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

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 (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 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