摘要:

CHEMICAL SOCIETY Volume 26 Pages 327406 October 1997 ISSN 0306-0012REVIEWS Issue 5 CSRVBR 26(5)327-406 Selection approaches to catalytic systems Paul A. Brady and Jeremy K. M. Sanders 327-336 The mechanistic and evolutionary basis of stereospecificity for hydrogen transfers in enzyme-catalyzed processes Kevin A. Reynolds and Koren A. Holand 337-344 Pentafluorophenylboranes: from obscurity to applications Warren E. Piers and Tristram Chivers 345-354 I I Molecular modelling of electron transfer systems by noncovalently linked porphyrin-acceptor pairing Takashi Hayashi and Hisanobu Ogoshi 355-364 Photo-induced electron and energy transfer in non-covalent1y bonded supramolecular assemblies Michael D. Ward 365-376 Asymmetric synthesis of building- blocks for peptides and peptidomimetics by means of the p-lactam synthon method Iwao Ojima and Francette Delaloge 37 7-3 86 Approaches to the synthesis of ingenol Sanghee Kim and Jeffrey Winkler 387-400 Hydrogen isotope exchange reactions involving C-H (D,T) 401-406 bonds Thomas Junk and W.James Catallo Articles that will appear in forthcoming issues Sandwich-type heteroleptic phthalocyaninato and porphyrinato metal complexes Dennis K. P. Ng and Jianzhuang Jiang Glycosylation employing bio-systems Vladimir KPen and Joachim Thiem Preparation of seven and larger membered heterocycles by electrophilic heteroatom cyclization Gerard Rousseau and Fadsi Homsi Molecular and chemical basis of prion-related diseases Sheila B.L. Ng and Andrew Doig Ultrasound in synthetic organic chemistry Timothy J. Mason The biomedical chemistry of technetium and rhenium Jonathan R. Dilworth and Suzanne J. Parrott Enzymes in organic synthesis: recent developments in aldol reactions and glycosylations Shuichi Takayama, Glenn J. McGarvey and Chi-Huey Wong Polymer-supported organic reactions: what takes place in the beads? Philip Hodge Equilibrium, frozen, excess and volumetric properties of dilute solutions Michael J. Blandamer New synthetic methods via radical cation fragmentation Mariella Mella, Maurizio Fagnoni, Mauro Freccero, Elisa Fasani and Angelo Albini Asymmetric synthesis of amino acids using sulfinimines (thiooxime S-oxides) Franklin A. Davis, Ping Zhou and Bang-Chi Chen Nonplanar porphyrins and their significance in proteins J. A. Shelnutt, Xing-Zhi Song, Jian-Guo Ma, Song-Long Jia, Walter Jentzen and Craig J. Medford Covalency in semiconductor quantum dots James R. Heath and J. J. Shiang Lanthanide(ir1) chelates for NMR biomedial applications Silvio Aime, Mauro Botta, Mauro Fasano and Enzo Terreno Chemical Society Reviews, 1997, volume 26

ISSN:0306-0012    

DOI:10.1039/CS99726FP015

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Chemical Society Reviews Editorial Board Jean-Pierre Sauvage (CNRS, Strasbourg) [Chair] Vincenzo Balzani (Bologna) Ed C. Constable (Basel) Chris Elschenbroich (Marburg) Tim C. Gallagher (Bristol) Editorial Office Martin Sugden (Managing Editor) David Bradley; Peter Whittington (Production) Debbie Halls (Editorial Secretary) http://chemistry .rsc.org/rsc tel: +44 (0)1223 420066 Chemical Society Reviews publishes concise, succinct and lightly referenced articles that provide an introductory overview to topics of current interest in chemistry. The articles appeal to the general research chemist as well as to the expert in the field and provide an essential starting point for further reading. Advanced undergraduates, postgraduates and experienced re- searchers should all benefit from reading Chemical Society RevieWS. Chemical Society Reviews (ISSN 0306-0012) is published bimonthly by the Royal Society of Chemistry, Thomas Graham House, Science Park, Cambridge, UK CB4 4WF.1997 subscription rate: 2 130 (USA $234). Customers in Canada will be charged the sterling price plus a surcharge to cover GST. Individuals can subscribe for 245 (USA $80) providing their institutional library takes a full price subscription. All orders accompanied by payment should be sent directly to The Royal Society of Chemistry, Turpin Distribution Services Ltd, Blackhorse Road, Letchworth, UK SG6 LHN. (NB Turpin Distribution Services Ltd., distributors, is wholly owned by the Royal Society of Chemistry.) Payment should be by cheque in pounds sterling payable on a UK clearing bank or in US dollars payable on a US clearing bank.Second class postage is paid at ZdenEk Herman (Prague) Horst Kunz (Mainz) John P. Maier (Basel) D. Mike P. Mingos (Imperial) Jeremy K. M. Sanders (Cambridge) Royal Society of Chemistry Thomas Graham House Science Park Cambridge UK CB4 4WF csr@rsc.org fax: +44(0) 1223 420247 The Editorial Board commissions articles that encourage international, interdisciplinary dialogues in chemical research. The Board welcomes any suggestions for new articles. A guide for authors and synopsis form can be found in the first issue of this year’s volume or on the RSC’s World-Wide Web home page (URL above).Alternatively, they can be requested from the Managing Editor, in paper or electronic form (postal and e- mail address above). Jamaica NY 1141-9998. Airfreight and mailing in the USA by Publications Expediting Services Inc., 200 Meacham Avenue, Elmont, NY I1003 and at additional mailing offices. US Postmaster: send address changes to Chemical Society Review, c/o Publication Expediting Services Inc., 200 Meacham Ave- nue, Elmont NY 11003. All dispatches outside UK by bulk airmail within Europe and Accelerated Surface Post outside Europe. 0The Royal Society of Chemistry, 1997 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form, or by any means, electronic, mechanical, recording, or otherwise, without the prior permission of the publishers. Typeset and printed in Great Britain by Black Bear Press Limited.

