residue was the likely candidate for this nucleophilic role was pinpointed by two further experiments. First, chemical modification of NPN43C9 with a variety of reagents was inhibitory only with diethyl pyrocarbonate (DEPC), a reagent specific for histidine residues. Secondly, molecular modelling of the antibody binding site region highlighted two histidine residues, one of which was suitably positioned for attack on the substrate carbonyl group.

SITE-DIRECTED MUTAGENESIS AND COMPUTER MODELLING

The use of site-directed mutagenesis and computer modelling enabled the ligand binding and catalytic residues to be identified (Stewart et al., 1994). A computer model of NPN43C9 with bound antigen identified specific residues as targets for site-specific mutagenesis, namely 5rL3*,HisLg1,ArgLg6,HisH35 and T ) T ~ Replacement ~~. of HisLg'by a glutamine generated a mutant devoid of catalytic activity but with an affinity for the hapten almost as high as for the parent antibody. This implicated HisLg1as the nucleophilic imidazole responsible for acyl-antibody intermediate formation. ArgLg6was also shown to be important for catalysis since, as predicted by modelling, its proximity to the carbonyl carbon suggested it should stabilize the anionic tetrahedral transition state. Mutation of ArgLg6to a neutral glutamine was found to destroy catalytic activity. Thus, the positively charged amino acid side-chain was assigned as flanking an oxyanion hole, polarizing the substrate carbonyl for nucleophilic attack, and stabilizing the anionic transition state by electrostatic interaction. The resultant mechanism for the hydrolysis of a p-nitrophenyl ester substrate is as follows. Substrate binding orientates the guanidinium cation of ArgLg6 towards the carbonyl group, locating the carbonyl carbon proximate to HisLg1.Attack of an imidazole nitrogen of HisLg1generates the acyl intermediate, assisted by coulombic interactions from The breakdown of the acyl-antibody intermediate involves attack by hydroxide and sequential release of antibody followed by phenol and acid products (Fig. 20).

CATALYTIC ANTIBODIES

At

--.-

285 HisL9' -Y- \

Fig. 20 The proposed catalytic mechanism for hydrolysis of ester substrate [Xf] showing proposed roles for active site residues ArgLg6and HisL".

In conclusion, NPN43C9 provides an excellent example of the application of standard techniques of physical organic chemistry in the characterization of an antibody both mechanistically and structurally.

6 Rescheduling the regio- and stereo-chemistry of parallel chemical reactions

The control of kinetic vs thermodynamic product formation can often be achieved by suitable modification of reaction conditions. A far more difficult task is to switch from the formation of a favoured major product to a disfavoured minor product, especially when the transition states for the two processes share most features in common. This challenge has been met by antibodies with considerable success, both for reaction pathways differing in regioselectivity and also for ones differing in stereoselectivity. In both situations, control of entropy in the transition state must hold the key.

G . BLACKBURN ETAL

286

DIELS-ALDER CYCLOADDITIONS

In the Diels-Alder reaction between an unsymmetrical diene and dienophile, up to eight stereoisomers can be formed (March, 1992a). It is known that the regioselectivity of the Diels-Alder reaction can be biased so that only the four ortho-adducts are produced (Fig. 21) through increasing the electronwithdrawing character of the substituent on the dienophile (Danishefsky and Hershenson, 1979). However, stereochemical control of the Diels-Alder reaction to yield the disfavoured exo-products in enantiomerically pure form has proved to be very difficult. Gouverneur et al. (1993) were interested in controlling the outcome of the reaction between diene [53] and N,N-dimethylacrylamide [54] (Fig. 22). They had shown experimentally that the uncatalysed reaction gave only two

re face

q+LR2

''0

7G<,

approach

approach

R2

re face

R'

OI,,

R'

si face

Q

si face

Fig. 21 Enantio- and diastereo-selectivity of the Diels-Alder reaction for orthoapproach.

(=+'!

CONH2

CONMe2 Q C O N M e 2

O Y()""CONMe2 NH

(571

I COfNa'

Fig. 22 The Diels-Alder cycloaddition between the dienophile [54] and diene [53] yields two diastereoisomers [55] and [56].Attenuated substrate analogues [57] and [58] were used in molecular orbital calculations of this reaction.

CATALYTIC ANTIBODIES

287

Table 3 Calculated activation energies of the transition structures relative to reactants for the reaction of acrylamide [57] with N-( 1-butadieny1)carbamic acid [58].

Calculated activation energy/kcalmol- ’ ~~~

Transition state geometry Ortho-endo Ortho-e.uo Meta-endo Meta-exo

RHF/3-21G

6-31G*/3-21G

27.3 28.84 29.82 30.95

40.80 42.70 42.88 43.94

stereoisomers; the ortho-endo (cis) [55] and the ortho-exo (trans) [56] adducts in an 85: 15 mixture. This experimental observation was underpinned by ab initio transition state modelling for the reaction of acrylamide [57] with N-butadienylcarbamic acid [58], which showed that the relative activation energies of the ortho-endo and ortho-exo transition states were of considerably lower energy than the meta-endo and meta-exo transition structures (not illustrated) (Table 3). The design of hapten was crucial for the generation of abzymes to deliver full regio- and diastereo-selectivity. Transition state analogues were thus devised to incorporate features compatible with either the disfavoured endo [59] or favoured ex0 [60] transition states (Fig. 23) (Appendix entry 17.5). Furthermore, because the transition state for Diels-Alder processes is very product-like, haptens [61] and [62] were developed to mimic a high-energy, boat conformation for each product, a strategy developed by Hilvert to minimize product binding to the abzyme (Hilvert et al., 1989). Two of the monoclonal antibodies produced, 7D4 and 22C8, proved to be completely stereoselective, separately catalysing the endo and the ex0 Diels-Alder reactions, with a k,,, of 3.44 X and 3.17 X min-’ respectively at 25°C. That the turnover numbers are low was attributed in part to limitations in transition state representation: modelling studies had shown that the transition states for both the ex0 and endo processes were asynchronous whereas both TSAs [61] and [62] were based on synchronous transition states (Gouverneur et al., 1993). In a further enterprise, compounds [63] and [64] (Fig. 24) were perceived as freely rotating haptens for application as TSAs for the same Diels-Alder addition. As expected, each proved capable of inducing both endo- and exo-adduct-forming abzymes. It can be noted that [63] produced more “exo-catalysts” (6 out of 7) whereas [64] favoured the production of “endocatalysts” (7 out of S), though it is difficult to draw any conclusion from this observation (Appendix entry 17.5) (Yli-Kauhaluoma er al., 1995).

G. BLACKBURN ETAL

288

endo-Transitionstate

(+L I R'

R2

-

endo-Hapten

exo-lkansition state

exo-Hapten

Fig. 23 Design of haptens [61] and [62], which are analogues of the favoured endo- [59] or disfavoured exo- [60] transition states, respectively.

[63]R = (CH2)&02H [64] R = 4-carboxyphenyl

Fig. 24 An alternative strategy for eliciting Diels-Alderase antibodies has employed freely rotating ferrocenes [63] and [64] as TSAs.

DISFAVOURED REGIO- AND STEREO-SELECTIVITY

Reversal of kinetic control in a ring closure reaction

In reactions where several different outcomes are possible, the final product distribution reflects the relative free energies of each transition state when the reaction is under kinetic control (Schultz and Lerner, 1993). Baldwin's rules predict that for acid-catalysed ring closure of the hydroxyepoxide 1651 the tetrahydrofuran product [66] arising from 5-exo-tet attack will be preferred

289

CATALMIC ANTIBODIES 5-ex0

6-endo

Fig. 25 The monoclonal antibody 26D9, generated to the N-oxide hapten [67], catalysed the 6-exo-tet ring closure of [65] regioselectively to yield the &favoured tetrahydropyran product [68]. This is a formal violation of Baldwin's rules, which predicts a 5-exo-tet spontaneous process to generate tetrahydrofuran derivative [66].

(Fig. 25) (Baldwin, 1976). By raising antibodies to the charged hapten [67], Janda and co-workers produced an abzyme which accelerated 6-ex0 attack of the racemic epoxide to yield exclusively the disfavoured tetrahydropyran product [68] and in an enantiomerically pure form (Appendix entry 14.1) (Janda et a!., 1993). This work reveals that an antibody can selectively deliver a single regio- and stereo-chemically defined product for a reaction in which multiple, alternative transition states are accessible and can also selectively lower the energy of the higher of two alternative transition states. Syn-elimination of P-fluoroketones The base-catalysed p-elimination of HF from the ketone [68] normally gives the favoured (E)-product [69] via a staggered conformation in the transition state. Hapten [70] is designed to enforce the syn-coplanar conformation of the phenyl and benzoyl functions in the transition state and so catalyse the disfavoured syn-elimination of [68] to give the (2)-a$-unsaturated ketone [71] (Fig. 26). Preliminary estimates of the energy difference between the favoured and disfavoured processes are close to 5 kcal mol-' (Cravatt et al., 1994), though this value is exceeded in the antibody-catalysed rerouting of carbamate hydrolysis from ElcB to BA,2 (Section 9, Appendix entry 5.3) (Wentworth etal., 1997). Antibody 1D4, raised to hapten [70]and used in 15% DMSO at pH 9.0 and 37"C, gave exclusively the (2)-product [71] with K , 212 PM and k,,, 2.95 X 10-3min-'. Under the same conditions, kobs is 2.48 X min-l for formation of [69] and immeasurably slow for the (undetectable) formation of [71] (Appendix entry 8.1). CATIONIC CYCLIZATIONS

The cationic cyclization of polyenes to give multi-ring carbocyclic compounds with many sterically defined centres is one of the more remarkable examples

G. BLACKBURN ETAL

290

Fig. 26 The elimination of HF from the P-fluoroketone [68] is catalysed by antibody 1D4, elicited to hapten [70], to form the disfavoured (2)-olefin [71]. This contrasts with the spontaneous process in which an anti-elimination reaction yields the (E)-a,punsaturated ketone [69]. The syn-eclipsed conformation of [70] is shaded.

98%

[731

Fig. 27 The N-oxide hapten [74] was used to elicit mAb 6D4 which catalysed the cyclization of [72] to form the cyclohexanol [73].

of regioselective and stereoselective enzyme control which has provided a major challenge for biomimetic chemistry (Johnson, 1968). It provides an excellent opportunity for the application of regio- and stereo-control by catalytic antibodies. Li et al. (1997) have discussed the use of catalytic antibodies to control the reactivity of carbocations. At an entry level, the acyclic olefinic sulfonate ester [72] is converted into the cyclic alcohol [73] (98%) using antibody 4C6 raised to hapten [73] with only 2% of cyclohexene produced (Appendix entry 15.1) (Li et al., 1994). Moving closer to a cationic transition state mimic, Hasserodt et al. (1996) used the amidinium ion [75] as a TSA for cyclization of the arenesulfonate ester [76]. One antibody raised to this hapten, 17G8, catalysed the conversion of substrate [76] into a mixture of the 1,6-dimethylcyclohexene [77] and 2-methylene-1-methylcyclohexane[78] (Fig. 28) (Appendix entry 15.3). By contrast, the uncatalysed cyclization of [76] formed a mixture of 1,2-

CATALYTIC ANTIBODIES

291

9:

I

Q o,

-

1771

17G8

AcHN

Fig. 28 Antibody 17G8 raised against TSA [75] catalysed the cyclization of [76] to give [77] and [78].

[83a] LG

=

AcNHC6H4S03

Fig. 29 Formation of isomeric decalins [71]-[73] by cyclization of a terpenoid alcohol catalysed by antibody HA5-19A4 raised to hapten [82]. The transition state [83a] has the leaving group in the equatorial position, as favoured by the Stork-Eschenmoser hypothesis.

dimethylcyclohexanols and a little 1,2-dimethylcyclohexene.Evidently, the antibody both excludes water from the transition state and also controls the loss of a proton following cyclization. While cyclopentanes have also been produced by antibody-catalysed cyclization (Appendix entry 15.2) (Li et al., 1996), much the most striking example of cationic cyclization by antibodies is the formation of the decalins [79], [80] and [81] (Fig. 29). The trans-decalin epoxide [82] (tll2 100 h at 37°C) was employed as a mixture of two enantiomeric pairs of diastereoisomers as a TSA to raise antibodies, among which HA5-19A4 emerged as the best catalyst for cyclization of substrate [83] (Appendix entry 15.4) (Hasserodt er al., 1997). Sufficient substrate [83] was transformed to give 10 mg of mixed products.

292

G. BLACKBURN ETAL

The olefinic fraction (70%) was predominantly a mixture of the three decalins [79], [80] and [81] in a 2: 3 :1 ratio and formed along with a diastereoisomeric mixture of cyclohexanols (30%).Moreover, the decalins were produced with enantiomeric excesses of 53%, 53% and 80%,respectively. It is significant that the (Z)-isomer of [83] is not a substrate for this antibody. Quite clearly, the antibody first catalyses ionization of the arenesulfonate to generate a carbocation. This process shows an ER of 3200 with K , 320 PM. The resulting cation can then either cyclize to decalins in a concerted process (as in transition state [83a]) or in two stepwise cyclizations. The formation of significant amounts of cyclohexanols seems to indicate that the latter is the case. Most interestingly, inhibition studies strongly suggest that the isomer of the haptenic mixture that elicited this antibody has structure [82], which would locate the leaving group in an axial position. This is contrary to the Stork-Eschenmoser concept of equatorial leaving group and presents a challenge for future examination (Eschenmoser et al., 1955; Stork and Burgstahler, 1955). It is an exciting prospect that catalysts of this nature may lead to artificial enzymes capable of processing natural and unnatural polyisoprenoids to generate various useful terpenes. 7 Difficult processes

As exponents of catalytic antibodies have become more confident of the power of abzymes, their attention has turned from reactions of moderate to good feasibility to more demanding processes. Their work has on the one hand tackled more adventurous stereochemical problems and on the other hand is attempting to catalyse reactions whose spontaneous rates are very slow indeed. Examples of both of these areas are discussed in this section. DIASTEREOISOMERIC RESOLUTION

Antibodies generally show very good recognition in favour of their antigens and against regio- or stereo-isomers of them. This results from a combination of the inherent chirality of proteins and the refined response of the immune system (Playfair, 1992). In extension, this character suggests that a catalytic antibody should be capable of similar discrimination in its choice of substrate and the transition state it can stabilize, as determined by the hapten used for its induction. As already shown above, the murine immune system can respond to a single member of a mixture of stereoisomers used for immunization (Section 6 ) . Such discrimination has been exemplified in antibody-catalysed enantioselective ester hydrolysis (Janda et al., 1989; Pollack et al., 1989; Schultz, 1989) and transesterification reactions (Wirsching et al., 1991; Jacobsen et al., 1992).

293

CATALYTIC ANTIBODIES

Table 4 Kinetic parameters for those antibodies raised against phosphonates [8&91] which effect the resolution of the fluorinated alcohols [84-87]. The configuration of the disastereoisomerically pure product from each antibody-catalysed process was shown to correspond to that of the antibody-inducing hapten. Alcohol product ~~

[MI [851 I861 ~371

Hapten ~

Configuration ~~

[881 ~ 9 1 [901 [911

k,,,"lmir~'

K,,,/pM

% eelde of product

0.88 0.91 0.94 0.86

390 400 410 380

99.0 98.5 98.5 98.0

~~

2R, 3R (+) 2s, 3s (-) 2R, 35' (+) 2S, 3R (-)

"At 25°C.

One study has made use of abzyme stereoselectivity to resolve the four stereoisomers (R,R', S,S', R,S' and S,R') of 4-benzyloxy-3-fluoro-3-methylbutan-2-01 [84-871 through the antibody-mediated hydrolysis of a diastereoisomeric mixture of their phenacetyl esters (Kitazume et al., 1991b). Antibodies were raised separately to each of four phosphonate diastereoisomers [88-911, corresponding to the four possible transition states for the hydrolysis of the four diastereoisomeric esters (Fig. 25) (Appendix entry 1.12). Each antibody operated on a mixture of equal parts of the four diastereoisomers as substrate to give each alcohol in -23% yield, with >97% eelde, and leaving the three other stereoisomers unchanged. By sequential action of the four antibodies in turn, the mixture of diastereoisomers could effectively be separated completely (Table 4).In a similar vein, Kitazume also resolved the enantiomers of 1,l,l-trifluorodecan-2-olwith 98.5 % enantiomeric excess (Appendix entry 1.11) (Kitazume et al., 1991a).

M6 F

1 --i-

\

Me I

Fig. 30 Four stereoisomeric alcohols [84-871were separated by selective hydrolysis of their respective phenacetyl esters using four antibody catalysts, each raised in response to a discrete stereoisomeric phosphonate hapten 188-911.

G. BLACKBURN ETAL

294

ACETAL AND GLYCOSIDE CLEAVAGE

Antibody-catalysed transformations for the synthesis, modification and degradation of carbohydrates are a subject of active investigation. Preliminary studies have reported antibody hydrolysis of model glycoside substrates (Yu et al., 1994) while the regio- and stereo-selective deprotection of acylated carbohydrates has been achieved using abzymes with moderate rate enhancements. Antibodies raised against the TSA [92] were screened for their ability to hydrolyse the diester [93]. One antibody, 17El1, used in a 20% concentration with respect to [93], effected hydrolysis exclusively at C-4. This process was fast enough to render spontaneous C-3 to C-4 acyl migration insignificant: i.e. no C3-OH product was detected. (This migration reaction is generally fast compared to chemical deacylation) (Fig. 31. Appendix entry 1.7) (Iwabuchi et al., 1994). In this context, the use of an antibody to cleave a trityl ether by an SN1process may have further applications (Appendix entry 7.1) (Iverson et al., 1990). Also, the objective of utilizing abzymes in the regioselective removal of a specified protecting group has been extended to show that an antibody esterase can have broad substrate tolerance (Appendix entry 1.18) (Li et a[., 1995b). The assault on the demanding task of glycosylic bond cleavage is making good, albeit slow, progress. As a first step, Reymond has described an antibody capable of catalysing the acetal hydrolysis of a phenoxytetrahydropyran, s-' at 24"C, and has a modest ER of 70 though it is slow, with k,,, 7.8 X (Appendix entry 7.4B) (Reymond et al., 1991). A general approach to the task has been to raise antibodies to TSAs related

AcHN

AcHN R = COCHzCHzSH ~~

Abzyme identity 17Ell

[921

~

~

Conditions

K , [93]

pH 8.2; 20°C

6.6 p~

kat

[931

0.182 min-'

KI [921 0.026 PM

Fig. 31 Antibody 17Ell raised against the TSA [92] was screened for its ability to hydrolyse diester [93] and, used in a 20% concentration with respect to [93], effected hydrolysis exclusively at C-4.

CATALYTIC ANTIBODIES

295

Fig. 32 Antibody AA71.17 raised against hapten [95] catalyses the hydrolysis of the aryl acetal [94].

by design to well-known inhibitors of glycosidases. Piperidino and pyrrolidino cations have high affinity for pyranosidases and furanosidases (Winchester and Fleet, 1992; Winchester et al., 1993) and can also be envisaged as components of a “bait and switch” approach to antibody production. Thus, Schultz has described the hydrolysis of the 3-indolyl acetal [94] (Fig. 32) by antibody AA71.17 raised to transition state analogue [95] (Appendix entry M is rather 7.2) (Yu et al.. 1994). Antibody AA71.17 has a good K , of 320 ~ L but slow in turnover, k,,, 0.015min-’ at 25°C. (By contrast, two other haptens based on a guanidino and a dihydropyran inhibitor did not elicit any antibodies that showed glycosidase activity.) More recently, Janda has described the production of a galactopyranosidase antibody in response to hapten [96]. This was designed to accommodate several features of the transition state for glycoside hydrolysis: notably a flattened half-chair conformation and substantial sp2character at the anomeric position. Some 100 clones were isolated in response to immunization with [96] and used to generate a cDNA library for display on the surface of phage (Appendix entry 7.3) (Janda et al., 1997). Rather than proceed to the normal screening for turnover, Janda then created a suicide substrate system to trap the catalytic species. Halazy had earlier shown that phenols with o- or p-difluoromethyl substituents spontaneously eliminate HF to form quinonemethides that are powerful electrophiles and that this activity can be used to trap glycosidases (Halazy et al., 1992). So, glycosylic bond cleavage in [97] (Fig. 33) results in formation of the quinonemethide [98] that covalently traps the antibody catalyst. By suitable engineering of a bacteriophage system, Janda was able to screen a large library of Fab fragment antibodies and select for catalysis. Fab 1B catalysed the hydrolysis at 37°C of p-nitrophenyl p-galactopyranoside with k,,, = 0.007 min-’ and K , = 530 p ~corresponding , to a rate enhancement of 70 000. Moreover, this activity was inhibited by hapten [96] with K , = 15 p ~ . By contrast, the best catalytic antibody, 1F4, generated from hapten [96] by classical hybridoma screening showed k,,, = lo-’ min-’ and K , = 330 p ~ a , rate enhancement of only 100. Clearly, this work both offers an exciting method for screening for

G. BLACKBURN E T A L

296

HO

CO-linker

OH

NH

[961

P71

CO-linker &-linker

[98]

CO-linker 60-linker

Fig. 33 Fragment antibody FablB is selected by suicide selection with substrate [97] from a library of antibodies generated to hapten [96]. The suicide intermediate is the o-quinonemethide [98].

antibodies that can lead to suicide product trapping and also appears to offer a general approach to antibodies with glycosidase activity.

PHOSPHATE ESTER CLEAVAGE

The mechanisms of phosphate ester cleavage vary significantly between monoesters, diesters, and triesters (Thatcher and Kluger, 1989). Each of these is a target for antibody cleavage and progress has been reported for all three cases. Phosphate monoesters This reaction is a particular challenge in light of the fact that phosphoryl transfers involving tyrosine, serine or threonine play crucial roles in signal transduction pathways that are control elements of many aspects of cellular physiology. The generation of an abzyme would provide an important biological tool for the investigation and manipulation of such processes. Schultz’s group employed an a-hydroxyphosphonate hapten [99] and subsequently isolated 20 cell lines of which 5 catalysed the hydrolysis of the model substrate p-nitrophenyl phosphate [loo] above background (Fig. 34) (Scanlan et al., 1991). Antibody 38E1 was characterized in more detail and kinetic parameters were afforded by Lineweaver-Burke analysis. This antibody exhibited 11 turnovers per binding site with no change in V,,,, and thus acted as a true catalyst. Moreover, examination of substrate specificity showed that catalysis was entirely selective for p-substituted species (Appendix entry 6.6). Phosphodiester cleavage Considering that the phosphodiester bond is one of the most stable chemical linkages in nature, its cleavage is an obvious and challenging target for antibody catalysis. In an attempt to model a metal-independent mechanism, a

CATALYTIC ANTIBODIES

Abzyme identity 38E1

297

Conditions pH 9.0

K , [lo01

155 ~

L M

k,," [lo01

Ki [991

0.0012 min-'

34 p~

"At 30°C.