ISSN:0306-0012    

DOI:10.1039/CS99726FX017

出版商:RSC    

年代:1997

数据来源: RSC

ISSN:0306-0012    

DOI:10.1039/CS99726BX019

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Selection approaches to catalytic systems CB2 1EW E-mail: jkms@cam.ac.uk The key feature of enzymic catalysis is recognition of the transition state. Synthesis of designed systems rarely leads to successful catalysts as the rules for conformation and intermolecular interactions are too imperfectly understood. This review describes several current ‘selection’ approaches to the generation of systems that can recognise transition state analogues. Examples covered include catalytic anti- bodies, ribozymes, imprinted polymers. combinatorial chem- istry, and thermodynamic templating. All have the potential to yield effective catalysts without prior design of every detail. 1 Introduction How can chemists produce new catalysts? In particular, can we mimic the astonishing efficiency and selectivity of enzymic catalysis and apply the same approach to important synthetic reactions for which there are no naturally occurring enzymes? In this review, we start from the principle that the key feature of enzymic catalysis is recognition of the transition state, and then describe several current approaches to the application of that principle.Our examples are taken from molecular biology, organic and inorganic synthesis, and polymer and solid-state chemistry. Many of the literature references are to reviews, rather than to original research papers. At one level, we already ‘understand’ enzyme action; as early as 1894, Emil Fischer proposed that catalysis arose as a result of a binding process.’ He termed this the ‘Lock and Key’ principle: the key (substrate) fits into the lock (enzyme) and is then modified in some way.More subtly, Pauling suggested in 1948 that an enzyme catalyses a reaction by selectively binding and stabilising the transition state more than the starting materials or products.2 This idea, which is essentially the view held today, is summarised in the energy profiles shown in Fig. 1. The enzyme Paul Brady was born in Manchester, UK in I972 but spent much of his youth living in Switzerland. He obtained his BA (Cheniisti-y) fi-om the University oj Cambridge in I993, cariying out sonie research with Ian Paterson and working at Ciba-Geigy and FMC Corporation. He remained in Cambridge to work with Jereniv Sanders, making oligosteroid receptors us the first step towards the thermodynamic templating approach that is described in this article.He was awarded his PhD for this work in I996 and will embark on a career in Patent Law in autumn 1997. Jeremy Sanders was born in Londoii in 1948, and obtained his BSc fr-om Impel-ial College, London, in 1969. He moved to Cambridge to M1oi-k for his PhD with Dudley Williams on Ianthanide shijt I-eagents. After Paul A. Brady E binds the substrate S to form a complex ES with binding free energy AGES.The two transition states for the reaction are TS,,, for the catalysed reaction and TS,,,,, for the reaction in the absence of catalyst. The corresponding activation energies for the reaction are AGCatand AG,,,,,.It is clear from this picture that AGETS,the free energy for binding the transition state to the enzyme, is larger than AGES. This is the origin of the rate enhancement. Despite understanding enzyme catalysis at this level of principle, we fall very short of the more severe, practical test- A GETSITSmt ES complex U EP complex Fig. 1 Free energy profiles for a catalysed and an uncatalysed reaction a postdoctoral year (1972-1973) in the USA, working on protein NMR, he joined the staff of the University of Cambridge; he wm appointed to his present Chair in 1996. He has been awarded the Meldolu Medal, Hickinbottom Award, two Pfizer Awards, the Loschmidt Prize and the Pedler Lectureship, and in I995 was elected a Fellow of the Royal Society.His research in- terests ure mainly centred on various aspects of moleculai. recogrzition, but also halve in- cluded chemical and biolo~gical applications of NMR spectros- copy. In 1996, he joined the Editorial Board of Chemical Society Reviews; his previous reviews in this journal hai>e covered lanthanide shift I-eu-gents, chlorophyll coordination chemistry and biodegradable spiders’ webs. spectroscopy,plastics and ,Jcrciii~K.1/1. Sanders Chemical Society Reviews, 1997, volume 26 327 that of designing and making our own artificial catalytic systems to rival natural enzymes Our ability to predict complex molecular shapes has improved with the development of computer modelling, but it is not yet at a level that enables us to predict accurately the three-dimensional structure of a protein or even of relatively small supramolecular complexes This is because such structures are the finely balanced outcome of many weak non-covalent binding interactions that are them- selves inadequately understood Transition states have only a transient existence, which means that good predictions about their molecular recognition properties are even more elusive As a result, most attempts to synthesise worhng catalysts by rational design of a particular structure have failed It is perhaps inevitable that we will make the wrong molecules in an effort to uncover the design rules for making the right ones Many scientists have therefore embarked on a different line of attack This focuses directly on finding systems that recognise transition states or their stdble analogues, and is in\pired by nature’s evolutionary methods instead of a single structure being designed, a large array of different molecules is created simultmeously From this pool of different binders or catalysts, the best one is selected Until the design rules are known, this \election approach should have an increased chance of succes$ since vastly more molecules are generated than in a design approach While this idea is conceptually straightfor- ward, putting the principles into practice has proved difficult In the following sections we describe the many selection ap- proacher that are being explored Despite the gred diversity in the systems and their methods of generation, all face the \ame challenge, which 13 to achieve efficient transition state stabili- sdtion 2 Catalytic antibodies3 A\ a large part of the catalytic activity of enzymes results from stronger binding of the transition state than of the starting material(\), any molecule that can bind the transition state species selectively should be a catalyst for that reaction With this in mind, Jencks suggested in 1969 that antibodies generated in the mammalian immune response should function as enzymes Antibodies are proteins that are produced in the body in response to an alien species, cdlled an antigen, and they bind to such a molecule or particle strongly and selectively 2.1 Antibody structure and generation Antibodies are proteins consisting of a number of domains (Fig 2) The variable domains consist of peptide chains ot diverse amino acid coompositions, giving rise to a binding site area of ca 20 X 20 A This site may bind small antigens by encapsulation or may present a more open recognition cleft for large antigens The different compositions of the variable regions are the source of antibody diversity Blood serum contains around lox different antibodies, which makes the likelihood of being able to bind a given antigen quite high Small molecules are often poor initiators of the immune response, or are simply broken down by metabolism, so antibodies to a desired antigen (termed hapten) are usually raised by covalently conjugating the hapten to a large protein molecule A sample of this antigen conjugate is then injected into the rat or mouse, usually repeatedly After around a week, antibodies to the hapten can be detected Repeated exposure to the antigen gives an enhanced response as improved, second generation antibodies are produced The spleen is then removed from the mouse and the antibodies are extracted A given immunisation leads to recognition of many different parts of the hapten and so leads to a diverse, ’polyclonal’, set of antibodies In order to study antibody-substrate interactions, a sample of pure single (monoclonal) antibody is required This is now possible using biotechnological manipulation md relatively large nmounts of a single monoclonal antibody can be prepared, albeit dt considerable co\t Antibodies have good bioavailabilty 328 Chemical Society Reviews, 1997, volume 26 Variable binding domains -\ / Fig.2 Schemdtic view of dn antibody which makes them potentially suitable for use as drugs, they are however readily degraded within the body, which could be a problem The first catalytic antibodies were reported simultaneously in 1986 by Schultz and by Lerner In each case, a tetrahedral phosphonate ester group was used as a transition state analogue (TSA) to mimic the tetrahedral shape and charge distribution of the oxy-anion intermediate in ester hydrolysis So, antibodies for the hydrolysis of substrate 1 were raised using TSA 2 These 0 Antibody O I N G O ~ O M ~ * 0,NGOH 1 Protein 2 antibodies bind anionic tetrahedral groups well and thus stabilise the transition state for the hydrolysis reaction, they showed enzyme-like kinetic behaviour.were inhibited by the hapten and accelerated the reaction by factors of up to lo4 As with enzymes, good substrate selectivity was observed anti-bodies raised to 2 show discrimination for 1 over the o-nitrophenyl substrate isomer Catalysis with turnover is relatively easy to achieve in hydrolysis reactions as the small product molecules bind to the enzyme less strongly than the larger starting material They are therefore reledsed into solution, allowing the enzyme to bind further substrate molecules Bond formation reactions are more difficult to catalyse as the products tend to be more strongly bound than the starting materials, resulting in inhibition of the catalyst A good example of a case where this problem has been overcome is Hilvert’s ‘Diels-Alderase’ antibody which accel- erates the reaction between the thiophene dioxide 3 and maleimide 4 The cycloaddition is followed by spontaneous elimination of sulfur dioxide from the initial product 5 to give 6, and atmospheric oxidation to the aromatic species 7 This change in geometry allows product release The transition state in both of thec,e steps resembles the high energy intermediate 5, CI antibody, which IS able to abstract the proton when substrate 10 is bound The strategy is effective and results in a 8 8 X 104-fold ...%o + rate increase for the elimination GEt CI A charged functional group present at the active site cdn do cl* Cl 0 3 4 5 6 CICI*o CI *NEt\ CI CI OO‘i Protein 8 7 and so the norbornene derivdtive 8 was a suitable hapten Antibodies raised to this hapten show multiple turnovers ( > 50) and a reasonable rate enhancement (an effective molarity of 100 M)The rate enhancements of 1 01-1 O4found in early systems are quite modect when compared with natural enzymes This results largely from the fact that most of the catalytic antibody (CAB) enhancement is due to transition state stabilisation by geometri- cal complementarity However, shape complementarity is only part of the key to enzyme catalysis It is also useful to have appropriately positioned catalytic groups present at the active site, and a more sophisticated approach is required to introduce these The first example of a designed hapten placing a desired catalytic group in an antibody was 9, which was used to 0 + 9 7 B-Antlbody 10 11 generate CABS for the catalysis of HF (3-elimination from 10 to give 11 It contains charged functionality to introduce oppo- sitely charged groups in the antibody The tertiary ammonium group in the hapten ensures placement of a basic group in the more than simply stdbilise a charge electrostatically Hapten 12, Protein 12 X -Antibody 13 “Antibody 02N 14 which is protonated at physiological pH, was u\ed for the generation of antibodies for the c [s-ti ans isomeriwtion ot the a,P-unsaturated ketone 13, it is believed that an anionic nucleophile in the active site enhance5 the catalysis ds shown (XCdt/kunCdt= I5 000) A similar ammonium cation was u5ed to induce a carboxyl group in the active site of dn antibody that catalyses an aldol addition,7 the carboxylate is thought to act ds a general base in the reaction, contributing to the 2 0 X 1 05-fold rate enhancement A similar general base mechanism was inferred for the antibody-catalysed decomposition of 15 to 16 The hapten 17 was used to generate antibodies containing a negatively charged group Substantial antibody catalysis was observed for the reaction (kLdt/XL,,,Cdf=: 108) and this was attributed mainly to the ‘precise positioning’ in the antibody of a carboxylate group thdt is able to act as a general base Catalysis was inhibited by added hapten and covalent modification confirmed that the active site did indeed contain the carboxyl group important for catdlysis However, the extraordinary efficiency of this antibody does not prove that it works in the way intended Hilvert’5 group had shown that the reaction is catalysed by acetate ion in a dipolar, aprotic solvent (acetonitrile) to almost the \ame level as by the antibody in water This is due to the activation of acetate ion once it is removed from a hydrogen bonding environment A\ the carboxylate function in the antibody is in a cavity, and hence in an aprotic environment, it is possible that the main part of the rate acceleration observed IS due to this medium effect This logic led Hollfelder et a1 to investigate other proteins with hydrophobic binding sites containing general base rev- dues They found that commercially available serum albumins catalysed the reaction with turnover numbers and rate accelera tions that are very similar to the antibodies rdised against the transition state analogue This result suggests that the preci5e positioning of the base in the antibody is much less important Chemical Society Reviews, 1997, volume 26 329 An tibody Antibody necessary, and there is clearly a need for methods of isolating an antibody directly based on catalytic ability rather than binding strength; such approaches are beginning to appear.' All the 53 selection approaches we describe here share other common I problems, and these are discussed at the end of the article.3 RibozymesI2 Effective enzymic catalysis has evolved over millions of years through mutation and selection. In order to use evolution in the laboratory to create a better catalyst or binder from a good one, 15 I a class of molecules is required that can be replicated under 'sloppy' conditions-that is, which can be amplified to give multiple copies of itself with some mutations. Every so often a mutation will lead to an increase in activity for a particular trait and so evolution occurs if the improved activity can be selected for.02NacN0-There is a widely held view that the present biological era in 16 than medium effects within the cavity.It is conceivable that similar effects have been overlooked in the past during attempts to rationalise results in an over-optimistic fashion. Nevertheless, there are now several examples of antibodies that catalyse inaccessible or otherwise unfavourable processes; a number of these involve redirecting the course of cationic cyclisation reactions.5 A recent example involving a change of mechanism, rather than product is the base-catalysed hydrolysis of carbamate 18 which normally occurs by the top route 0 R H I 20 H Protein-N H P- / KN/R H 19 illustrated (initial deprotonation of the NH group followed by expulsion of ArO-) but use of the TSA 19 led to antibodies that catalyse a pathway involving initial nucleophilic attack by hydroxide to give the intermediate 20.1° Note that, even when racemic haptens are used, the resulting antibodies are generally highly enantioselective because one individual molecule of an arbitrary chirality has always initiated the response leading to a particular monoclonal antibody.The success rate for generating and selecting good hapten- binding antibodies is high, but catalytic activity is quite rare and even the best antibodies do not rival enzymic efficiency. Expansion of the repertoire of available functional groups is 330 Chemical Society Reviews, 1997, volume 26 which information flows from DNA to RNA to protein was preceded by a simpler system with fewer components: a chain of RNA (ribonucleic acid) is simultaneously capable of storing information in its sequence and folding into a three-dimensional structure that carries out some vital catalytic roles in the cell, so it seems feasible that there may have been an 'RNA world' from which ours evolved.Ribozymes are RNA molecules which have some natural catalytic activity: for example the Tetra-hymena ribozyme can cleave certain RNA sequences with a specificity similar to that of RNA processing enzymes. The approach to carrying out an evolution process in the laboratory is generally as shown in Fig. 3. A random library of RNA molecules of a suitable size (generally >50 bases) is taken. A subset of these is selected on the basis of the desired function.This subset is then amplified by reverse transcription to cDNA and then transcribed using RNA polymerase under sloppy conditions (a fault every ca. 100-1000) bases to obtain a new set of RNAs based on the active molecules. Random library of oligo-RNAs New library biased towards selected Select for a function After ca. 10 molecules cycles, library f \ evolved to desired function J Transcribe to RNA under Reverse transcribe sloppy conditions to cDNA Fig. 3 Schematic sequence for the evolution of an RNA library Conceptually, this sequence is straightforward and it has been used to good effect to isolate RNAs with desired binding properties: the selection step takes place by affinity chromato- graphy so that molecules which bind the given substrate well are selectively retained on a column.Given the successful use of transition state analogues in the generation of catalytic antibod- ies, it was originally hoped that evolution of a binder to a transition state analogue would result in a ribozyme with catalytic activity. This has generally proved not to be the case, with a few exceptions. This lack of success is probably due to the lack of functional diversity resulting from the availability of only four building blocks in RNA molecules. Thus, a more dynamic selection process, screening for catalytic activity rather than binding has been employed. The first example of an in vztro evolved ribozyme was that of Beaudry and Joyce 13 Starting from the Tetrahymena ribozyme, which has RNA cleavage activity, a ribozyme with DNA cleavage activity was evolved The selection procedure em- ployed was a DNA cleavage reaction, in which part of the cleaved DNA remained bound to the 3' terminal of the ribozyme The DNA involved was designed to be a primer sequence for cDNA (complementary DNA) synthesis After incubation of the ribozymes with the cleavage DNA, they were amplified by reverse transcription into cDNA followed by transcription to RNA with an RNA polymerase Only ribo- zymes containing the primer were reverse transcribed and amplified, so the amplified ribozyme pool was enhanced in DNA-cleaving molecules Repetition of this procedure ten times resulted in the generation of ribozymes which accelerated the DNA cleavage reaction by a factor of 100-fold over the wild type More recently, Tsang and Joyce have improved this acceleration to 10s using a more stringent selection proce- dure l4 A pool of RNA polymerase ribozymes was more recently evolved by Bartel and Szostak They screened for the ability of a ribozyme to join a substrate to itself Ribozymes that performed this reaction successfully were isolated from the pool by affinity column chromatography for the substrate and then amplified After ten rounds of evolution, a fairly homologous ribozyme family which showed a rate acceleration of 7 X 106 for RNA polymerisation was obtained from a starting pool of over 10l5different random RNA sequences This is not actually catalysis, but rather templated ligation, and still falls a long way short of the rate acceleration of 3 X lo1' achieved by the enzyme Similar procedures have now been used to evolve ribozymes for RNA hydrolysis, polynucleotide kinase, self- alkylation and self-acylation The first example of a ribozyme obtained as a strong binder to a transition state analogue was for the isomerisation of biphenyl 21 to its diastereomer 22 l6 Ribozymes were selected J 21 22 I I COOH COOH 24 25 for their ability to bind the transition state analogue 23 attached to a support An initial library of 1015 195-mers was passed though the affinity matrix and good binders were retained These were subsequently washed off and amplified This selection was carried out seven times to afford ribozymes that showed a rate enhancement of 90-fold and catalytic turnover Hapten 23 has also been used to generate catalytic antibodies for the same reaction 17 These show greater activity than the ribozyme (kca&,,,,t = 2000), perhaps illustrating the better resolution, and hence better binding, available to an oligopep- tide with 20 potential building blocks than to an oligonucleotide with four Using a similar procedure, TSA 24 has been used to select ribozymes that insert copper into the porphyrin 25 Ix Analogue 25 is an N-alkylated porphyrin, which had previously been shown to be a good structural analogue of the trdnsition state for metal insertion Ribozymes clearly show potential in the field of catdlysi\, but it is likely that their structural make-up does not give enough diversity and specificity for them to rival peptide-bdsed catalysts They can rival peptides in schemes involving nucleotides as substrates, but fare less well in other dreds Their evolutionary capacity should go some way to offset thir deficiency if appropriate selection procedures for catdlytic activity can be developed 4 Imprinted polymersiy Antibodies lack features that are important for many practical applications, such as thermal stability and chemical robustners They have a short lifetime and are expensive to produce, so a synthetic analogue would be of much interest Copying their mode of production synthetically is clearly not practical, but molecular-sized cavities can be generated in the solid state by polymerisation in the presence of a guest template This is known as molecular imprinting or 'footprinting' The process leading to an imprinted polymer is illustrated schematically in Fig 4 a bulk polymerisable monomer is mixed with a binding monomer and the guest molecule Then polymerisation 1s initiated and the polymer forms around the imprint molecule After removal of the guest, the polymer should contain cavities of the correct size and shape for the imprint concerned u-1I I Add monomers ~ ~-KZ Catalytically active Fig.4 Schematic sequence for the generation of dn imprinted polymer Whilst conceptually very simple. this approach hd\ been difficult to realise In practice Dickey pioneered the method in the 1940s by preparing silica gels in the presence of methyl orange and homologues, and demonstrating selective adsorp- tion of the appropriate dye 2o However, the silicas were not very stable, and lacked a diversity of functional groups to participate in binding or catalysis In a carbon-based polymerisation Chemical Society Reviews, 1997, volume 26 331 scheme, any functional groups can be attached to monomer units, but they need to be appropriately positioned.There are two general strategies to accomplish this, the self-assembly method and the controlled distance method. An example of a self-assembly directed polymerisation is the generation of a chiral stationary phase for separation of f3-adrenergic blockers by polymerisation of methacrylic acid 26 and a crosslinker 27 in the presence of a (Sj-(-)-timolol 0 0 0 U 26 27 28 template 28.*' Monomer 26 forms noncovalent linkages (mainly hydrogen bonds) with the template in organic solution and this self-assembly holds the monomer units in appropriate positions during the polymerisation process.The template is removed from the finished polymer by washing with acetic acid. The polymer generated in this way has a three-dimensional network, with template-complementary binding sites. When used in powdered form as a stationary phase for chromatog- raphy, the polymer enables easy enantioselective separation of (S)-(-j-timolol from a racemic mixture. 29 Polymerise * 30 The controlled distance approach uses a monomer that is a conjugate of the template and recognition site.An example is illustrated in Fig. 5. The conjugate monomer 29 contains two benzylic groups for polymerisation spanned by an aromatic diimine. The monomer is co-polymerised with the inert scaffolding monomer 30 and when reaction is complete, the central aromatic spacer is removed by washing with acid to give polymer 31. This polymer has a cavity with amine binding groups situated in the correct positions and orientations to recognise the dialdehyde guest 32. The controlled distance method is more likely to be successful than the self-assembly directed method described above if the intermolecular forces are quite weak. On the other hand, as hydrogen bonds and van der Waals forces are strongly dependent on bond length and the controlled method may not optimise these lengths, the self- assembly method could potentially create a better fit if the forces are strong enough.At this stage, neither method has shown itself to be generally superior. Thus, it is now possible to create polymers incorporating custom-designed binding sites. Such polymers have been used for selective extractions and as chiral stationary phases. From this position, footprinted polymers can progress to the more ambitious challenge of catalysis. Maier has recently generated a silica-based polymer using the controlled distance approach to form cavities capable of catalysing the transesterification of 33 to 35 iia intermediate 34:**the silyl ether monomer unit 36 contains a phosphonate side chain analogous to those used in catalytic antibodies for mimicking ester hydrolysis.The side chain is thermally cleavable and after integration into the polymer is removed by heating. The polymer is produced by co- polymerisation of Si(OEtj4 and 1 mol% of 36. At pH 7 no transesterification is observed in the absence of catalyst, but on addition of the imprinted polymer rapid reaction ensues. No detailed kinetics have been reported, but selectivity was Hydrolyse template 32 \ /31 Fig. 5 The controlled distance approach to polymer generation 332 Chemical Society Reviews, 1997, volume 26 0 L-33 34 /IC2H50H 0 r-. OSIOE~~ 35 36 observed in experiments where hexanol and phenylethanol compete for the ester, there is tenfold selectivity for the alkyl alcohol, despite the fact that both alcohols react at virtually the same rate in the presence of sulfuric acid as catalyst Catalysis by the imprinted polymer was achieved by using cavities that were the correct shape to bind the transition state well As with antibodies, better rate acceleration would be expected if catalytic functional groups were positioned in the correct positions in the cavity This has recently been achieved by Shea z3 for the dehydrofluorination of 10, a polymer was prepared in the presence of transition state analogue 37 The template was expected to position the amino-substituted acrylamide 38, present at 3 mol% in the polymerisation mixture, 0 37 38 in such a way that there would be a basic amino group in position in the cavity to remove a proton from the substrate 10 A rate acceleration of 13 times relative to reaction in free solution was observed This is significantly less than the catalytic antibody’s I O5-fold increase and suffered from product inhibition but the approach does show some promise There have also been attempts to impart reaction stereo- and regio-specificity using imprinted polymers A good example is in the reduction of steroid 3-and 17-ketones with LiA1H4 reported by Bystrom et a1 24 Divinyl benzene was co-polymerised with steroid 39, which contains a vinyl group linked to the 17(3 position After polymerisation, the steroid- polymer ester linkage was cleaved with LiAlH4 and the steroid template was washed away to leave steroid-shaped cavities containing specifically positioned hydroxy groups Hydride was then attached to the hydroxy groups by treating with further LiA1H4 When the steroid diketone 40 was added to the polymer, it was reduced specifically at the 17 position to give product 41 with good stereochemical control ((3 a = 80 5) This contrasts with the reduction in solution which has a 99% preference tor the 3 position Using a polymerisation template 42 with the vinyl group attached at the 3a position, a polymer is generated which effects the reduction specifically at the 3 position The cholestanol product obtained in this way was predominantly the less readily available 3a-isomer (a (3 = 72 28) in contrast to the solution reaction (a = 10 90) Specifically imprinted polymers can now be prepared with some reliability By comparison with their biological rivals, catalytic antibodies, they offer many advantages They are relatively simple and cheap to prepare and can be formed rapidly (2-3 days) They show good mechanical, chemical and thermal stability.can be re-used almost indefinitely, and can be 03 39 -O71 0 42 40 41 stored for years without loss of activity On the other hand, the observed binding has not approached that of catalytic antibodie\ in either specificity or strength In chiral separation?, selectiv- ities have usually been less than tenfold This repreqents a binding energy difference of ca 5 kJ mol-I, which is less than a single hydrogen bond, implying that there is a non-integral number of binding interactions at each we How is this possible’ One answer lies in the heterogeneity of binding sites a small fraction of the binding sites may possess cpecificity similar to antibodies, but the majority are non-specific z5 Polymer systems currently in use tend to possess floppy backbones that lack any structural or bonding features leading to useful secondary structure, so there is little to prevent formation of an ill-defined continuum of cavity shapes and sizes The presence of many and diverse non-?elective binding sites is the main factor limiting the applications of imprinted polymers Heterogeneity has the further consequence that the precise interactions responsible for binding cannot be eluci- dated This is unsatisfactory from an intellectual point of view and means that systematic changes to improve the activity are not easy to envisage 5 Combinatorial chemistry26 One promising approach to the synthesis of large collections of diverse molecules is combinatorial chemistry This allows vast ‘libraries’ of different molecules to be synthesised simul-taneously Strictly speaking, these are not self-designed systems in the same sense as the others described here, but the idea of selection from diversity is similar In molecular biology, it ha? long been common practice to use molecular Iibrdries, and techniques for generating similar libraries of organic com-pounds have been developed over the past few years The key problem IS how to separate and identify a compound with the desired properties although sub-milligram quantities of pure compounds can be characterised by NMR, a synthetic library of around 100000 compounds will contain only tiny amount? of each The first approach to this problem was spatial segregation Geysen synthesised different oligopeptides on polyethylene pins arranged in a matrix, whereby each rod was dipped into different reagents The same conceptual operation can also be employed on a smaller scale by use of photolithography Chemical Society Reviews, 1997, volume 26 333 However, there is a practical limit to the size of library which can be generated by these methods: it is not feasible to have, for example, 106 pins on a support.A larger number of compounds can be generated by the split synthesis method, where the target molecules are bound to small beads. These can be as small as 50 pm in diameter, so in theory 600000 beads, each carrying a unique compound, could be present in a 1 cm3 sample. The beads are separated into equally sized portions and each portion is treated with a different building block reagent in the first step. The beads are then recombined and mixed before being separated again after which a second building block is added. This process can be repeated as many times as required. Each bead in the final library has a product from a specific reaction sequence bound to it as shown in Fig. 6.All the molecules on a given bead are the same and thus, by an appropriate screening method, an active bead can be selected. To identify the molecule on the bead, a method of tagging or a deconvolution process has to be used. Round 1 Round 2 Round 3 /A1-B1-C1 A1 -B1 -C2 ,Al-BI A1 -B1 -C3 AI-Bl-C4 ‘A1 -B1 -C5 ‘AI-85 ‘A5 A5-B5 A5-B 5-C 5 No. of species in library 25 125 Fig. 6 Schematic of a split synthesis protocol involving five different reagents at each step These methods allow efficient generation and screening of relatively complex libraries of compounds. Most of the effort to date has concentrated on the generation of binders for drug discovery and this is becoming more successful with the extension of the methods to more diverse sets of molecules.However, the application to the generation of catalysts is still underdeveloped. The methods used for the generation of catalysts can be the same as those in combinatorial synthesis of drug leads, but a more demanding selection procedure is necessary to identify an active catalyst. Although many libraries have been prepared, they have been of limited structural diversity and topology as they have generally been prepared in a linear fashion and using a limited set of reactions: many reactions that work well in solution are much less effective when applied to beads, and there is a great need for more efficient solid-state organic chemistry. The combinatorial approach will have still greater potential when more methods have been developed for its application to broader and more diverse sets of molecules.Many in the pharmaceutical industry are now attacking this task. Combinatorial chemistry covers a wider range of approaches than is often realised. Menger has recently investigated polymer-bound polyamides as catalysts for the hydrolysis of a phosphodiester.27 Various amounts of a range of different acids were condensed with polyallylamine as shown in Fig. 7. Solutions of the combinatorially generated polymers were then screened for catalytic activity in the hydrolysis reaction. In the presence of Zn2+ ions, one particular polymer was found to increase the rate of the reaction by a factor of 3 X 104.This is five times more effective than the rate acceleration achieved by a catalytic antibody generated for the same reaction. No rate 334 Chemical Society Reviews, 1997, volume 26 R’COOH + R2COOH + R3COOH +R4COOH + wn 1 R’ AO Fig. 7 Menger’s combinatorial synthesis of polymeric catalysts for the hydrolysis of a phosphodiester enhancement was found for polymers bearing only a single substi tuent. A similar approach yielded polymeric reducing agents.28 These catalysts were easy to prepare and show good activity, but as with imprinted polymers, their structure is heterogeneous and unknown, and physical selection of the active sequences appears impractical so the key structural features leading to success are also unknown.A combinatorial approach can also be taken to the optimisa- tion of catalysts previously found to be active. For example, polyoxometalate oxidation catalysts2y and metal-promoted insertions of carbenes into C-H bonds30 have each been optimised combinatorially. 6 Thermodynamic tempIating3’ In the imprinted polymer approach, the binding of the hapten stabilises a particular spatial arrangement of the building blocks during the assembly process: the best final cavities arise because they were thermodynamically stabilised at some point in their creation. However, heterogeneity is inevitable when irreversible, kinetically controlled reactions are used for bond creation: ‘incorrect’ bond formation cannot be proof-read and corrected.The problem is exacerbated in the polymer approach which prevents separation of successful from unsuccessful cavities. To avoid these problems, several groups have been working on a new approach that will utilise reversible, thermodynamically controlled bond making to generate hosts. Aliseev and Nelen have developed a protocol for the selection of the best of three isomeric hosts. An affinity chromatography column with a guest attached is connected to an equilibration chamber and the host molecules are circulated around the system. The reaction employed was a photochemical cisltrans isomerisation, and after several cycles, the amount of the strongly binding isomer increased from 3 to 85%, due to its retention on the high affinity column.32 The more difficult process of carrying out the equilibration in the same reaction vessel as the binding process has recently been explored by Huc and Lehn.33 Several components which were able to bond to each other by the reversible formation of imines were mixed together.In the presence of carbonic anhydrase the equilibrium distribution of imines was shifted towards the guest structure that fitted best into the enzyme binding site. In this way a potential inhibitor for the enzyme was generated combinato- rially from a ‘virtual’ combinatorial library. The same con- ceptual process can be envisaged the other way around, i.e. the combinatorial formation of a host in the presence of a template: binding of this guest should stabilise a particular product thermodynamically in a mixture and hence increase the amount of that product formed. Thermodynamic templating has been used for a variety of purposes,31,34.35 but not, as far as we know, for generating catalysts in the way proposed here.This concept is shown pictorially in Fig. 8. If several building blocks are assembled in the presence of a guest, then at least one combination would be expected to bind the guest. This product should be reduced in energy (AHblnd,”&relative to non-binding products and should steroid that binds best to that particular metal best.38 Binding preferences were confirmed by electrospray mass spectrome- try39 In order to obtain good resolution in the binding of potential guests, a wide repertoire of building blocks will be Isdolationand required.The variety of building blocks has recently been Release extended to alkaloid derivatives 43 and the xanthene derivative Fig. 8 Thermodynamic templating for lead generation: the host synthesis reaction takes place in the presence of the template, favouring a host which binds the template well therefore be preferred if its assembly proceeds in a reversible and thermodynamically controlled fashion. Non-binding hosts produced should on average be proof-read and recycled into other products assuming that all possible hosts can be accessed without kinetic barriers. Any strongly binding product will become more concentrated in the reaction mixture through a process of thermodynamic templating and after isolation it could be identified as the best of all possible hosts.Furthermore, use of a different template should allow isolation of a different host from the same reaction mixture. These principles have been utilised in the generation of inorganic coordination complexes such as Lehn's self sorting helices,-76 but as these compounds are held together only by relatively weak interactions in solution, they lack the robust character of covalent molecules. The approach outlined in Fig. 8 depends on finding a reaction that is fast even in the absence of an excess of one reactant, and on devising suitable building blocks. Our work in this area began with transesterification of steroids (Fig. 9).37 The starting materials are furnished with an alcohol and a methyl ester group, so removal of methanol from the reaction mixture leads to a dynamic mixture of the cyclic compounds shown.When the reaction is carried out in the presence of metal ions as templates, the distribution of rings is shifted towards the macrocyclic 0 0 Linear0 Intermediates I 44, which also undergo efficient reversible cyclisation when Meoq HOF I C02MeMeowHO 43 44 placed under the transesterification condition~.~O When rather rigid building blocks are utilised, the distribution of products obtained is small, but for more flexible starting materials a combinatorial mixture containing many different products can be ~btained.~' A major attraction of this approach is that the products are discrete compounds that can be characterised and modified rationally.Issues that will become important in the future development of these concepts include the nature of the guest (a transition state analogue or a drug for which a receptor is required) and whether the guest should be in free solution or attached to a solid support as shown in Fig. 8. In solution, it will mimic the real binding process required better than when on a solid support, but the latter will allow for easier isolation of a binding product. The current building blocks are quite large, so the resolution they are able to achieve at the binding site is not very high. Either a wide diversity of such building blocks or smaller units will be required to improve this.MeOH t 0 KOMe, toluene Dicyclohexyl 18-crown-6 HO Fig. 9 The thermodynamically controlled cyclisation of steroid derivatives by transesterification Chemical Society Reviews, 1997, volume 26 335 A similar approach has been common for some time in the synthesis of zeolites. It is now possible to design structure- directing templates for the synthesis of microporous qolids, and it is surely only a matter of time before a transition state analogue is used to generate a catalytic zeolite.42 Hill’s ‘self- repairing’ polyoxyanion oxidation catalysts43 are superficially similar in concept in the sense that catalyst synthesis and catalysis occur simultaneously, but it does not appear that teinplating by the substrate plays a major r6le 7 Conclusions These selection methods have made considerable progress and are now in reguldr use for lead generation Each method has inherent problems.there is a numerical and analytical limit to the diversity accessible by library methods; imprinted polymers have heterogeneous and unknown binding sites, antibodies are expensive, inefficient, and their peptidic nature brings problems of thermal and enzymic instability, and ribozymes, in addition to expense and instability, may lack the structural diversity necessary for effective host generation The more recent idea of thermodynamic templating combines several of the best features of the other techniques but it too will require a wide range of building blocks and careful hapten design in order to fulfil its potential In addition to the specific problems peculiar to each of these techniques, all sharc a geilera! problem: our understanding of transition state analogues and hapten/template design needs to improve in order to take full advantage of all the potentially powerful techniques described in this article, and ways need to be found for inducing catalysts for multi-step reactions.So, despite the fears expressed by some chemists, this type of approach does not give the intellectual responsibility for design away to the molecules themselves. it shifts the challenge to devising better ways of selecting the right molecule However, given the relatively short history of the ‘selection approach’, major progress has been made In just ten years, catalysts for a variety of reactions have been created and the number of active systems developed in this way may exceed those generated by more conventional designed approaches We believe that selection approaches to self-designed systems will be ctrategically important in decades to come 8 Acknowledgements We thank the EPSRC and Rhbne-Poulenc Rorer for financial support, and numerous colleagues for stimulating discussions.9 References 1 E Fischer, Bei Dtsth Cheni Ges , 1894, 27, 2985 2 L Pauling, Nuture. 1948, 161, 707 3 N R Thomas, Nat Piod Rep, 1996, 13,479 4 P G Schultz and R A Lerner. Science, 1995, 269, 1835 5 T Li R A Lerner and K Janda,Atc Cheni Res, 1997,30, 115 6 W P Jencks, CatulyJis in Chemistiy and Enzymoloqy, McGraw Hill NY, 1969, page 16 7 T Koch, J L Reymond and R A Lerner, J Am Chem Soc , 1995,117, 9383 8 S N Thorn, R G Daniels.M -T M Auditor and D Hilvert, Narure, 1995,373, 228 9 F Hollfelder, A J Kirby and D S Tdwfik, Nature, 1996, 383, 60 10 P Wentworth Jr , A Datta, S Smith, A Marshall, L J Partridge and G M Blackbum .I Am Chem Soc 1997.119,2315 II K D Janda, L -C Lo, C -H L Lo, M -M Sim. R Wang, C -H Wong and R A Lemer, Science, 1997, 275, 945 12 J R Lorsch and J W Szostak, Acc Chem Res , 1996, 29, 103 13 A A Beaudry and G F Joyce, Science, 1992,257,635 14 J Tsang and G F Joyce, Biochemistry, 1994, 33, 5966 15 E H Ekland, J W Szostak and D P Bartel, Science, 1995, 269, 364 16 J R Prudent, T Uno and P G Schultz, Science, 1994, 264, 1924 17 T Uno, J Ku, J R Prudent, A Huang and P G Schultz, J Am Chem Sot , 1996, 118, 3811 18 M M Conn, J R Prudent and P G Schultz, J Am Chenr Soc , 1996, 118, 7012 19 G Wulff, Angeu Chem Int Ed Enql 1995, 34, 1812 20 F H Dickey, Pioc Nut1 Acad Sir USA, 1949,35, 227 21 M Kempe and K J Mosbach, Chtomutuyr , 1995, 694, 3 22 J Heilmann and W F Maier, Angeu Chem Int Ed Engl , 1994,33, 47 1 23 K J Shea and J V Beach, J Am Chem Soc , 1994.116, 379 24 S Bystrom, A Borje and B Akermark, J Am Chem SOL, 1993, 115, 208 1 25 L I Anderssen, R Muller, G Vlatakis and K Mosbach, PIOCNarI Atud Scr USA, 1995,92,4788 26 G Lowe, Chem Soc Re1 , 1995,24,309 27 F M Menger, A V Eliseev and V A Migulin, J Org Chem, 1995, 60,6666 28 F M Menger, C A West and J Ding, Chem Commun , 1997,633 29 C L Hill and R D Gall, J Mol Catalysis, 1996, 114, 103 30 B M Cole, K D Shimizu, C A Krueger, J P A Hamty, M L SnapperandA H Hoveyda,Anyew Chem Int Ed Enpl, 1996, 35, 1668 31 S Anderson, H L Anderson and J K M Sanders, Acc Chem Res, 1993, 26,469 32 A V Eliseev and M I Nelen, J Am Chem Soc , 1997,119, 1147 33 I Huc and J-M Lehn Pioc Nut/ Acud Sci USA, 1997,94, 2106 34 W C Still, P Hauck and D Kempf, Tefiahedrun Lett, 1987, 28, 2817 35 J T Goodwin and D G Lynn, .I Am Chem Soc , 1992, 114, 9197 36 V C M Smith and J -M Lehn, Chem Commun , 1996,2733 37 P A Brady, R P Bonar-Law, S J Rowan, C Suckling and J K M Sander\, Chem Commun , 1996, 221 38 P A Brady and J K M Sanders, J Chem Soc Peikin Trans I, 1997, 3237 39 P A Brady and J K M Sanders, to be submitted 40 S J Rowan, D G Hamilton, P A Brady and J K M Sanders, J Am Chem Soc , 1997.119, 2578 41 S J Rowan and J K M Sanders, Chenz Comniun , 1997, 1407 42 D W Lewis, D J Willock, C R A Catlow, J M Thomas and G J Hutchings, Nufuie, 1996, 382, 604 41 C L Hill and X Zhang, Nature, 1995, 373. 324 Received, 4th Apt 11 1997 Accepted, 231-d May 1997 336 Chemical Society Reviews, 1997, volume 26