Fig. 34 (above) Antibody 38E1, generated from the a-hydroxyphosphonate hapten

[99], catalysed the hydrolysis of p-nitrophenyl phosphate [loo].

nucleotide analogue [1011 comprising an 0-phosphorylated hydroxylamine moiety was chosen by Sakurai and co-workers (Sakurai et al., 1996). This hapten design aims to represent the geometry and spatial constraints in the phosphate linkage so as to retain the stereoelectronic configuration of the phosphorus atom, and finally to act as a simple model of a dinucleotide. To this end, the retention of the phosphate backbone seeks to facilitate the formation of an oxyanion hole in which the electrophilicity of the phosphorus centre is increased in the bound substrate, whilst the positive charge on the hapten is designed to elicit an anionic amino acid in the abzyme combining site, to act either as a general base to activate a nucleophilic water molecule or as a nucleophile operating directly at the phosphorus centre. More details are awaited from this work. The classic case of assisted hydrolysis of phosphate diesters is neighbouring

298

G. BLACKBURN ETAL

Fig. 35 Hapten anti-[102] was used to generate 25 mAbs from which 2G12 proved to catalyse the hydrolysis of the phosphate diester [103].

group participation by a vicinal hydroxyl group, specifically the phosphate ester of a 1,2-diol, which provides a rate acceleration greatly in excess of lo6 (Westheimer, 1968). While vanadate complexes of 1,Zdiols have been explored as pentacoordinate species for inhibiting enzymes, they are toxic and are too labile for use in murine immunization (Crans et al., 1991). Janda has found a solution to this problem through the use of pentacoordinate oxorhenium chelates (Weiner et al., 1997). Hapten [lo21 was employed as separate diastereoisomers to generate 50 monoclonal antibodies which were screened for their ability to hydrolyse uridine 3'-(p-nitrophenyl phosphate) [lo31 (Fig. 35) (Appendix entry 6.4). The most active of three antibodies, 2G12, had k,,, 1.53 X lop3s-' at 25°C and K , 240 p ~giving , k,,,lk,,,,, 312. A more favourable expression for protein catalysts working at substrate concentrations below K , is k,,,/(K, X k,,,,,) (Radzicka and Wolfenden, 1995), which is 1.3 X lo6M - ~and compares favourably to the value for RNase A of 10" M - ~for the same substrate. The TSA anrZ-[102] proved to be a powerful inhibitor for 2G12 with Ki estimated at 400 nM. Evidently, this is a system with scope alike for improvement in design and for broader application. It is clearly one of the most successful examples of antibody catalysis of a difficult reaction. Phosphotriester hydrolysis Catalysis of this reaction was first exhibited by antibodies raised by Rosenblum et al. (1995). More recently, Lavey and Janda (1996a) have explored the generation of abzymes capable of catalysing the breakdown of poisonous agrochemicals. Wenty-five mAbs were raised against the N-oxide hapten [lo41 of which two were found to be catalytic. The hapten was designed to generate antibodies for the hydrolysis of triester [lo51 using the "bait and

299

CATALYTIC ANTIBODIES

Abzyme identity

Substrate

Conditions

[I051 [1061

pH 8.1, 25°C

15C5

3H5

pH 9.15

Km

kc,,"

87 p~ 0.0027 min-' 5.05 mM 0.0020 min-'

Ki [lo41 -

0.98 p~

"Measuredat pH 8.25 and 25°C.

Fig. 36 The N-oxide [lo41 was used as hapten to raise mAbs to catalyse the hydrolysis of both triesters [lo51 and [106].

switch" methodology: cationic charge on the nitrogen atom targeted to induce anionic amino acids to act as general base catalysts; partial negative charge on oxygen to encourage the selection of antibody residues capable of stabilizing negative charge in the transition state (Fig. 36). Antibody 15C5 was able to catalyse the hydrolysis of the triester [lo51 with min-' whilst a second antibody from the same immunization k,,, 2.65 X programme was later found to hydrolyse the acetylcholinesterase inhibitor Paraoxon [lo61 with k,,, = 1.95 X lo-' min-' at 25°C (Appendix entry 6.2) (Lavey and Janda, 1996b). Antibody 3H5 showed Michaelis-Menten kinetics and was strongly inhibited by the hapten [104]. It exhibited a linear dependence of the rate of hydrolysis on hydroxide ion concentration, suggesting that 3H5 effects catalysis by transition state stabilization rather than by general acid/base catalysis. Phosphate ester hydrolysis is one of the most demanding of reactions for catalyst engineering. The progress made so far with catalytic antibodies is highly promising and appears to be competitive with studies using metal complexes if only because they can deliver turnover while metal complexes have for the most part to solve the problem of tight product binding.

AMIDE HYDROLYSIS

While ester, carbonate, carbamate and anilide hydrolyses have been catalysed effectively by antibodies, the difficult tasks of hydrolysis of an aliphatic amide or a urea remain largely unsolved. Much of this problem hinges on the fact that breakdown of a TI' is the rate-determining step, as established by much

G. BLACKBURN E T A L

300

kinetic analysis and, more recently, by computation. Teraishi has computed that C-N bond cleavage for the TI- for hydrolysis of N-methylacetamide (or aminolysis of acetic acid) lies some 18 kcal mol-’ above that for C-O(H) bond cleavage (Teraishi et al., 1992). Clearly, protonation of nitrogen has to be an essential feature of the breakdown of such intermediates and for anilides that is hardly a practical proposition under ambient conditions. To date only two investigations of this problem have shown any success. A group at IGEN raised antibodies to the dialkylphosphinic acid [107]. These were screened for their ability to hydrolyse four alkyl esters and four primary amides at pH 5.0,7.0, and 9.0. Just one out of 68 antibodies, 13Dl1, hydrolysed the C-terminal carboxamide stereospecifically of only the ( R ) substrate [108], which was rendered visible by the attachment of a dansyl

R

f [ 1071

[lo81 X = MI, [lo91 X = OMe

fluorophore to support hplc analysis of the course of the reaction (Appendix entry 5.1) (Martin et al., 1994). At pH 9.0 and 37”C, 13Dll showed K , 432 p~ and k,, 1.65 X lO-’s-’, a half-life of 42 d. This activity was fully inhibited by . the dansyl group proved to be an hapten [lo71 with Ki = 14 p ~Unexpectedly, essential component of the substrate. Even more unexpectedly the antibody did not hydrolyse the corresponding methyl ester [109]. Whereas most amide substrates for catalytic antibodies have been activated by the use of aromatic amines (Appendix entries 5.3,5.4), Blackburn chose to explore hydrolysis of an aliphatic amide, activated through halogenation in the acyl moiety (Shen, 1995; Datta et al., 1996). Chloramphenicol [110] was selected as substrate on account of its dichloroacetamide function and the tetrahedral intermediate for hydrolysis was mimicked by the neutral sulfonamide [ l l l ] and the zwitterionic “stretched transition state analogue” aminophosphinic acid [112]. Antibodies produced to each of these haptens proved too weak to hydrolyse chloramphenicol at a rate sufficiently above background (koH = 1.3 X M-’ s-l) for further study. However, a switch to the more active amide, trifluoramphenicol [113] (k, = 6 X lO-’s-’, kOH= 6.3 X lo-’ M-’ s-’ at 37”C), enabled useful data to be obtained for antibody 2B5 which showed Michaelis-Menten kinetics with k,,, = 2 X s-’ and K , = 640 p~ at pH 7.0, 37°C. Once again, the use of kcat/(& X k,,,,,) gives a more favourable value for the ER of 5200. The high K,,, is likely to be a consequence of exchanging the dichloroacetamide moiety

CATALYTIC ANTIBODIES

301

O2N

O2N

[ I 101

0

O

[113]

OH

OH

in the hapten for the tnfluoroacetamide group in the substrate and could presumably be improved by redesign of the hapten. The low rate of turnover achieved clearly indicates the difficult task ahead for antibody cleavage of a peptide based on tetrahedral intermediate mimicry alone. By contrast, the reverse reaction, that of amide synthesis, has proved to be a good target for antibody catalysis and a range of different enterprises have been successful (Appendix entries 18.1-18.4). It would appear here that little more is needed than a good leaving group and satisfactory design of a TSA based on an anionic tetrahedral intermediate (Benkovic et al., 1988; Janda et al., 1988a; Hirschmann et al., 1994; Jacobsen and Schultz, 1994). 8 Reactive immunization

A novel approach for the induction of catalysis in antibody binding sites is a strategy dubbed “reactive immunization” (Winching et al., 1995). This system uses haptens of intermediate chemical stability as immunogens. After the first stimulation of the mouse B cells to generate antibodies, one of the products of in vivo chemical transformation of the original hapten is then designed to act as a second immunogen to stimulate further mutational development of antibodies that will be better able to catalyse the desired reaction. The system seems well designed to achieve the benefits of a neutral and a charged hapten within the same family of monoclonal antibodies. An organophosphate diester [114], was chosen as the primary reactive immunogen. Following spontaneous hydrolysis in vivo it becomes a stable monoester transition state analogue [115], which in turn gives a new challenge to the immune system (Fig. 37) (Appendix entry 2.14). Cross-reactivity has been established as an advantage in this process since heterologous

G . BLACKBURN E T A L

302

R

J,~F;-*, H

Ar

[ I 141

spontaneous > R

bi vivo

J JF-o-

/“A,

N

o

N

[ I IS]

H

-

STABLE IMMUNOGEN TSA

REACTIVE IMMUNOCEN Ar = 4-(methylsulfonyl)phenyl

a R = represents position of linker to which carrier protein is attached b R = H in free hapten

JNrn 11 161

H

SUBSTRATE

Fig. 37 Antibodies raised simultaneously against the reactive and stable immunogen shown above were capable of efficient “turnover” of the related aryl ester substrate [116] (Ab 49H4: K , = 300 p ~k,,, ; = 31 rnin-’ at 22°C).

immunization with both diary1 ester [114] and the corresponding monoaryl ester gave cross-reacting serum with enhanced affinity for the monoaryl TSA. This further promotes the induction of active-site amino acids capable of acting as nucleophiles or general acid/base catalysts. In practice, reactive immunization with [114a] generated 19mAbs, 11 of which were able to hydrolyse substrate [114b]. The most efficient abzyme, SP049H4, was analysed kinetically using radioactive substrates. It was established that 49H4 had undergone reactive immunization, since it was able to turn over the aryl carboxylate aryl [116] very effectively with K , = 300 p ~k,,,; = 31 min-’. A similar approach has been used by Lerner and Barbas to induce catalytic antibodies mimicking type I aldolases. The reaction scheme is shown in Fig. 38: the aim here was to induce an enamine moiety which can achieve catalysis through lowering the entropy for bimolecular reaction between ketone substrate and aldol acceptor. Compound [117] is a 1,3-diketone which acts to trap the “critical lysine”, forming the vinylogous amide [118], which can be monitored spectrophotometrically at 318 nm (Appendix entry 16.2) (Lerner and Barbas, 1996). Screening for this catalytic intermediate by incubation with hapten facilitated the detection of two monoclonal antibodies with k,,, = 2.28 X lo-’ M-’ min-’. Furthermore, kcat/(K,X k,,,,,) is close to lo9, making these antibodies nearly as efficient as the naturally occurring fructose 1,6-bisphosphate aldolase. Studies on the stoichiometry of the reaction by titration of antibody with acetylacetone indicated two binding sites to be involved in the reaction. The antibodies generated in this programme were initially found to accept a broad range of substrates including acetone, fluoroacetone, 2-butanone, 3-pentanone and dihydroxyacetone. The list has now been expanded to include

CATALYTIC ANTIBODIES

303

Fig. 38 The mechanism by which an essential Lys residue in the antibody combining site is trapped using the 1,3-diketone [117] to form the covalently linked vinylogous amide [118].

hundreds of different aldol condensations. However, a more remarkable property of 38C2 emerged when it was screened for its ability to catalyse an intramolecular Robinson annulation reaction (Fig. 39). Ab38C2 accepts equally well both the (R)-(-) and (S)-(+) enantiomers of 2-(3-oxobutyl)-2-methylcyclohexanone [ 1191 and converts them stereospecifically into the respective stereoisomer of l-methyldecal-5-en-3-one: (R)-isomer k,,, = 0.126 min-I, K , = 2.45 mM; (S)-isomer k,,, = 0.186 min-l, K , = 12.4 m~ at 25°C (Appendix entry 16.2) (Zhong et al., 1997). While this example of the Robinson annulation is clearly not enantioselective, the same antibody converts the meso-ketone [120] into the WielandMiescher (WM) decalenedione product: k,,, = 0.086 min-' and K , = 2.34 mM at 25"C, parameters that give an impressive ER of 3.6 X lo6. Good evidence suggests that the mechanism of the reaction involves the formation of a ketimine with the s-amino group of a buried lysine residue in the antibody, as shown in Fig. 39. Most significantly, the reaction delivers the (S)-( +)-WM product in 96% ee (by polarimetry) and in 95% ee by nmr and hplc analysis for a 100mg scale reaction. A recent report tells that this antibody is to be made commercially available at a cost of $100 for 10mg. The realization of that objective would mark the start of a new era of application of abzymes to organic stereoselective synthesis. Finally, the whole process of reactive immunization opens up the opportunity of using mechanism-based inhibitors as haptens, capable of actively promoting a desired mechanism by contrast to their conventional use as irreversible enzyme inhibitors.

G. BLACKBURN ETAL

304

(S)-(+)-WM-ketone

Substrate

k,,/min-' ~~

w-(+)-[1 191

(w-( -)-[1191 [I201

KJmM ~

Ab

kunca,Imh-'

ER

nd nd

nd nd

~~

0.186 0.126 0.086

12.4 2.45 2.34

2.4 X

3.6 X lo6

Fig. 39 Robinson annulation of cyclohexanones [119] and [120] catalysed by antibody Ab38C2 (Zhong et al., 1997). 9

Medical potential of abzymes

DETOXIFICATION BY CATALYTIC ANTIBODIES

The idea that abzymes might be used therapeutically to degrade harmful chemicals in homo offers a new route to the treatment of victims of drug overdose. Landry's group have produced antibodies to catalyse the hydrolysis of the benzoyl ester of cocaine [121] yielding the ecgonine methyl ester [122] and benzoic acid, products which retain none of the stimulant or reinforcing properties of the parent drugs. The transition state for this cleavage was mimicked by the stable phosphonate monoester [123] which led to a range of antibodies of which 3B9 and 15A10 were the most effective (Fig. 40) (Appendix entry 1.3) (Landry et al., 1993).

PRODRUG ACTIVATION BY CATALYTIC ANTIBODIES

Many therapeutic agents are administered in a chemically modified form to improve features such as their solubility characteristics,ease of administration and bioavailability (Bowman and Rand, 1988). Such a "prodrug" must be designed to break down in vivo to release the active drug, sometimes at a

305

CATALYTIC ANTIBODIES

,CONH-linker

Abzyme identity 3B9 15A10

Conditions

K , [121]

pH 7.7 pH 8.0

490 p~

k,,," [121]

0.11 min-l 2 2 0 ~ ~2.3 min-'

Ki [I231 <2 p~ -

"Temperature not defined.

Fig. 40 Hapten [123] was used to raise an antibody 3B9 capable of the hydrolysis of cocaine [ 1211 to the alcohol [122] thereby effecting cocaine detoxification.

particular stage of metabolism or in a particular organ. The limitation that this imposes on the choice of masking function could be overcome by the use of an antibody catalyst for unmasking the prodrug which could, in principle, be concentrated at a specified locus in the body. Such selectivity could have implications in targeted therapies. Antibody-mediated prodrug activation was first illustrated by Fujii's group using antibodies raised against phosphonate [124] to hydrolyse a prodrug of chloramphenicol [125] (Fig. 41). Antibody 6D9 was shown to operate on substrate [126] to release the antibiotic [125] with an ER of 1.8 X lo3 (Appendix entry 1.8) (Miyashita et al., 1993). Furthermore, Fujii showed unequivocally that antibody-catalysed prodrug activation is viable by demonstrating inhibition of the growth of B. subtilis by means of the ester [126] only when antibody Mab 6D9 was present in the cell culture medium. The antibody-catalysed hydrolysis was unaffected by chloramphenicol at 10 mM and thus did not suffer from product inhibition, supporting the multiple turnover effect seen in the growth inhibition assay. Campbell and co-workers also succeeded with this type of strategy by eliciting antibody 49.AG.659.12 against a phosphonate TSA [127], designed to promote release of the anti-cancer drug 5-fluorodeoxyuridine from a D-valyl ester prodrug [128] (Fig. 42) (Appendix entry 1.10) (Campbell et al., 1994). This catalyst was able to bring about inhibition of the growth of E. coli by the release of the cytotoxic agent 5-fluorodeoxyuridine in vitro. Much the most developed example of prodrug activation comes from our own laboratory. The cytotoxicity of nitrogen mustards is dependent on substitution on the nitrogen atom: electron-withdrawing substituents

306

G. BLACKBURN ETAL

~~

Abzyme identity 6D9

Conditions

K m [126]

pH 8.0; 30°C

64 FM

kcat

[126]

0.133 min-'

Ki [I241 0.06 p~

Fig. 41 The monoclonal, 6D9, raised against phosphonate 11241 catalysed the hydrolysis of one possible regio-isomer [125] of a phenacetyl ester prodrug derived from chloramphenicol [126].

Abzyme identity 49.AG.659.12

Conditions pH 8.0; 37°C

Km

[I281

218 PM

[I281

Ki [I271

0.03 min-'

0.27 KM

kcat

Fig. 42 Prodrug [128] is an acylated derivative of the anticancer drug 5-fluorodeoxyuridine. Antibody 49.AG.659.12, raised against phosphonate 11271 was found to activate the prodrug [128] in v i m , thereby inhibiting the growth of E. coli.

CATALYTIC ANTIBODIES

[I321 R 2 = 0 [I331 R 2 = E t

307

C02

11341

Fig. 43 Carbamate prodrugs [l29a,b] are targets for abzyme cleavage to release a mustard [130] of enhanced cytotoxicity.ElcB hydrolysis of aryl carbonates involves the anion [131]. TSAs [132] and [133] were used to generate antibodies to catalyse a BA,2 mechanism for hydrolysis whose kinetic behaviour was evaluated with ester [134].

deactivate and electron-releasing substituents activate a bifunctional mustard. Thus, cleavage of a carbamate ester of a phenolic mustard can enhance its cytotoxicity to establish the carbamate as a viable prodrug for cancer chemotherapy (Blakey, 1992; Blakey et al., 1995). So the target for prodrug activation is defined as an aryl carbamate whose nitrogen substituent is either an aryl [129a] or alkyl [129b] moiety. Aryl carbamates are known to cleave by an ElcB process with a high dependency on the pK, of the leaving phenol (p- = 2.5). By contrast, aryl N-methylcarbamates are hydrolysed by a BA,2 process with a much lower dependency on leaving group (p" = 0.8) (Williams and Douglas, 1972a,b). Given the electron-releasing nature of the nitrogen mustard function (a -0.5), the kinetic advantage of antibody hydrolysis via the BAc2pathway coupled to the proven ability of antibodies to stabilize tetrahedral transition states led to the formulation of TSAs 11321 and [133]. Siting the linker in the locus of the nitrogen mustard was designed (i) to minimize potential alkylation of the antibody by the mustard function and (ii) to support mechanistic investigations by variation of the p-substituent of the aryl carbamate with little or no change in K,, both features that were realized in the outcome. Both of these TSAs generated large numbers of hybridomas including many catalysts capable of carbamate hydrolysis. A mechanistic analysis of antibody DF8-D5 showed it to cleave , p-nitrophenyl carbamate [134] with k,,, = 0.3 s-', K , = 120 p ~ and k,,,/k,,,,, = 300 at 14°C (Appendix entry 4.3) (Wentworth et al., 1997). This is some tenfold more active than a carbamatase antibody generated by Schultz

-

308

G. BLACKBURN ETAL

to a p-nitrophenyl phosphonate TSA but with a similar ER (Appendix entry 4.1) (Van Vranken et al., 1994). Most significantly, variations in the p-substituent in substrates for DF8-5 hydrolysis identified a Hammett po value' of 0.526 to establish the BA,2 nature of the reaction. For the p-methoxyphenyl carbamate substrate (4 = -0.3) the apparent ER is 1.2 X lo6. Given that there is a lo8 difference in rate for the ElcB and BA,2 processes for the p-nitrophenyl carbamate [134], the data show that antibody DF8-D5 has promoted the disfavoured BAc2 process relative to the spontaneous ElcB cleavage by some 13 kcal mol-l. Lastly, it is noteworthy that DF8-D5 was raised against the phosphonate diester [133a] as hapten, which raises the possibility that it may be an unexpected product of reactive immunization (Section 8). The medical potential of such carbamatases depends on their ability to deliver sufficient cytotoxic agent to kill cells. Antibody EA11-D7, raised against TSA [133b] proved able to hydrolyse the prodrug [129b] with K , = 201 p~ and k,,, = 1.88min-' at 37°C (Appendix entry 4.2) (Wentworth et al., 1996). Ex vivo studies with this abzyme and human colonic carcinoma (LoVo) cells led to a marked reduction in cell viability relative to controls. This cytotoxic activity was reproduced exactly by the Fab derived from EA11-D7 and was fully inhibited by a stoichiometric amount of the TSA [132b]. Using , 70% of cells were killed in a 1h incubation with EA11-D7 at 0.64 p ~some prodrug [129b] and the antibody transformed a net 4.18pmol of prodrug delivering more than 2 X IC,, of the cytotoxic agent [130]. This performance is, however, well behind that of the bacterial carboxypeptidase CPG2 used by Zeneca in their ADEPT system (Bagshawe, 1990; Blakey et al., 1995), being lo3 slower than the enzyme and 4 X lo4 inferior in selectivity ratio. Nonetheless, it is the first abzyme system to show genuine medical potential and will stimulate further work in this area. CELL VIABILITY AS AN ABZYME SCREEN

A report from Benkovic describes a new method of selecting Fabs from the whole immunological repertoire in order to facilitate a metabolic process (Smiley and Benkovic, 1994). A cDNA library for antibodies was raised against hapten [137] and then expressed in a particular strain of E. coli devoid of any native orotic acid decarboxylase (OCDase) activity (Fig. 44). The bacteria were then established in a pyrimidine-free medium where only those bacteria could grow which expressed an antibody capable of providing pyrimidines essential for DNA synthesis, and hence bacterial growth. Six colonies expressing an active antibody fragment were found viable in a screen of 16 000 transformants (Appendix entry 9.2). The remarkable feature of this 5 A pla- plot would have an even flatter slope.