ISSN:0306-0012    

DOI:10.1039/CS9972600327

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

The mechanistic and evolutionary basis of stereospecificity for hydrogen transfers in enzyme-catalysed processes Baltimore, MD 21202, USA f7 Depurtment of Chemistry, Gettyshurg College, Gettyshurg, PA 17325, USA The origin of stereospecificity in enzyme-catalysed processes is attributed to either mechanistic imperatives or to the evolutionary origin of an enzyme. Recent analysis of the syn-anti dichotomy for hydratase-dehydratase enzymes has clearly demonstrated that it is a historical contingency that has played a significant role in determination of stereo- specificity. Historical contingencies also appear to play a significant role in (a) the selection of one out of a possible eight different stereochemical courses available for an enoyl thioester reductase, and (b) the discrimination between diastereotopic hydrogens in coenzyme B12-dependent re- arrangements.1 Introduction Enzymes catalyse reactions with remarkable degrees of re-giospecificity and stereospecificity. Often enzyme-catalysed processes contain some cryptic stereospecificity that can only be solved by either stable or radioactive isotope labelling experiments. Experiments carried out by Frank Westheimer and Birgit Vennesland in the 1950s are now used as the bio- chemistry textbook example of the use of such an approach to investigate the stereospecificity of the reaction catalysed by yeast alcohol dehydrogenase (YADH).' In this study it was demonstrated that a hydrogen from the pro-4R position of NADH was transferred to the Re face of acetaldehyde.In the t Pi esmt uddress Institute of Structural Biology and Drug Discovery, Suite 2 12B, Virginia Commonwealth University, 800 East Leigh Street, Richmond, Virginia 232 19, USA. Keirin Reynolds was born in Oxford, England. He obtained a chenzistry degree in 1984 and then worked with John A. Robinson as a PhD student at the University of Southampton ujhere he studied isobufyryl CoA mutase. He then moved to he United States for postdoc- w-a1 work on amino acid race- Pases with Heinz G. Floss at be University of Washington. n 1989, he moved to the Jniversity of Maryland at Bal- ;more as Associate Professor If Pharmaceutical Sciences. In ddition to his interests in en- yme stereochemistry he has esearch groups investigating spects of fatty acid and poly- etide biosynthesis in strepto- fycetes.reverse reaction the pro-R hydrogen at C 1 position in ethanol is transferred to the Re face of NAD+ (Scheme 1). In one direction of the reaction the enzyme discriminates between the diaster- eotopic hydrogens attached at a C4 of the cofactor. In the opposite reaction direction the YADH discriminates between two diastereotopic hydrogens attached to C1 of the substrate ethanol. The basis for the high degree of stereospecificity in the YADH case is the particular arrangement of the substrate and cofactor at the active site. The X-ray crystal structure of YADH SI R Acetaldehyde NADH, H+ OH (i.'... I R Ethanol NAD' Scheme 1 Stereochemical course of the reaction catalysed by the yeast alcohol dehydrogenase Koren Holland was born in Massachusetts in 1963.After tuking her chemistry degree at Skidmore College in 1985, she moved to the University of Maryland at College Park. Under the direction of John Kozurich she obtained her PhD in hioorganic chenzistry in 1990. She receilscd a postdoctoral research fellow>- ship from the National Insti- tutes of Health and moved to the Johns Hopkins University, De-partment of Chemistry, und worked under the direction oj Craig Townsend until 1992. She is currently an Assistunt Professor of Chemistry at Gct- tysburg College with research interests in secondary metu-bolic pathways. Kevin A. Reynolds Koren A.Holland Chemical Society Reviews, 1997, volume 26 337 with ethanol and NAD+ at the active site is clearly consistent with the observed stereospecific course of the enzyme-catalysed reaction .2 2 Interpretation of stereochemical diversity Intriguingly, other alcohol dehydrogenases differ from the YADH and catalyse reactions by transferring the opposite diastereotopic hydrogen of NAD(P)H. Furthermore, an analysis of all dehydrogenases has revealed that almost a 50/50 mix of enzymes use the pro-4R versus the pro-4S of NAD(P)H. Clearly in approximately half of the cases studied the orientation of the NAD(P)+ ring with respect to substrate is different to that observed for the YADH (Scheme 2).3 Mechanistic and historical theories concerning the origin of this stereochemical diversity for dehydrogenases have been presented.- These have been extensively reviewed previously and are summarized below.HReI hydride transfer A. syn B.anti Scheme 2 The conformation of syn and anti NAD(P)H The mechanistic model holds that the two heterotopic hydrogens of NAD(P)H have a slightly different redox potential depending upon the syn and anti orientation (Scheme 2) and that this potential should closely match that of the substrate in order for the enzyme to be a good catalyst.4 This argument suggests that the differing stereospecificities of dehydrogenases have evolved in order to meet a mechanistic imperative associated with the different substrates. Thus, this model is consistent with Bentley 's first rule that enzymes accepting the same substrates generally have the same stereospecificity.7 However, the mechanistic model presents no clear explanation for exceptions to this rule (the stereospecificity of 3-hydroxy-3-methylglutaryl CoA reductases from Acholeplasma laidlawii and yeast for example are different).6 An additional weakness of the mechanistic model is that it hinges on the hypotheses that evolutionary selection pressures have produced enzymes with catalytic optimality.It has been argued more recently, however, that modern protein diversity only represents a very limited exploration of different amino acid sequences space and that this exploration is limited by the success of earlier motifs.8 Therefore, the reaction pathways of many enzyme-catalysed processes may represent a local rather than a global local optimum. The suggestion that some extant proteins do not utilize the most efficient reaction pathway is consistent with an historical interpretation of dehydrogenase stereospecificity.This model proposes that stereospecificity is not a functional trait but is a vestige of an arbitrary choice made early and then retained during the evolutionary process. The stereospecificity of an enzyme-catalysed process is, therefore, maintained during evolution and is a reflection of an enzyme's heritage. An historical model, like the mechanistic model, can rationalize Bentley's first rule with the presumption that all enzymes from all organisms acting on a particular substrate have descended from a single ancestral enzyme.Such a presumption now appears to be quite reasonable; malate dehydrogenase is an example of an enzyme which has both a highly conserved stereospecificity and amino acid sequence (greater than 50% amino acid sequence identity is shared between malate dehydrogenase cloned and sequenced from mammals, plants and microorganisms).9 For exceptions to Bentley 's first rule (such as HMG CoA reductase) the historical model assumes that either (a) these enzymes have independent pedigrees, or (b) the stereospecificity of these dehydrogenases has drifted (whereas for instance it is conserved for the malate dehydrogenases).6 The discussions surrounding functional and historical models have been extended to other enzyme-catalysed processes which discriminate between diastereotopic hydrogens.It has been suggested that some of these reactions may also be controlled by mechanistic imperatives and that any historical model to explain all of the observed data would be extremely complex.6 In this review article we analyse recent stereochemical and protein sequence analyses for three enzyme classes; dehydratase- hydratases, enoyl thioester reductases, and coenzyme B 12-dependent mutases. The data presented clearly show that in all three cases the stereospecificity of the enzyme class appears to be based on an historical contingency. We aim to stimulate the reader to consider the ramifications of these results for the controversy surrounding the basis of dehydrogenase ster-eospecificity .3 Hydratase-dehydratase enzymes There are two classes of hydratase-dehydratase enzymes; those that catalyse the addition of water in a syn fashion and those that do so in an anti fashion. This dichotomy could reflect functional advantages. All hydratase-dehydratase reactions where a proton is abstracted a to a carboxylate group proceed with anti chemistry (seven examples, including fumarate hydratase, aconitate hydratase and enolase are known). 10 In nonenzymatic reactions the anti elimination is favoured over the syn elimination as the latter requires an eclipsed geometry in the transition state. Accordingly, the stereospecificity of an anti elimination reaction may serve as a selected function which either constrains divergent evolution or encourages convergent evolution.However, there is a group of enzymes that produce syn elimination reactions of water (including enoyl coenzyme A hydratases, fatty acid synthases and p-hydroxydecanoyl thio- ester dehydratases). A functional theory for determining stereospecificity for hydratase-dehydratases would require that a two step syn elimination is specifically favoured for the substrates used by these enzymes. In this class of enzymes the proton is abstracted a to a carbonyl group of a thioester or a ketone, and it is suggested that acidity of this proton may alter the preferred pathway; the a-protons of thioesters are more acidic than those of either carboxylate salts or free acids and may tilt the stereochemistry towards syn elimination in which an economical single acid-base group would interact with both the a-proton and the p-leaving group.3J1 Thus a functional theory exists for the stereochemical dichotomy of the hydra- tase-dehydratases as well as the dehydrogenases.This functional theory has been investigated by comparing the chemical stereoselectivity for a non-enzymatically catalysed reaction with the analogous enzyme-catalysed process. lo It was shown that conjugate addition of water or fumarate was slightly biased towards anti addition (1.3 :1). In contrast, addition of water to the a$-unsaturated thioester of crotonic acid produces a substantial 4.3 : 1 bias towards anti addition.These observa- tions are opposite to the enzymatic syn-anti dichotomy, where fumarate hydratase and enoyl CoA hydratase give anti and syn addition-elimination pathways, respectively. Therefore, the enoyl CoA hydratase utilizes the syn pathway for reasons other than a mechanistic imperative. A historical theory would argue that each stereochemical class of hydratase-dehydratase has originated from at least two different ancestral progenitors; enzymes catalysing syn elimina-tion pathways are not related to those which catalyse anti eliminations. This theory would also predict that once a syn elimination pathway was adopted it would be subsequently conserved during the evolution. An enzyme such as the enoyl 338 Chemical Society Reviews, 1997, volume 26 CoA hydratase would, therefore, have optimized this syn pathway within the confines of a conserved active site structure.The difference between syn and anti elimination pathways involving unhindered acyclic substrates is relatively small (less than 3 kcal mol-1) and thus may not be sufficient to deter an enzyme from adopting a syn elimination pathway. Interestingly, the difference in reduction potential of the NAD(P)H in the syn and anti form is even less, approximately 1.3 kcal mol-I (I cal = 4.184 J).* The dehydroquinate dehydratases (DHQase) provide addi- tional compelling evidence for a historical contingency in the stereochemistry of hydratase-dehydratase enzymes. This en- zyme catalyses a dehydration reaction of dehydroquinic acid to dehydroshikimic acid (Scheme 3).The type I DHQases catalyse this reaction with removal of the pro-2R hydrogen, whereas the type I1 DHQases remove the pl-0-2s hydrogen.12 This ster- eochemical dichotomy for the same substrate cannot be readily explained by a mechanistic imperative (the substrate is the same for both types of DHQase). An historical contingency would argue that the two enzyme types are unrelated. Indeed, there is no clear amino acid sequence homology between the type I and the type I1 dehydratases.’3 All type I1 dehydroquinate dehy- dratases that have been cloned and sequenced show strong homology at the level of the predicted amino acid sequence. The same is true for all of the type I dehydroquinate dehydratases. Two distinct families of dehydroquinate dehydratases have evolved that catalyse a reaction with a different mechanism and a different stereospecificity. 3 The stereospecificity of the DHQases reflects the evolutionary origins, or pedigree, of the enzyme.HO Type I: syn elimination %OH -0ec -02c%oHOH HReo )HR~+,OH- 0 Hs:, OH- 3-dehydroshikimate Scheme 3 Stereochemical course of the reaction catalysed by Type I and Type I1 dehydroquinate dehydratases 4 The enoyl thioester reductases This class of enzymes is involved in the conversion of an enoyl thioester to an acyl thioester (Scheme 4).14 There are three cryptic stereochemical details for these reactions: (a) the transfer of either the pro-4R or pro-4S hydrogen of NADPH to the substrate, (b) addition of this hydrogen to either the Re or Si face of the (3-carbon of the substrate, and (c) the addition of a solvent hydrogen to either the Re or Si face of the a-carbon of the substrate.Accordingly, there are a total of eight different stereospecific courses that an enoyl thioester reductase can follow (assuming that the regiospecificity of the process is always addition of the solvent hydrogen at the a-carbon) (Scheme 4). Until recently, only four of these stereochemical outcomes had been observed with the enoyl thioester reductases (Table 1). A general pattern had emerged in which the nucleotide specificity, either pro-4R or pi-0-4s. appeared to determine the stereospecificity of the hydrogen addition at the (3-carbon of the fatty acid, pro-3R or pro-3S respectively.15 If this observation were applicable for all enoyl thioester re- ductases, no additional stereochemical outcomes would have been observed.Recently, however, it has been shown that a crotonyl CoA reductase (CCR) from Streptumyces cullinus converts crotonyl CoA to butyryl CoA with addition of thepro-4S hydrogen of the Stereochemical Outcome R,*SR Vlll Ha Scheme 4 Eight possible stereochemical outcomes for a reaction catalysed by an enoyl thioester reductase. In all outcomes the solvent hydrogen is added to the cu-carbon (an additional eight stereochemical outcomes would be possible if addition to the fi-carbon was allowed. SR represents either coenzyme A (SCoA) or an acyl carrier protein (SACP).nucleotide to the Re face of the (3-carbon and solvent hydrogen to the Re face of the a-carbon.I4 With the inclusion of this novel stereochemical course five of the eight stereochemical out-comes possible for enoyl CoA thioester reductases have now been observed (Table 1). As the nucleotide stereospecificity does not determine the stereospecificity of hydrogen addition at the (3-carbon, it has been suggested that eventually examples of all eight possible stereochemical outcomes may be observed for these enzymes.14 Enoyl thioester reductases represent an unusual case where both a functional and a historical model must be invoked to interpret the observed regio- and stereo-specificities.l4 In all cases examined to date the solvent hydrogen is added to the a-carbon. This observation is associated with a mechanistic imperative associated with the polarization of the a,(3-double bond. In contrast, there is no clear mechanistic imperative for the tremendous diversity in the stereospecifity of these enzymes. As the reduction potential of acyclic enoyl thioesters is unlikely to vary significantly, the dichotomy in nucleotide specificity is likely determined by a historical contingency. This argument can be extended to include the additional ster-eochemical details of enoyl thioester reductases; for instance the addition of the nucleotide hydrogen to the Re or Si fxe of the (3-carbon of the substrate has no clear mechanistic advantage.Thus, the five different stereospecificities of the enoyl thioester reductases may correlate with a minimum of five different evolutionary origins (i.e. the enzymes that exhibit different stereospecificities are not related, while those exhibiting the same stereospecificities may or may not be related). An alternative interpretation is that the enoyl thioester reductases share a common evolutionary origin, but that their ster-Chemical Society Reviews, 1997, volume 26 339 Table 1 Stereochemical outcome of the reaction catalysed by various enoyl reductases‘ Attack of hydrogen Stereospecificity Stereochemical Enzyme source of NAD(P)H c-3 C-2 Type of addition outcome Brevihactel-ium ammoniagene.+‘ pl-o-4s Si Si anti I1 Yeast!’,‘Rat”.( pru-4s pro-4R Si Re Si Si anti sYn 11 IV Chicken” pro4R Re Si syn IV E.coli (chain elongation) pro-4R Re Re anti 111 E. cdi” pro4 Si Re SYn I S. cdlinus (ChcA)” pi.o-4s Si Si anti I1 S. collinus (CCR)h,” pro-4s Re Re anti V 0 Only enoyl thioester reductases where all three cryptic stereochemical details have been elucidated are included. Roman numerals are used to correlate the observed stereochemical outcomes with those depicted in Scheme 4.Most of the enzymes are enoyl ACP reductases involved in de novo fatty acid synthesis. The exceptions are the crotonyl CoA reductase (Ccr) and I-cyclohexenylcarbonyl CoA reductase (ChcA) of S. collinus and the E. coli enoyl thioester reductase involved in fatty acid elongation.I,Denotes enzymes whose predicted amino acid sequence is available. Denotes two groups of enzymes that have significant amino acid sequence similarity within the group. The remaining enzymes have no significant amino acid sequence similarity with each other. (1 Findings from this work. eospecificity has diverged. It has been noted that the tremen- dous diversity observed in both the predicted amino acid sequence and putative evolutionary origin of enoyl thioester reductases presents clear evidence contrary to the latter of these proposals.14 Evidence consistent with the former of these proposals has been provided by examination of the ster-eospecificity and amino acid sequence of the enoyl thioester reductases listed in Table 1.Enoyl thioester (ACP) reductases are involved in fatty acid biosynthesis. The Escherichia coli enoyl ACP reductase involved in de no130 synthesis exhibits a different stereospecif- icity to the eukaryotic enoyl ACP reductases (Table 1). Consistent with this observation, the predicted amino acid sequence data of fabZ (the gene that putatively encodes this enzyme) shows similarity to some alcohol dehydrogenases, but not with other enoyl ACP reductase.16 In contrast, the enoyl ACP reductase sites in the chicken and rat fatty acid synthases (FASs) exhibit 83% identity (overall these two synthases exhibit 67% identity), and catalyse reactions with the same stereospecificity (Table 1). The yeast and Brevibacteriurn ummoniugenes have significant sequence homology (30 and 46% identity along the entire sequence for FASl and FAS2, respectively)17 and the respective enoyl ACP reductases catalyse reactions with identical stereospecificities (Table l).l8 Finally, it has been noted that the yeast-B.ammoniagenes and the chicken-rat FAS enoyl ACP reductases which catalyse reactions with different stereochemical outcomes have no clear amino acid sequence ~imilarity.1~ It has even been suggested that the yeast and chicken FAS either diverged from a common evolutionary pathway at a very early time, or were obtained via a different evolutionary pathway. 19 The crotonyl CoA reductase (CCR) of S. collinus has a different stereospecificity to other enoyl thioester reductases where the overall stereospecificity is kn0wn.1~ Analysis of the amino acid sequence of this CCR with these enoyl thioester reductases, where the sequence is known, reveals no obvious similarity.In fact CCR is not related to enoyl thioester reductases at all, but rather to alcohol dehydrogenase members of the quinone oxidoreductase superfamily. Finally, the stereospecificity of the 1 -cyclohexenylcarbonyl CoA reductase (ChcA) of S. collinus matches that observed for the yeast-B. ammoniagenes FAS enoyl ACP reductase. Sig-nificantly, the amino acid sequence of ChcA shows no sequence similarity with other enoyl thioester reductases but rather with members of the short-chain alcohol dehydrogenase super- These observations for CCR, ChcA and the FAS enoyl ACP reductases are all consistent with the suggestion that an observation of diversity in stereospecificity indicates unrelated enzymes while identical stereospecificity may indicate related 340 Chemical Society Reviews, 1997, volume 26 enzymes.The comparison of the yeast FAS enoyl ACP reductase and ChcA, however, demonstrates that the enzymes need not be related. This historical model for enoyl thioester reductase stereo- specificity has been presented only very recently, and is based on the availability of both a complete stereochemical analysis of each enzyme-catalysed reaction and the predicted amino acid sequence.14 However, the principles of this model have been used for more than a decade to make predictions regarding relationships between enoyl thioester reductases based simply on a comparison of just one or two stereochemical details.Subsequent analyses have always proved consistent with these initial predictions. For instance, collected findings with fungal systems have shown that within an organism the enoyl thioester reductase reactions that occur in both fatty acid and polyketide bio- synthesis proceed with opposite stereospecificities from the perspective of solvent hydrogen addition at the a-carbon. The data have often been used to argue that different enzymes catalyse the reductions in these processes.21 Furthermore, it has been argued that these enzymes would have to be unrelated. All available data supports this hypothesis.14 The one exception to the findings with fungal polyketide and fatty acid biosynthesis has been the long standing studies of the biosynthesis of averufin and fatty acids in Aspergillus parasiticus. In this case the first three enoyl thioester reduction steps in both processes proceed with solvent hydrogen addition to the same face of the growing polyketide and fatty acid.This observation has been used to argue that the first three addition reactions in the polyketide biosynthetic process cannot be catalysed by a polyketide synthase (which has a different stereospecificity) but rather by a fatty acid synthase (i.e conserved stereospecificitiy may indicate that a reaction is carried out by a related enzyme).21 This prediction has recently been substantiated by genetic analysis of the averufin polyketide synthase (PKS) gene cluster which reveals a FAS-like gene essential to the biosynthetic process.22 In S.collinus it has been shown that a pathway from shikimic acid to cyclohexanecarboxylic acid (CHC) proceeds with three separate enoyl thioester reductions, each proceeding in an anti fashion with addition of solvent hydrogen to the Si face of the a-carbon (this in vivo analysis did not allow the nucleotide specificity to be determined) (Scheme 5).23 The enzymes which catalyse these reactions, therefore, may be similar. In fact the possibility that one enzyme is responsible for catalysing all three reductions was raised.23 Recent enzymatic and genetic analysis of ChcA has proved consistent with this prediction.20 In S.hygroscopicus it has been shown that a pathway from shikimic acid to dihydroxycyclohexanecarboxylic acid (DHCHC) proceeds with two separate enoyl thioester reduc- COOH COSCoA FOSCoA ments: methylmalonyl CoA mutase, isobutyryl CoA mutase, glutamate mutase and methyleneglutarate mutase. Methylmalonyl CoA mutase catalyses the conversion of succinyl CoA to methylmalonyl CoA by removal of the pro-3R OH OH OH Shikimic Acid I COSCoA YOSCoA COSCoA COSCoA COSCoA COSCoA CHC Scheme 5 Conversion of shikimic acid to cyclohexanecarboxylic acid (CHC) in S. collinus. The three a$-enoyl thioester reduction steps have been shown to be catalysed in viwu by I-cyclohexenylcarbonyl CoA reductase (ChcA).tions (Scheme 6).Z4 The stereospecificity of both of these reductions differs from that observed in the CHC pathway. Thus it has been predicted that the enoyl thioester reductases involved in DHCHC biosynthesis are unrelated to the ZhcA protein. All data obtained to date are in full agreement with this proposal. In addition it has been noted that the stereochemical outcome of hydrogen of succinyl CoA (Scheme 7). A high degree of homology between the predicted amino acid sequences of the human, mouse and bacterial (Propionibacteriumshermanii and S. cinnamonensis) methylmalonyl CoA mutases suggests that these enzymes have evolved from one common ancestor.27 Isobutyryl CoA mutase catalyses the conversion of butyryl CoA to isobutyryl CoA with the abstraction of the pro-% hydrogen (Scheme 7).As shown in Scheme 7 the absolute ster-eochemistry for this hydrogen abstraction is the same as that observed for methylmalonyl CoA mutase. As there is no obvious mechanistic rationale for the stereospecific selection of either the ~~-0-3sor the pro-3R hydrogen in these reactions it is reasonable that such a choice is based on an historical contingency. Thus, methylmalonyl CoA mutases and isobutyryl CoA mutase may be related, Professor J. A. Robinson has determined that the isobutyryl CoA mutase and large subunit of methylmalonyl CoA mutases of S. cinnamonensis share 44% identity (personal communication). -ooc, 0 OCH3 #H CH~ \!'/-C-SCOA I1 the two enoyl thioester reductions in the DHCHC pathways differs, suggesting that they must be carried out by different enzymes.24 In fact it has been suggested that the first reduction may not even be carried out with the substrate activated as a thioester. 146COOH -&Ha &HaI HO" , OH OH OH OH OH Reductase 1 OH Shikirnic Acid , -c Hb..fili: -HbbHaCOR OHAH Reductase2 bH DHCHC Scheme 6 Conversion of shikimic acid to dihydroxycyclohexanecarboxylic (DHCHC) acid in S.hygroscopic-us Taken together these observations are all consistent with the following conclusions; (a) enoyl thioester reductases that have different stereospecificities have unrelated amino acid se-quences and (b) the stereospecificity of an enoyl thioester reductase is not a functional trait and reflects the evolutionary origin of the protein.Finally, there is no evidence that stereospecificity of enoyl thioester reductases can diverge. 5 Coenzyme BI2-dependent rearrangements A number of coenzyme B 12-mediated rearrangements also involve the stereospecific removal of a heterotopic hydro- gen.25.26 For the purposes of this review we have only considered mutases involved in carbon skeleton rearrange- 0 y43 I --ooc, ooc Scheme 7 Stereochemical course for the coenzyme-B I 2-dependent re-arrangements catalysed by (a)methylmalonyl CoA mutase, (h)isobutyryl CoA mutase, (c)2-methyleneglutarate mutase, and (d)glutamate mutase. *H represents the hydrogen which is abstracted in the rearrangements. Glutamate mutase catalyzes the interconversion of glutamate and (2S,3S)-3-methylaspartate.This reaction proceeds with abstraction of the pro-4S hydrogen and thus differs from the reactions catalysed by either isobutyryl CoA mutase or methylmalonyl CoA mutase (Scheme 7).26 Methyleneglutarate mutase catalyses the interconversion of 2-methyleneglutarate and (R)-3-methylitaconate (2-methylene-3-methylsuccinate). In this reaction the pro-4R hydrogen methyleneglutarate is removed (Scheme 7).26 It has been suggested previously that the stereochemical difference in the proton abstraction for these two enzymes may have arisen due to small changes in the active site geometry with respect to the carboxylate binding functional- ities.26 In other words the stereospecificity of these coenzyme B I 2-dependent rearrangements has diverged.In apparent agree- ment with this proposal is the amino acid similarity between the cobalamin-binding domains of the methyleneglutarate mutase and glutamate mutase.28 An alternative explanation for the diversity in stereospecif- icity for these two enzymes is that it reflects diversity in evolutionary origin. As such, glutamate mutase should be Chemical Society Reviews, 1997, volume 26 341 unrelated to methyleneglutarate mutase, while the latter may be related to isobutyryl CoA and methylmalonyl CoA mutase. Such a proposal might seem to contradict the observed sequence similarity between the cobalamin-dependent domains of these enzymes. However, it has been noted that while these enzymes share a homologous region of about 110 residues in the cobalamin binding domain, there are no homologies outside of this region.2' This domain is located in the C-terminus of 2-methyleneglutarate mutase and methylmalonyl CoA mutase and in a separate subunit of glutamate mutase.The larger catalytic subunit of this latter enzyme shows no significant similarity to any known protein (including methyleneglutarate mutase). These and additional observations have led to the proposal that the different mutases have not diverged from a common ancestor but rather have acquired their cobalamin- binding domain by a gene fusion event.25 Thus this aspect of the stereochemical diversity of mutases, like the dehydratases-hydratase and enoyl thioester reductases, may be driven by a historical contingency in which the stereospecificity does not readily diverge.Proponents of the functional model for stereospecificity in dehydrogenases have contended that stereospecificity can diverge during evolution.2 It has been suggested that changes in the amino acid residues that interact with the nicotinamide ring may allow a 180" rotation around the glycosidic bond and thus alter the ster- eospecificity (Scheme 2). A double mutant of yeast alcohol dehydrogenase has been created to demonstrate that with the appropriate substitutions the enzyme specificity can be from Re specific to Si specific.' The corresponding mutant transferred the pr-o-4R hydrogen of NADH 850000 times for every transfer of the pro-4s hydrogen (as compared to 7 0000000 000 to 1 in the native protein).While there is a dramatic drop in the ratio of the transfers of the different heterotropic hydrogens, it is still clear that the mutant YADH retains significant stereospeci- ficity. No clear evidence that dehydrogenase stereospecificity can change during evolution has been presented thus far. The stereospecificity of the coenzyme B 12-dependent mutases by comparison is very fragile. When methylmalonyl CoA is presented with succinyl CoA containing a heavy isotope at the pro-3R position (the labilisable position) the enzyme will actually alter the steric course and abstract the pro-3S hydrogen.29 A similar situation has been reported for the stereospecificity of the glutamate mutase reaction.3O Thus a simple isotopic substitution can alter the stereospecificities of both of these mutases.Despite this fragility it appears that all methylmalonyl CoA mutases and the related isobutyryl CoA mutases have retained the same stereospecificity during evolu- tion. Similarly there is no evidence to indicate that either the respective stereospecificities of glutamate mutase and 2-methy- leneglutarate have changed. The abstracted hydrogen in a coenzyme B 12-dependent rearrangement is replaced by a migrating group (COSCoA in the case of methylmalonyl CoA mutase). While this review has been limited to discussion of stereospecificity to discrimination between heterotopic hydrogens, it is noted that the related isobutyryl CoA mutase and methylmalonyl CoA mutase both carry out this replacement with retention of configuration, while glutamate mutase and methyleneglutarate mutase do so with inversion of configuration (Scheme 7).26 Thus, only the reactions catalysed by related mutases proceed with identical stereochemical courses.Recently, a number of laboratories have identified a gene meaA from both methylotrophs and streptomycetes that appears to encode a novel coenzyme B 12-dependent mutase. The rneaA gene product shares global homology with both methylmalonyl CoA mutase and isobutyryl CoA mutase.31 While the substrate for this enzyme remains undetermined it is reasonable to predict that the stereochemical course of the rearrangement will exhibit the same characteristics as these mutases.6 Concluding comments There are four interrelated observations that can be drawn from the analyses of the stereospecificities and amino acid sequences of the three classes of enzymes described herein. (a) Enzymes within a specific class that are related at the amino acid level all catalyse their respective reactions with the same stereospecificity . This observation supports the hypothe- sis that once stereochemistry is set it is retained in the process of evolution. (b) If the enzymes within a specific class are unrelated at the amino acid level it is possible that they may catalyse reactions with opposite stereochemistry . However, as there are a limited number of stereochemical courses available for enzymes (a total of eight for enoyl thioester reductases and only two for hydratase-dehydratases) there will be examples where unre- lated enzymes share the same stereochemical course.In either case the stereospecificity is dictated by an historical con-tingency and not a mechanistic imperative. (c) Enzymes that catalyse similar or analogous reactions with different stereochemistries have unrelated amino acid se-quences. The observation that enzymes can converge to catalyse analogous reactions with opposite stereospecificities is incon- gruous with mechanistic imperatives and fully aligned with historical interpretations of the origins of stereospecificity. (d) Although probability predicts that the same stereo-chemical course will be shared by unrelated enzymes [conclu- sion (b)], it appears that enzymes that catalyse analogous reactions with the same stereochemistries are often related.This observation holds true for each of the following: the dehy- droquinate dehydratases, each of the coenzyme B 12-dependent mutases and the enoyl thioester reductases involved in fatty acid biosynthesis. Historical contingencies have now been shown to be the origin of the stereospecificities of a number of enzyme reactions which discriminate between diastereotopic hydrogens. As discussed above the historical model for interpreting ster-eospecificity makes the following assumptions (i) that enzymes from all organisms catalysing the same reaction with the same stereospecificity are homologous, and (ii) that stereospecificity is conserved during evolution.Critics of this model note that in cases where enzymes catalyse the same reactions with opposite stereochemistry the following additions or changes must be made; (1) there are separate ancestral genes for these enzymes or (2) that stereospecificity can diverge for some enzymes and not others. It now appears that the former of these modifications is consistent with the data available to date. It has been over fifteen years since the functional imperative was proposed for dehydrogenase stereospecificity . Most of these interpretations were carried out without any significant protein sequence information. The rapidly growing protein sequence database should now provide an opportunity to revisit this controversial issue especially in the light of recent developments demonstrating the role of historical contingencies for stereospecificity in enoyl thioester reductases (which also use a nucleotide cofactor) and hydratase-dehydratases (where there was a substantially stronger case for a mechanistic imperative).7 References 1 H. F. Fisher, P. Ofner, E. E. Conn, B. Vennesland and F. Westheimer, J. Biol. Chem., 1953, 202, 687. 2 E. G. Weinhold, A. Glasfield, A. D. Ellington and S. A. Benner, Proc. Natl. Acad. Sci. USA, 1991, 88, 8420. 3 K. R. Hanson and I. A. Rose, Acc. Chem. Res., 1975, 8, I. 4 S. A. Benner, Experientia, 1982, 38, 633. 5 N. J. Oppenheimer, J. Am. Chem.Soc., 1984,106, 3032. 6 A. Glasfield, G. F. Leanz and S. A. Benner, J. Biol. Chem., 1990, 268, 11692. 7 R. Bentley, in Molecular Asymetry in Biology, ed. R. Bentley, Academic, New York, 1970, vol. 2, p. 1. 8 R. L. Dorit, L Schoenbach and W. Gilbert, Science, 1990,250, 1377. 9 C. R. Goward and D. J. Nicholls, Protein Science, 1994, 3, 1883. 342 Chemical Society Reviews, 1997, volume 26 10 J. R. Mohrig, K. A. Moerke, D. L. Cloutier, B. D. Lane, E. C. Person and T. B. Onasch, Science, 1995, 269, 527. 11 J. A. Gerlt and P. G. Gassman, J. Am. Chem. SOC.,1992, 114, 5928. 12 A. Schneier, J. Harris, C. Kleanthous, J. R. Coggins, A. R. Hawkins and C. Abell, Bioorg. Med. Chem. Letts., 1993, 3, 1399. 13 J. R. Bottomley, A. R. Hawkins and C.Kleanthous, Biochem. J., 1996, 319, 269. 14 H. Liu, K. K. Wallace and K. A. Reynolds, J. Am. Chem. Soc., 1997, 119,2973 15 V. E. Anderson and G. G. Hammes, Biochemistry, 1984, 23, 2088. 16 H. Bergler, P. Wallner, A. Ebeling, B. Leitinger, S. Fuchsbichler, H. Aschauer, G. Kollenz, G. Hogenauer and F. Tumowskey, J. Biol. Chem., 1994,86,4387. 17 G. Meuer, G. Biermann, A. Schutz, S. Harth and E. Schweizer, Mol. Gen. Genet., 1992, 232, 106. 18 M. C. O’Sullivan, J. Schwab, T. M. Zabriskie, G. L. Helms and J. C. Vederas, J. Am. Chem. Soc., 1991, 113, 3997. 19 S. Chang and G. G. Hammes, Proc. Nutf. Acud. Sci. USA., 1989, 86, 8373. 20 P. Wang, C. D. Denoya, M. R. Morgenstem, D. D. Skinner, K. K. Wallace, R. DiGate, N. Banavali, G. Schuler, M. K. Speedie and K. A. Reynolds, J. Bacteriol., 1996, 178, 6873. 21 K. Arai, B. J. Rawlings, Y. Yoshizawa and J. C. Vederas, J. Am. Chem. Soc., 1989, 111, 3391. 22 C. M. H. Wantanabe, D. Wilson, J. E. Linz and C. A. Townsend, Chem. Biol., 1996, 3, 463. 23 B. S. Moore, H. Cho, R. Casati, E. Kennedy, K. A. Reynolds, J. M. Beale and H. G. Floss, .I. Am. Chem. SOC.,1993, 115, 5254. 24 K. K. Wallace, K. A. Reynolds, K. Koch, H. A. I. McArthur, R. Wax and B. S. Moore, J. Am. Chem. Soc., 1994, 116, 11600. 25 E. N. G. Marsh, BioEssuys, 1995, 17, 5. 26 W. Buckel and B. Golding, Chem. SOC.Rev., 1996, 25, 329. 27 A. Birch, A. Leistner and J. A. Robinson, J. Bucteriol., 1993, 175, 3511. 28 B. Beatrix, 0.Zelder, D. Liner and W. Buckel, Eur. J.Biochenz., 1994, 221, 101. 29 K Wolfe, M. Michenfelder, A. Konig, W. E. Hull and J. Retey, Emr. J. Biochem., 1986, 156, 545. 30 G. Hartrampf and W. Buckel, FEBS Lett., 1984,171, 73. 31 L. V. Chistoserdova and M. E. Lidstrom, Microhiol., 1996, 142, 1459. Receirved, 29th April 1997 Accepted, 30th June 1997 Chemical Society Reviews, 1997, volume 26 343