309

CATALYTIC ANTIBODIES

*

a:

[I351

A

Ribose-5-P

Antibody Catalysis Uracil

HN

O

i

I I.,

PRTase Y

A

-0

I

Ribose-5-P

KEY UMP PRTase - phosphoribosyl transferase ODCase - OMP decarboxylase orotidine 5'-phosphate OMP UMP uridine 5'-phosphate

\

h y A

[ 1371

DNA Biosynthesis

Fig. 44 Pathways for uridylate biosynthesis. Mutants lacking enzymes PRTase or ODCase can complete a route to UMP provided by an antibody orotate decarboxylase in conjunction with the naturally occumng uracil PRTase. Decarboxylation of orotic acid [135] is thought to proceed through the transition state [136], for which the hapten [137] was developed (Smiley and Benkovic, 1994).

system is that OCDase, which catalyses the decarboxylation of orotidylic acid to UMP (Fig. 44),is thought to be at the top end of performance by any enzyme in accelerating this decarboxylation by some lOl7-f0ld (Radzicka and Wolfenden, 1995). This example of antibody catalysis illustrates the ability of abzymes to implant cell viability in the face of a damaged or deleted gene for an essential metabolic process. The medical opportunities for applications of such catalysis are clear. There is sufficient encouragement in these examples to show that out of all the prospects for the future development of catalytic antibodies, those in the field of medicine, where selectivity in transformation of unusual substrates may be of greater importance than sheer velocity of turnover of substrate, may well rank highest. 10 Industrial potential of abzymes

In view of the tremendous interest in biocatalysis, it is not surprising that only a decade after the dCbut of catalytic antibodies a vast literature has developed

310

G. BLACKBURN E T A L

documenting them. The field of research workers is international, with groups from three continents showing activity in the area. The potential of “designer enzymes” is already becoming a reality in relation to both the chemical industry and the pharmaceutical field. Over 70 different chemical reactions, ranging from hydrolyses to carboncarbon bond-forming reactions, have been catalysed by antibodies and their application to general synthetic organic chemistry seems promising. Typical lie in the range 10 p~ to 1 mM and binding selectivity Michaelis constants (K,,,) for the TSA over the substrate is in the range 10-105-fold. It therefore appears that antibodies have fulfilled expectations that they would be capable of comparable substrate discrimination to enzymes but over a wider range of substrate types than anticipated, and especially effective when programmed to a designated substrate. The range of reactions that may be catalysed by antibodies appears to be limited only by a sufficient knowledge of the transition state for any given transformation combined with synthetic accessibility to a stable TSA. On the other hand, abzymes are generally able to accelerate reactions by at most lo7 times the rate of the spontaneous process. It has to be said that scientists at large are looking for a major step forward in antibody catalysis to achieve rate accelerations up to lo9 that would establish abzymes as a feature of synthetically useful biotransformations. At the same time, it is essential to demonstrate that product inhibition is not an obstacle to the scaled-up use of abzymes. In relation to synthesis to deliver usable amounts of product, Lerner has shown that stereoselective reactions can be performed on a gram scale, as in the enantioselective hydrolysis of a 2-benzylcyclopentenyl methyl ether to the corresponding (S)-2-benzylcyclopentanoneof high ee (Appendix entry 7.4A) (Reymond et al., 1994). In addition, Janda has described an automated method of transposing antibody-catalysed transformations of organic molecules onto the multigram scale by employment of a biphasic system. The viability of this system was demonstrated by an epoxide ring-closing antibody, 26D9, to transform 2.2 g of substrate, corresponding to a turnover of 127 molecules per catalytic site in each batch process. This proves that the abzyme does not experience inhibition by product (Appendix entry 14.1) (Shevlin et ul., 1994). It would appear that an improvement in abzyme performance of little more than two orders of magnitude is needed before catalytic antibodies can be put to work in bioreactors and participate in kilogram scale production. Lastly, two technical features of antibody production may be valuable for the future production of cheaper abzymes of commercial value. First, the use of polyclonal catalysts, primarily from sheep, has had a tough early passage but now appears to be established for a wide range to transformations (Gallacher et al., 1990, 1992; Stephens and Iverson, 1993; Tbbul et al., 1994; Basmadjian et al., 1995; Wallace and Iverson, 1996). While these catalysts may not lend themselves to detailed examination by physical organic chemistry, they have

CATALYTIC ANTIBODIES

311

the potential to deliver catalysis at a much lower cost. Secondly, as science becomes “greener” and animal experimentation is more tightly regulated, approaches to screening antibodies with in vitro libraries may become a more important component of this field of work. Thomas has made a beginning with in v i m immunization and shown that useful catalysis can be identified (Stahl er al., 1995). While there are some limitations in this system, notably the relatively low substrate affinity of antibodies generated in this way, it is capable of refinement and may become a useful component of future abzyme selection systems. 11 Conclusions

On an evolutionary time scale, abzyme research is just reaching adolescence (Thomas, 1996), yet already over 80 different antibody-catalysed chemical reactions have been catalogued during its first decade of life. The details uncovered concerning the mechanisms of abzyme-catalysed reactions have been richer than expected. The diversity of “designer” catalysts has been explored, with the potential impact on the field of medicine and the production of fine chemicals being implicated. However, the immaturity of antibody catalysis has been exposed by its inefficiency, which, in spite of intense research efforts to improve all aspects of abzyme generation, continues to hinder wide-scale acknowledgement of its contribution to biocatalysis, particularly from under the shadow of powerful enzymes. There now exists sufficient literature about catalytic antibodies, not only in terms of their kinetic behaviour (Appendix) but also through structural information derived from X-ray crystallographic data (Golinelli-Pimpaneau et al., 1994; Haynes et al., 1994; Zhou et al., 1994) and 3-D modelling of protein sequences (Roberts er al., 1994), that it has become possible to speculate on a more general basis concerning the scope, limitations and realistic future of the field (Stewart and Benkovic, 1995; Kirby, 1996). In terms of transition state stabilization, catalytic antibodies have been shown to recognize features of the putative transition state structure encoded by their haptens with affinity constants in the nanomolar region, whereas it has been estimated that enzymes can achieve transition state complementarity with association M to deliver rate accelerations of up to lOl7-f0ld constants of the order of (Radzicka and Wolfenden, 1995). The whole subject of binding energy and catalysis has been authoritatively and critically reviewed by Mader and Bartlett (1997), with especial focus on the relationship between transition state analogues and catalytic antibodies. Enzymes have evolved to interact with every species along the reaction pathways that they catalyse, whereas our manipulation of the immune system is still relatively simplistic, using a single hapten to stimulate a full, often multistep, reaction sequence of catalysis. The serendipity that may be involved in the isolation of an efficient antibody

312

G.BLACKBURN ETAL

catalyst is now well appreciated, while recent studies have shown that non-specific binding proteins such as BSA may display catalysis approaching the level of abzymes, albeit without any substrate selectivity (Hollfelder et al., 1996). All of this serves to emphasize the fact that protein recognition of discrete high-energy reaction intermediates does not necessarily translate into efficient protein catalysis. However, the improvements in hapten design and antibody generation strategies described above are being used to highlight more intricate catalytic features. Charged and nucleophilic active-site residues (Zhong et al., 1997), substrate distortion (Datta et al., 1996; Yli-Kauhaluoma et al., 1996), desolvation and proximity effects have all now been identified as components of antibody-mediated catalysis. Using structural information available for an ever-increasing number of catalytic antibodies, manipulation of the antibody combining site is now attainable using procedures such as chemical modification (Pollack and Schultz, 1989; Schultz, 1989) and mutagenesis (Jackson et al., 1991; Stewart et al., 1994; Kast et al., 1996) to pinpoint or to improve the action of abzymes. The semi-rational design of antibody catalysts using a combination of such techniques is also supporting the systematic dissection of these primitive protein catalytic systems so as to provide valuable information concerning the origin and significance of catalytic mechanisms employed in enzymes. If the ultimate worth of antibody catalysts is to be more than academic, then the key must be found in their programmability. Here, the capabilities of abzymes such as their promotion of disfavoured processes and selectivity for substrates and transformations for which there are no known enzymes may offer prospects more significant than the further chasing after enzyme performance. After all, the tortoise has a stable ecological niche!

Appendix. Catalogue of antibody-catalysed processes For key to references via entry numbers, see p. 382 HYDROLYTIC AND DISSOCIATIVE PROCESSES

1. Aliphatic ester hydrolysis

I

Reactiodconditions

Haptenlcomments/Ki

4.4 x lo)

I

W

nr,not reported.

I

Reactiodconditions

KnJ

CLM

Hapten/comments/Ki

7G12 3G2

0 OH

TSA 7G12: Ki 1.9 X lo-'

~ L M

3G2: Ki 4.7 X lo-'

~ L M

1.3 X 10'

7.0 X lo-* 3.7 x lo?

5.4

3.3 x 10-2

1.7 X lb

w

G 8 k -

Cocaine

I

3B9.pH 7.7 15A10, pH 8.0 plyclonal

$--.

+

BzOH

- u

15A10

4.9 x 16

1.1 x 10-1

5.4 x 1 6

2.2 x 16

2.3

2.3 X 104

R , = ( a-Id,NHCO(C H,hCO$. R, = Me, R, = H

TSA 3B9 Ki 2.0 PM

nr

N

N

Vaccine immunogen Polyclonal R, =Me,R,= Me,R3=NH-IR

R, = M e , R 2 = D T , R , = H R , =DT,R,= Me,R3=H

L4

2.9 X 16

+ TSA +

nr, not reported.

H

K i 4.0 PM

2.0

N

316

g

I

2,' I

X

s1

2 ?

- +

N

0 +

X

z

G. BLACKBURN E T A L

CATALYTIC ANTI BOD1ES

I 2

I

3

317

318

Yz

I

I

X

El

2

I

z

G. BLACKBURN ETAL

b 2

CATALYTIC ANTIBODIES 319

320

g

i 8

U

D

8 I

2

2

I?

G. BLACKBURN ETAL

I

2R, 3R

3.9 x 1 6

8.8 x 10-1

nr

2s, 3 s

4.0 X 1 6

9.1 X lo-'

nr

3s

4.1 X 1 6

9.4 x 10-1

nr

3R

3.8 X 1 6

8.6 X lo-'

nr

Ll2

Me Me F IgG, pH 7.3.25'C

TSA 1.13

pH 8.0 4OC Kinetic resolution 30°C Kinetics

80% ee

1.3 x 103

TSA

Ki 2.8 )*-M 40% ee

nr, not reported.

2.0

2.4 X 1 6

w

N N

Reactiodconditions

d%

H

pH 9.0 21oc

2H6 2H6-I 21H3 21H3-I

Rester to R-alcohol Rester to R-alcohol S e t 0 S-alcohol Sester to S-alc~hol

kcatl

min-1

HaptedcommentslK,

L14 2H6-I 2H6

$

T

k

H

4.0 X lb 2.2 x

21H3

\

H

Entry

21H3-1

4.6

8.3 X 104

lb 4.0

7.2 X 104

3.9 x 102

9.0 X lo-'

2.0 x 102 6.0 X lo-'

1.6 X

lb

1.1x 103

I = immobilized antibody

TSA 2H6: K i 2.0 p . ~ 21H3: Ki 1.9 X lo-' p . ~

0 m

Fi

x m

c 50

z

3 ft

CATALYTIC ANTIBODIES

13

N

X

B o! 4

x

x

5 % " 2

N

ccp

324

x

x

3 2 I

N

x

x f

x

$I

I

-

22:

x

I

N

2

X

I

N

2 2

P +

2

m

-2 t __c

I

G. BLACKBURN ETAL

m

e

X

El

\o

09

2

d

c!

X

8a

CATALYTIC ANTIBODIES

a ri

z

*o

'a

2 Q 6

326

I

‘

5

0 . /

P

3”

G.BLACKBURN FTAL

vi

CATALYTIC ANTIBODIES

x

x

2 3 5 x

3 2 3 2

X

s s

-8 L

0,

H

Reactiodconditions

RUIQ 0

0

e

HaptedcommentslK,

H4-o

YH

30C6: Ki 8.3 X 10' p~

(zinc 84A3

kcad - 1

1.1 x 103

5.0 x 10-3 (aPP.)

3.5

2.1

2.4 X 102

2.0 x 10-3 (aPP-)

dependent)

NH

:21A6: K, 6.0 FM

Bait and Switch (BS)

0

2.1

3

3

4.9

0

2.5

P=!

I 4Nn wq

m

KD2-260, pH 6.0, 2OoC 7K16.2, pH 7.5.30"C +

0

0

m v)

AcOH

7K16.2

OH

H

3.7 x lc?

7.2 X lo-'

5.3 x 10'

1.5 X lc? (estimate based on pH rate

d

7K16.2: Ki 1.4 X lo2 p.M

TS A -

NPN43C9

2.7 x 104

2.8

profile) NPN43C9, pH 9.3.25"C Fab- 1D, pH 7.2

""Q

OH

nr, not reported.

Fab-1D

1.1 x 1 6

2.5 X 10'

N

W

hJ (D

330

a

o! rl

f

p I

t j +

I

&

4 ri

X

5 2

I

"t,

G. BLACKBURN ETAL

i

E N

X f

s zi2 ..

9

00

m

I

-i'

0

2.11

H6-32

8.5 X 1 6

7.1 X lo-'

2.4 X 1 b

3

5 0

D

5

0 2

W

0

0 rn

H6-32: K i 3.6 X 10' p~ 0 2

H

I TQJJm2H H6-32, pH 7.8,25"C H5-38, pH 7.8.25"C H7-59, pH 7.8.25OC

v)

+

OM

e3

H5-38

8.7 X 1 6

1.o

3.3 x 103

H7-59

1.1 x lb

4.9 x 10-1

1.6 X lb

0 2

H

+

0 2

go"

H5-38: Ki 5.0 p~

-+

0

0 2

0

H7-59 Ki 2.3 X lo2 p~ nr, not reported.

w

2

332

1

z

I

i

sX

s

G. BLACKBURN ETAL

2.14

SQ Me

H

3.0 X l@ 3.1 X 10'

pH 8.0

6.7 X lb

SQ Me

Reactive Immunization (RI)

2.15 Semisynthetic antibodies pH 7.0 10°C

c5

MOPC315

Nucleophilic thiol groups were introduced into a 2,4-dinitrophenyl ligand specific antibody binding site by chemical modification

K i(DNP-Gly) 2.5 X lo-'

nr, not reported.

FM

1.2

8.7 x 10-1

6.0 X 10''

334

I

sX 4

c

"

x

I

x

ElEl

f $ Z

5

G +

0"

+ =

s"

G. BLACKBURN ETAL

X

2 1 N

\ /+

-2

-v+

b

5 do p d

0"

$3

v,

Y

X

5

CATALYTIC ANTIBODIES

72 4 X

2

0

\ /

63

4

-

c.

0

h

f rn

\ /

63 0"

I

+P

4 p 0"

Q

L,

335

a

O=(z'

N

d

U

Q

p

G. BLACKBURN ETAL

2 X

2

- +

z

p

x

i"

p 0"

U

9

i

2

CATALYTIC ANTIBODIES

17 I W

sX

2 N

X

2

3

337

338

Y,

j.

?

2 X

N

0

2 X v!

N

n.

G. BLACKBURN ETAL

ip 6 'ZI

CATALYTIC ANTIBODIES

2 h

f

Reactiodconditions

VIP

1

+

Entry 5.6

Autoantibodies

Gln'g#W'7

VIP(1-16)

HaptenlcommentslK,

Fab pH 8.5 38OC VIp(17-28)

Human serum IgG fraction was found to hydrolyse vasoactive intestinal polypeptide (W). Unknown immunogen

3.8 X

1.6 X 10'

Autoantibodies

t Tg-fragments

Human serum IgG fraction was found to hydrolyse thyroglobulin (Tg). Unknown immunogen

5.7

3.9 x 10-2 3.9 x 10-3

N

~~

Autoantibodies Bw-EAR-MCA

I

Boc-EAR

BJP-B6

+

MCA

Bence Jones proteins (BJPs) (monoclonal antibody light chains) isolated from the urine of multiple myeloma patients, were found to hydrolyse peptide methylcoumarin amide peptide-MCA substrates

1.5 x 10'

3.3 x 10-2

N

5.8

CATALYTIC ANTIBODIES

I

a”

.to

341

342

3

G. BLACKBURN ETAL

+ 0

ad-9 b

X

x

'?:

CATALYTIC ANTIBODIES

I

I

a,

i

X

z I

N

X

2

2

z

+ I

d

m

*

S

3:

I

5

sX *V

9

E Reactiodconditions

Haptedcomments/Ki

Entry

6.5

I

Plasmid DNA (pUC18) Fab fragment froman IgG purified from human sera pH 7.5, 3OoC

Autoantibodies

Human serum IgG fraction (Fab) was found to hydrolyse DNA. Unknown immunogen

4.3 x 10'

1.4 x 10'

nr

Nicked DNA

6.6

38E1 pH 9.0

1.6 X 1 6

1.2 x 10-3

8.0 X lb

n

m

k

7c

-

m

c n Z

3

e

CATALYTIC ANTIBODIES

N

r-'

X

rl

'g

x

345

2

e

5a U

.L

5

I

Reactiodconditions

HaptedcommentslKi

5.3 x 102 7.0 x 10-3

7.0 X 104

8

0

in vitro Chemical Selection

Ki1.5 X 10’ FM

I

A En01 Ether Hydrolysis

0

I

B Acetal Hydrolysis

%% ee

9-40

A

3.4 x ioz

5.7 x 10-3

2.5 x 103

B

1.0 x ioz

4.7 x 10-3

7.0 X 10’

C

5.0 X 10’

7.2 X

6.0 X 102

2.3 X I d

lo-’ 1.0 X lo-*

2.5 X 10’

1.5 X loe3 4.4 x 102

4.3 x 102

CATALMIC ANTIBODIES

w

d

I

I

0

347

348

3

~

m

X

2

4

i

G.BLACKBURN ETAL

1 ': X 2

CATALYTIC ANTIBODIES

2

vB

349

W

CJl

0

Reactiodwnditions

HaptedcommentdK,

Entry

Selenoxide Elimination

M

8.5

1.5

1.8 X 10-1 1.6 X 1 6

21D8

1.7 X 1 6

1.7X 10'

1.9 x 104

n

25E10

2.6 X 1 6

2.3 X 10'

2.3 X 10"

0

NO2

pH 8.0

M

9. Decarboxylations

p$ Br

1

S4H

21D8.pH 8.0, 20°C

25E10,pH 8.0.20T

F 7c

O 2 W C N+ OH

coz

Medium Effect 21D8: K i 6.8 X p . ~

25E10:Ki 2.4 X

m C 3l

z

m

-i

p . ~

b

r'

CATALYTIC ANTIBODIES

$1 $1

P

4

X 'c!

m

X

s

351

Reactiodconditions

e

HaptedmmmentslKi 1.4 X 105

kcaJ -

I

2.8 X lo-'

10. Cycloreversions Retro DieLs-Alder Reaction

10.1

1.3 X 102 &'2m2"

'& tie

+

I

Me

HNO

he

Heterologous Immunization

(with hydroxylated form) Ki(pHY.O)9.0 x lo-' (LM

7.3 x 10-2

2.3 X l@

A

6.5

B

2.8

A = R I =OH, R, = H, B = R, = NHCH ,C02Me, R, = H

R= Me cis, syn A. 15F1-3B1, pH 7.5, 2OoC,300 nm B. UD4C3.5, pH 7.5,25"C, 300 nm

R = H trans, syn

BS 15FI-Bl: K i < 1.0 p~ 0

nr, not reported.

0

UD4C3.5: Kd (fluorescence quench) 5.4 X lo-* pM

1.2 X

16

4.7 x 10-1

2.2 x 16

3.8

X

16

11. Retro aldol reactions Reactiodconditions

I

'F h - 1

Hapten/commentslKi

Retro-Aldol Condensation

# /

OH

29C5.1 pH 5.0

kcuJKn

1.3 X 102 M - ~h-' ( k , and K, not measured separately)

n

m

kx

m t

n Z

3

e

CATALYTIC ANTIBODIES

355

INTRAMOLECULAR PROCESSES

12. Isomerizations

Reactiodconditions

I

kcat1

&-1

Haptedcomments/Ki

Entry

l2.1

Peptidyl-prolyl cidtrans isomerization

1.0 x 102

6.6

2.7 X 10'

l2.2

cis-trans Isomerization

2.2 x 102

DYJlO-4 25°C

NO2

BS Ki 6.7 FM

4.8

1.5 x 104

CATALYTIC ANTIBODIES

N

sX

u ! 3

B

m I

2

X

z N

sX

2

13. Rearrangements Reactiodconditions Cope Rearrangement

H9

kcad k",,

Hapten/commentslKi

13.1

pdcoR 5.3 x 103

OH

I

AZ-28

TSA

K i3.0 X lo-* to 1.6 X lo-'

JLM

CATALYTIC ANTIBODIES

u

3

0

X

4

2

f

"*

I

3

I:

359

W

0 0

h

HaptedcommentslK,

k - d1 w.3

Peptide Bond Rearrangement and Succinimide Hydrolysis

fl

H

1.9 X l@ 7.2 x 10-3 H

RG2-23C7

8.3 x lo-'

3.6 X lo1

7.0 X lo1 Jlr

n

;

0 iT

m

c n Z

3 E

CATALYTIC ANTIBODIES

I

vi

X

2

3 X

v-l

B

r:

W

a"

a"

W Q)

N

14. Epoxide opening k d min- 1

Reactiodconditions

141

1. anti-Baldwin Ring Closure

H

g.:, -gMe

1.

3.6 X 1 6

9.0 X

lo-'

nr

2.

2.0 x 1 6

9.0 X lo-'

nr

2. Oxepane Synthesis

TSA

pH 6.6

n W

! i x W

78% ee

c R z

3 b I-

15. Cationic cyclization TXI-4C6

TX1-4C6, pH 7.0, (biphasic)

I

TMl-87D7, pH 7.0, (biphasic

2%

98 %

TM1-87D7:

90%

10 %

2.5 X 10'

2.0 x 10-2

N

5.8 X 10'

1.3 X lo-'

N

1.0 x 16

2.1 x 10-2

N

3.1 X 10'

1.0 x 10-2

nr

TSA

TX1-4C6: Ki 1.0 (LM TM1-87D7: K,1.4 (LM

TM1-87D7

C=R=H

TSA A: Ki 1.0 FM B: Ki 1.0 JLM C: Ki 1.0 (LM nr, not reported.

nr

FPhMe2

TXlK:

A = R = cis Me

2.3 x 16 2.0 x 10-2

15.1

-

364

b

Y

8

3 X

2 N

X

N I

c1

2

X

1

2

i

G. BLACKBURN ETAL

B

16. Aldol reactions

NHC a ink \Ar

1.4x 1 0 - ~ 2.0 x 105

A

3.6 X 10'

B

4.7 x 16 4.9 x 1 0 - ~ 3.6 X 10"

16.1

0

3 3 0

78H6 pH 7.5

9

Aldol and Disfavoured elimination

0

5W P rn

cn B

Ho'

BS BIMOLECULAR ASSOCIATIVE AND SUBSTITUTION PROCESSES 16. Aldol reactions continued

I I

Reactiodconditions

HaptedcommentslK,

Entry

Aldol and Retroaldol Reaction with a Range of Aldehydes and Ketones

16.2

&+

38C2 Aldol

1.7 X 10'

6.7 x 10-3

2.9 x 104

Retro-aldol

5.4 x 10'

4.4x 10-3

nr

Reactive Immunization W

% nr, not reported.

366

3

3

E

ti

G.BLACKBURN ETAL

CATALYTIC ANTIBODIES

g s

B

-

0 2 "

+

367

368

a

3

8

El

x "! *

5 X

5

c

2

L

1 +

8

G. BLACKBURN ETAL

CATALYTIC ANTIBODIES

m 1

2

X

2

8 X

X

2

m I

2

2

I

X

0

x

P

m I

X

2

c!

?4

369

18. Acyl transfer reactions Reactiodwnditions

HaptedwmmentslK,

Amide Formation

UT

Entry 2.2 x 103 (aPP-1

18.1 1.1x 10’ (app.1

NHCOCH 3

H PhOH

+

la2

B PH

I

24B11 pH 7.0

H7

0

+

0 OH

a w

Lactone

4.9 x 103

Aniline

1.2 x 103

n W

HDZ

TSA Ki7.5 X lo-’

1.6 X 10’

! i x W

p~

c a

z

- 3 b

CATALYTIC ANTIBODIES

X

5 9 3

N I

X

3

2

0

e

'T U

371

372

U I

,?

0

G. BLACKBURN ETAL

CATALYTIC ANTIBODIES

s

rl

I

X

El

r!

N

z

t

h

-

I

0"

x %

+ a

373

0

@ a

8

0

CI

d

19. Animation reactions Reactiodconditions Oxime Formation

20A2F6 Ketone ([NHZOH] 20 IILM)

+

19.1

2.7 X lb (app.1

P

0 2

20AFzF6, pH 7.3.25T 1 4 3 ~ ~ ~ 6.5, 1 2wc ; ~

0 2

43D43Dl2 Ketone ([NHZOH] 20 mM)

9.4 x 16 (aPP-)

I

ZOAFzR? syn : mri 9:1 43D4-3D12 syn : anti 1:9

D:-m2H

0 2

TSA l9.2

Aldimine Formation

3.9 x lb 1.6 X 16

1.5 x 101

CATALYTIC ANTIBODIES

375

376 G. BLACKBURN ETAL

0

3

"&F

Conjugate Addirion

Enone

20.2

6.4 X 10' 2.1 x 10-2

3.0 X lo-'

1.4 X 1 6

TSA

-

Ki 1.1x 10' FM CN

Porphyrin Metalation Porphyrin +

M%

7G12-AlO-GI-Al2 pH 8.0 Porphyrin M"

nr, not reported.

Zn2+

4.9 x 10'

5.2 x 10-~2.6 X lC?

cu 2+

5.0 x 10'

8.4 x 10-5

1.7 X 103

20.3

378 3

d

Q 6

X

2

N

d

I

a@a

I

'

?2, I

dc

G. BLACKBURN ETAL

I

I

213

1

8.4 x 10-4

a+H,CN

20B11 pH 6.6

66% A y ee

Linke 9

r

bPP.> A

r

TSA

214 2.4 x 104

Metal cofactor complex nr, not reported.

4.0 X 1 6

2.4 X 10'

22. Reductions Reactiodconditions

HaptedcommentslK,

kcat1

h

-

1

Entry 22.1

1.0 x 10-1 (aPP-1

TSA Kd6.1X 10-1p . ~ 22.2

nr Safranine T [O]

Cofactor complex Safranine T [R]

Kd8.0 x

p.M

-

223

Sulfite

3.0 x 103

6.0 X

a0n

Ho

lo-'

0

Ho

0

22.4

R=Et R = isopropyl R=Bn

R = Et (50 mM NaBH,CN)

5.2 X 10' (app.1

NaBH3CN (0.15 mM R = Et)

5.7 x 104 bPPJ

NaCNEiH 37839.3 pH 5.0 22°C

TSA

Kd3.3 X lo-' p~ nr, not reported.

-

382

G. BLACKBURN ETAL

HYDROLYTIC AND DISSOCIATIVE PROCESSES

1. Aliphatic ester hydrolysis 1.1 (Shen et al., 1992); 1.2 (Tanaka et al., 1996); 1.3 3B9 (Landry et al., 1993), 15A10 (Yang et al., 1996), polyclonal (Basmadjian et al., 1995); 1.4 (Nakatani et al., 1993); 1.5 (Ikeda et al., 1991); 1.6 (Fujii et al., 1991); 1.7 (Iwabuchi et al., 1994); 1.8 (Miyashita et al., 1993); 1.9 (Kitazume et al., 1994); 1.10 (Campbell et al., 1994); 1.11 (Kitazume et al., 1991a); 1.12 (Kitazume et al., 1991b); 1.13 (Ikeda and Achiwa, 1997); 1.14 Ab (Janda et al., 1989), Ab-I (Janda et al., 1990a); 1.15 (Pollack et al., 1989); 1.16 A (Tawfik et al., 1993), B (Tawfik et al., 1997); 1.17 (Janda et al., 1994); 1.18 (Li et al., 1995b); 1.19 (Izadyar et al., 1993).

2. Aryl ester hydrolysis 2.1 (Suga et al., 1994b); 2.2 (Tawfik et al., 1990); 2.3 pH 8.7 (Guo et al., 1994), pH 9.5 (Zhou et al., 1994); 2.4 (Khalaf et al., 1992);2.5 A (Martin et al., 1991), B (Durfor et at., 1988);2.6 30C6 (Janda ef at., 1990b), 84A3 (Wade etal., 1993), 27A6 (Janda et al., 1991~); 2.7 KD2-260 (Ohkubo et al., 1993), 7K16.2 (Shokat et al., 1990); 2.8 NPN43C9 (Gibbs et al., 1992a),Fab-1D (Chen et al., 1993);2.9 (namontano et al., 1986); 2.10 (Tramontano et al., 1988); 2.11 (Suga et al., 1994a);2.12 (Janda et al., 1991b); 2.13 (Napper et al., 1987);2.14 (Winching et al., 1995); 2.15 (Pollack et al., 1988). 3. Carbonate hydrolysis 3.1 A (Shokat et al., 1990), B (Jacobs et al., 1987), C (Spitznagel et al., 1993); 3.2 (Pollack et al., 1986); 3.3 PCA270-29 (Gallacher et al., 1991), N-CAT 2-6 (Stahl et al., 1995); 3.4 (Wallace and Iverson, 1996). 4. Carbamate ester hydrolysis 4.1 (Van Vranken et al., 1994);4.2 (Wentworth et al., 1996);4.3 (Wentworth et al., 1997).

5. A m i d e hydrolysis 5.1 (Martin et al., 1994); 5.2 (Iverson and Lerner, 1989); 5.3 (Janda et al., 1988b); 5.4 (Benedetti et al., 1996); 5.5 (Gallacher et al., 1992); 5.6 (Paul et al., 1989); 5.7 (Li et al., 1995a);5.8 (Paul et al., 1995). 6. Phosphate ester hydrolysis 6.1 (Rosenblum et al., 1995); 6.2 15C5 (Lavey and Janda, 1996a), 3H5 (Lavey and Janda, 1996b); 6.3 (Brimfield et al., 1993); 6.4 (Weiner et al., 1997); 6.5 (Shuster et al., 1992; Gololobov et al., 1995); 6.6 (Scanlan et al., 1991).

CATALYTIC ANTIBODIES

383

7. Miscellaneous hydrolyses 7.1 37C4 (Iverson et al., 1990), polyclonal (Stephens and Iverson, 1993); 7.2 (Yu et al., 1994); 7.3 (Janda et al., 1997); 7.4 A (Reymond et al., 1992, 1993, 1994), B (Reymond etal., 1991), C (Shabat et al., 1995), D (Sinha etal., 1993b), E (Sinha et al., 1993a). 8. Eliminations

8.1 (Cravatt et al., 1994); 8.2 43D4-3D21 (Shokat et al., 1989), 20A2F6 (Uno and Schultz, 1992); 8.3 (Yoon et af., 1996); 8.4 (Thorn et al., 1995); 8.5 (Zhou et al., 1997).

9. Decarboxylations 9.1 21D8 (Lewis et al., 1991), 25E10 (Tarasow et al., 1994); 9.2 (Smiley and Benkovic, 1994); 9.3 (Bjornestedt et al., 1996); 9.4 (Ashley et al., 1993). 10. Cycloreversions 10.1 (Bahr et al., 1996); 10.2 A (Cochran et al., 1988), B (Jacobsen et al., 1995). 11. Retro-aldol reactions

11.1 (Flanagan et al., 1996); 11.2 (Reymond, 1995). INTRAMOLECULAR PROCESSES

12. Isomerizations 12.1 (Yli-Kauhaluoma et al., 1996); 12.2 (Jackson and Schultz, 1991); 12.3 (Khettal et al., 1994); 12.4 (Uno et al., 1996). 13. Rearrangements 13.1 (Braisted and Schultz, 1994; Ulrich et al., 1996); 13.2 1F7 (Hilvert et al., 1988; Hilvert and Nared, 1988), llF1-2Ell (Jackson et al., 1988); 13.3 2B4 (Gibbs et al., 1992b), RG2-23C7 (Liotta et al., 1993); 13.4 (Chen et al., 1994); 13.5 (Willner et al., 1994). 14. Epoxide opening 14.1 1 (Janda et al., 1993; Shevlin et al., 1994), 2 (Janda et al., 1995).

384

G. BLACKBURN E T A t

15. Cationic cyclization 15.1 TX1-4C6 (Li etal., 1994),TMI 87D7 (Li et al., 199%); 15.2 (Li et al., 1996); 15.3 (Hasserodt et al., 1996); 15.4 (Hasserodt et al., 1997). 16. Aldol reactions 16.1 (Koch et al., 1995). BIMOLECULAR ASSOCIATIVE AND SUBSTITUTION REACTIONS

16. Aldol reactions continued 16.2 (Wagner er al., 1995); 16.3 A (Reymond and Chen, 1995a), B (Reymond and Chen, 1995b; Zhong et al., 1997). 17. Diels-Alder cycloaddition 17.1 (Hilvert et al., 1989); 17.2 (Braisted and Schultz, 1990); 17.3 (Suckling et al., 1993); 17.4 rrans (Meekel et al., 1995), cis (Resmini et al., 1996); 17.5 7D4 and 22C8 (Gouverneur et al., 1993), 4D5 and 13G5 (Yli-Kauhaluoma et al., 1995). 18. Acyl transfer reactions 18.1 (Janda et al., 1988a); 18.2 (Benkovic et al., 1988); 18.3 (Jacobsen and Schultz, 1994); 18.4 (Hirschmann et al., 1994; Smithrud et al., 1997); 18.5 (Jacobsen et al., 1992); 18.6 A (Wirsching et al., 1991), B (Ashley and Janda, 1992); 18.7 (Gramatikova and Christen, 1996). 19. Amination reactions 19.1 (Uno et al., 1994); 19.2 (lbbul et al., 1994); 19.3 (Cochran et al., 1991). 20. Miscellaneous 20.1 (Li et al., 1995d); 20.2 (Cook et al., 1995); 20.3 (Cochran and Schultz, 1990a); for a second example see Kawamura-Konishi et al. (1996). REDOX REACTIONS

21. Oxidations 21.1 (Hsieh et al., 1994); 21.2 (Keinan er al., 1990); 21.3 (Koch et al., 1994);21.4 (Cochran and Schultz, 1990b).

CATALYTIC ANTIBODIES

22.

385

Reductions

22.1 (Nakayama and Schultz, 1992); 22.2 (Shokat et al., 1988); 22.3 (Janjic and Tramontano, 1989); 22.4 (Hsieh et al., 1993).

References Addadi, L., Jaffi, E. K. and Knowles, J. R. (1983). Biochemistry 22,4494 Alberg, D. G., Lauhon, C. T., Nyfeler, R., Fassler, A. and Bartlett, P. A. (1992). J. Am. Chem. SOC.114,3535 Albery, W. J. (1993). Adv. Phys. Org. Chem. 28, 139 Albery, J. and Knowles, J. R. (1976). Biochemistry 15, 5631 Albery, J. and Knowles, J. R. (1977). Angew. Chem. Int. Ed. Engl. 16,285 Ashley, J. A. and Janda, K. D. (1992). J. Org. Chem. 57,6691 Ashley, J. A., Lo, C.-H. L., McElhaney, G. P., Wirsching, I? and Janda, K. D. (1993). J. Am. Chem. SOC. 115,2515 Bagshawe, K. D. (1990). Biochem. SOC. Trans. 18,750 Bahr, N., Giiller, R., Reymond, J.-L. and Lerner, R. A. (1996). 1 Am. Chem. SOC.118, 3550 Baldwin, J. E. (1976). J. Chem. SOC., Chem. Commun. 738 Bartlett, I? A. and Giangiordano, M. A. (1996). J. Org. Chem. 61, 3433 Bartlett, €? A. and Johnson, C. R. (1985). J. Am. Chem. SOC. 107,7792 Bartlett, P. A. and Lamden, L. A. (1986). Bioorg. Chem. 14, 356 Bartlett, I? A. and Marlowe, C. K. (1983). Biochemistry 22,4618 Basmadjian, G. I?, Singh, S., Sastrodjojo, B., Smith, B. T., Avor, K. S., Chang, F., Mills, S. L. and Seale, T, W. (1995). Chem. Pharm. Bull. 43, 1902 Benedetti, F., Berti, F., Colombatti, A., Ebert, C., Linda, P. and Tonizzo, F. (1996). J. Chem. SOC., Chem. Commun. 1417 Benkovic, S. J., Napper, A. D. and Lerner, R. A. (1988).Proc. Natl. Acad. Sci. USA 85, 5355 Benkovic, S. J., Adams, J. A., Borders Jr., C. L., Janda, K. D. and Lerner, R. A. (1990). Science 250,1135 Benkovic, S. J., Adams, J., Janda, K. D. and Lerner, R. A. (1991). In Catalytic Antibodies, Ciba Foundation Symposium 159. Wiley, Chichester, p. 4 Bjornestedt, R., Zhong, G., Lerner, R. A. and Barbas 111, C. F. (1996). J. Am. Chem. SOC.118, 11 720 Blakey, D. C. (1992). Acta Oncol. 31,91 Blakey, D. C., Burke, P. J., Davies, D. H., Dowell, R. I., Melton, R. G., Springer, C. J. and Wright, A. F. (1995). Biochem. SOC. Trans. 23, 1047 Blokzijl, W. and Engberts, J. B. E N. (1994). ACS Symp. Ser. 568,303 Bowman, W. C. and Rand, M. J. (1988). Textbook of Pharmacology, 2nd edn. Blackwell Scientific, Oxford Braisted, A. C. and Schultz, F! G. (1990). J. Am. Chem. SOC.112,7430 Braisted, A. C. and Schultz, I? G. (1994). J. Am. Chem. SOC.116, 2211 Brimfield, A. A., Lenz, D. E., Maxwell, D. M. and Broomfield, C. A. (1993). Chem.-Biol. Interact. 87, 95 Burton, D. R. (1990). TIBS 15,64 Campbell, D. A., Gong, B., Kochersperger, L. M., Yonkovich, S., Gallop, M. A. and Schultz, I? G. (1994). J. Am. Chem. SOC.116,2165 Charbonnier, J.-B., Carpenter, E., Gigant, B., Golinelli-Pimpaneau, B., Eshhar, Z., Green, B. S. and Knossow, M. (1995). Proc. Natl. Acad. Sci. USA 92, 11 721

386

G. BLACKBURN ETAL

Charbonnier, J.-P., Golinelli-Pimpaneau, B., Gigant, B., Tawfik, D. S., Chap, R., Schindler, D. G., Kim, S.-H., Green, B. S., Eshhar, Z. and Knossow, M. (1997). Science 275,1140 Chen, Y., Reymond, J.-L. and Lerner, R. A. (1994). Angew. Chem. Int. Ed. Engl. 33, 1607 Chen, Y.-C. J., Danon, T., Sastry, L., Mubaraki, M., Janda, K. D. and Lerner, R. A. (1993). J. Am. Chem. SOC. 115,357 Chook, Y. M., Ke, H. M. and Lipscomb, W. N. (1993). Proc. Natl. Acad. Sci. USA 90, 8600 Chook, Y. M., Gray, J. V., Ke, H. M. and Lipscomb, W. N. (1994). J. Mol. Biol. 240, 476 Ciobanu, M. and Matsumoto, K. (1997). Liebigs Ann. Chem. 623 Cochran, A. G. and Schultz, F! G. (1990a). Science 249,781 Cochran, A. G. and Schultz, l? G. (1990b). J. Am. Chem. SOC.112,9414 Cochran, A. G., Sugasawara, R. and Schultz, F! G. (1988). J. Am. Chem. SOC. 110, 7888 Cochran, A. G., Pham, T., Sugasawara, R. and Schultz, F! G. (1991). J. Am. Chem. SOC. 113, 6670 Cook, C. E., Allen, D. A., Miller, D. B. and Whisnant, C. C. (1995). J. Am. Chem. SOC. 117,7269 Crans, D. C., Felty, R. A. and Miller, M. M. (1991). J. Am. Chem. SOC.113,265 Cravatt, B. F., Ashley, J. A., Janda, K.D., Boger, D. L. and Lerner, R. A. (1994). J. Am. Chem. SOC.116,6013 Danishefsky, S. and Hershenson, E M. (1979). J. Org. Chem. 44,1180 Datta, A., Partridge, L. J. and Blackburn, G. M. (1996). European Network on Antibody Catalysis 1993-95. ECSC-EC-EAEC, Brussels Davidson, M. M. and Hillier, I. H. (1994). J. Chem. SOC.,Perkin Trans. 2 1415 Dell, C. F! (1997). Contemp. Org. Synth. 4, 87 Durfor, C. N., Bolin, R. J., Sugasawara, R. J., Massey, R. J., Jacobs, J. W. and Schultz, F! G. (1988). J. Am. Chem. SOC. 110,8713 Eschenmoser, A., Ruzicka, L., Jeger, 0. and Arigoni, D. (1955). Helv. Chim. Acta 38, 1890 Evans, M. G. and Polanyi, M. (1935). Trans. Faraday SOC.31,875 Eyring, H. (1935). Chem. Rev. 17,65 Fenniri, H., Janda, K. D. and Lerner, R. A. (1995). Proc. Natl. Acad. Sci. USA 92, 2278 Flanagan, M. E., Jacobsen, J. R., Sweet, E. and Schultz, F!G. (1996).J. Am. Chem. SOC. 118,6078 Fujii, I., Lerner, R. A. and Janda, K. D. (1991). J. Am. Chem. SOC. 1l3,8528 Fujii, I., Tanaka, F., Miyashita, H., Tanimura, R. and Kinoshita, K. (1995).J. Am. Chem. SOC.117,6199 Gallacher, G., Jackson, C. S., Topham, C. M., Searcey, M., nrner, B. C., Badman, G. T. and Brocklehurst, K. (1990). Biochem. SOC. Trans. 18,600 Gallacher, G., Jackson, C. S., Searcey, M., Badman, G. T., Goel, R., Topham, C. M., Mellor, G. W. and Brocklehurst, K. (1991). Biochem. J . 279,871 Gallacher, G., Searcey, M., Jackson, C. S. and Brocklehurst, K. (1992). Biochem. J . 284, 675 Gibbs, R. A., Posner, B. A., Hpula, D. R., Dodd, S. W., Finkelman, M. A. J., Lee, T. K., Wroble, M., Whitlow, M. and Benkovic, S. J. (1991). Proc. Natl. Acad. Sci. USA 88, 4001 Gibbs, R. A., Benkovic, P. A., Janda, K. D., Lerner, R. A. and Benkovic, S. J. (1992a). J. Am. Chem. SOC. 114,3528