ISSN:0306-0012    

DOI:10.1039/CS9972600337

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Pentafluorophenylboranes: from obscurity to applications T2N IN4 Pentafluorophenyl substituted boranes and borates are important as co-catalysts in metallocene-based industrial processes for the homogeneous polymerization of olefins. Although first prepared in the early 1960s, the remarkable properties of tris(pentafluoropheny1)borane have only re- cently been exploited for applications in catalysis. Spurred by these developments, the related compounds bis(penta- fluoropheny1)borane and salts of the tetrakis(pentafluor0- pheny1)borate anion have also found application in olefin polymerization and other fields. In this article, we trace the rise of pentafluorophenyl boron compounds from curiosities to important commodities. 1 Introduction and historical background Boron trihalides BX3 (X = F, C1) are quintessential Lewis acids.In both inorganic and organic textbooks they are used as models to explain the concept of Lewis acidity. The extremely moisture-sensitive and volatile nature of these three-coordinate boron compounds, however, makes them difficult to handle. In the 1950s these disadvantages led to investigations of per- fluoroalkyl boranes, e.g. B(Rf)3 (Rf = perfluoroalkyl), in which the highly electronegative Rf ligand was expected to confer strong Lewis acidity on the boron centre and the B-C bonds were predicted to be less hydrolytically sensitive. However, perfluoroalkylboranes exhibit a fundamental drawback viz. thermal instability resulting from fluorine migration to boron with the elimination of a perfluoroalkene or difluorocarbene.Consequently, they can only be obtained in conjunction with n-bonding ligands, e.g. B(CF3)(NMe2)2. Tristram Chivers was born in Bath, England in 1940 and obtained his BSc, PhD (with Professor R. D. Chambers) and DSc degrees )*om the University of Durham. He was a post- doctoral fellow for one year (19644.5) at the University of Cincinnati and returned to the UK as an SRC Fellow at the University of Sussex. He emigrated to Canada in 1967 as a Post-doctoral Teaching Fellow at the University of British Columbia. He moved to The University of Calgary in 1969 and served as department head from 1977-1 982. He was pro- moted to Full Professor in 1978. In 1991, he was elected to a Fellowship in the Royal Soci- ety of Canada.Other awards include the Alcan award of the Canadian Society for Chem-istry (1987) and the Royal Soci- ety of Chemistry award for main group element chemistry (1993). He has been the senior editor of the Canadian Journal of Chemistry since 1993. lristram C’hivers The first reports of pentafluorophenyl-boron compounds date back to the early 1960s when the groups of Chambers and Massey reported the syntheses of X2B(C6F5) (X = ClF), C1B(C6F5)2 and B(C6F5)3. Early work in this area has been reviewed.’ Although the strong Lewis acid properties of B(C6F5)3 were recognised at that time, they were not exploited (vide infra). Indeed CdFS-borane chemistry lay dormant until the 1980s when two examples of the applications of C6F5B reagents were described.Paetzold and co-workers utilized the combination of the strongly electron-withdrawing influence of C6F5 group attached to boron and the kinetically stabilising effect of a bulky But group on nitrogen to generate the first monomeric iminoborane C6F5B=NBuf, which quickly di-merizes at room temperature (eqn. 1). The monomeric imino- ?SF5 1 c6F5 B lBut-C~F~B=NBU~-BUN/\B-N 500-6OO’C 25 ‘C 0NBU‘ (1) CI’ ‘SiMe3 -Me3SiCI \/? borane undergoes cycloaddition reactions with nitriles or isonitriles and representative examples are given in Scheme 1 .* The second application of C6F5B compounds involved the surprising use of B(C6F5)3 as a C6F5-transfer agent in the synthesis of pentafluorophenyl-xenon compounds [eqn.(2)].3 XeF2 + B(C6FS)3 [XeC6F51+ [BF2(C6Fdd- (2) Warren Piers was born in Edmonton, Alberta in 1962 but moved to the Vancouver area where he grew up. He obtained his BSc degree at the University of British Columbia in 1984 and continued there as an NSERC Postgraduate Scholar under the tutelage of Professor Michael Fryzuk, graduating with a PhD in 1988. He then spent two years at the California Institute of Technology as an NSERC and Killam Postdoctoral Fellow with Professor John Bercaw. In 1990, he joined the Chemistry and Biochemistry Department at the University of Guelph as an Assistant Professor. In 1995, he moved back to Alberta to join the Chemistry Department at the University of Calgary as an Associate Professor.Honours awarded include the Province of Ontario’s John C. Polanyi Prize in Chemistry (1991) and an Alfred P. Sloan Foundation Research Fellow- ship (1996-1 998). Warren Piers Chemical Society Reviews, 1997, volume 26 345 But Scheme 1 The transfer of a C6Fs group from boron to another main group element has also been observed very recently in the decomposi- tion of the adduct formed between a bis(amino)silylene and B(C6F5)3. This adduct was prepared to demonstrate the Lewis base properties of the silylene (eqn. 3).4 But But BU' BU' 20'c Bmonths1 Bu' But In the 1990s C6FS-boron compounds emerged from relative obscurity to a position of prominence as a result of a number of seminal publications including the discoveries of (a) the exceptional hydroborating ability of HB(C6Fs)2; (b)the effec- tiveness of B(C6F&, in combination with group 4 metallocene alkyls, as an initiator for olefin polymerisations; (c) the advantages of B(C6Fs)3 as a Lewis acid catalyst for a variety of organic transformations and (4the use of [B(C6Fs)4]- as a non- coordinating counter-ion.In this review the highlights of these recent findings will be discussed, with emphasis on the preceding themes. 2 Bis(pentafluorophenyl)borane,HB(C,jF& 2.1 Synthesis, properties and hydroboration chemistry The monomeric chloroborane ClB(C6F5)2135 is a convenient starting material for preparing multigram quantities of the highly electrophilic secondary borane [HB(C6F&]2.6 Tradi- tional routes to boranes (e.g.reduction with LiAlH4) cannot be employed for this particular borane since it is essential to avoid donor solvents such as diethyl ether or THF, which react with HB(c6Fs), (see below). This problem can be surmounted by using H- sources such as Schwartz's reagent ([Cp2Zr(Cl)H],), stannanes or silanes which are compatible with non-polar hydrocarbon solvents. By far the most convenient reagent for this reaction is Me2Si(C1)H [eqn. (4)]. It is inexpensive, volatile, Me2Si(CI)HCI * (4) -Me2SiC12 96% n easily dried and may be employed as the solvent for conversion of C1B(C6F5)2 to [HB(C6F&I2, which precipitates as a microcrystalline white powder; a simple and satisfying reaction, the yield is virtually quantitative. Although C1B(C6Fs)2 is monomeric, the lack of x-bonding from the hydride ligand imposes a dimeric structure on 346 Chemical Society Reviews, 1997, volume 26 [HB(C6F5)& in the solid state.This is evidenced by the H-B stretch at 1550 cm-1 in the IR spectrum, characteristic of a B-(p-H)z-B function, and an X-ray structural analysis of the molecule.-f- In benzene or toluene solution, the major species is probably also a dimer since an intense peak at 18.0 ppm in the lB { lH}NMR spectrum is in the region typical of neutral, four- coordinate boron centres. A second peak at 60.1 ppm, in the spectral region associated with neutral three-coordinate boron atoms, is strong evidence that [HB(C6Fs)2]2 dissociates to the monomer in dilute solutions (ca.0.02 M) to the extent of ca. 10% in these solvents. The use of [HB(C6F5)2]2 in other common solvents is not advised on the basis of the reagent's reactivity with even weak donors (Scheme 2). Not surprisingly, reaction of HB(C6FS)2 F -PhAD, (C6F5)ZB-O-PhDH:(c6F5)2 Scheme 2 with water or alcohols leads to loss of dihydrogen and production of the borinic acid or esters; to the extent that many polar solvents are difficult to dry, hydrolysis to give the borinic acid is a common occurrence. Loss of water to give the anhydride occurs upon attempted sublimation of the borinic acid. The borane forms strong adducts with Lewis base donors such as trimethylphosphine, THF and diethyl ether.In the case of the ethereal solvents, at temperatures necessary to induce dissociation, C-0 bond cleavage competes effectively with desired additions of HB(C6F5)2 across carbon-carbon multiple bonds. Kinetic studies suggest the reaction with THF is more complex than concerted addition of B-H across the C-0 bond. The rate for the ring-opening of THF is not first order in adduct concentration but dependent only on the initial concentration of the THF-HB(C6Fs)2. Furthermore, the addition of small amounts of HB(C6Fs)2 accelerates product formation. These results imply a rate law of the form shown in eqn. (5)and that rate = kobs [THFaHB(C6FS)21init. [HB(C6FS)21 (5) attack by free HB(C6F&, present in catalytic amounts by dissociation, on the R20-HB(C6F5)2 adduct is the key step leading to products.In support of this model, opening of the selectively labelled styrene oxide shown in Scheme 2 by DB(C6Fs)2, gives the threo stereochemistry in the product, providing strong evidence against concerted addition of B-D across the substituted C-0 bond, which would lead to the erythro product as shown. 1' D. J. Parks, W. E. Piers and G. P. A. Yap, unpublished results. erythro threo The accessibility of a highly electrophilic monomer in non- coordinating solvents imbues this reagent with exceptional hydroboration capabilities, at least in terms of rate, in comparison to more conventional reagents for this well-developed reaction. In all substrate types that have been examined, save one, hydroboration reactions were complete within seconds of adding the organic molecule to a suspension of [HB(C6Fs)2], in toluene.The only class of substrates which does not undergo rapid hydroboration with HB(C6F5)2 are tri- or tetra-substituted olefins with one or more strongly electron- withdrawing substituent. A case in point are the products of internal alkynes monohydroborated with HB(C6F5)2; these vinyl boranes do not add a second equivalent of borane (eqn. 6). In addition to the convenience of high rates, bis(pentaflu0- ropheny1)borane offers comparable or better regio- and chemo- selectivity than other hydroboration reagents. For example, styrenes are hydroborated exclusively to the 2-borylethyl isomers and terminal alkynes can selectively be mono-hydroborated with HB(C6F& (Scheme 3); other HBRR' boration of pendant unsaturated functions.8 During the course of these studies it became apparent that the borane reagent reacts with zirconium-carbon bonds at rates competitive to hydroboration; consequently, we examined the reactions of HB(C6F5)2 with Cp2ZrR2 (R = CH3,g CH2C6H5, CHZSiMe3) in some detail.Two competing reaction paths were observed (Scheme 4). Alkyl-hydride exchange between boron and zirconium pro- duced 'Cp;?ZrR(H)' and RB(C6F5)2; the zirconocene alkyl hydride complexes immediately consumed another equivalent of HB(C6F5)2 yielding stable dihydrido-bis(pentafluor0-pheny1)borates as the zirconium-containing products.This pathway dominated when R was sterically bulky. The second mode of reactivity involved Lewis acid-induced loss of RH and production of borane-stabilized alkylidene derivatives. This path is related to that followed in the reaction between Cp2TiMe2 and A1Me3, leading to the well known CpZTi=CH2 synthon, Tebbe's reagent. c{ '"cp ' PSF5 F6F5 HYB(c6F5)2R t Scheme 3 reagents cannot generally hydroborate terminal alkynes without a certain amount of dihydroboration. Although selective monohydroboration with HB(C6F5)2 is possible, treatment of terminal alkynes with two equivalents of reagent does lead rapidly to products of dihydroboration, the bidentate Lewis acids RCH2CH[B(C6F5)2]2 (eqn. 7). Such compounds have potential application in the context of olefin polymerization, discussed in more detail below.The use of HB(C6FS)Z offers a convenient route to this family of Lewis acids without the side reactions encountered in other synthetic protocols.7 2.2 Reactions of HB(C6Fs)z with dialkyl zirconocenes The original impetus for making HB(C6F5)2 was to employ it to incorporate highly Lewis acidic boron centres into the ligand framework of early transition metal metallocenes via hydro-Scheme 4 The borane-stabilized methylidene complex arising from reaction of Cp2ZrMe2 and HB(C&& in hexanes is a highly reactive species. In the absence of coordinating Lewis bases, the material decomposes rapidly in benzene solution. Furthermore, the compound reacts with another equivalent of HB(C,FS), to produce an intriguing complex in which the Zr=CH2 unit is complexed by two HB(C6F5), fragments.An X-ray crystal structure of this compound (Fig. 1) revealed that bonding contact between the methyleqe carbon and zirconium is maintFined, [Zr-C = 2.419(4) A compared with a distance of 2.27 A in Cp2Zr(CH3)2] rendering this carbon five-coordinate. The bonding in this compound is viewed to occur between the well established frontier orbitals of the Cp2Zr2+ fragment and the dianionic ligand [(CsFs)2B(H)CH2B(H)(C6F&]2-, essen-tially a chelating hydrido borate species. An extended Huckel treatment of an idealized model for the compound provided a rationale for the unusual bonding environment about the Chemical Society Reviews, 1997, volume 26 347 methylene carbon.Due to the greater electronegativity of carbon vs. boron, significant charge localization on the methylene carbon in the free ligand is available for donation to the vacant 2a1 orbital on the zirconium centre.1° Fig. 1 ORTEP representation of the molecular structure !f Cp2Zr(+ CH(C6H5)[(pH)B(ChF5)2]}. Selected bond distances (A): Zr-C( l), 2.341( 10); Zr-C(2), 2.583; Zr-C(3), 2.737; Zr-HB, 2.02(7); C(l)-C(2), 1.456(12); C(1)-B, 1.54S( 15); B-HB, 1.35(8). Selected bond angles ("): Zr-C( 1)-B, 87.6(6); Zr-C( l)-C(2), 82.2(6); C(2)-C( 1)-B, 133.5; Zr-HB-B, 107(4); C( 1)-B-HB, 102.6; C( l)-C(2)-C(3), 117.3(9); C( l)-C(2)-C(7), 126.3(9). Reprinted with permission from Angew. Chem., Znt.Ed. Engl., 1995,34, 1230. Related compounds that also contain ligands with penta- coordinate carbons were formed in the reactions of phosphine- stabilized zirconocene olefin compounds (eqn. 8).11 One R R = H, CH2CH3, C6H5 Cp2Zr~BrC6F5)2(8)H. R equivalent of borane complexes the phosphine ligand while the second adds across one of the Zr-C bonds of the coordinated olefin. 13C NMR data and an X-ray structure for the compound where R = H established that bonding contact between Zr and Cp was maintained in the products. Chemically intriguing as these compounds are, they are characterized by rather strong metal-borate interactions which dampen further reactivity. While there is considerable potential for the application of HB(C6FS)z in organic synthesis and catalytic olefin polymerization, as we see in the next section, the related, fully substituted tris(pentafluorophenyl)borane, B(C6F5)3, has already had a tremendous impact in both of these arenas.3 Tris(pentafluorophenyI)borane,B(C6F& 3.1 Properties Tris(pentafluoropheny1)borane was first reported in 1964 by Massey and Park.' Their papers on the compound and its chemistry were part of a series entitled PerfluorophenylDerivatives of the Elements and, as the title suggests, were 348 Chemical Society Reviews, 1997, volume 26 primarily focused on the fundamental chemistry associated with boron compounds with fluorinated substituents. After the early Massey and Park publications, interest in B(C6Fs)3 waned until very recently when Ewen12 and Marksl3 independently showed that this Lewis acid is a very effective initiator for olefin polymerization reactions when combined with group 4 metal-locene alkyls.Citations of the early Massey and Park papers (excluding self-citations) numbered only 25 in the 25 years after their publication; since 1990, these papers have been cited in over 60 papers and at least 57 patents have been granted for processes which employ B(C6F5)3 (Fig. 2). Figures for the industrial usage of B(C6F5)3 are difficult to obtain but several companies now sell the material (cost is ca. $4500 per kg). Citations of J. Organomet. Chem. 1964, 2, 245 Citations of J. Organomet. Chern. 1966,5. 218 301 Q Patents UsingTris-(Pentafluoropheny1)borane 1965-69 1970-79 1980-85 1986-89 1990-94 1995-96 Fig.2 Citations of Massey and Park's original papers on B(C6F& and patents using tris(pentafluoropheny1)borane Tris(pentafluoropheny1)borane was originally prepared by treatment of pentafluorophenyllithium with BC13 at -78 "C. Perhaps a safer alternative, especially for larger scale syntheses, is to employ the Grignard reagent, pentafluoro-phenylmagnesium bromide, and a boron trihalide, since the lithium reagent has a tendency to eliminate lithium fluoride explosively at ambient temperatures. The Grignard reagent, on the other hand, can be refluxed in toluene without incident. Compared to the boron trihalides, tris(pentafluor0-pheny1)borane is in many ways an ideal boron-based Lewis acid, possessing both high Lewis acid strength and stability. It is remarkably thermally robust; while many fluorinated boron compounds tend to decompose by eliminating BF3, B(C6Fs)3 is stable for several days at temperatures up to 270 "C.It is also resistant towards oxidation by molecular oxygen and is water tolerant. Hydrolysis to eliminate C6F5H does occur, but only very slowly; in fact, some organic reactions catalysed by B(C6F5)3 can be carried out in aqueous solution. The compound forms a strong and stable adduct with water which can be isolated as a trihydrate (H20)2H20.B(C6F5)3.l4 In terms of Lewis acid strength, it was surmised by Massey and Park that B(C6F& is a slightly better Lewis acid than BF3 but not as strong an acceptor as BC13, as judged by spectroscopic comparisons between amine adducts of a number of these boron-based Lewis acids.Finally, it can be prepared base-free (although it is extremely hygroscopic in this form) and it is soluble in many non-coordinating organic solvents. 3.2 Reactions of B(C,F& with zirconium alkyls These properties led Marks and Ewen to test B(C6F5)3 as a co- catalyst for group 4 metallocene-based homogeneous olefin polymerization systems. In the late 1980s it became increas- ingly apparent that the role of the traditional co-catalyst for this reaction, a complex species prepared from trimethylaluminium and water known as methylaluminoxane (MAO), was to serve as both an alkylating agent and a powerful Lewis acid capable of abstracting a methyl group from the metal centre and creating the active cationic organometallic species responsible for olefin enchainment. Thus, it was predicted that treatment of dimethyl zirconocenes ('pre-alkylated' catalyst precursors) with a Lewis acid more powerful than the resulting Cp2Zr+CH3 cation would lead to highly active olefin polymerization catalysts.Tris(penta- fluoropheny1)borane was tagged as an excellent candidate for playing this role. This expectation turned out to be well-founded. As shown in eqn. (9), treatment of a variety of dimethyl zirconocenes with + or pentane Cp' = q5-C5H5 B(c6F5)3 q5-l ,2-Me&H3 q5-CsMes q5-l,3-(SiMe3)2C5H3 B(CSF5)3 leads to what Marks described initially as 'cation-like' compounds of general formula [Cp',ZrCH3]+[CH3B(C6F5)31-[Cp' as defined in eqn.(9)]. Several products were structurally characterized, providing insights into the nature of ion-ion interactions in these compounds. 1s In most instances, the ions are linked to one another through an unsymmetrical bridging methyl group most strongly associated with the boron atom; an ORTEP diagram of Cp*2Zr'CH3.-H3CB-(C6F5)3 (Cp* = CsMe5) is shown in Fig. 3 as a representative example. The Zr-v-CH3 ?eparations range from 2.549(3) to 2.640(7) and are ca.0.3 A longer than terminal Zr-C bonds. The p-CH3- B distances, on the other hand, are normal, illustrating the ionic character of these compounds. In structures where the p-CH3 hydrogens were located crystallographically.C-H agostic type interactions between two of the CH bonds were implicated while the near linearity of the ZPV-H~C-B vector, which ranges from 161.8" to 176.6O, suggests a significant dipole-dipole component to the ion-ion interaction in these compounds. The static situation arrested in the solid state structures of these compounds is only part of the story. In solution, a 'tug-of- war' between the opposing Lewis acids Cp2Zr+CH3 and B(C6F5)3 for the methide group is being waged. Through careful analysis of the NMR spectra of the complexes (1,2-Me2C5H3)2M+CH3e-H3CB-(C6F5)3Zr, Hf) at(M = various temperatures, Deck and Marks were able to distinguish two dynamic processes that exchange the methyl groups in the metallocene wedge and the ring methyl substituents (Scheme 5).