CATALYTIC ANTIBODIES

387

Gibbs, R. A., Taylor, S. and Benkovic, S. J. (1992b). Science 258,803 Gigant, B., Charbonnier, J.-B., Eshhar, Z., Green, B. S. and Knossow, M. (1997). Proc. Natl. Acad. Sci. USA 94, 7857 Golinelli-Pimpaneau, B., Gigant, B., Bizebard, T., Navaza, J., Saludjian, F,! Zemel, R., Tawfik, D. S., Eshhar, Z., Green, B. S. and Knossow, M. (1994). Structure 2, 175 Gololobov, G. V., Chernova, E. A., Schourov, D. V., Smirnov, I. V., Kudelina, I. A. and Gabibov, A. G. (1995). Proc. Natl. Acad. Sci. USA 92, 254 Gouverneur, V. E., Houk, K. N., De Pascual-Teresa, B., Beno, B., Janda, K. D. and Lerner, R. A. (1993). Science 262,204 Gramatikova, S. I. and Christen, P. (1996). J. Biol. Chem. 271,30583 Grossberg, A. L. and Pressman, D. (1960). J. Am. Chem. SOC.82,5478 Guo, J., Huang, W. and Scanlan, T. S. (1994). J. Am. Chem. SOC.116,6062 Halazy, S., Berges, V., Erhard, A. and Danzin, C. (1992). Bioorg. Chem. 18, 330 Hammond, G. S. (1955). J. Am. Chem. SOC.77,334 Hasserodt, J., Janda, K. D. and Lerner, R. A. (1996). J. Am. Chem. SOC.118,11654 Hasserodt, J., Janda, K. D. and Lerner, R. A. (1997). J. Am. Chem. SOC.119,5993 Haynes, M. R., Stura, E. A., Hilvert, D. and Wilson, I. A. (1994). Science 263,646 Hendry, P. and Sargeson, A. M. (1990). Prog. Inorg. Chem. 38,201 Hilvert, D. and Nared, K. D. (1988). J. Am. Chem. SOC.110,5593 Hilvert, D., Carpenter, S. H., Nared, K. D. and Auditor, M.-T. M. (1988). Proc. Nufl. Acad. Sci. USA 85, 4953 Hilvert, D., Hill, K. W., Nared, K. D. and Auditor, M.-T. M. (1989). J. Am. Chem. SOC. 111,9261 Hirschmann, R., Smith 111, A. B., Taylor, C. M., Benkovic, P. A., Taylor, S. D., Yager, K. M., Sprengeler, F! A. and Benkovic, S. J. (1994). Science 265, 234 Hollfelder, F., Kirby, A. J. and Tawfik, D. S. (1996). Nature 353, 60 Houk, K. N., Gonzalez, J. and Li, Y. (1995). Acc. Chem. Res. 28, 81 Hsieh, L. C., Yonkovich, S., Kochersperger, L. and Schultz, I? G. (1993). Science 260, 337 Hsieh, L. C., Stephans, J. C. and Schultz, P. G. (1994). J. Am. Chem. SOC. 116, 2167 Huse, W. D., Sastry, L., Iverson, S. A., Kang, A. S., Alting-Mees, M., Burton, D. R., Benkovic, S. J. and Lerner, R. A. (1989). Science 246, 1275 Ikeda, K. and Achiwa, K. (1997). Bioorg. Med. Chem. Lett. 7,225 Ikeda, S., Weinhouse, M. I., Janda, K. D. and Lerner, R. A. (1991). J. Am. Chem. SOC. 113,7763 Iverson, B. L. and Lerner, R. A. (1989). Science 243,1184 Iverson, B. L., Cameron, K. E., Jahangiri, G. K. and Pasternak, D. S. (1990). J. Am. Chem. SOC. 112,5320 Iwabuchi, Y., Miyashita, H., Tanimura, R., Kinoshita, K., Kikuchi, M. and Fujii, I. (1994). J. Am. Chem. SOC. 116,771 Izadyar, L., Friboulet, A., Remy, M. H., Roseto, A. and Thomas, D. (1993). Proc. Natl. Acad. Sci. USA 90,8876 Jackson, D. Y. and Schultz, P. G. (1991). J. Am. Chem. SOC.113,2319 Jackson, D. Y., Jacobs, J. W., Sugasawara, R., Reich, S. H., Bartlett, P. A. and Schultz, l? G. (1988). J. Am. Chem. SOC.110,4841 Jackson, D. Y., Prudent, J. R., Baldwin, E. P. and Schultz, €! G. (1991). Proc. Natl. Acad. Sci. USA 88,58 Jacobs, J., Schultz, P. G., Sugasawara, R. and Powell, M. (1987). J. Am. Chem. SOC.109, 2174 Jacobs, J. W. (1991). Bio/Technology 9,258 Jacobsen, J. R. and Schultz, P. G. (1994). Proc. Natl. Acad. Sci. USA 91, 5888

388

G. BLACKBURN ETAL

Jacobsen, J. R., Prudent, J. R., Kochersperger, L., Yonkovich, S. and Schultz, P. G. (1992). Science 256,365 Jacobsen, J. R., Cochran, A. G., Stephans, J. C., King, D. S. and Schultz, P. G. (1995). J. Am. Chem. SOC. 117,5453 Janda, K. D., Lerner, R. A. and Ramontano, A. (1988a). J. Am. Chem. SOC. 110, 4835 Janda, K. D., Schloeder, D., Benkovic, S. J. and Lerner, R. A. (1988b). Science 241, 1188 Janda, K. D., Benkovic, S. J. and Lerner, R. A. (1989). Science 244,437 Janda, K. D., Ashley, J. A., Jones, T. M., McLeod, D. A., Schloeder, D. M. and Weinhouse, M. I. (1990a). J. Am. Chem. SOC.112,8886 Janda, K. D., Weinhouse, M. I., Schloeder, D. M., Lerner, R. A. and Benkovic, S. J. (1990b). J. Am. Chem. SOC.112,1274 Janda, K. D., Ashley, J. A., Jones, T. M., McLeod, D. A., Schloeder, D. M., Weinhouse, M. I., Lerner, R. A., Gibbs, R. A., Benkovic, F! A., Hilhorst, R. and Benkovic, S. J. (1991a). J. Am. Chem. SOC.113,291 Janda, K. D., Benkovic, S. J., McLeod, D. A., Schloeder, D. M. and Lerner, R. A. (1991b). Tetrahedron 47,2503 Janda, K. D., Weinhouse, M. I., Danon, T., Pacelli, K. A. and Schloeder, D. M. (1991~). J. Am. Chem. SOC.1l3,5427 Janda, K. D., Shevlin, C. G. and Lerner, R. A. (1993). Science 259,490 Janda, K. D., Lo, C.-H. L., Li, T., Barbas 111, C. F., Winching, P. and Lerner, R. A. (1994). Proc. Natl. Acad. Sci. USA 91, 2532 Janda, K. D., Shevlin, C. G. and Lerner, R. A. (1995). J. Am. Chem. SOC.117,2659 Janda, K. D., Lo, L.-C., Lo, C.-H. L., Sim,M.-M., Wang, R., Wong, C.-H. and Lerner, R. A. (1997). Science 275, 945 Janjic, N. and Tramontano, A. (1989). J. Am. Chem. SOC. 111,9109 Jencks, W. F! (1969). In Catalysis in Chemistry and Enzymology, McGraw-Hill, New York, p. 163 Jencks, W. F! (1975). Adv. Enzymol. Relat. Areas Mol. Biol. 43,219 Johnson, W. S. (1968). Acc. Chem. Res. 1.1 Kast, P., Asif-Ullah, M., Jiang, N. and Hilvert, D. (1996). Proc. Natl. Acad. Sci. USA 93, 5043 Kawamura-Konishi, Y., Hosomi, N., Neya, S., Sugano, S., Funasaki, N. and Swuki, H. (1996). J. Biochem. 119,857 Keinan, E., Sinha, S. C., Sinha-Bagchi, A., Benory, E., Ghozi, M. C., Eshhar, Z. and Green, B. S. (1990). Pure Appl. Chem. 62,2013 Kemp, D. S., Cox, D. D. and Paul, K. G. (1975). J. Am. Chem. SOC.97,7312 Khalat A. I., Proctor, G. R., Suckling, C. J., Bence, L. H., Irvine, J. I. and Stimson, W. H. (1992). J. Chem. SOC., Perkin Trans. 1, 1475 Khettal, B., De Lauzon, S., Desfosses, B., Ushida, S. and Marquet, A. (1994). C.R. Acad. Sci. Paris, Life Sci. 317, 381 Kirby, A. J. (1980). Adv. Phys. Org. Chem. 17,183 Kirby, A. J. (1996). Acta Chem. Scand. 50,203 Kitazume, T., Lin, J. T., Takeda, M. and Yamazaki, T. (1991a). J. Am. Chem. SOC.1l3, 2123 Kitazume, T., Lin, J. T., Yamamoto, T. and Yamazaki, T. (1991b). J. Am. Chem. SOC. 113, 8573 Kitazume, T., Tsukamoto, T. and Yoshimura, K. (1994). J. Chem. SOC., Chem. Commun. 1355 Koch, A., Reymond, J.-L. and Lerner, R. A. (1994). J. Am. Chem. SOC.116,803 Koch, T., Reymond, J.-L. and Lerner, R. A. (1995). J. Am. Chem. SOC.117,9383

CATALYTIC ANTIBODIES

389

Kohler, G. and Milstein, C. (1975). Nature 256, 495 Kohler, G. and Milstein, C. (1976). Eur. J. Zmmunol. 6, 511 Kohler, G., Howe, S. C. and Milstein, C. (1976). Eur. J. Zmmunol. 6, 292 Landry, D. W., Zhao, K., Yang, G. X.-Q., Glickman, M. and Georgiadis, M. (1993). Science 259, 1899 Lavey, B. J. and Janda, K. D. (1996a). Bioorg. Med. Chem. Lett. 6, 1523 Lavey, B. J. and Janda, K. D. (1996b).J. Org. Chem. 61, 7633 Lee, A. Y., Stewart, J. D., Clardy, J. and Ganem, B. (1995). Chem. Biol. 2, 195 Lerner, R. A. and Barbas 111, C. F. (1996). Acta Chem. Scand. 50,672 Lewis, C., Kramer, T., Robinson, S. and Hilvert, D. (1991). Science 253, 1019 Li, T. and Janda, K. D. (1995). Bioorg. Med. Chem. Len. 5,2001 Li, T., Janda, K. D., Ashley, J. A. and Lerner, R. A. (1994). Science 264,1289 Li, L., Paul, S., wtyulkova, S., Kazatchkine, M. D. and Kaveri, S. (1995a).J. Zmmunol. 3328 Li, T., Hilton, S. and Janda, IS. D. (1995b).J. Am. Chem. SOC.117,2123 Li, T., Hilton, S. and Janda, K. D. (199%). J. Am. Chem. SOC.117,3308 Li, T., Janda, K. D., Hilton, S. and Lerner, R. A. (1995d). J. Am. Chem. SOC. 117, 2367 Li, T., Janda, K. D. and Lerner, R. A. (1996). Nature 379, 326 Li, T., Lerner, R. A. and Janda, K. D. (1997). Acc. Chem. Res. 30,115 Lienhard, G. E. (1973). Science 180, 149 Liotta, L. J., Benkovic, I? A., Miller, G. P. and Benkovic, S. J. (1993). J. Am. Chem. SOC. 115,350 Liotta, L. J., Gibbs, R. A., Taylor, S. D., Benkovic, P. A. and Benkovic, S. J. (1995). J. Am. Chem. SOC.117,4129 LofAs, S. and Johnsson, B. (1990). J. Chem. SOC.,Chem. Commun. 1526 Mader, M. M. and Bartlett, P. A. (1997). Chem. Rev. 97,1281 March, J. (1992a). In Advanced Organic Chemistry. Wiley, New York, p. 839 March, J. (1992b). In Advanced Organic Chemistry. Wiley, New York, p. 217 Martin, M. T., Napper, A. D., Schultz, P. G. and Rees, A. R. (1991). Biochemistry 30, 9757 Martin, M. T., Angeles, T. S., Sugasawara, R., Aman, N. I., Napper, A. D., Darsley, M. J., Sanchez, R. I., Booth, P. and Titmas, R. C. (1994). J. Am. Chem. SOC.116, 6508 Meekel, A. A. I?, Resmini, M. and Pandit, U. K. (1995). J. Chem. Soc., Chem. Commun. 571 Miller, G. P., Posner, B. A. and Benkovic, S. J. (1997). Bioorg. Med. Chem. 5,581 Miyashita, H., Karaki, Y.,Kikuchi, M. and Fujii, I. (1993). Proc. Natl. Acad. Sci. USA 90,5337 Nakatani, T., Hiratake, J., Shinzaki, A., Umeshita, R., Suzuki, T., Nishioka, T., Nakajima, H. and Oda, J. (1993). Tetrahedron Lett. 34, 4945 Nakayama, G. R. and Schultz, F! G. (1992). J. Am. Chem. SOC.114,780 Napper, A. D., Benkovic, S. J., Tramontano, A. and Lerner, R. A. (1987). Science 237, 1041 Ohkubo, K., Urata, Y., Seri, K.-I., Ishida, H., Sagawa, T., Nakashima, T. and Imagawa, Y. (1993). Chem. Lett. 1075 Padlan, E. A., Cohen, G. H. and Davies, D. R. (1985). Ann. Znst. Pasteur Zmmunol. C136,271 Page, M. I. and Jencks, W. P. (1971). Proc. Natl. Acad. Sci. USA 68, 1678 Patten, P. A., Gray, N. S., Yang, P. L., Marks, C. B., Wedemayer, G. J., Boniface, J. J., Stevens, R. C. and Schultz, P. G. (1996). Science 271, 1086 Paul, S., Volle, D. J., Beach, C. M., Johnson, D. R., Powell, M. J. and Massey, R. J.

390

G. BLACKBURN ETAL

(1989). Science 244, 1158 Paul, S., Li, L., Kalaga, R., Wilkins-Stevens, P., Stevens, E J. and Solomon, A. (1995). J. Biol. Chem. 270,15257 Pauling, L. (1947). Chemical Achievement and Hope for the Future. Silliman Lecture. Yale University Press, Yale Pauling, L. (1948). Nature 161, 707 Playfair, J. H. L. (1992). Immunology at a Glance, 5th edn. Blackwell Scientific, Oxford Pollack, S. J. and Schultz, P. G. (1989). J. Am. Chem. SOC.111, 1929 Pollack, S. J., Jacobs, J. W. and Schultz, I? G. (1986). Science 234, 1570 Pollack, S. J., Nakayama, G. R. and Schultz, P. G. (1988). Science 242, 1038 Pollack, S. J., Hsiun, P. and Schultz, P. G. (1989). J. Am. Chem. SOC.111, 5961 Raasch, M. S. (1980). J. Org. Chem. 45,856 Radzicka, A. and Wolfenden, R. (1995). Science 267, 90 Raso, V. and Stollar, B. D. (1975). Biochemistry 14, 584 Reich, H. J. (1979). Acc. Chem. Res. 12, 22 Resmini, M., Meekel, A. A. P. and Pandit, U. K. (1996). Pure Appf. Chem. 68,2025 Reymond, J,-L. (1995). Angew. Chem. Int. Ed. Engl. 34,2285 Reymond, J.-L. and Chen, Y. (1995a). Tetrahedron Lett. 36, 2575 Reymond, J.-L. and Chen, Y. (1995b). J. Org. Chem. 60,6970 Reymond, J.-L., Janda, K. D. and Lerner, R. A. (1991). Angew. Chem. Znt. Ed. Engl. 30,1711 Reymond, J.-L., Janda, K. D. and Lerner, R. A. (1992). J. Am. Chem. SOC.114,2257 Reymond, J.-L., Jahangiri, G. K., Stoudt, C. and Lerner, R. A. (1993). J. Am. Chem. SOC.115,3909 Reymond, J.-L., Reber, J.-L. and Lerner, R. A. (1994). Angew. Chem. Int. Ed. Engl. 33,475 Roberts, V. A., Stewart, J., Benkovic, S. J. and Getzoff, E. D. (1994). J. Mol. Biol. 235, 1098 Rosenblum, J. S., Lo, L.-C., Li, T., Janda, K. D. and Lerner, R. A. (1995). Angew. Chem. Int. Ed. Engl. 34, 2275 Sakurai, M., Wirsching, P. and Janda, K. D. (1996). Bioorg. Med. Chem. Lert. 6, 1055 Sauer, J. (1966). Angew. Chem. Int. Ed. Engl. 5, 211 Scanlan, T. S., Prudent, J. R. and Schultz, P. G. (1991). J. Am. Chem. SOC.113,9397 Schultz, P. G. (1989). Acc. Chem. Res. 22, 287 Schultz, P. G. and Lerner, R. A. (1993). Acc. Chem. Res. 26,391 Sergeeva, M. V., Yomtova, V., Parkinson, A., Overgaauw, M., Pomp, R., Schots, A., Kirby, A. J. and Hilhorst, R. (1996). Isr. J. Chem. 36, 177 Shabat, D., Itzhaky, H., Reymond, J.-L. and Keinan, E. (1995). Nature 374, 143 Shen, J.-Q. (1995). PhD Thesis, Sheffield University Shen, R., Priebe, C., Patel, C., Rubo, L., Su, T., Kahn, M. and Sugasawara, R. (1992). Tetrahedron Lett. 33, 3417 Shevlin, C. G., Hilton, S. and Janda, K. D. (1994). Bioorg. Med. Chem. Lett. 4, 297 Shokat, K. M., Leumann, C. J., Sugasawara, R. and Schultz, P. G. (1988). Angew. Chem. Int. Ed. Engl. 27, 1172 Shokat, K. M., Leumann, C. J., Sugasawara, R. and Schultz, P. G. (1989). Nature 338, 269 Shokat, K. M., KO, M. K., Scanlan, T. S., Kochersperger, L., Yonkovich, S., Thaisirvongs, S. and Schultz, P. G. (1990). Angew. Chem. Znt. Ed. Engl. 29, 1296 Shuster, A. M., Gololobov, G. V., Kvashuk, 0. A,, Bogomolova, A. E., Smirnov, I. V. and Gabibov, A. G. (1992). Science 256, 665 Sinha, S. C., Keinan, E. and Reymond, J.-L. (1993a). J. Am. Chem. SOC.115, 4893

CATALYTIC ANTIBODIES

391

Sinha, S. C., Keinan, E. and Reymond, J.-L. (1993b). Proc. Natl. Acad. Sci. USA 90, 11 910 Srniley, J. A. and Benkovic, S. J. (1994). Proc. Natl. Acad. Sci. USA 91, 8319 Smithrud, D. B., Benkovic, P. A., Benkovic, S. J., Taylor, C. M., Yager, K. M., Witherington, J., Philips, B. W., Sprengeler, P. A., Smith 111, A. B. and Hirschmann, R. (1997). J. Am. Chem. SOC. 119,278 Spitznagel, T. M., Jacobs, J. W. and Clark, D. S. (1993). Enzyme Microb. Technol. 15, 916 Stahl, M., Goldie, B., Bourne, S. I? and Thomas, N. R. (1995). J. Am. Chem. SOC. 117, 5164 Stephens, D. B. and Iverson, B. L. (1993). Biochem. Biophys. Res. Commun. 192, 1439 Stewart, J. D. and Benkovic, S. J. (1995). Nature 375, 388 Stewart, J. D., Roberts, V. A., Thomas, N. R., Getzoff, E. D. and Benkovic, S. J. (1994). Biochemistry 33, 1994 Stork, G. and Burgstahler, A. W. (1955). J. Am. Chem. SOC.77,5068 Suckling, C. J., Tedford, C., Bence, L. M., Irvine, J. I. and Stimson, W. H. (1993). J. Chem. SOC.,Perkin Trans. 1 1925 Suga, H., Ersoy, O., Tsumuraya, T., Lee, J., Sinskey, A. J. and Masamune, S. (1994a). J. Am. Chem. SOC. 116,487 Suga, H., Ersoy, 0..Williams, S. F., Tsumuraya, T., Margolies, M. N., Sinskey, A. J. and Masamune, S. (1994b). J. Am. Chem. SOC. 116,6025 Summers, R. (1983). Ph.D. Thesis, Harvard University Sutton, I? A. and Buckingham, D. A. (1987). Acc. Chem. Res. 20,357 Tanaka, F., Kinoshita, K., Tanimura, R. and Fujii, I. (1996). J. Am. Chem. SOC. 118, 2332 Tarasow, T. M., Lewis, C. and Hilvert, D. (1994). J. Am. Chem. SOC. 116, 7959 Tawfik, D. S., Zemel, R. R., Arad-Yellin, R., Green, B. S. and Eshar, Z. (1990). Biochemistry 29, 9916 Tawfik, D. S., Green, B. S., Chap, R., Sela, M. and Eshhar, Z. (1993). Proc. Natl. Acad. Sci. USA 90,373 Tawfik, D. S., Lindner, A. B., Chap, R., Eshhar, Z. and Green, B. S. (1997). Eur. J. Biochem. 244,619 Teraishi, K., Saito, M., Fujii, I. and Nakamura, H. (1992). Tetrahedron Lett. 33, 7153. Thatcher, G. R. J. and Kluger, R. (1989). Adv. Phys. Org. Chem. 25,99 Thomas, N. R. (1996). Nat. Prod. Rep. 13,479 Thompson, R. C. (1973). Biochemistry l2,47 Thorn, S. N., Daniels, R. G., Auditor, M.-T. M. and Hilvert, D. (1995). Nature 373, 228 Tramontano, A., Janda, K. D. and Lerner, R. A. (1986). Science 234, 1566 Tramontano, A., Ammann, A. A. and Lerner, R. A. (1988). J. Am. Chem. SOC. 110, 2282 Tubul, A., Brun, P.,Michel, R., Gharib, B. and De Reggi, M. (1994). Tetrahedron Lett. 35, 5865 Ulrich, H. D., Driggers, E. M. G. and Schultz, P. G. (1996). Acta Chem. Scand. 50, 328 Uno, T. and Schultz, I? G. (1992). J. Am. Chem. SOC.114,6573 Uno, T., Gong, B. and Schultz, l? G. (1994). J. Am. Chem. SOC. 116,1145 Uno, T., Ku, J., Prudent, J. R., Huang, A. and Schultz, I? G. (1996). J. Am. Chem. SOC. 118,3811 Van Vranken, D. L., Panomitros, D. and Schultz, P. G. (1994). Tetrahedron Lett. 35, 3873

392

G. BLACKBURN ETAL

Wade, W. S., Ashley, J. A., Jahangiri, G. T., McElhaney, G., Janda, K. D. and Lerner, R. A. (1993). J. Am. Chem. SOC. 115,4906 Wagner, J., Lerner, R. A. and Barbas 111, C. F. (1995). Science 270, 1797 Wallace, M. B. and Iverson, B. L. (1996). J. Am. Chem. SOC.118,251 Weiner, D. P.,Wiemann, T., Wolfe, M. M., Wentworth Jr., F! and Janda, K. D. (1997). J. Am. Chem. SOC.119,4088 Wentworth Jr., P.,Datta, A., Blakey, D., Boyle, T., Partridge, L. J. and Blackburn, G. M. (19%). Proc. Natl. Acad, Sci. USA 93, 799 Wentworth Jr., F!, Datta, A., Smith, S., Marshall, A., Partridge, L. J. and Blackburn, G. M. (1997). J. Am. Chem. SOC.119,2315 Westerick, J. 0. and Wolfenden, R. (1972). J. Biol. Chem. 247,8195 Westheimer, F. H. (1968). Acc. Chem. Res. 1, 70 Williams, A. and Douglas, K. T. (1972a). J. Chem. SOC.,Perkin Trans. 2 1455 Williams, A. and Douglas, K. T. (1972b). J. Chem. SOC.,Perkin Trans. 2 2112 Willner, I., Blonder, R. and Dagan, A. (1994). J. Am. Chem. SOC. 116,3121 Winchester, B. and Fleet, G. W. (1992). Glycobiology 2,199 Winchester, B., Aldaher, S., Carpenter, N. C., Dibello, I. C., Choi, S. S., Fairbanks, A. J. and Fleet, G. W. J. (1993). Biochem. J . 290,742 Wirsching, l?, Ashley, J. A., Benkovic, S. J., Janda, K. D. and Lerner, R. A. (1991). Science 252, 680 Wirsching, F!, Ashley, J. A., Lo, C.-H. L., Janda, K. D. and Lerner, R. A. (1995). Science 270,1775 Wolfenden, R. (1969). Nature 223,704 Xue, Y. F. and Lipscomb, W. N. (1995). Proc. Natl. Acad. Sci. USA 92,10595 Yang, G., Chun, J., Arakawa-Uramoto, H., Wang, X., Gawinowicz, M. A., Zhao, K. and Landry, D. W. (1996). J. Am. Chem. Soc. 118,5881 Yli-Kauhaluoma, J. T., Ashley, J. A., Lo, C.-H., 'hcker, L., Wolfe, M. M. and Janda, K. D. (1995). J. Am. Chem. SOC.117,7041 Yli-Kauhaluoma, J. T., Ashley, J. A., Lo, C.-H. L., Coakley, J., Wirsching, I? and Janda, K. D. (1996). J. Am. Chem. Soc. 118,5496 Yoon, S. S., Oei, Y., Sweet, E. and Schultz, l? G. (1996). J. Am. Chem. SOC. 118, 11 686 Yu, J., Hsieh, L. C., Kochersperger, L., Yonkovich, S., Stephans, J. C., Gallop, M. A. and Schultz, l? G. (1994). Angew. Chem. Int. Ed. Engl. 33,339 Zewail, A. H. and Bernstein, R. B. (1988). Chem. Eng. News 66,24 Zhong, G., Hoffmann, T., Lerner, R. A., Danishefsky, S. and Barbas, C. F. (1997). J. Am. Chem. SOC.119,8131 Zhou, G. W., Guo, J., Huang, W., Fletterick, R. J. and Scanlan, T. S. (1994). Science 265, 1059 Zhou, Z . S., Jiang, N. and Hilvert, D. (1997). J. Am. Chem. SOC.119,3623