*6 The first exchanges only the diastereotopic ring sub- stituents and involves separation of the anion from the cation (i.e.cleavage of the Zr-p-H3C interaction), flipping of the terminal Zr-CH3 group from one side of the metallocene girdle Ion pair symmetrization B(CbF& Dissociation/reabstraction F(2) Fig. 3 ORTEP representation of the molecular structure of Cp*,Zr+CH3-H3CB-(C6F&. Selected bond distances (A):Zr-C(2 I), 2.223(6); Zr-C(40), 2.640(7); B-C(40), 1.66( 1). Selected bond angle: Zr- C(40)-B, 176.6(4). Reprinted in part with permission from J. Am. Chem. Soc., 1994, 116, 10 015. Copyright 1994 American Chemical Society. to the other followed by reassociation of the ion pair. This ion- pair symmetrization process is accelerated in more polar solvents and dominates for M = Zr.The second distinct process permutes both the diastereotopic ring sites and the bridging/ terminal zirconocene methyl groups. As shown in Scheme 5, this exchange process involves initial dissociation of free B(C6F5)3 through P-H~C-B bond cleavage followed by abstrac- tion of the other zirconocene methyl group. Interestingly, this dissociation-reabstraction operation is the predominant process for M = Hf. Probably the comparative M-C bond strengths, which are ca. 10 kcal mol-I (1 cal = 4.184 J) greater for M = Hf, govern the relative importance of these two processes in cation-like compounds formed upon treatment of CpzM(CH3)Z with B(C6F&. These studies provided the first quantitative data for ion-pairing in a metallocene catalyst, a very important contribution since it is becoming clear that the extent of such interactions has a profound effect on the properties of the catalyst.The seminal work by Marks described above has spurred several researchers to examine reactions of other zirconium alkyl compounds with B(C6Fsj3, mostly with a view towards the Scheme 5 Chemical Society Reviews, 1997, volume 26 349 development of olefin polymerization catalysts (Scheme 6). Alkyl removal is not limited to methyl groups; benzyl ligands are readily abstracted, often producing zirconium cations stabilized by multihapto benzyl ligands as in the example (a) discovered by Bochmann et a1.'7 Methyl group abstraction from the more sterically encumbered complex Cp3ZrCH3 is also facile and leads to the [cp3Z~][[cH3B(c6Fs),]- system, which decomposes upon attempted isolation (b).l8 It may, however, be trapped with donor ligands including carbon monoxide; the resulting product is an example of a stable 'non-classical' [v(CO) = 2150 vs.2143 cm-1 for free CO] o-only carbonyl complex. Acetylide groups are also susceptible to removal, although the initial product undergoes further reaction by insertion of the abstracted CzC triple bond into the cationic acetylide complex, forming the zwitterionic species shown in reaction (c)." Zwitterionic complexes that are active olefin polymerization catalysts are accessible from reactions of other organozircon- ium derivatives with B(C6F5)3 (d-0.The zircona-cyclopentadiene complex of example (d) does not react by abstraction of a zirconium-carbon (3 bond; rather, the borane attacks the n-system of one of the Cp donors, inducing a proton transfer from the Cp ring to the 0-bound organic ligand in the metallocene girdle. The resulting cationic centre is stabilized through a dative fluorine to zirconium linkage, a type of ion-ion interaction which has been observed in other systems as well. Attack of x-coordinated ligands bound to zirconium in the metallocene wedge also produces structurally interesting zwit- terionic species. The zwitterion formed from the zirconocene butadiene complex (e) is again stabilized by a CF-+Zr interaction.20 A different family of zwitterionic compounds is formed upon reaction of 'tuck-in' zirconocenes with 0\A F F FRFCP* FyB-F ,,,/c F f F B(C6Fs)3.21 In example (f), the abstracted methylene group, which bears a fractional negative charge, remains in close contact with the zirconium centre to at least partially com-pensate its electrophilicity.A priori, zwitterionic catalyst systems offer several potential advantages over prevailing metallocene systems. Ideally, they are single component systems which do not require further co- catalyst to reach an active state. Thus, they provide the advantages of neutral single component systems while retaining the activity levels of cationic group 4 catalysts.If properly designed, activity damping ion-ion interactions can be weak- ened, leading to more active catalysts. For example, the intramolecular ion-ion interaction in the product of reaction (f) is substantially weaker than the intra-ion p-methyl interactions discovered by Marks et al. (vide supra).]4Furthermore, the solubility properties of zwitterionic compounds are potentially superior to those of non-zwitterionic catalysts. Although significant advances have been recently disclosed regarding the design and synthesis of zwitterionic catalysts, careful polymeri- zation studies to evaluate the advantages of such catalysts over conventional metallocene catalyst systems have yet to be done. 3.3 Reactions of B(C6F& with other transition metal alkyls If B(C6Fs)3 can remove alkyl groups from Lewis acidic metals like zirconium, it should be capable of abstracting alkyls from less Lewis acidic metals to generate reactive cationic species.This premise has not been tested extensively, although reports along these lines have begun to appear. Puddephatt et al. showed22 that platinum(I1) cations can be generated via methide II 1. CpsZrCH32x C6F.5 -F Scheme 6 350 Chemical Society Reviews, 1997, volume 26 abstraction from the complex [(dbbipy)Pt(CH&] (dbbipy = 4,4'-di-tert-butyl-2,2'-bipyridine) [eqn. (1 O)]. The cations were trapped as ethylene or carbon monoxide adducts and they are closely related to Brookhart's nickel and palladium-based cations, which are an important new class of ethylene polymerization catalyst.23 Brookhart's method of generating these cations employs the acid [HOEt*]+[B(Arf)4] -, where Arf = 3,5-(CF&C6H3; Puddephatt's results suggest that B(C6F& may also be a suitable initiator for these catalysts.In another study, Green et found that the reactions of B(C6F5)3 with metal-methyl compounds can be much more complex than simple methide abstraction. For reaction with Cp(C0)2FeCH3, the preferred site of attack by the borane appears to be one of the carbonyl ligands. Formation of the final observed product [eqn. (ll)] was accompanied by loss of fluoroboranes FnB(C6F5)3 --n and likely arose from a multistep process involving Lewis acid-induced CO migratory insertion followed by rearrangement of this initially formed product and the insertion of a B(C6F&-derived C6F4 fragment into the iron- acyl carbon bond.This reaction illustrates that B(C6F5)3 can be a source of C6Fs groups to transition metals as well as main group elements (see Introduction). Indeed, transfer of C6F5 from boron to zirconium does occur, albeit slowly, in the cation- like complexes discussed above. l5 Half sandwich group 4 alkyl derivatives also react with B(C6F5)3 to produce catalyst systems capable of polymerizing olefins. Perhaps the most interesting system is that derived from Cp*Ti(CH3)3 (Cp* = pentamethylcyclopentadienyl) and B(C6F5)3 and studied by the groups of Baird2s and Zambelli26 An equimolar mixture of these two constituents yields the active species most often depicted as the loosely associated complex shown in eqn.(12), although intimate structural details are not as plentiful as those for the bis(Cp)metallocene systems. Dissociation of this ion pair into discrete ions probably occurs to a degree, depending on the medium. The olefin chemistry associated with this complex is extensive. Reactions with olefins such as ethylene and propyl- ene are generally accepted to proceed viamechanisms similar to those found in metallocene chemistry. The mechanism(s) involved in styrene polymerizations, on the other hand, have been the subject of some controversy. Baird et aE. originally proposed that the primary mechanism involved in the produc- tion of syndiotactic polystyrene with this catalyst initiator was carbocationic rather than the coordination polymerization mechanism common to metallocene systems.25 This conclusion was based partially on an observation which resulted from improper post-treatment of the polymer and was subsequently shown by both the Baird and Zambelli groups to be in However, there is convincing evidence that the Cp*Ti(CH&- B(C6F5)3 system can serve as a carbocationic initiator for the production of atactic polystyrene in reactions carried out at lower temperatures.27h Furthermore, for monomers more com- monly polymerized by cationic mechanisms (e.g.vinyl ethers and N-vinylcarbazoles), the CP*T~(CH~)~-B(C~F~)~ complex indeed serves as a carbocationic initiator for their polymeriza- tion.28 Clearly, this catalyst system is a most versatile polymerizing agent.A final example of methide abstraction occurs not from a transition metal but from the complex Cp2AlCH3 to produce the aluminocenium cation [Cp2Al]+[MeB(C,F,),]- [eqn. ( 13)], a potent cationic initiator for isobutylene polymerization.29 This is a remarkable reaction because it suggests that B(C6F5)3 is, at least under these conditions, a more powerful Lewis acid than the cationic aluminium species. The aluminocenium compound was found to be stable in both dichloromethane and toluene solutions at low temperatures, under which conditions it rapidly polymerized isobutylene to high molecular mass polymers with narrow pol y dis pe rsi t ies .These diverse examples of methide abstraction by B(ChF5)3 are portents of a nearly unlimited future for exploration of reactions of this type. Although there are a few alternative reaction paths possible (such as C6F5 transfer reactions), in most cases, highly reactive cationic metal species are obtained. 3.4 Lewis acid catalysis with B(C6F& Many organic transformations are catalysed by Lewis acids such as BF3, AlC13 or SnC14. While effective, these catalysts suffer from some serious disadvantages. They are often most conveniently handled as adducts with a weak Lewis base, which attenuate the catalyst's Lewis acidity; they are extremely water- sensitive, producing hydrogen halides harmful to acid-labile substrates; the element-halogen bonds are reactive towards some potential substrates and therefore limit their usage.By contrast, B(C6F& is water-tolerant and has generally un-reactive B-C bonds. Given the rise to prominence of B(C6FS)3 in recent years, as well as its commercial availability, it is not surprising that researchers have begun to examine its effective- ness as a Lewis acid catalyst for a variety of organic reactions. The most comprehensive studies have been carried out by Yamamoto and co-workers, who have surveyed a variety of Lewis acid-catalysed organic reactions employing B(C6F& as the catalyst (Table 1).3" Specifically, they demonstrated several examples of B(C6F5)3 catalysed addition of silyl enol ethers to aldehydes, alkyl chlorides and a,@unsaturated ketones (entries 14).Various aldol type reactions of ketene silyl acetals with N-benzylimines were also effectively catalysed by B(C6Fs)3. In a comparative study, the authors also showed B(C6F& to be superior to conventional Lewis acid mediators of this reaction. Less extensively examined were allylsilation and Diels-Alder reactions, although the examples investigated gave impressive results. In general, reactions proceeded at convenient rates with low catalyst loadings under extremely mild conditions. The catalyst is water-tolerant but reactions proceed more efficiently if anhydrous borane is employed (entry l), presumably because the substrate need not displace water from the borane.Although not necessary for activity, if water is rigorously excluded, Chemical Society Reviews, 1997, volume 26 351 Table 1 Lewis acid catalysed organic reactions using B(C6F&n Si(C1)H no reaction was observed. These observations sug- gested that a different hydrosilation pathway from that predicted E Reactants Catb Product tlYieldc was responsible for the observed turnovers. Some form of % (after work-up) h/% Lewis acid catalysis seemed the only reasonable alternative, so the reaction was carried out using B(CGF& as the catalyst, with OH 0 excellent results (Table 2). ’ 6 2’td up,6/96Ph Ph 12/84 Table 2 Tris(pentafluoropheny1)borane catalysed hydrosilation of aromatic carbonyl functions E Substrate Cat Product TON Yield Yo % xs$ Bub NH 0 8/99 Bz .uOBut22/89 Ph 6’ 6 phAsiMe35 phu2/84 A variety of aromatic aldehydes, ketones and esters were ACHO07‘ 5 12199s CHO u Taken from Ref.30. Unless otherwise indicated, non-anhydrous B(C6F5)3was used. TimeEsolated yield. Anhydrous grade B(C6F5)3 used. e Work-up without acid. f One example reported. x exo:endo = 88: 12. F Scheme 7 products sensitive to acid can be isolated (entry 2) and acid- labile substrates can be activated (entry 5). Tris(pentafluoropheny1 jborane is also a very effective cat- alyst for the hydrosilation of carbony131 functions . Initially, it was expected that HB(C6F5)2 would serve as a mediator of this exothermic reaction via the pathway depicted in Scheme 7.When a 1 : 1 mixture of acetophenone and Me2Si(CIjH was subjected to 10 mol% of HB(C6;& in [2H6Jbenzene, hydro- silation of the substrate was observed, albeit with a slow turnover frequency. Surprisingly, separate experiments showed that when compounds ROB(C6F5)2 were treated with Me2- 352 Chemical Society Reviews, 1997, volume 26 hydrosilylated efficiently in the presence of 2 mol% B(C6F5)3 using one equivalent of Ph3SiH as the reducing agent; isolated yields of the silyl ethers were excellent. In the presence of more than one equivalent of silane, further (albeit slower) reduction of the carbonyl substrate was observed. Ester reductions were very rapid at room temperature and > 90% selective (vs.over-reduction) for the acetal products which could subsequently be converted to aldehydes; other products observed were ethers and silyl ethers stemming from competitive deoxygenation of the acetal products during the later stages of the reaction.Methods for selective reduction of esters to aldehydes under such mild conditions are rare and potentially valuable. An intriguing and counterintuitive mechanism appears to be operative in this hydrosilation reaction. In other Lewis acid mediated hydrosilation reactions, it is normally assumed that the role of the Lewis acid lies in activation of the carbonyl function via coordination with an oxygen lone pair for subsequent reaction of the adduct with silane to effect hydrosilation. Kinetic experiments on the B(C6Fs)3-catalysed hydrosilations, however, showed that the substrate most basic towards B(C6F5)3 (benzaldehyde) was reduced the slowest and the least basic substrate (ethyl benzoate) was hydrosilated the fastest.Furthermore, substrate dependence studies showed that higher concentrations of carbonyl compound inhibited the hydrosilation reaction. These results indicate that the borane activates the silane rather than the carbonyl substrate in the key step of the hydrosilation mechanism. Although equilibria strongly favour the formation of RC(X)=O.B(C6F5)3 adducts, the small amounts of free borane present interact with silane; free carbonyl substrate then reacts with the borane-silane complex to effect hydrosilation. Thus, the puzzling observation that less basic ester substrates are hydrosilated faster than aldehydes or ketones is accommodated. 4 Tetrakis(pentafluoropheny1)borate as a non-coordinating anion The success of B(C6F& as a co-catalyst capable of generating ‘cation-like’ metallocene compounds is to some degree tem- pered by the interactions (albeit weak) between the abstracted methyl group and the zirconium centre To eliminate this form of ion-ion contact, the reagent triphenylcarbenium tetra-ki\(pentafluorophenyl)borate may be employed as an abstract- ing agent for activating metallocenes Conveniently synthesized from LIB(C~F~)~ 1 and Ph3CCl,3* this reagent rapidly removes methyl groups from zirconium based metallocenes to produce highly active catalysts [eqn (1 4)] The tetrakis(pentafluor0- Cp’2M(CH3)2 + [Ph3CI”B(C6Fs>41---.$ [CP’~MCH,]+[B(C~F~)~]-+Ph3CCH3 (14) M = Ti, Zr, Hf Cp’ = a cyclopentadienyl donor pheny1)borate ion is amongst the least coordinating counter- anions in the pantheon of such species available and it is this low basicity to which high polymerization activities may be attributed The structure of [Cp*2ThCH3lf[B(C6F5)4]- reveals only tenuous interactions between two fluorine atoms of one Cc,F5 ring and the thorium centre 33 For example, the closest Th-F contacts are outside calculated distances based on atomic radii, and the corresponding C-F distances are normal Furthermore, all four C6F5 rings were found to be magnetically equivdlent by I9F NMR spectroscopy, even at low temperatures Thir compound exhibited significantly higher activities towards ethylene polymerization and hex- 1-ene hydrogenation than [Cp*zThCH3]+IB(C6H5)4]- whose anion is more coordinat-ing Spurred in part by these innovations in metallocene ion generation, the [Ph3C]+[B(C6F5)4]- reagent has recently reju- venated the rearch for bonu fide silylium cations in the condensed phase 34 As Lambert’s review conveys, this area of silicon chemistry has a long and colourful history, which has at times been controversial 35 While the existence of R?Si+ has been confirmed in the gas phase, many attempts to generate there species in solution have failed or been the subject of considerable debate To the extent that generation of a planar, three-coordinate silylium ion in the condensed phase is likely unattainable, the use of the [B(C6F5)4]- anion has allowed for the isolation and characterization of compounds with relatively high silylium ion character For example, treatment of EtTSiH with [Ph7C]+[B(C6F5)4]-in toluene leads to the toluene solvated \pecies shown in eqn (15) The crystal structure of the Et [B(C6F5)41 Et‘--SI+-Et E~~SIH+ [Ph3C]’[B(C6F,),] -(15)1 Qlproduct36 revealed that the ‘[Et,Si]+’ fragment was completely separated from the borate counterion, however, one molecule of toluene coordinates to the silicon and largely delocalizes the positive charge away from silicon 2%i NMR studies suggest that other members of this family of compounds, with bulkier groups on the silicon, are solvated to a lesser degree as evidenced by downfield shifted resonances The [B(C,F,),]-ion is therefore a very effective non-coordinating anion-it is less basic than even toluene towards the silicon centre While sufficiently non-coordinating anions are available, it is impossible to circumvent solvation com- pletely.whether attempting to generate free silylium ions or naked metallocenium ions in 3olution The debate continues on the legitimacy of describing these compounds as silylium ions, but clearly the tetrakis(pentafluoropheny1)borate ion has al- lowed researchers to come closer to such species than ever before In solution chemistry, perhaps the best measure of such things is made by observation of the comp~und~s chemical behaviour Lambert and Zhao, for example, have shown recently that reaction of [Et3Si C6Hs]+[B(C6F5)4]- with 1,1-diphenylethylene leads to the stable carbocation shown in eqn (16),37demonstrating that the solvated silylium compound does serve at least as a synthetic equivalent of ‘Et3Si+’ \ 5 Concluding remarks The chemistry of pentafluorophenyl substituted-boron com-pounds remains an intensely active area of research in both academic and industrial laboratories around the world For the moment, the major area of application is in the area of olefin polymerization, but the potential for the use of these compounds in Lewis acid-catalysed processes is enormous given their availability in industrial quantitities Further applications for HB(C6F5)2 and B(C6F5)? as well as the development of new highly electrophilic boranes3s and non-coordinating borate\39 will continue to be the focus of ongoing and future investiga- tions The emergence of tris(pentafluoropheny1)borane from obFcu- rity to an industrially important commodity provides d nice illustration of how curiosity-driven research can lead to important applications It also shows how these applications can take many years to develop and why they require the contributions of dozens of individual researchers to germinate Who could have predicted the emergence of B(C6F5)3 as an important component of polyolefin producing catalysts at the time it was first reported? Similarly, who can fore\ee what technologies will emanate from today’s cunosity-driven inves- tigations? The expectations of society notwithstanding, not all basic research can lead to useful technologies but, as a whole, the enterprise of fundamental research worldwide can lead to success stories such as that of the pentafluoropheny 1 borane5 6 References 1 (a)R D Chambers and T Chivers Oigunomet Chrm Re1 1966 1 279 (h)S C Cohen and A G Massey Ad1 Fliiorrne Chmi 1970 6 149 2 C Klofkom M Schmidt T Spaniol T Wagner 0 Cownor and P Paetzold Chenz Bet 1995 128 1037 3 (LI)D Naumann and W Tyrra J Cheni SOL Cheni Conlriiidn 1989 47 (h)H J Frohn and S Jakobs J Cheni Soc Cheni Coniniuri 1989 625 4 N Metzler and M Denk Chem Comnzun 1996 2657 5 R E v H Spence W E Piers L R MacGillivray dnd M J Zawor otko Acta Ciystalloqi Sect C 1995 51 1688 6 D J Parks R E v H Spence and W E Piers Anqm Chfm /nt Ed Eiiql 1995 34 809 7 L Jid X Yang C Stem and T J Mark\ Oiqunonicrollrt 7 1994 13 3755 8 R E v H Spence and W E Piers 01qanonietullic 5 1995 14 4617 9 R E v H Spence D J Parks W E Piers M MdcDondld M J Zaworotho and S J Rettig Anyew Cheni lilt Ed EnqI 1995 34 1230 10 U Radius S J Silverio R Hoffmdnn and R Cleiter 01qatiomc tullrt r 1996 15 3737 II Y Sun W E Piers and S J Rettig Oiqanometallics 1996 15 41 10 Chemical Society Reviews, 1997, volume 26 353 12 J.A. Ewen and M. J. Edler, CA-A, 1991, 2, 027; 145. Chem. Abstr., 1991,115, 136998g; 2.56893. 13 X. Yang, C. L. Stern and T. J. Marks, J. Am. Chem. Soc., 1991, 113, 3623.14 A. R. Siedle, R. A. Newmark and W. M. Lamanna, Organometallics, 1993, 12, 1491. 15 X. Yang, C. L. Stem and T. J. Marks, J. Am. Chem. Soc., 1994, 116, 10 015. 16 P. A. Deck and T. J. Marks, J. Am. Clzem. Soc., 1995, 117, 6128. 17 M. Bochmann, S. J. Lancaster, M. B. Hursthouse and K. M. A. Malik, Organometallics, 1994, 13, 2235. 18 T. Brackemeyer, G. Erker and R. Frohlich, Organometallics, 1997, 16, 531. 19 B. Temme, G. Erker, R. Frohlich and M. Grehl, Angew>. Chem., Int. Ed. Engl., 1994, 33, 1480. 20 B. Temme, J. Karl and G. Erker, Chem. Eur. J., 1996, 2,919. 21 Y. Sun, R. E. v. H. Spence, W. E. Piers, M. Parvez and G. P. A. Yap, J. Am. Chem. Soc., 1997, 119, 5132. 22 G. S. Hill, L. M. Rendina and R.J. Puddephatt, J. Chem. SOC.,Dalton Trans., 1996, 1809. 23 L. K. Johnson, S. Mecking and M. Brookhart, J. Am. Chem. Soc., 1996, 118, 267. 24 M. L. H. Green, J. Haggitt and C. P. Mehnert, J. Chem. Soc. Chem. Commun., 1995, 1853. 25 R. Quyoum, Q. Wang, M.-J. Tudoret, M. C. Baird and D. J. Gillis, J. Am. Chem. Soc., 1994, 116, 6435 and references therein. 26 C. Pellecchia, P. Longo, A. Proto and A. Zambelli, Makromol. Chem., Rupid Commun., 1992, 13, 265. 27 (a) C. Pellecchia, D. Pappalardo, L. Oliva and A. Zambelli, J. Am. Chem. Soc., 1995, 117, 6593; (b) Q. Wang, R. Quyoum, D. J. Gillis, M. J. Tudoret, D. Jeremic, B. K. Hunter and M. C. Baird, Organome-tallies, 1996, 15, 693. 28 Q. Wang and M. C. Baird, Macromolecules, 1995, 28, 8021. 29 M. Bochmann and D. M. Dawson, Angew. Chem. Int. Ed. Engl., 1996, 35, 2226. 30 K. Ishihara, N. Hanaki, M. Funahashi, M. Miyata and H. Yamamoto, Bull. Chem. Soc. Jpn., 1995, 68, 1721. 31 D. J. Parks and W. E. Piers, J. Am. Chem. SOC., 1996, 118,9440. 32 J. C. W. Chien, W. M. Tsai and M. D. Rausch, J. Am. Chem. Soc., 1991, 113,8570. 33 X. Yang, C. L. Stem and T. J. Marks, Urganometallics, 1991, 10, 840. 34 J. B. Lambert, L. Kania and S. Zhang, Chem. Rev., 1995,95, 1 191. 35 See M. Arshadi, D. Johnels, U. Edlund, C.-H. Ottosson and D. Cremer, J. Am. Chem. Soc., 1996, 118,5120 (and references therein) for a taste of the discussion. 36 J. B. Lambert, S. Zhang, C. L. Stern and J. C. Huffman, Science, 1993, 260, 1917. 37 J. B. Lambert and Y. Zhao, J. Am. Chem. SOC.,1996, 118, 7867. 38 Y.-X. Chen, C. L. Stem, S. Yang and T. J. Marks, J.Am. Chem. Soc., 1996, 118, 12 451. 39 L. Jia, X. Yang, C. L. Stem and T. J. Marks, Organometallics, 1997,16, 842. Received, 7th April 1997 Accepted, 14th May 1997 354 Chemical Society Reviews, 1997, volume 26