Author Index Numbers in italic refer to the pages on which references are listed at the end of each chapter Abboud, J.-L., 174,243 Abe, A., 11, 83 Achiwa, K., 382,387 Adams, G., 135,140 Adams, J. A., 283,385 Addadi, L., 268,385 Agawa, T., 37,83, 84 Ahlberg, I?, 207,231,247 Akabori, S., 6,80 Alam, I., 72,82 Alberg, D. G., 256,385 Alberti, A., 94, 132, 137 Albery, J., 255,385 Albery, W. J., 255,385 Aldaher, S., 295,392 Ali, S. F., 153, 165, 170, 171, 172, 174, 195,234,248 Allen, D. A., 384, 386 Allen, L. L., 203,244 Allouche, A., 95,137 Al-Obaidi, N., 45,80 Alting-Mees, M., 282, 387 Amadelli, R., 118, 137 Aman, N. I., 300,382,389 Amin, M., 225, 227, 228,243 Amman, D., 58,80 Ammann, A. A., 279,382,391 Ando, T., 153, 170, 173, 180, 181, 182, 183,234,243,248 Angeles, T. S., 300, 382,389 Antoni, G., 181,245 Apeloig, Y., 203, 244 Arad-Yellin, R., 382, 391 Arafa, E. A., 70, 80 Arakawa-Uramoto, H., 382,392 Arata, R., 11,83 Arigoni, D., 292,386 Arimura, M., 135,139 Ashan, M., 170,172,243 Ashley, J. A., 278, 283, 287, 289, 290, 292, 301, 312, 382, 383, 384,385,386, 388,389,392

Ashwell, M., 234,243 Asif-Ullah, M., 261, 312,388 Attridge, C. J., 210,246 Atwood, J. L., 22,83 Auditor, M.-T. M., 264, 266, 270, 272, 287, 383,384, 387,391 Avor, K. S., 310,382,385 Axelsson, B. S., 153, 167, 168, 181, 185, 188,208,234,241,243,245,246 Azzaro, M. E., 176,243 Bach, R. D., 237,243 Baciocchi, E., 120,137 Badman, G. T., 310,382,386 Badoz-Lambling,J., 6, 23, 82 Bagshawe, K. D., 308,385 Bahr, N., 383,385 Baldwin, E. P., 312,387 Baldwin, J. E., 289,385 Bancroft, E. E., 116, 117,137,141 Barbas 111, C. F., 302, 312, 382, 383, 384, 385,388,389,392 Barbe, J.-M., 58, 83 Bard, A. J., 84, 94, 95, 103, 117, 118, 129, 80,137,139 Barnes, J. A., 148, 154, 155, 157, 161, 194,243 Bartell, L. S., 146, 152, 205,243 Bartlett, P. A., 256,268,269,279,311, 383,385,387,389 Bartmess, J. E., 202, 203,244 Bartocci, C., 118, 137 Basmadjian, G. l?, 310, 382,385 Baum, M. W,, 238,246 Baumann, H., 95,137 Beach, C. M., 382,390 Beaumont, P. C., 133,140 Becker, A. R., 153,244 Beer, P. D., 2, 9, 11, 15, 19,21,27,30, 37, 39, 41, 45, 49, 50, 51, 54, 55, 58, 62, 66,70,72,77,80,81,82,84 Belevskii, V. N., 95, 102,140 393

394

Bell, A. P., 21,82 Bell, R. P., 211,213,217,243 Bence, L. H., 382,388 Bence, L. M., 384,391 Bender, M. L., 181,188,243 Benedetti, F., 262, 382,385 Benkovic, P. A,, 270,272,277,283,301, 382, 383,384,387,388,389,391 Benkovic, S. J., 261, 264, 265, 270, 272, 277,279,280,281,282,283,284,292, 301,308,309,311,312,382,383,384, 385,387,388,389,390,391,392 Bennet, A. J., 234,243 Beno, B. R., 242,243 Beno, B., 286, 287, 384,387 Benory, E., 385,388 Berg, H., 189,243 Berg, U., 189,246 Berges, V.,295,387 Bergson, G., 206,234,235,236,243,244 Berke, C. M.,203,247 Bernotas, R. C., 135,140 Bernstein, R. B., 256,392 Berti, F., 262, 382,385 Bertran, J., 174,243 Bethell, D., 44,82 Bhaskar Maiya, G., 58,83 Bierbaum, V. M., 150,244 Biernat, J. E, 6, 82 Bietti, M., 120,137 Bigeleisen, J., 144,243 Bizebard, T., 311,387 Bjornstedt, R., 383,385 Black, K. A., 242,244 Blackburn, B. J., 93, 136, 139 Blackburn, C., 9, 11,49,81 Blackburn, G. M., 289,300,307,308, 312,382,386,392 Blakey, D. C., 307,308,382,385,392 Blanchard, J. S., 214,243 Blandamer, M. J., 170, 172,243 Blokzijl, W., 270,385 Blonder, R., 383,392 Blount, H. N., 94, 95, 102, 116, 117,137, 139,140,141 Bobbitt, J. M., 103, 137 Bocchi, V., 37,82 Boger, D. L., 289,383,386 Bogomolova, A. E., 382,391 Bolin, R. J., 382,389 Bommuswamy, J., 243,243 Boniface, J. J., 263,390

AUTHOR INDEX

Booth, I?, 300,382,389 Borders, Jr., C. L., 283,385 Bordwell, F. G., 131,137 Bougoin, M., 11,82 Boujlel, K., 130,140 Bourne, S. P., 311,382,391 Bowman, W. C., 304,385 Boyd, R. J., 148,243 Boyer, M., 94,129, I37 Boyle, T., 308, 382,392 Braden, M. L., 237,243 Braisted, A. C., 270, 383, 384,385 Brandl, M., 209,245 Bray, T.M., 104,139 Brede, O., 104,114, 133,141 Breuer, E., 131,140 Bright, J. I?, 45,80 Brimtield, A. A., 382,390 Brocklehurst, K., 310,382,386 Broomfield, C. A., 382,390 Brown, H. C., 176,243 Brown, K. C., 220,243 Brown, R. S., 231,232,233,246,247 Brown, W. G., 198,246 Brownawell, M. L., 234,246 Brun, E,310,384,391 Bnozka, Z., 70,83 Buckingham, D. A., 276,391 Budac, D., 106,125,137 Buddenbaum, W. E., 144,165,243 Buettner, G. R., 129,137 Buist, G. J., 181,188,243 Burdon, J., 105,137 Burgstahler, A. W., 292,391 Burke, F! J., 307,308,385 Burlitch, J. M., 210,246 Burrows, C. J., 30,83 Burton, D. R., 254,282,386,387 Cameron, K. E., 294,383,387 Campbell, D. A., 305,382,386 Carassiti, V., 118, 137 Carloni, I?, 94,107,131,132,135,137 Carpenter, E., 263,386 Carpenter, N. C., 295,392 Carpenter, S. H., 270, 383,387 Carr, A. A., 135,140 Casnati, A., 37,82 Cato, S. J., 187,245 Cerri, V., 95,137 Cha, Y.,228,243 Chamulitrat, W., 134, 135,138

AUTHOR INDEX

Chandra, H., 95, 115, 137 Chang, F., 310, 382, 385 Chanon, M., 119,189,138,243 Chap, R., 260,263,382,386,391 Charbonnier, J.-P., 263,264,386,387 Charlot, G., 6, 23,82 Chaumette, J. L., 70, 82 Chen, G., 104,139 Chen, Y.-C. J., 277, 382,383, 384, 386, 390 Chen, Z., 15, 30, 37, 39, 41, 49, 62, 66, 70,72, 77,81,82,84 Cheng, J.-P., 131,137 Chiang, Y., 205,245 Chignell, C. F., 135, 140 Childs, W. V., 116,137 Chodkowska, A., 123,140 Choi, S. S., 295, 392 Chook, Y. M., 268,386 Christen, I?, 384,387 Christensen, L., 116, 137 Chu, S. Y. F., 22, 48, 82 Chun, J., 382,392 Church, D. F., 106,140 Ciobanu, M., 270,386 Clardy, J., 268,271,389 Clark, D. S., 382,391 Cleary, T. P., 25,83 Cleland, W. W., 144, 214,216,243 Coakley, J., 312, 383, 392 Cochran, A. G., 383,384,385,386 Cohen, G. H., 263,389 Collins, E. M., 37,82 Collins, J. B., 203,247 Colombatti, A., 262, 382,385 Comelli, E., 37, 82 Connor, H. D., 95,140 Cook, C. E., 384,386 Cook, K. S., 179,180,245 Cook, P. F., 214, 216, 243 Cooper, S. R., 36,82,84 Coulter, G. A., 95,139 Courbis, I?, 95, 102, 138 Coward, J. K., 169,247 Cox, D. D., 273,388 Cramer, C. J., 187,243 Crane, C. G., 27,81 Crans, D. C., 298,386 Cravatt, B. F., 289, 383,386 Craze, G.-A., 154,243 Crescenzi, M., 120,137 Crook, S. W., 169,181, 182,247

395

Crossland, I., 240,244 Crowe, D. B., 19,81 Crozet, M. I?, 94,129,137 Cupertino, D. C., 44,82 Cyr, M. J., 58,83 Czamik, A. W., 50,82 Daasbjerg, K. L., 98, 99,139 Dagan, A., 383,392 Daniels, M. W., 179, 180,245 Daniels, R. G., 264, 266, 383,391 Danishefsky, S., 286, 312, 384,386,392 Danks, I. P., 21,82 Danks, J. P., 11,27,81 Danon, T., 264,265,277,382,386,388 Danzin, C., 295,387 Darsley, M. J., 300, 382,389 Datta, A., 289, 300, 307, 308, 312,382, 386,392 Davidson, M. M., 268,386 Davies, A. G., 119, 237 Davies, D. H., 307, 308,385 Davies, D. R., 263,389 Davies, M. J., 112, 134,137 Davis, E. R., 116,139,141 De Lauzon, S., 383,388 De Pascual-Teresa, B., 286,287, 384,387 De Reggi, M., 310,384,391 De, G., 58, 82 Deffner, U., 103,140 DeFrees, D. J., 202,203,204,244 Delgado, M., 36, 82 Dell, C. I?, 270,386 Dent, S. W., 62, 84 D e h y , C. H., 150,244 Desfosses, B., 383,388 Devlin, 111, J. L., 203,248 Dewar, M. J. S., 187, 193, 244 Dibello, I. C., 295,392 Dickson, C. A. P., 62,81 Dietrich, B., 50, 82 Diodone, R., 16, 18,83 Djeghidjegh, N., 130,137 Dodd, S. W., 283,387 Domarev, A. N., 95, 102,140 Donnelly, M. F., 146, 246 Dougherty, G., 44,82 Dougherty, R. C., 193,244 Douglas, K. T., 307, 392 Dowd, W., 195,247 Dowell, R. I., 307, 308, 385 Drake, D. A., 205,245

396

Drew, M. G. B., 15, 19,30, 37,41, 54, 55, 58,72,81 Driggers, E. M. G., 270, 383, 391 Dubose, C. M., 112,140 Dudman, C., 21,81,82 Durfor, C. N., 382,389 Dusemund, C., 69,82 Eberson, L., 94, 95,96, 97, 98, 99, 101, 102,103,105,106,107,108,109,110, 111, 112, 113, 116, 117, 118,119, 120, 122, 123, 124, 125, 127, 128,130, 131, 132,135,137,138,139 Ebert, C., 262, 382,385 Ebine, S., 6, 80 Echegoyen, L., 2,6, 35, 36,82,83 Edison, D. H., 220, 246 Edmonds, T. E., 2,82 Ekstrom, M., 121,138 El Badre, M. C., 130,137,145 El Nasr, M. M. S., 186, 245 El-Desoky, H., 16,83 Emir, B., 130, 145 Engberts, J. B. F. N., 120,270,137,385 Erhard, A., 295,387 Eriksen, T., 122,139 Ersoy, O., 264,265,266,382,391 Eschenmoser, A., 292,386 Eshhar, Z., 260,263,264,311,382,385, 386,387,388,391 Evans, C. A., 91,125,138,139 Evans, J. C., 132,198,138,244 Evans, M. G., 255,386 Eyring, H., 255,386 Fabbi, M., 37, 82 Fahey, R. C., 146, 152,247 Fainzil'berg, A. A., 123,140 Fairbanks, A. J., 295,392 Fang, Y.-R., 164, 165, 166, 167, 168, 170, 171, 172, 183, 185, 190, 191, 234,244, 245,248 Fasella, E., 120,137 Fassler, A., 256,385 Faulkner, L. R., 84,80 Feity, R. A., 298,386 Fenniri, H., 260,386 Ferguson, G., 37,82 Finkelman, M. A. J., 283,387 Fipula, D. R., 283,387 Fischer, R. G., 105,141 Fisher, R. D., 195,247

AUTHOR INDEX

Flack, S. S., 70, 82 Flanagan, M. E., 383,386 Fleet, G. W., 295,392 Flesia, E., 94, 129,137 Fletcher, N., 62,81 Fletterick, R. J., 277,311, 382,392 Flores, C. L., 103,137 Fok, N. V., 95,137 Ford, D. A., 58,83 Forrester, A. R., 93,130,138 Forschult, S., 123,139 Fox, M. A., 117,119,138 Francl, M. M., 203,244 Frejaville, C., 95,137 Friboulet, A., 382,387 Frieden, C., 213,214,245 Friedman, D. S., 203,244 Frisone, G. J., 198,244 Fry, A., 167,169,181,182,244,247 Fu, E., 33,82 Fugihira, M., 37, 84 Fuji, I., 262, 263, 294, 300, 305, 382, 386, 387,389,391 Funasaki, N., 384, 388 Funderburk, L. H., 212,229,245 Furuta, H., 58, 83 Gabibov, A. G., 382, 391 Gadru, K., 119,138 Gale, l? A., 27, 37, 39, 41,72,77,81, 82 Gallacher, G., 310, 382,386 Gallo, R., 189,243 Gallop, M. A., 294, 295, 305, 382, 383, 386,392 Ganem, B., 268,271,389 Gamer, C. D., 33,83 Gatto, V. J., 25, 36,82, 83 Gawinowicz, M. A., 382,392 Gawlita, E., 185, 187, 188,245 Gellman, S. H., 70,83 Georgiadis, M., 304, 382,389 Getzoff, E. D., 284,311,312,390,391 Gharib, B., 310, 384,391 Ghozi, M. C., 385,388 Giangiordano, M. A., 279,385 Gibbs, R. A., 270,277,283,382,383, 387,388,389 Gierasch, L. M., 116,141 Gigant, B., 263,264,311,386,387 Gilbert, B. C., 112, 134,137 Gilbert, J. C., 95, 103, 129,137 Glad, S. S., 159, 162, 163,244

AUTHOR INDEX

Glickman, M., 304, 382, 389 Goel, R., 382,386 Gokel, G. W., 22, 35, 36,82,83 Gold, V., 198,244 Goldie, B., 311, 382, 391 Goli, D. M., 35,83 Golinelli-Pimpaneau,B., 263, 311,386, 387 Gololobov, G. V., 382,391 Gong, B., 305, 382, 384, 386,391 Gonzalez, J., 256,387 Gonzalez-Lafont, A., 148, 149,248 GonzBlez-Luque, R., 97, 119, I38 Goodin, R. D., 95, 103, 129,137 Goodnow, T. T., 22, 83 Gordon, M. E., 210,246 Goto, T., 11, 83 Goulden, A. J., 62, 66,81,84 Gourdon, A., 49,82 Gouverneur, V. E., 286,287,384,387 Govindan, C. K., 106,140 Grabowski, J., 170, 172, 175,248 Gramatikova, S. I., 384,387 Granell, J., 33, 82 Gray, C. H., 169, 247 Gray, J. V., 268, 386 Gray, N. S., 263,390 Graydon, A. R., 54,66,81,84 Greci, L., 94, 107, 131, 132, 135,137 Green, B. S., 260, 263, 264,311, 382, 385, 386,387,388,391 Green, M. L. H., 33,82 Greenstock, C. L., 133, 138 Crier, D., 203,247 Grieve, A,, 49, 62,81,84 Griller, D., 133, 139 Gronchi, G., 95,102,137,138 Gronert, S., 150,244 Grossberg, A. L., 264,387 Guilard, R., 58, 83 Gullberg, F,! 181,245 Culler, R., 383, 385 Guo, J., 277, 311, 382, 387,392 Guo, X., 234,243 Gust, D., 58,83 Gustafson, S. M., 242,244 Gustowski, D. A,, 35, 36,82,83 Gutsche, C. D., 37, 72,82, 83 Habata, Y., 6, 80 Hage, J. I?, 133,138 Haggitt, J., 49, 81

397

Hagi, A,, 134, 135,139 Hagiwara, T., 134,139 Haire, D. L., 91, 104,139 Hakke, L., 198,244 Halazy, S., 295,387 Halder, E., 58, 80 Halevi, E. A., 144, 205,244 Hall, C. D., 21, 22, 48, 81, 82 Halldin, C., 181, 245 Hamilton, E., 211, 246 Hammel, K. E., 128,136,138,139 Hammerich, O., 94, 116, 138 Hammond, G. S., 256,387 Hammond, I? J., 21, 81,82 Handoo, K. L., 119,138 Hanna, I? M., 134,135,138 Hantschmann, A., 123,140 Harris, J. M., 153,179,181,244 Harris, S. J., 37, 82 Hartman, J. R., 36,82 Hartshorn, M. F!, 102,103,112,113, 122, 123,124,125,132,138 Hartshorn, S. R., 146, 195, 217,219,244, 247 Hasserodt, J., 290,291,384,387 Hay, D. A., 135,140 Haynes, M. R., 270,271,311,387 Hazlewood, C., 51, 55,81 He, G. H., 11,83 Healy, E. F., 187,244 Heath, J. A., 37, 39,81,82 Hehre, W. J., 198, 200, 202, 203, 204, 244,247,248 Heinze, J., 16,83 Hendry, F!, 276,387 Hennig, H., 123,140 Hepburn, S. P., 93,130,138 Hershenson, F. M., 286,386 Hershey, N. D., 190,247 Hesek, D., 11,50, 51, 55, 58, 62,81,84 Hilhorst, R., 273, 283,388,390 Hilinski, E. F., 119,139 Hill, H. A. O., 136,139 Hill, J. W., 167, 244 Hill, K. W., 272,287,384,387 Hillborn, J. W., 106, 139 Hillier, I. H., 268,386 Hilton, S., 294, 310, 382, 383, 384,389, 390 Hilvert, D., 254, 261, 266, 270, 271, 272, 273, 287, 311, 312, 383, 384,387,388, 389,391,392

398

Hinton, R. D., 132, 133,139 Hiratake, 382,389 Hirschmann, R., 272, 301, 384,387, 391 Hodacova, J., 50, 51, 62, 81 Hoffman, R., 203,244 Hoffman, T., 303,304,312,384,392 Hogg, J. L., 144,244 Hollfelder, E, 312, 387 Holm, T., 211,244 Holm, T., 240,244 Holmgren, S. K., 70,83 Holtz, D., 202,244 Hosomi, N., 384,388 Houk, K. N., 242,243,244,247 Houk, K. N., 256, 286, 287,384,387 Howe, S. C., 256,389 Hsieh, L. C., 294,295,383,385,387,392 Hsiun, I?, 292,382,390 Hu, W. P., 148,149,150, 151,244 Huang, W., 277,311,382,383,387,392 Hui, J. F., 11, 82 Humphrey, J. S., Jr., 198, 199,247 Humski, H., 147, 244 Huse, W. D., 282,387 Huser, M., 58,80 Huskey, W.I?, 214,215,223,244 HussCnius, A,, 234,244 Iannone, A., 128,140 Ide, H., 135,139 Ikeda, S., 382, 387 Ikeda, T., 11,83 Imagawa, Y.,382,389 Irvine, J. I., 382, 384,388,391 Ishida, H., 382,389 Itzhaky, H., 383,390 Iversen, I? E., 131, I39 Iverson, B. L., 275,294, 310, 382, 383, 387,391,392 Iverson, S. A., 282,387 Iwabuchi, Y.,294,382,387 Iwasaki, T., 131,141 Izadyar, L., 382, 387 Jackson, C. S., 310, 382,386 Jackson, D. Y.,264, 269,312, 383,387 Jackson, S. K., 132, I38 Jacobs, J, W., 253,256,257,269,279, 280,382,383,387,388,389,390 Jacobsen, J. R., 292,301,383,384,388 Jaeger, C. D., 117, 118,139 Jaffi, E. K., 268,385

AUTHOR INDEX

Jagessar, R., 58,81 Jagow, R. H., 146,152,247 Jahangiri, G. K., 294,383,387 Jahangiri, G. T., 278,382,392 Janda, K. D., 253, 256,260,264,265, 270, 277, 278, 281,283, 286, 287, 289, 290,291,292,294, 295,297,298, 299, 301,310,312,382,383,384,385,386, 387,388,389,390,391,392 Janjic, N., 385, 388 Janzen, E. G., 91, 93, 94, 95, 102, 104, 105, 111, 112, 116, 117, 121, 123, 125, 129, 132, 133, 135, 136,137, 139, 140, 141 Jeger, O., 292,386 Jencks, W. P., 154,193,256,268,279, 244,245,388,389 Jenson, F., 159, 162, 163, 244 Jewett, J. G., 199,247 Ji, G.-Z., 131,137 Jiang, N., 261, 273, 312, 383,388, 392 Jiang, W., 191, 192, 195,244 Johnson, C. R., 268,385 Johnson, D. R., 382,390 Johnson, W. S., 290,388 Johnsson, B., 260,389 Johnston, L. J., 104, 113,141 Johnstone, R. A. W., 12,83 Jones, C. J., 45,80, 81 Jones, T. M., 283, 382,388 Jonsson, M., 122,139 Kadish, K. M., 58,83 Kahn, M., 382,390 Kai, Y.,37, 83 Kaifer, A. E., 2, 6, 22, 35, 36, 82,83 Kalaga, R., 382,390 Kaldor, S. B., 220,244 Kalvoda, R., 30, 84 Kalyanaraman, B., 127, 128, 136,138, 139 Kanehisa, N., 37,83 Kang, A. S., 282,387 Kang, C., 133,140 Kaplan, E., 146,152,177,244 Karaki, Y.,305,382,389 Karelson, M. M., 187,245 Kasai, N., 37,83 Kashyap, R. P., 37,72,83 Kast, P,, 261,312,388 Katritzky. A. R., 187,245 Katz, H. E., 50, 83

AUTHOR INDEX

Kaushal, P., 124,132,139 Kaveri, S., 382, 389 Kawamura-Konishi, Y., 384,388 Kazatchkine, M. D., 382,389 Ke, H. M., 268,386 Keefe, A . D., 11, 50, 80, 81 Keinan, E., 383,385,388,390,391 Kemp, D. S., 273,388 Kennedy, C. H., 136,140 Kersten, P. J., 128, 139 Kessik, M. A., 195,247 Khalaf, A. I., 382,388 Khettal, B., 383,388 Kibukawa, 11,83 Kiick, D. M., 241,245 Kikuchi, M., 294, 305, 382,387,389 Kilburn, J. D., 70, 82 Kim, C.-K., 148, 151, 152, 153,156,243, 248 Kim, J. K., 202,203,244 Kim, S.-H., 263,386 Kimura, E., 30,83 Kimura, K., 11,83 Kimura, T., 181, 182,243 King, D. S., 383,388 Kingston, J. E., 30, 55, 81 Kinnear, K. I., 70,80 Kinoshita, I., 8, 83 Kinoshita, K., 263,294,382,386,387,391 Kirby, A. J., 154, 272, 273, 311, 312, 243, 387,388,390 Kirino, Y., 133, 139 Kirk, T. K., 128, 136,138, 139 Kirsch, J. F., 144, 245 Kirschenbaum, L. J., 127, 140 Kitazume, T., 262, 293, 382, 388, 389 Klamt, A., 187,245 Klink, F. W., 116, 137 Klinman, J. l?, 228, 231,243,245 Kluger, R., 209, 296,245,391 Knier, B. L., 154, 245 Knight, J. W., 237, 243 Knossow, M., 263,264,311,386,387 Knowles, J. R., 255, 268, 385 Knubley, R. J., 37, 81 KO, M. K., 262,382,390 Koch, A.. 384,389 Koch, T., 385,389 Kochersperger, L., 262,292,294,295, 305,382, 383, 384, 385,386,387,388, 390,392 Kochi, J. K., 119, 139

399

Kocian, O., 77, 81 Kodama, M., 30,83 Koelling, J. G., 176, 243 Koenig, T., 211,245 Koh, H. J., 170, 172, 173, 183, 184,245 Kohler, G., 256, 260, 389 Kolpin, C. F., 166, 137 Konaka, R., 136,140 Kos, A. J., 203,246 Koseki, Y., 11, 83 Koshy, K. M., 170, 172, 195,245 Kossai, R., 130,140 Kotake, Y., 121, 133,139,140 Kraeutler, B., 117, 139 Kramer, T., 273, 383,389 Kraiicler, B., 58, 80 Krawiec, M., 72, 83 Kresge, A. J., 205,245 Ku, J., 383, 392 Kurz, J. L., 179, 180, 186, 213, 214, 245 Kvashuk, 0. A., 382,391 Kwan, T., 133,139 Lagercrantz, C., 123, 124, 132,139 Lai, Z. G., 154, 190, 191, 192, 245,248 Lamden, L. A,, 256,385 Lancaster, J. E., 105,141 Landry, D. W., 304,382,389,392 Lanfredi, A. M. M., 37,82 Langley, G. J., 70, 82 Lkngstrom, B., 153, 167, 168, 181, 185, 188,208, 234,241,243,245 Larsson, B., 110,138 Lauhon, C. T., 256,385 Lavey, B. J., 298,299,382,389 Le Noble, W. J., 189,245 Ledwith, A., 94,137 Lee, A. Y., 268, 271, 389 Lee, B. C., 170, 172, 173, 183,184,245 Lee, B.-S., 170, 172, 173, 183, 184, 245 Lee, I., 153, 163, 170, 172, 173, 183, 184, 245 Lee, J., 264, 265, 382, 391 Lee, T. K., 283,387 Leffek, K. T., 174, 175, 178,245 Lenz, D. E., 382,390 Lerner, R. A., 253,256,260, 264,265, 270, 275, 277,278, 279, 281, 282,283, 286, 287, 288, 289, 290, 291, 292, 294, 295,298, 301, 302,303,304,310,312, 382,383, 384, 385,385,386,387,388, 389,390,391,392

400

Leumann, C. J., 264,383,385,390 Levi, B. A., 202,203,244 Lewis, C., 273, 383,389,391 Lewis, D. E., 180,181,220,217,247 Lewis, E. S., 212, 229,245 Li, J.-Z., 58, 82 Li, L., 382,390 Li, T., 264, 270, 290, 291, 294, 298, 382, 384,388,389,390 Li, Y., 256, 387 Lienhard, G. E., 279,389 Lin, C.-R., 132,139 Lin, J. T., 293, 382, 388 Lin, S., 229, 230,245 Lind, J., 122, 124, 132,138,139 Linda, P.,262, 382,385 Lindemann, B., 58,80 Lindner, A. B., 382,391 Liotta, L. J., 270, 383,389 Lipscomb, W. N., 268,386,392 Liu, J. I., 136,139 Liu, W.-Z., 131,137 Llobet, A., 133,138,140 LO,C.-H. L., 260,287,295, 301, 312, 382, 383,384,385,388,392 LO,G. Y.-S., 198, 244 LO,L.-C., 260,295,298,382,383,388,390 Lockhart, J. C., 70, 80 LofAs, S., 260, 389 Lorentzon, J., 97, 119,138 Lowe, N. D., 33,83 Lowe, V. J., 33, 82 Lowry, T. H., 206,245 Lubienski, M., 21, 82 Lumry, R. W., 98,140 Lund, H., 98,99,131,139 Lund, T., 98, 99, I39 Lynch, V.,58,83 Lynn, B. C., 22,83 Maccoll, A., 165, 171, 186,245 MacCullough, J. J., 108,138 MacLean, J. W., 174,175,245 Madden, K. P., 133,139 Mader. M. M., 311,389 Madhavan, S., 185,186,188,247 Madigan, E., 37,82 Mai, V. A., 210,246 Main, B., 19,81 Makino, K., 126, 134, 135, 139 Maldotti, A., 118, 137 Malmborg, P., 181,245

AUTHOR INDEX

Marcek, J. E, 30,83 March, J., 256,286,389 Marcus, R. A., 96,139 Marder, S. R., 33,82 Margolies, M. N., 266, 382,391 Marks, C. B., 263,390 Marlowe, C. K., 279,385 Marquet, A., 383,388 Marriott, P. R., 133,139 Marshall, A., 289,307,382,392 Martin, M. T., 300, 382, 389 Masamune, S., 264,265,266, 382,391 Masnovi, J. M., 119, 139 Mason, R. P.,95,127,129,134,135,136, 138,139,140 Massey, R. J., 382,389 Masui, M., 131, 140 Matheson, A. E, 175, 178,245 Matsson, O., 153, 167, 168, 181, 185, 188, 189,205, 206, 207, 208,234, 235, 236, 241,243,244,245,246,248 Matsuda, T., 11, 83 Matsumoto, K., 270,386 Mattay, J., 119,139 Mattioli, M., 120,137 Maxwell, D. M., 382,390 McAleer, J. F., 9, 11, 27, 49, 81 McCafferty, G., 36, 82 McCay, P. B., 104,116,139 McCleverty, J. A., 45, 80, 81 McDonald, G. J., 176,243 McDonald, R., 231, 233,247 McElhaney, G., 278,382, 383,385,392 McElhill, E. A., 202,246 McIntire, G. L., 94, 95, 102, 116,140, 141 McIver, R. T., Jr., 202,203,244 McKervey, M. A,, 37,82 McKillop, A., 37, 83 MacLennan, D. J., 144,222,245 McLeod, D. A., 283,382,388 Medina, J. C., 22, 83 Meekel, A. A. P., 384,389,390 Meinert, R., 133, 140 Mel’nikov, M. J., 95, 137 Melander, L., 144,198,205,206,208, 209,216,220,234,245,246 Mellor, G. W., 382, 386 Melton, R. G., 307, 308, 385 Merchhn, M., 97,119,138 Merenyi, G., 122, 124, 132,138,139 Michel, R., 310, 384, 391

AUTHOR INDEX

Milakofsky, L., 195, 247 Miller, A. R., 189,245 Miller, D. B., 384, 386 Miller, G. P.,261, 383,389 Miller, M. M., 298,386 Mills, S. L., 310, 382,385 Milstein, C., 256,260,389 Mischke, S., 210, 246 Misik, V., 127, 140 Miyashita, H., 263, 284, 305, 382, 386, 387,389 Moet-Ner (Mautner), M., 198,246 Moffatt, J. R., 153, 244 Monta, Y., 37, 84 Moody, J. D., 58, 83 Moore, A. L., 58, 83 Moore, T. A., 58, 83 Moran, M. B., 37,82 More O'Ferrall, R. A., 193, 205,246 Moreno, S. N. J., 129,140 Morgan, C. R., 35,83 Mori, G., 37, 82 Morita, Y., 37, 83 Morris, R. A., 148,149, 150,248 Mortimer, R. J., 62, 77, 81, 84 Mossoba, M. M., 126,139 Motallebi, S., 231, 233, 247 Mottley, C., 95, 127, I40 Mousset, G., 95, 102, 130,137,138 Mubaraka, M., 277, 382,386 Mulliken, R. S., 198, 246 Murikama, A., 134, 135, 139 Murr, B. L., 146,246 Murray, C. J., 228, 243 Nadvi, N. S., 234,246 Nagorski, R. W., 231,232,233,246,247 Nagren, K., 181,245 Nakamura, H., 262, 300,391 Nakashima, T., 382, 389 Nakatani, T., 382, 389 Nakayama, G. R., 274,275,382,385, 389,390 Nakayama, H.. 382.389 Napper, A. D., 279; 300, 301,382,384, 385.389 Nared, K. D., 270,272,287, 383,384, 387 Nasr, M. M., 179, 180,245 Navaza, J., 311,387 Neta, P., 114, 126, 133, 140 Netherton, L. T., 169, 181, 182, 247

401

Neya, S., 384,388 Ng, K.-M., 119, 137 Nieduzak, T. R., 135, 140 Nilsson, M., 95, 107, 108, 109, I38 Nishi, M., 134, 135, 139 Nishi, Y., 125,139 Nishioka, T., 382,389 Nomura, E., 37,83 Notario, R., 174,243 Nutter, Jr., D. E., 116, 139 Nyberg, K., 105, 106, 116,138,139 Nyburg, S. C., 21,22,82 Nyfeler, R., 256,385 O'Dea, J. J., 6, 88, 83 O'Leary, M. H., 185, 186, 188,246,247 Occhialini, D., 98, 99, I39 Oda, J., 382,389 Oei, Y., 383,392 Oertel, U., 95, I37 0gaard Madsen, J., 211,240,244 Ogden, M. I., 19, 30, 37, 70, 81 Ohkubo, K., 382,389 Ohkuma, T., 133,139 Ohiweiler, D. F., 135,140 Okhlobystina, L. V., 123, 140 Oppenheimer, N. J., 214,216,243 Ori, A,, 75,83 Osbourne, R., 154,243 Osteryoung, J., 6, 88, 83 Overgaauw, M., 273,390 Owens, M., 37,82 Ozaki, S., 131, 140 Pacelli, K. A., 264, 265, 382, 388 Padlan, E. A., 263, 389 Page, M. I., 268,389 Paley, M. S., 179, 181,244 Palys, M., 70, 83 Pandit, U. K., 384,389, 390 Paneth, l?, 185, 186, 187, 188, 234,244, 246,247 Panomitros, D., 308,382,392 Pantano, J. E., 179, 180, 245 Park, K. H., 234,246 Parker, V. D., 94, 104, 113, 138,140, 141 Parkin, D. W., 234,246 Parkins, A. W., 21, 22, 82 Parkinson, A., 273, 390 Parsons, I. W., 105, 137 Partridge, L. J., 289, 300, 307, 308, 312, 382,386,392

402

Pascal, R. A., 210,238,246 Paschkewitz, J. S., 148, 149, 150,248 Pasternak, D. S., 294, 383,387 Patel, C., 382, 390 Patten, F! A., 263,390 Paul, K. G., 273,388 Paul, S., 382,389,390 Pauling, L., 255,390 Paulson, J. F., 148,149, 150,248 Pedersen, S. U., 98, 99, 139 Perkins, M. J., 91, 123,133,137,139,140 Persson, J., 167, 168, 170, 185, 189,208, 245,246,248 Persson, O., 102,105,108, 112,113, 122, 123,124,125,132,138 Pham, T. V., 164, 165, 166, 167, 170, 171, 172,183,197,246,248 Pham, T., 384,386 Philips, B. W., 212, 384,391 Pilgrim, A. J., 15, 30, 81 Pincock, J. A., 106, 139 Pinnick, H. R., Jr., 170, 172,247 Playfair, J. H. L., 292,390 Plenio, H., 16, 18, 83 Pochini, A., 37, 82 Poirier, R. A., 148, 155, 156, 157, 158, 159,160,161,165,168,179,191,246 Polanyi, M., 255,386 Pollack, S. J., 253, 256, 274, 275, 292, 312,382,390 Pollack, S. K., 202, 203,244 Pomp, R., 273,390 Pople, J. A., 203,244 Posner, B. A., 261,283,387,389 Pospisil, L., 30, 84 Powell, M., 382,388,390 Powers, E. L., 133,140 Poyer, J. L., 104,116,132,139,140 Prasthofer, T. W., 179, 181,244 Pressman, D., 254,387 Price, R. C., 225, 227, 228,243 Priebe, C., 382,390 Proctor, G. R., 382,388 Pross, A., 193, 246 Prudent, J. R., 292,296,312,382,383, 384,387,388,390,392 Pryor, W. A., 106,140 Queen, A., 198,244 Raasch, M. S., 270,390 Radner, F., 102,103,123,132,138

AUTHOR INDEX

Radom, L., 203,244 Radzicka, A., 289, 309, 311,390 Raimondi, L., 242,247 Rajemann, M., 189,243 Rand, M. J., 304,385 Rapp, M. W., 170,172,195,247 Raso, V,, 256,390 Rawle, S. C., 36,82 Reber, J.-L., 310, 383,390 Reddy, F! A., 37,83 Redman, C., 133,140 Rees, A. R., 382,389 Rehm, D., 103,140 Rehorek, D., 111,112,121,123,140 Reich, H. J., 273, 390 Reich, S. H., 269, 383,387 Reinhammar, B., 128,139 Reinhoudt, D. N., 70, 83 Reinke, L. A., 104,139 Reitstoen, B., 104,140 Remy, M. H., 382,387 Rentzepiz, F! M., 119,139 Reppond, K. D., 169,181,182,247 Resmini, M., 384,389,390 Reszka, K., 135,140 Reuwer, J. F., Jr., 181, 219,223,247 Reymond, J.-L., 294, 310, 383, 384, 385, 385,386,389,390,391 Reynolds, W. L., 98,140 Richardson, K. S., 206,245 Ridgway, C., 77, 81 Rieke, C. A., 198,246 Riesz, €!, 126,127,139,140 Rimland, A., 181,245 Rios, A. M., 35, 83 Roberts, B. I?, 124,132,139 Roberts, J. D., 202, 246 Roberts, V. A., 284, 311, 312, 390,391 Robertson, R. E., 170,172,195,198, 244,245 Robinson, M. J. T., 234, 246 Robinson, S., 273, 383,389 Rodgers, L. R., 238,246 Rojas, M. T., 22, 83 Romano, F. J., 220,243 Roos, B. O., 97, 119,138 Rose, M. E., 12,83 Rosenblum, J. S., 298, 382,390 Roseto, A,, 382,387 Rowlands, C. C., 132,138 Rubo, L., 382,390 Rudkevich, D. M., 70, 83

AUTHOR INDEX

Rudzinski, J., 185, 186, 188,247 Rumyantseva, G. V., 136,140 Rusterholz, B., 58,80 Ruzicka, L., 292,386 Sagawa, T., 382,389 Saito, M., 262, 300,391 Saji, T., 6, 8, 23, 83 Sakamoto, H., 11,83 Sakamoto, Y., 6,80 Sakurai, M., 297,390 Salam, S. S., 45, 80, 81 Salem, L., 203,244 Saludjian, P., 311, 387 Salvetter, J., 123,140 San Filippo, J., Jr., 234,246 Sanchez, R. I., 300, 382,389 Sandanayake, K. R. A. S., 69,82 Sang, H., 132,140 Sargeson, A. M., 276,387 Sastrodjojo, B., 310, 382, 385 Sastry, L., 277,282, 382,386,387 Sata, M., 6, 80 Satish, A. A., 131,137 Sauer, J., 270, 390 Saunders, G. C., 33,82 Saunders, W. H., Jr., 144, 165, 171, 181, 182, 193, 198, 205, 206,207, 208, 209, 216, 217, 218, 219, 220, 222, 223, 224, 225, 226, 227,228,229, 230, 234,243, 245,246,247,248 Savage, P. B., 70,83 SavCant, J.-M., 98, 140 Sawyer, D. T., 133,138,140 Sayo, H., 131,140 Scanlan, T. S., 262, 277, 296, 311, 382, 387,390,392 Schaad, L. J., 181, 219, 223,247 Schepp, N. P., 104, 113,141 Schindler, D. G., 263, 386 Schleyer, I? vR., 203,244,246 Schloeder, D. M., 264,265,281,283,382, 388 Schmidtchen, F. F!, 50,83 Schots, A., 273,390 Schowen, B. K., 169,247 Schowen, R. L., 169,214,215,223,244, 247 Schriver, G. W., 203,247 Schulthess, P., 58,80 Schultz, F? G., 253,256, 262, 263, 264, 265,269, 270, 274, 275,288,292, 294,

403 296,301, 305,308, 312, 382, 383,384, 385,385,386,387,388, 389, 390,391, 392 Schultz, R. A., 35, 83 Schuster, G. B., 125,140 Schuurmann, G., 187,245 Scott, J. M. E., 170, 172,243 Seale, T. W., 310, 382, 385 Searcey, M., 310,382,386 Seely, G. R., 58, 83 Sela, M., 260, 382,391 Seltzer, S., 211, 246 Sendijarevic, V., 147, 244 Sergeeva, M. V., 273,390 Seri, K.-I., 382, 389 Sessler, J. L., 58,83 Seyferth, D., 210,246 Sgarabotti, F,! 94, 107, 131, 132, 135,137 Shabat, D., 383,390 Shafer, S. G., 153, 244 Shaik, S. S., 98, 193,138,246 Shapiro, B. I., 123, 140 Sharpe, N. W., 21,82 Shen, J.-Q., 262, 300, 390 Shen, R., 382,390 Shetty, R. V., 94, 95, 102, 133, 140 Shevlin, C. G., 289, 310, 383, 388,390 Shi, Z., 148,243 Shine, H. J., 94,234, 137,246 Shiner, V. J., Jr., 144, 145, 146, 147, 153, 165, 170, 172, 195, 198, 199,217, 219, 243,244,247 Shinkai, S., 69, 75, 82, 83 Shinzaki, A., 382, 389 Shokat, K. M., 262, 264,265, 382, 383, 385,390,391 Shono, Y., 11, 83 Shuster, A. M., 382, 391 Sikanyika, H., 9, 11, 49, 81 Sim, M.-M., 260,295,383,388 Simon, W., 58, 80 Simonet, J., 95, 102, 130, 137, 138,140 Sims, L. B., 169, 180, 181, 182, 217,220, 247 Singh, S., 310, 382, 385 Singleton, D. A., 238, 242, 243, 247 Sinha, S. C., 383, 385,388,391 Sinha-Bagchi, A., 385,388 Sinnott, M. L., 234,243 Sinskey, A. J., 264, 265, 266, 382,391 Sjoberg, S., 206,234,235,236,243 Slae, S., 205,246

404

Slawin, A. M. Z., 11,81 Slebocka-mlk, H., 231,232,233,246,247 Smid, J., 11,82 Smiley, J. A., 308, 309, 383,391 Smirnov, I. V., 382,391 Smith 111, A. B., 272, 301, 384,387,391 Smith, B. T., 310, 382,385 Smith, D. K., 54,81 Smith, P. J., 145, 179, 180, 195,247 Smith, S., 289, 307, 382,392 Smithrud, D. B., 272, 384,391 Sohn, D. S., 170,172,173,183,184,245 Sola, M., 174,243 Soloman, A., 382,390 Sosonkin, I. M., 95, 102,140 Spencer, I?, 30,81 Spitznagel, T. M., 382, 391 Sprengeler, P. A., 272,301,384,387,391 Springer, C. J., 307, 308,385 Sridhar, R., 133,140 Sridharan, S., 170, 172, 175,248 Stahl, M., 311, 382,391 Stasicka, Z., 123, 140 Steenken, S., 120, 133,137,140 Stein, A. R., 234,236,247 Stell, J. K., 112, 134,137 Stephans, J. C., 294, 295, 383, 385,387, 388,392 Stephens, D. B., 310,383,391 Stem, M. J., 215,247,248 Stevens, F. J., 382, 390 Stevens, R. C., 263,390 Stewart, D. R., 72,83 Stewart, J. D., 268, 271, 280, 281, 284, 311,312,389,390,391 Stewart, J. J. I?, 187,244,245 Stimson, W. H., 382,384,388,391 Stipa, P., 94, 107, 131, 132, 135,137 Stivers, E. C., 181, 219, 223, 247 Stokes, S. E., 50, 51, 54, 55, 66,81 Stollar, B. D., 256,390 Stoke, K., 129,140 Storer, J. W., 242,247 Stork, G., 292,391 Stoudt, C., 383,390 Streitwieser, A., Jr., 146, 152,203, 204, 205,247 Strogov, G. N., 95, 102,140 Stronks, H. J., 94, 95, 102, 116,139, 140, 141 Stura, E. A., 270,271, 311,387 Su,T., 382,390

AUTHOR INDEX

Subramanian, Rm., 217, 218,219,224, 247 Suckling, C. J., 382, 384,388,391 Suga, H., 264,265,266,382,391 Suga, K., 37,84 Sugano, S., 384,388 Sugasawara, R., 264,269,300,382,383, 384,385,386,387,388,389,390 Sulochana Wijesundera, W. S., 234,243 Summerhays, K. D., 202,203,244 Siimmermann, W., 103,140 Summers, R., 256,391 Sunko, D. E., 198,200,203,204,244, 247 Surzur, J. M., 94, 129,137 Sutin, N., 96, 139 Sutton, I? A., 276,391 Suzuki, H., 384,388 Suzuki, S., 146, 152,247 Suzuki, T., 382,389 Svard, H., 181,245 Svensson, J. O., 102,103, 123,138 Swain, C. G., 181, 190,219,223,247 Swann, B. F?, 37,83 Sweet, E., 383,386,392 Symons, M. C. R., 95,114,115,126,137, 140 Syrkin, Ya. K., 123, 140 Szele, I., 200, 247 Szemes, F., 62,66,81,84 Szylhabel-Godala, A., 185, 186, 187, 188, 245,247 Taagepera, M., 202,203,244 Taft, R. W., 202,203,244,248 Takeda, M., 293,388 Tamm, T., 187,245 Tanabe, H., 170,173,182,183,234,243 Tanaka, F., 263, 382, 386,391 Taniguchi, H., 37, 133,83,139 Tanimura, R., 263,294, 382,386,387, 391 Tarasow, 383,391 Tawfik, D. S., 260,263,311,312,382, 386,387,391 Taylor, C. M., 272, 301, 384,387, 391 Taylor, E. C., 37,83 Taylor, S. D., 270, 301, 384,387, 389 Taylor, S., 270, 383,387 Tedford, C., 384,391 Tencer, M., 234,236,247 Terabe, S., 136,140

405

AUTHOR INDEX

Teraishi, K., 262, 300,391 Thaisirvongs, S., 262, 382,390 Thatcher, G. R. J., 296, 391 Thibblin, A., 207,231,247 %el, W., 187, 244 Thomas, A. A., 238,247 Thomas, C. E., 135, 140 Thomas, D., 382,387 Thomas, N. R., 284, 312,382,391 Thompson, R. C., 278,311,391 Thorn, S. N., 264,267,383,391 Thornalley, P. J., 136, 139 Thornton, E. R., 146, 152, 177, 193, 198, 244,247 Tilset, M., 104,140 Timpe, H. J., 95, 137 Titmas, R. C., 300, 382,389 Todo, P., 95,102,138 Tomasi, A., 128,140 Tonizzo, F., 262, 382,385 Topham, C. M., 310,382,386 Tordo, P.,94, 95,129,137,138 Tramontano, A., 253, 256, 279, 301, 382, 384,385,388,389,391 TrCmillion, B., 6, 23, 82 Truhlar, D. G., 148, 149, 150, 151, 187, 243,244,248 Tsuge, O., 131,141 Tsukamoto, T., 262, 382,389 Tsumuraya, T., 264,265,266,382,391 Tubul, A., 310, 384, 391 Tucker, J. H. R., 21, 22,82 Tucker, L., 287, 384,392 Tucker, S. C., 148, 149,248 Xddenham, M., 33,82 Turner, B. C., 310,386 Turner, P., 231, 233, 247 lhyrikov, V. A., 123, 140 Tyutyulkova, S., 382,389 Ugozzoli, F., 37,82 Ulrich, H. D., 270,383,391 Umeshita, R., 382, 389 Ungara, R., 37, 82 Uno, T.,264, 265, 383, 384, 391, 392 Urano, S., 131, 141 Urata, Y., 382, 389 Ushida, S., 383,388 Utley, J. H. P., 107, 116, 118, 138 Van Hook, W. A., 144,247 Van Sickle, D. E., 204, 205,247

Van Vranken, D. L., 308,382,392 Vaz, R., 135,140 Verboom, W., 70,83 Vicek, A. A., 30, 84 Viggiano, A. A., 148, 149, 150, 248 Vila, F., 95,137 Visser, H. C., 70,83 Vitullo, V. P., 170, 172, 175,248 Volke, J., 30,84 Volle, D. J., 382, 390 Wada, E., 180,243 Wada, F., 11,83 Wade, W. S., 278,382,392 Wagner, C. K., 238,246 Wagner, J., 384,392 Wallace, M. B., 310, 382,392 Walter, T. H. 116,141 Wan, P., 106,125,137 Wang, H., 104,141 Wang, R., 260,295, 383,388 Wang, X., 382,392 Wang, Y., 148, 155, 156, 157, 158, 159, 160,161, 165,168, 179, 191,246 Washburne, S. S., 210,246 Waszczylo, Z., 153, 170, 171, 172, 183, 248 Watkin, D. J., 36, 82 Watson, W. H., 37, 72, 83 Wayner, D. D. M., 104, 113,141 Wear, T., 62, 66, 81, 84 Webb, R. L., 202,246 Webster, M., 70, 82 Wedemayer, G. J., 263,390 Weightman, J. S., 62, 84 Weinberg, N., 148,243 Weiner, D. P., 298,382,392 Weinhouse, M. I., 264,265,283,382, 387,388 Weller, A., 103, 140 Wentworth Jr., P., 289,298,307,308, 382,392 Westaway, K. C., 144, 145, 146, 148, 153, 154, 155, 156, 157, 158, 159, 160, 161, 164,165,166, 167,168, 169,170,171, 172, 174, 179, 180, 183, 185, 190, 191, 192, 193, 195, 196, 197, 234,244,245, 246,247,248 Westerick, J. O., 278,392 Westheimer, F. H., 206, 208, 220, 298, 248,392 Whisnant, C. C., 384,386

406

Whitlow, M., 283,387 Whitwood, A. C., 112, 134,137 Wiebe, R. H., 133,138 Wiemann, T., 298,382,392 Wilczewski, T., 6, 82 Wilkins-Stevens, P., 382,390 Williams, A,, 307, 392 Williams, D. J., 11, 81 Williams, I. H., 147, 148, 154, 155, 157, 161, 194, 200,201,243,248 Williams, S. F., 266, 382,391 Willner, I., 383,392 Wilson, I. A., 270, 271, 311,387 Wilson, J. C., 169, 181, 182,247 Winchester, B., 295,392 Wirsching, P., 292,297, 301, 312, 382, 383,384,385,388,390,392 Witherington, J., 272, 384, 391 Wolf, J. F., 203, 248 Wolf, R. E. J., 36, 82,84 Wolf, R., 211,245 Wolfe, M. M., 287, 298, 382, 384, 392 Wolfe, S., 148, 151, 152, 153, 156,243, 248 Wolfenden, R., 278, 279, 298, 309, 311, 390,392 Wolfsberg, M., 144,202,203,215,243, 244,247,248 Wong, C.-H., 260, 295, 383,388 Wong, K. H., 11,82 Workentin, M. S., 104, 113,141 Wright, A. F., 307,308,385 Wright, D. R., 179,180,245 Wroble, M., 283,387 Xu, Q. Y. Y.,58,83 Xue, Y. F., 268, 392 Yagbasan, R., 36,82 Yager, K. M., 272,301,384,387,391 Yamamoto, K., 210,246 Yamamoto, T., 293,382,388

AUTHOR INDEX

Yamataka, H., 153, 170, 173, 180, 181, 182, 183,234,243,248 Yamazaki, T., 293, 382,388 Yang, G. X.-Q., 304,382,389 Yang, G., 382,392 Yang, J., 16, 83 Yang, P. L., 263,390 Yarkov, S. P., 95,102,140 Yatsunami, T., 30, 83 Yli-Kauhaluoma, J. T., 287, 312, 383, 384,392 Yomtova, V., 273,390 Yonkovich, S., 262, 292, 294, 295, 305, 382, 383, 384, 385, 386, 387,388, 390, 392 Yoo, H., 36,82 Yoon, S. S., 383,392 Yoshida, K., 94, 141 Yoshimura, K., 262, 382,389 Yu, J., 294,295, 383,392 Yu, R.-Q., 58, 82 Zemel, R., 311, 382, 387,391 Zerner, M. C., 187,245 Zewail, A. H., 256, 392 Zhang, B.-L., 238,246 Zhang, Y., 135, 234, 139,243 Zhang, Z., 131,137 Zhao, X. G., 148,149,248 Zheng, G. D., 58,82 Zheng, G., 104,141 Zhong, G., 303,304,312,383,384,385, 392 Zhou, G. W., 277,311,382,392 Zhou, K., 304,382,389,392 Zhou, Z. S., 273,382,383,392 Zoebisch, E. G., 187,244 Zompa, L., 30,84 Zubarev, V. E., 95, 104, 114, 133, 137, 141 Zweig, A., 105, 141

Cumulative Index of Authors

Ahlberg, I?, 19,223 Albery, W. J., 16, 87; 28, 139 Allinger, N. I., 13, 1 Anbar, M., 7, 115 Arnett, E. M., 13, 83; 28, 45 Ballester, M., 25, 267 Bard, A. J., 13, 155 Beer, l? D., 31, 1 Baumgarten, M., 28, 1 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 Bethell, D., 7, 153; 10, 53 Blackburn, G. M., 31, 249 Blandamer, M. J., 14,203 Bowden, K., 28,171 Brand, J. C. D., 1,365 Brandstrom, A., 15, 267 Brinkman, M. R., 10,53 Brown, H. C., 1, 35 Buncel, E., 14, 133 Bunton, C. A., 22, 213 Cabell-Whiting, P. W., 10, 129 Cacace, F., 8, 79 Capon, B., 21, 37 Carter, R. E., 10, 1 Chen, Z., 31. 1 Collins, C. J., 2, 1 Cornelisse, J., 11, 225 Crampton, M. R., 7,211 Datta, A., 31, 249 Davidson, R. S., 19, 1; 20, 191 Denham, H., 31,249 Desvergne, J. P., 15, 63 de Gunst, G. P., 11,225 de Jong, F.. 17, 279 Dosunmu, M. I., 21, 37

Eberson, K., 12, 18,79; 31, 91 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 Gale, l? A., 31, 1 Gilbert, B. C., 5,53 Gillespie, R. J., 9, 1 Gold, V.,7, 259 Goodin, J. W., 20, 191 Gould, I. R., 20, 1 Greenwood, H. H., 4,73 Hammerich, O., 20, 55 Harvey, N. G., 28,45 Hasegawa, M., 30, 117 Havinga, E., 11, 225 Henderson, R. A., 2 3 , l Henderson, S., 23, 1 Hibbert, F., 22, 113; 26, 255 Hine, J., 15, 1 Hogen-Esch, T. E., 15, 153 Hogeveen, H., 10,29,129 Huber, W., 28, 1 Ireland, J. F., 12, 131 Iwamura, H., 26, 179 Johnson, S. L., 5,237 Johnstone, R. A. W., 8, 151 Jonsall, G., 19, 223 JosC, 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., 25,99 407

Kochi, J. K., 29, 185 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 Ledwith, A., 13, 155 Lee, I., 27, 57 Le Fi?vre, R. J. W., 3, 1 Liler, M., 11, 267 Long, F. A., 1,1 Liining, U., 30,63 Maccoll, A., 3, 91 Mandolini, L., 22, 1 Matsson, O., 31, 143 McWeeny, R., 4, 73 Melander, L., 10, 1 Mile, B., 8, 1 Miller, S. I., 6, 185 Modena, G., 9,185 More O’Ferrall, R. A., 5, 331 Morsi, S. E., 15, 63 Miillen, K., 28, 1 Nefedov, 0. M., 30,l Neta, F!, 12, 223 Nibbering, N. M. M., 241 Norman, R. 0. C., 5,33 Nyberg, K., 12, 1 Okamoto, K., 30,173 Olah, G. A., 4, 305 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 Pletcher, D., 10, 155 Pross, A., 14, 69; 21, 99

408

Ramirez, F., 9. 25 Rappoport, Z., 7, 1; 27, 239 Reeves, L. W., 3,187 Reinhoudt, D. N., 17,279 Ridd, J. H., 16, 1 Riveros, J. M., 21, 197 Robertson, J. M., 1,203 Rose, P. L., 28, 45 Rosenthal, S. N., 13, 279 Ruasse, M.-F., 28,207 Russell, G. A., 23, 271 Samuel, D., 3, 123 Sanchez, M. de N. de M., 21,37 SandstrBm, J., 25, 1 SavCant, J.-M., 26, 1 Savelli, G., 22,213 Schaleger, L. L., 1, 1 Scheraga, H. A., 6,103 Schleyer, €! von R., 14, 1 Schmidt, S. I?, 18,187 Schuster, G. B., 18, 187; 22,311

CUMULATIVE INDEX OF AUTHORS

Scorrano, G., 13, 83 Shatenshtein, A. I., 1, 156 Shine, H. J., 13, 155 Shinkai, S., 17,435 Siehl, H.-U., 23,63 Silver, B. L., 3, 123 Simonyi, M., 9, 127 Sinnott, M. L., 24,113 Stock, L. M., 1, 35 Sustmann, R., 26,131 Symons, M. C. R., 1,284 Takashima, K., 21, 197 Takeuchi, K., 30,173 Ta-Shma, R., 27, 239 Tedder, J. M., 16, 51 Tee, 0. S., 29, 1 Thatcher, G. R. J., 25, 99 Thomas, A., 8, 1 Thomas, J. M., 15, 63 Tonellato, U., 9, 185 Toullec, J., 18, 1 'Iiidos, F., 9, 127 lbmer, D. W., 4, 31

Turro,N. J., 20, 1 Ugi, I., 9, 25 Walton, J. C., 16, 51 Ward, B., 8, 1 Watt, C. I. F., 24,57 Wentworth, I?, 31,249 Westaway, K. C., 31, 143 Westheimer, F. H., 21, 1 Whalley, E., 2, 93 Williams, A., 27, 1 Williams, D. L. H., 19, 381 Williams, J. M. Jr, 6, 63 Williams, J. O., 16,159 Williams, R. V., 29, 273 Williamson, D. G., 1,365 Wilson, H., 14, 133 Wolf, A. I?, 2,201 Wyatt, P. A. H., l2,131 Zimmt, M. B., 20, 1 Zollinger, H., 2, 163 Zuman, I?, 5 , l

Cumulative Index of Titles Abstraction, hydrogen atom, from 0-H bonds, 9, 127 Acid solutions, strong, spectroscopic observation of alkylcarbonium ions in, 4, 305 Acid-base behaviour in macrocycles and other concave structures, 30,63 Acid-base properties of electronically excited states of organic molecules, 12, 131 Acids and bases, oxygen and nitrogen in aqueous solution, mechanisms of proton transfer between, 22, 113 Acids, reactions of aliphatic diazo compounds with, 5, 331 Acids, strong aqueous, protonation and solvation in, 13, 83 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 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, 156 Anions, organic, gas-phase reactions of, 24, 1 Antibiotics, p-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 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 Bromination, electrophilic, of carbon-carbon double bonds: structure, solvent and mechanisms, 28,207 I3C 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 having aryl substituents, structure and reactivity of, 22, 311 Carbocation rearrangements, degenerate, 19, 223 Carbon atoms, energetic, reactions with organic compounds, 3,201 Carbon monoxide, reactivity of carbonium ions towards, 10,29 Carbonium ions (alkyl), spectroscopic observation in strong acid solutions, 4, 305 Carbonium ions, gaseous, from the decay of tritiated molecules, 8, 79 Carbonium ions, photochemistry of. 10, 129 Carbonium ions, reactivity towards carbon monoxide, 10,29 409

410

CUMULATIVE INDEX OF TITLES

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, W, 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 Chemically induced dynamic nuclear spin polarization and its applications, 10,53 Chemiluminescence of organic compounds, 18,187 Chirality and molecular recognition in monolayers at the air-water interface, 28,45 CIDNP and its applications, 10,53 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 Crystallographic approaches to transition state structures, 29, 87 Cyclodextrins and other catalysts, the stabilization of transition states by, 29, 1 D20-H20 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 aprotic 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 the mechanisms of enzyme action, 1947-1963,21, 1 Displacement reactions, gas-phase nucleophilic, 21, 197 Double bonds, carbon<arbon, electrophilic bromination of structure, solvent and mechanism, 28,171 Effective charge and transition-state structure in solution, 27, 1 Effective molarities of intramolecular reactions, 17, 183 Electrical conduction in organic solids, 16, 159

CUMULATIVE INDEX OF TITLES

411

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 by, 1, 284 Electron spin resonance studies of short-lived organic radicals, 5,23 Electron storage and transfer in organic redox systems with multiple electrophores, Electron transfer in the thermal and photochemical activation of electron donor-acceptor complexes in organic and organometallic reactions, 29, 185 Electron-transfer reaction, free radical chain processes in aliphatic systems involving an, 23,271 Electron-transfer reactions in organic chemistry, 18, 79 Electron-transfer, single, and nucleophilic substitution, 26, 1 Electron transfer, spin trapping and, 31, 91 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 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 probolem of, 11, 1 Enzyme action, catalysis by 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 on nmr spectra of, 23, 63 Equilibrium constants, NMR measurements of, as a function of temperature, 3, 187 Ester hydrolysis, general base and nucleophilic catalysis, 5, 237 Ester hydrolysis, neighbouring group participation by carbonyl groups in, 28, 171 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 molecules, structure of electronically, 3, 365 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, identification by electron spin resonance, 1,284 Free radicals and their reactions at low temperature using a rotating cryostat, study of, 8, 1 Gaseous carbonium ions from the decay of tritiated molecules, 8, 79 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 General base and nucleophilic catalysis of ester hydrolysis and related reactions, 5, 237

412

CUMULATIVE INDEX OF TITLES

H20-D20 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 Hydrated electrons, reactions of, with organic compounds, 7, 115 Hydration, reversible, of carbonyl compounds, 4, 1 Hydride shifts and transfers, 24, 57 Hydrocarbons, small-ring, gas-phase pyrolysis of, 4, 147 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 carbonxarbon u-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, 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 of organic reactions in water and aqueous mixtures, 14,203 Kinetics, reaction, polarography and, 5, 1 P-Lactam antibiotics, the mechanisms of reactions o& 23, 165 Least nuclear motion, principle of, 15, 1 Macrocycles and other concave structures, acid-base behaviour in, 30, 63 Macromolecular systems in biochemical interest, 13CNMR spectroscopy in, 13,279

CUMULATIVE INDEX OF TITLES

413

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 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, l2 1 Mechanisms and reactivity in reactions of organic oxyacids of sulphur and their anhydrides, 17, 65 Mechanisms, nitrosation, 19, 381 Mechanisms of proton transfer between oxygen and nitrogen acids and bases in aqueous solutions, 22, 113 Mechanisms of reaction in solution, entropies of activation and, 1, 1 Mechanisms of reaction of p-lactam antibiotics, 23, 165 Mechanisms of solvolytic reactions, medium effects on the rates and, 14, 10 Mechanisms, organic reaction, isotopes and, 2, 1 Mechanistic applications of the reactivity-selectivity principle, 14, 69 Mechanistic studies, heat capacities of activation and their use, 5, 121 Medium effects on the rates and mechanisms of solvolytic reactions, 14, 1 Meisenheimer complexes, 7, 211 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 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 Non-planar and planar aromatic systems, 1, 203 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 Nucleophilic aromatic photosubstitution, 11, 225 Nucleophilic catalysis of ester hydrolysis and related reactions, 5,237

414

CUMULATIVE INDEX OF TITLES

Nucleophilic displacement reactons, gas-phase, 21,197 Nucleophilicity of metal complexes towards organic molecules, 23, 1 Nucleophilic substitution in phosphate esters, mechanism and catalysis of, 25, 99 Nucleophilic substitution, single electron transfer and, 26, 1 Nucleophilic vinylic substitution, 7, 1

0 - H bonds, hydrogen atom abstraction from, 9, 127 Oxyacids of sulphur and their anhydrides, mechanisms and reactivity in reactions of organic, 17, 65 Oxygen isotope exchange reactions of organic compounds, 3,123 Perchloro-organic chemistry: structure, spectroscopy and reaction pathways, 25,267 Permutational isomerization of pentavalent phosphorus compounds, 9,25 Phase-transfer catalysis by quaternary ammonium salts, 15,267 Phosphate esters, mechanism and catalysis of nucleophilic 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 Protic and dipolar aprotic solvents, rates of bimolecular substitution reactions in, 5, 173 Protolytic processes in H20-D20 mixtures, 7, 259 Protonation and solvation in strong aqueous acids, 13, 83 Protonation sites in ambident conjugated systems, 11, 267 Proton transfer between oxygen and nitrogen acids and bases in aqueous solution, mechanisms of, 22, 113 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 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 studies 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

CUMULATIVE INDEX OF TITLES

41 5

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 indices in conjugated molecules, 4,73 Reactivity, organic, a general approach to: the configuration mixing model, 21, 99 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 systems, organic, with multiple electrophores, electron storage and transfer in, 28, 1 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 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 Solvent-induced changes in the selectivity of solvolyses in aqueous alcohols and related mixtures, 27, 239 Solvent, protic and dipolar aprotic, rates of bimolecular substitution-reactions in, 5, 173 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 of alkylcarbonium ions in strong acid solutions, 4, 305 Spectrocopy, 13CNMR, in macromolecular systems of biochemical interest, 13,279 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 and mechanisms in carbene chemistry, 7,153 Structure and mechanism in organic electrochemistry, 12, 1 Structure and reactivity of carbenes having aryl substituents, 22, 31 1 Structure of electronically excited molecules, 1, 365 Substitution, aromatic, a quantitative treatment of directive effects in, 1, 35 Substitution, nucleophilic 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

416

CUMULATIVE INDEX OF TITLES

Superacid systems, 9, 1 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 Topochemical phenomena in solid-state chemistry, 15, 63 Itansition states, the stabilization of by cyclodextrins and other catalysts, 29, 1 Transition-state structure in solution, cross-interaction constants and, 27, 57 Transition state structures, crystallographic approaches to, 29,87 Transition-state structure in solution, effective charge and, 27, 1 Transition state structure, secondary deuterium isotope effects and, 31,143 Transition-state theory revisited, 28, 139 nitiated molecules, gaseous carbonium ions from the decay of, 8,79 Tritium atoms, energetic reactions with organic compounds, 2,201 'hmstile rearrangements in isomerization of pentavalent phosphorus compounds, 9, 25 Unsaturated compounds, basicity of, 4, 195 Vinyl cations, 9, 185 Vinylic substitution, nuclephilic, 7, 1 Volumes of activation, use of, for determining reaction mechanisms, 2, 93

Water and aqueous mixtures, kinetics of organic reactions in, 14, 203

This Page Intentionally Left Blank

I S B N 0-12-033531-X