ISSN:0306-0012    

DOI:10.1039/CS9972600345

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Molecular modelling of electron transfer systems by noncovalently linked porphyrin-acceptor pairing In the respiratory system, electron carriers such as ubiqui- none and cytochrome c play an important role in the electron transfer (ET) reaction between oxidoreductases embedded in the mitochondrial membrane. This review focuses on a strategy for constructing porphyrin-electron acceptor pairs via specific interactions and the evaluation of ET in these systems. Particularly, the molecular recognition of ubiqui- none analogues by porphyrin host molecules and the mimicking of protein-protein complexation using myoglo- bin reconstituted with a synthetic porphyrin are reviewed, and the characteristics of the photoinduced ET within noncovalently linked complexes are discussed.1 General introduction Electron transfer (ET) reactions in biological systems provide one of the most fascinating fields of current interdisciplinary research. It is well known that ET occurring within mitochon- dria and chloroplasts plays a crucial role in respiratory oxidative phosphorylation and photosynthesis, in which many oxido- reductases are embedded in the membrane and moveable electron carriers are organized. These two systems produce the high energy ATP through multi ET reactions and proton flux by action of the membrane-bound enzymes. Therefore, it is of particular interest for chemists and biochemists to elucidate the ET systems, both experimentally and theoretically. Although, in general, the natural system is quite complicated, we know that simple synthetic models focusing on the function iPresent add)esses: T.H., Department of Chemistry and Biochemistry, Graduate School of Engineering. Kyushu University, Higashi-ku, Fukuoka 8 12-81, Japan; H.O., Fukui National College of Technology, Geshi, Sabae 916, Japan. Takashi Hayashi wias horn in 1962 in Osaka. He received his PhD at Kyoto Univei-sity working under the supervision of Y. Ito. He ulas then an assistant professor (I 990-1 997) at Kyoto Univei-sity. In July of 1997 he nzoved to Kyushu Uni\wsity as an associate professor. He stayed at I. Bertini’s laboratory at University of Floi-ence (Italy) in 1986. From I995 to 1996, he M’US uiorking with C.-H. Wong at the Scripps Research In-stitute (USA) as u visiting i-e-searcher.His current interests are hioorganic and bio-organic chemistry, particu-larly, molecular recognition and porphyr-in chenaistry. of the target enzyme often give us good information with regard to the reaction mechanisms and/or some structural features of important components. Over the last decade, biomimetic model systems have been suitable for the discussion of the ET mechanism. After elucidation of the crystal structure of thc bacterial photosynthetic reaction centre,’ a great number of covalently linked donor-spacer-acceptor models have been reported in the literature, some of them giving valuable insight into the ET mechanism through focusing on donor-acceptor distances and their relative special organization, features of the spacer, driving force, and solvent effect.In particular, model\ based on covalently linked donor-acceptor system5 such a\ porphyrin-spacer-porphyrin or porphyrin-spacer-quinone suggest that the unique organization of the characteristic chromophores found in the photochemical reaction centre is responsible for the highly efficient charge separation.’-i In contrast, noncovalently linked donor-acceptor complexes are essential models towards the understanding of the ET reaction in the respiratory system, in which electron transport is accomplished by specific mobile carriers such as ubiquinone and cytochrome c. However, studies on noncovalently linked donor-acceptor systems with specific interactions IYU molec-ular recognition remain quite limited.4 Thus, this has stimulated the design of new donor-acceptor complexes based on weak noncovalent interactions such as hydrogen bonds, salt bridges and van der Wads contacts and the subsequent investigation of the long-range ET reaction.2 The respiratory system in the mitochondrial membrane Fig. 1 shows the schematic pathway of ET from NADH to O2 in the respiratory mitochondrial membrane, where two electron carriers control the overall rate of ET process i!ia specific interaction with the membrane-bound oxidoreductases. Ubiqui- Hisanobu Ogoshi u,as horn in 1934. He r*ecei\ied his ducation fr-on1 Kyoto Unilw-sity. In 3963-1965. he MYIS a I-esear-ch associate ofIllinois Institute of Technology (USA).Aftet- that, he joined the staff of Departnicnt of Synthetic- Ckemisti-y at Kyoto University. He nioiw-l to Na-gaoka Institute of Technology as a full professor in 1980 and t-etui-nedto Kyoto Uni12ersity in 1988.He organized tlw Sei*entli Inter*nationul Svniposiuni on Moleciilai. Recognition and ln-c-lusion at Kyoto in 1992. In April of 1997, he moved to the Fukui National College of Technology as a president and became an emeritus projessor of Kyoto Universitv. Hisanobii Ogoshi Chemical Society Reviews, 1997, volume 26 355 NADH NAD+) 2H' +i/202 H2O Fig. 1 A schematic diagram of the mitochondrial electron transport chain indicating the ET pathway. Complex I1 is not shown. The shape of each complex is simplified and different from the real structure in order to emphasize the interaction sites for electron carriers.Proton flux through the membrane is omitted. none has two methoxy groups at the 2- and 3-positions on the quinone ring and a long hydrophobic isoprenoid chain which makes it soluble in the inner mitochondrial membrane lipid bilayer. It was thought that electrons are smoothly shuttled from NADH-coenzyme Q reductase (Complex I) to ubiquinone and from reduced ubiquinone, ubiquinol, to Coenzyme Q-cytochrome c reductase (Complex 111) via specific binding sites in each complex, although no X-ray structure has ever been solved with satisfactory resolution. Enzyme studies using synthetic ubiquinone or ubiquinol analogues have provided us with helpful information about the specific binding sites for electron carrier in the huge Complex I or Complex 111, suggesting the importance of structural features of quinone substituents (alkyl and alkoxy groups) upon binding with proteins.5 Cytochrome c is another important electron carrier between Complex I11 and cytochrome c oxidase (Complex IV).This haemoprotein, one of the well-known cytochromes, is almost spherical and contains 104 amino acid residues. One of the significant features regarding the structure of cytochrome c is that several lysines are distributed on the surface of the protein near the haem concave. In fact, according to the X-ray study of cytochrome c-cytochrome c peroxidase (CCP) complex, pos- itive residues of cytochrome c, Lys73 and Lys87, form special ion pairs with negative residues of CCP, Glu290 and Asp34, respectively.6 In the respiratory system, it is known that Complexes I11 and IV have binding sites including acidic residues for highly basic cytochrome c.Furthermore, the high resolution X-ray structures of Complex IV from Paracoccus denifrificans and bovine heart have recently been reported.7 According to these analyses, the binding site of Complex IV contains several acidic residues which could interact with lysine residues on the surface of cytochrome c. 3 Model studies on noncovalently linked donor-acceptor complexes In nature, organization is quite often the result of supramole-cular assembly based on noncovalent intermolecular interac- tions via specific molecular recognition of protein-protein, protein-sugar, DNA-receptor, antigen-antibody etc.Thereby host-guest chemistry and supramolecular chemistry based on molecular recognition processes have grown very rapidly along the elucidation of biological systems. In the case of ET systems aimed at mimicking complex natural processes, the design of donor-acceptor assemblies based on weak intermolecular interactions (for example, hydrogen bonding, salt bridge, and/or van der Waals contact) has been regarded as a suitable approach. We and several other groups have focused on the respiratory system, particularly, ET via electron carriers for a few years, and begun to construct the ET model using noncovalently linked donor-acceptor systems.Currently, our research can be divided into two approaches: preparation of a new ubiquinone receptor based on a functional porphyrin and investigation of its molecular recognition behaviour and mimicking of protein- protein complexation using a reconstituted myoglobin whose surface shows a specific binding interface. In this review, we wish to describe the strategy of molecular recognition for electron carriers and the ET reaction between photoexcited porphyrin and electron carriers. 4 Porphyrin-ubiquinone analogue complex via hydrogen bonding To our knowledge, examples of noncovalently linked porphy- rin-quinone pairs aimed at ET model systems are very few. In 1990, Ogoshi and co-workers have reported the cofacial interaction of functional porphyrin, 5,15-cis-bis(2-hydroxy- 1-naphthyl)octaethylporphyrin (2)and several quinones by use of two-point hydrogen bonds between carbonyl oxygen of quinone and phenolic protons of porphyrin.8 Sessler and co- workers have presented a sideways-linked porphyrin-quinone complex via base pairing and demonstrated the photoinduced ET reaction.9 The previous porphyrin host mole~ules,8,9,~0 however, are limited to quinone guests, since almost all host molecules can only bind p-benzoquinone or duroquinone analogues.Fur- thermore, although the ET studies in the porphyrin-quinone pairings showed the quinone quenching properties and allowed estimation of the forward ET rate through fluorescence lifetime measurements, there has been no direct evidence of any separated-charge state as a product of ET reaction, except in one report.9~ Recently, we have prepared a new host porphyrin for ubiquinone, the latter being an important electron carrier in the respiratory system." In the following sections, we briefly present the specific interactions between this functional por- phyrin and some ubiquinone analogues as well as the ET reaction occurring within this noncovalently linked complex.356 Chemical Society Reviews, 1997, volume 26 4.1 Molecular recognition of ubiquinone analogues When we design a host molecule for a target guest, the following factors seem to be important: (i) preorganization of the host providing sufficient affinity with the guest molecule structure and (ii) the multifunctionalization of the host molecule to increase the specificity of host-guest complex.Considering these two factors, the porphyrin ring is an attractive potential host molecule, its relatively flat and rigid structure being suitable as a receptor framework and allowing easy introduction of various functional groups. '2 Furthermore, porphyrin has unique photoreactivity characteristics and redox behaviour, thus we adopted it as a host molecule for ubiquinone analogues. Although the structure of the ubiquinone binding site is not known with accuracy, we have postulated that several protein residues such as tyrosine interact with ubiquinone through hydrogen bonds with phenolic protons. Consequently, we have prepared the rneso-a,cw,a,a-tetra(2-hydroxy-1-naphthyl)-por-phyrin (1) as a host molecule.lla Host 1 was obtained from the condensation of pyrrole and 2-methoxy-1-naphthylaldehyde and the following deprotection of methyl groups in sufficient yield.The obtained atropisomeric mixture was easily separated by classical silica gel chromatog- raphy and the atropisomerization derived from internal rotation about C(meso)-C(ipso) bonds was not observed even in boiling toluene. The interaction between 1 and quinone was monitored by NMR, IR and UV-VIS spectroscopic measurements, which show the cofacial 1 : 1 complex in chloroform and toluene. The summary of spectroscopic features is shown in Fig. 2. The OH proton signal of 1 was shifted downfield by 1.91 ppm and the OH stretching frequency of 1 was replaced b:, a new absorption at 3449 cm-1 in the presence of 1 equiv.of tetramethoxy- p-benzoquinone (4c), which indicates that the four OH groups of 1 enter hydrogen bonding interaction with 4c. The upfield shift of methoxy protons of 4c is due to the porphyrin ring f,-4.90 ppma t HO -kT 1 r3.98ppma Me0 OMe 4c + 3.04 ppma 3449 cm-lb porphyrin-quinone: 3.35 We 0--0:2.69 -3.18 A" Fig. 2 Structural data of 1and 1.4~.(a)IH NMR chemical shifts in CDCI3. (b) IR data of OH stretching in CHC13. (c) X-Ray crystal structure analysis. current. X-Ray crystal structure analysis shows the cofacial porphyrinquinone assembly with a cofacial distance of 3.35 AJC Binding constants were determined from UV-VIS spectro-photometric titration of various quinones with CHC13 solution of 1, using the spectral change from 550 to 700 nm with several isosbestic points (Fig.3). Selected binding constants and their II n 550 600 650 700 hl~~ Fig. 3 UV-VIS spectral change of 1 on addition of ubiquinone (4h) in CHCl? at 298 K. thermodynamic parameters are summarized in Table 1. The binding affinity of 1 for p-benzoquinone (4a) is similar to that of 2. The latter has two hydroxy groups as an interaction site, suggesting that the two quinone carbonyl oxygens interact with the two hydroxy groups of 2-naphthol in 1 or 2. Furthermore, in the case of 2,3-dimethoxy-5-methyl-p-benzoquinone(4b) and 4c, the affinity of 1 for the quinone guest increases with the number of methoxy substituents on the quinone ring, whereas the presence of those methoxy groups has no effect on the interaction with 2. According to the thermodynamic parameters obtained from variable-temperature experiments, a large en- thalpic gain is observed upon formation of the 1.4b and 1.4~ complexes as compared to formation of the 2.4b or 2.4~ complexes, although there is no marked change in the entropic Table 1 Binding constants (K,) and thermodynamic parameters (AGO, AH", TAS") on complexation between porphyrin and quinone at 298 K" Host 1 (4 points)hc 2 (2 points)c'c K, = 3.0 X 10M-I K, = 5.5 X 10 M-I AC" = -2.0 kcal mol-I AGO = -2.4kcalmol-I AH" = -6.5 kcal mol-1 AH" = -5.6kcalmol-I TAS" = -4.5 kcalmol-I TAS" = -3.3 kcal mol- 0 4a K, = 5.4 X 1o2M-l K, = 3.5 X 10M-I AGO = -3.7 kcal mol-1 AGO = -2.1 kcal mol -I *OMe OMe AH" = -7.2 kcal mol-1 AH" = -6.Okcalmol-I TAS" = -3.5 kcal mol-1 TAS" = -4.0 kcal mol- I 4b 0 K, = 2.0 x 104~-1 K, = 7.8~~~ AGO = -5.8 kcal mol-I AGO = -1.2k~almol-~ AHo = -10.5kcalmol-1 AH" = -4.3 kcal mol-I Me0 OMe TAS" = -4.7 kcalmol-l TAS" = -3.1 kcalmol-1 0 4c Precision is generally estimated to be *lo%.Measured by UV-VIS titration in CHC13.' Ref. 1l(a).d Measured by 'H NMR titration in CDCl?. e Ref. 8. Chemical Society Reviews, 1997, volume 26 357 term. These results indicate that methoxy groups also act as interaction sites for the four 2-naphthols of 1.Fig. 4 illustrates the comparison of binding constants between 1 and several ubiquinone analogues having two methoxy groups. Compared to 2,5-dimethoxy-p-benzoquinone(4e) and 2,6-dimethoxy-p-benzoquinone (40, 2,3 -dimethoxy -p-benzoquinone (4d) shows a large enhancement of its binding constant with 1. The host 1also binds native ubiquinone (4h) with a binding constant 1of I .6 X lo2~-(M = mol dm--3) at 298 K in CHC1-1, although the long isoprenoid chain seems to bring about some steric repulsion. Furthermore, in solid state, the hydroxy groups of all naphthol groups interact with all theooxygen atoms of 4c, showing an 0-0 distance of 2.68-3.21 A.’ It is suggested that two adjacent methoxy substituents at the 2- and 3-positions on the quinone ring cooperatively act as the effective third and/or fourth interaction sites ilia bifurcated hydrogen bonding.Thus, porphyrin 1 having four naphthol groups as binding sites seems to be a suitable host molecule for ubiquinone analogues. 0 0 0 0 0 0 4d 4e 4f K, = 8.3 x lo2 M-’ Ka= 1.9 x lo2 M-’ K, = 1.7 x 1O2 M-’ 0 0 I 00 4g 4h K, = 10 M-’ Ka= 1.6 x lo2 M-’ Fig. 4 Comparison of binding constants between 1 and dimethoxy-p-benzoquinones in CHC13at 298 K 4.2 Photoinduced ET in porphyrin-quinone pairing via multi-hydrogen bonding It is of particular interest to investigate the photoinduced ET by use of porphyrin-quinone assembly. The sufficient affinity of porphyrin 1 for ubiquinone analogues encouraged application of these complexes to the study of the ET reaction in this system.We used the zinc complex of 1 and nzeso-a,a,a,a-tetra(7-hydroxy- l-naphthy1)porphyrin (3)13a as probes of ET reaction (Fig. 5). From CPK molecular models the distances between the porphyrin pl?ne and quinone of 1 and 3 are estimated to be 3.5 and 6.0 A, respectively. Each zinc metallated porphyrin also has a good affinity for 4c in toluene; Ka(1Zn.4c) = 2.5 X 105 M-1, Ka(3Zn-4c)= 2.2 X 10-1M-I at 298 K. Dramatic changes in the fluorescence spectra of 1Zn and 3Zn have been observed upon addition of quinone and the Stern- Volmer plots indicate efficient fluorescence quenching of 1Zn and 3Zn in the low concentration range of quinone ([4c]o < 5 x 10-3 M) as shown in Fig.6. The findings demonstrate that static quenching of fluorescence occurs through ET from the photoexcited state of 1Zn to 4c in the cofacial mode. In contrast, no fluorescence quenching of zinc(r1) me~o-a,a,a,a-tetra(2- methoxy-I -naphthyl)porphyrin (5Zn) was detected in the same concentration range as 4c. The time-resolved fluorescence measurements of 1Zn and 3Zn in toluene reveal monoexponential decay with a lifetime of 2.0 k0.1 and 2.2 k0. I ns, respectively. The transient absorption spectra of 1Zn or 3Zn excited by dye laser pulse (570 nm, FWHM d 3 ps) display a broad excited state absorption in the 615-750 nm region, which is assigned to a zinc porphyrin 358 Chemical Society Reviews, 1997, volume 26 hv/ hv Fig.5 Schematic complexation of 1Zn-4c and 3Zn.4~as an ET model system 4 1 I I i 0 1 2 3 [~CIx lo3/M Fig. 6 Stern-Volmer plots for the fluorescence quenching of porphyrin 1Zn(o),3Zn (e),and 5Zn (W) with quinone 4c in toluene at 293 K: excitation at 538, 541 and 540 nm for lZn, 3Zn and 5Zn, respectively; emission at 645,644 and 645 nm for lZn, 3Zn and 5Zn, respectively; [lZn] = 7.42 x M, [3Zn] = 7.36 X 10-5 M, [5Zn] = 7.20 x 10-5 M singlet excited state. There was no substantial change in the spectra recorded within 50 ps delays after excitation by a laser pulse, which is consistent with the lifetime of the fluorescence. Upon addition of 4c to 1Zn and 3Zn solutions, the transient absorption spectra show a significant time-dependent positive absorption at 640 nm as shown in Fig.7. The spectra are characteristic of zinc porphyrin cation radical. The decays of the radical species as an intermediate of ET were fitted to monoexponential curves, which leads to the charge recombina- tion constants of (1. I f0.4) X 1011 s-1 for lZn.4~and (6.8 + 0.3) X 1010 s-1 for 3Zn.4~.In contrast, forward ET rate constants were estimated to be 24 X loll s-l (limited by the pulse width and signal-to-noise ratio) for both samples. 13h Heitele, Michel-Beyerle, Staab and co-workers have reported the rate constants of charge separation and recombination in a series of covalently bridged porphyrin-quinone systems. l4 For example, in the case of quinone-capped porphyrin 6, the rate 11.5 ps 14.0 ps 16.5 ps 24.0 ps 444.0ps 610 660 710 hInm Fig.7 Transient spectra after excitation at 570 nm (~3ps laser flash) in the presence of quinone 4c (8.94 X lo-' M) in toluene, showing the disappearance of the cation radical species. (a) 1Zn (1.69 X M); (b) 3Zn (2.36 X M) 6 --.-&--substituents, respectively, as recognition sites to form Watson- Crick hydrogen-bonding interaction. Upon addition of 8 to the CH2C12 solution of 7, the quenching of zinc porphyrin fluorescence was observed and the fluorescence decay profile became biphasic with two exponential components of lifetimes comparable with the estimated value from 'H NMR titration. These results indicate that singlet ET from photoexcited 7 to 8 takes place with a rate constant of ca.8 X lo8 s-l through exothermic process by 0.50 eV, although the porphyrin cation radical was not detected by picosecond laser flash photolysis. They have also prepared a well-organized energy-transfer system in noncovalently linked porphyrin-porphyrin complex by guanine-cytosine base pairing. 15 The derived rate constants in both systems are not markedly dependent on the concentra- tion of acceptors and the nature of interaction sites, indicating electron and/or energy transfer from donor to acceptor in the intra-aggregated complex formed by multipoint hydrogen bonding. 4.4 Mechanistic study on photoinduced ET in noncovalently linked donor-acceptor systems via hydrogen bonding A slightly different approach to understanding the ET mecha- nism in noncovalently linked donor-acceptor systems has been taken by Nocera and co-workers.They have reported the intra- aggregated ET system of zinc porphyrin 9-dinitrobenzoic acid 10 array mediated by hydrogen bonding.16 The binding constants for 9-10 and deuterated 9-10 pairs were estimated to I R' H Chemical Society Reviews, 1997, volume 26 359 be 698 and 316 M-1, respectively, by static fluorescence quenching. One of the interesting findings in this work is the isotropic effect on ET reaction from photoexcited 9 to 10. The comparison of forward and charge recombination ET between protonated and deuterated carboxylic acid interfaces indicates the isotope effect of kH/kD = 1.7 and 1.6, respectively; rate constants of forward and charge recombination processes between 9 and deuterated 10 are 3.0 X 1Olo and 6.2 X lo9 SKI, respectively.Their results indicate that the hydrogen bonding linkage between donor and acceptor may play a role in the ET studies of the ET pathway in native and/or modified proteins have been reported by several groups.18 At the same time, a number of groups have also begun to focus on the inter- molecular ET in a noncovalent protein-protein complex such as cytochrome c-cytochrome c peroxidase (CCP) pairing. Partic- ularly, Hoffman1*,*9 McLendon,18-19 Millettl9 and KostiC19 have demonstrated the relationship between interaction mode and ET rate in protein-protein complexes.To understand the ET mechanisms in self-associated protein-protein complexes, Rodgers has presented a mimetic system where a highly anionic process. uroporphyrin is bound to a positive cytochrome c via electro-static docking forces.*O These studies have suggested that the 0-H molecular recognition of redox proteins and their conforma- or D----0\die* tional interconversion upon binding seem to dominate not only 0----H or D-0 the specificity of ET but also the ET rate. One major goal using -.. 10 a simple and efficient model is to understand how the ET is / controlled by specific interface with the surface of protein. Thus it has been of particular interest to design artificial ET models 9 Therien and co-workers have recently presented significant ET model compounds in which zinc(1r) porphyrin and iron(m) porphyrin are linked by hydrogen-, 0-and n-bonds, respec- tively.I7 The singlet ET rate constants for three systems were determined by time-correlated single photon counting spec- troscopy and the rate constant for hydrogen-bonding system 11 (kET = 8.1 X lo9 s-l) is comparable to that for the n-bond system 12 (kET = 8.8 X 109 s-I), whereas the ET in system 13 composed totally of a-symmetry bonds is slower (kET = 4.3 X 109 s-I) than that in 11 mediated by hydrogen bonds.The difference in the rate constants in these systems must derive from the magnitude of donor-acceptor electronic coupling, and they suggest that the electronic coupling modulated by a hydrogen-bond interface in 11is greater than that provided by an interface composed of o-bonds in 13.The comparison of the ET reaction in their systems predicts that the hydrogen-bond interface in protein could mediate ET as efficiently as a covalent bridge. 5 Photoinduced ET in a protein complex by reconstituted zinc myoglobin The next stage in the study of the respiratory system model is the construction of an artificial protein-electron carrier to investigate the role of molecular recognition and binding between electron donor and acceptor in biological ET systems. Over the past few years, elegant work focusing on mechanistic based on modified proteins. 5.1 Photoinduced ET from zinc porphyrin to a linked quinone in protein Myoglobin is one of the most popular haemoproteins as an oxygen-storage protein and it is well characterized in terms of both primary and tertiary structures, thereby facilitating model building.Unlike cytochrome c, the haem as a prosthetic group in myoglobin is not covalently bound to the protein core. This feature enables the replacement of native haem with an artificial metalloporphyrin, thus it facilitates the preparation of reconsti-tuted myoglobin having a unique function. In the last decade, we have reported various myoglobins reconstituted with artificial haems and studied their properties, showing that some modification of the peripheral propionate at the 6-or 7-position of the haem structure induces no serious problem concerning protein stabilization.21 According to this result, we have prepared a new zinc porphyrin bearing a quinone component as an electron acceptor 14 (ZnP-Gly,-Q, n = 0-2) and incorpor- ated 14 into apoprotein to obtain the reconstituted myoglobin rMb(l4), which was characterized by TOF-mass, NMR, and UV-VIS spectroscopic measurements.” The photoinduced ET from zinc porphyrin to quinone in rMb(l4) was monitored by fluorescence spectroscopy of the zinc porphyrin moiety in buffer solution (pH 7.0).As compared with the reference protein, rMb(lS), the fluorescence intensity of rMb(l4, n = 2) was weak and a shorter fluorescence lifetime of 0.6 ns was observed, whereas the fluorescence quenching of R’ -[spacerj-R’ = OMe, R2 = H 360 Chemical Society Reviews, 1997, volume 26 rMb(l5) in the presence of quinone 4b was not detected and the single exponential curve of fluorescence decay with T = 1.9 ns was observed.These results demonstrate that the singlet ET occurs from photoexcited zinc porphyrin to quinone in protein and the forward rate constant is estimated to be 1.1 X lo9 s-* from the following equation: kET = l/~,-l/~,where TI and T~ represent the fluorescence lifetime of rMb(l4) and rMb(l5). The result obtained indicates that the myoglobin reconstituted with synthetic zinc porphyrin is a suitable photodonor. Thus, we have taken a new approach to construct a noncovalently linked protein-acceptor model as shown in the following section. 5.2 Artificial recognition between reconstituted myoglobin and methyl viologen According to the X-ray crystal structure analysis and a series of kinetic studies on cytochrome c receptors, it is found that the binding sites of cytochrome c on the surface of the receptor protein consist of several anionic residues forming special pairs with a few specific lysines of cytochrome c.6 On the basis of this concept, we have designed the novel zinc myoglobin rMb( 16) having a negative recognition site for positively charged electron acceptors such as methyl viologen dication (18), as shown in Scheme 1.23 Artificial prosthetic group, zinc porphy- rin 16, which has eight carboxylates at the terminal of 6-and 7-propionates of mesoporphyrin, was easily inserted into horse heart apomyoglobin by routine methodology. The stability of the myoglobin obtained rMb(l6) was similar to that of the reference protein rMb( 17) reconstituted with mesoporphyrin zinc complex 17, and both characteristic visible absorption CH2COOH I CONH-CHCOOH CONH-CHCOOHI CHZCOOH yH2COOH CONH-CHCOOH 16 CONH -CHCOOH I CH2COOH 17 Scheme 1 Chemical Society Reviews, 1997, volume 26 361 spectra and CD spectra were comparable.These features indicate that serious structural changes of the reconstituted protein do not result from the insertion of artificial porphyrin 16 into apomyoglobin. A study of the fluorescence quenching of rMb(l6) by 18was monitored at 580 and 635 nm wavelengths in different solvent conditions. The Stern-Volmer plots in Fig.8 show that there is static quenching in a low concentration of quencher 18 (<6 X 10-3 M). Furthermore, efficient quenching is found at high pH and/or low ionic strength. In contrast, no fluorescence quench- ing of rMb(l7) was observed upon addition of 18 (< 10-2 M). These results indicate the ground-state complexation between rMb(l6) and 18 vzu electrostatic recognition as shown in Scheme 2, in which donor and $cceptor are separated by an edge-to-edge distance of 13-17 A based on CPK models. \ I I h0 c 3.0 2.0 -1.o 13 2 0.0 , 2.0 4.O 60 [quencher181 x 103/M Fig. 8 Stem-Volmer plots for the fluorescence quenching of porphynn rMb(l6) and rMb( 17) in the presence of 18at various pH at 25 "C The solid and dashed lines correspond to the data obtained in 10 and 100 mM phosphate buffers, respectively.rMb(l6): pH = 5.76 (D),pH = 6.10 (m), pHpH = 641 (0).= 7.08 ( A), pH = 7 00 (+) [rMb(l6)] = M rMb(l7): pH = 700 (0) [rMb(l7)] = M. The changes of fluorescence emission were monitored at 584 nm (Lex = 543 nm). Scheme 2 5.3 Singlet ET between reconstituted myoglobin and methyl viologen The measurement of fluorescence lifetime of rMb(l6) and rMb(l7) gave a monoexponential curve with a lifetime of 2.0 f 0.1 ns, whereas, in the presence of 18, fluorescence decay of rMb(l6) exhibits a biphasic curve expressed by two lifetimes as shown in Fig. 9; TL = 2.0 f 0.1 and tS = 0.4 L-0.1 ns, respectively. The longer and shorter lived components of the decay curve are assignable to free rMb(l6) and the rMb(16).18 complex, respectively, and the relative contribution of the shorter one, T~,increases with increasing concentration of 18, which also supports the stable complexation between rMb(l6) and 18 during the ET process.From the lifetime results the rate constant of forward singlet ET (k,,S) from photoexcited rMb(l6) to 18 is 2.1 X 109 s-l derived as ketS = l/.cS -l/tL. 362 Chemical Society Reviews, 1997, volume 26 0 10 ns Fig. 9 Fluorescence decay profiles obtained from rMb(l6) solution in the presence of 18at room temperature in pH 7 0, 10 mM phosphate buffer with excitation wavelength at 543 nm [rMb( 16)] = 2 8 X 10-6 M and [18] = 1.4 X 10-2 M 'cs = 0 4 ns (As = 0.76), 'cL = I 8 ns (AL = 0 24) Transient absorption spectra of rMb(l6) in the presence of 18 were monitored by a streak camera after a 50 ps, 532 nm laser flash.A strong absorption band centred at 455 nm can be assigned to the superposition of excited singlet and triplet states and cation radical species of zinc porphyrin in myoglobin. These data support direct evidence of ET from rMb(l6) to 18 and can lead to the kinetic analysis of the intermediate species; the charge recombination of cation radical species was directly monitored with a rate constant of k,, = 3.3 X 108 s-1.24 In contrast, the transient absorption spectra for rMb(l7) in the presence of 18 show no clear decay of absorption in the same time range. According to the previous literature,*5 the photo- induced ET from triplet state of zinc myoglobin to 18 is diffusion controlled, whereas no such singlet ET reaction has been reported for zinc myoglobin.To our knowledge, the present work is the first example of a singlet ET between zinc myoglobin and electron acceptor via intermolecular interaction. Our results suggest that fast and/or efficient ET in biological systems is regulated by specific binding sites. At present, we are preparing a new myoglobin reconstituted with zinc porphyrin, having a different type of interface at the propionate terminus to compare the binding modc of 18 with rMb(l6). Recent work shows that rMb(l6) with eight carboxylate groups as an interface is one of the best receptors for 18, thus, the degree of localization of anionic charges on the surface of protein should be quite important for the recognition of 18 in aqueous solution.26 5.4 Future directions: protein-protein complexation One of our aims is to construct artificial systems where reconstituted myoglobin interacts with redox protein such as cytochrome c in order to elucidate the recognition event at a molecular level.Upon the binding of methyl viologen with the present reconstituted myoglobin, the recognition behaviour should be attributable to point-to-point interaction by use of electrostatic forces. In contrast, it is expected that the essential feature of protein-protein complexation might depend on face- to-face recognition mode with multipoint interaction. There- fore, considering the design of an artificial protein-protein complex, it is necessary to note the distribution of special residues which play a binding-domain role with partner protein.Although clear evidence of the interaction between rMb( 16) and cytochrome c has never been obtained by spectroscopic measurements, we have recently found a new myoglobin rMb(l9) which has sufficient affinity for cytochrome c.27 Functional groups of 19, which are designed to match the position of particular lysines in cytochrome c, are slightly different from those of 16. We will report the novel protein- protein complex and its ET behaviour within the specific interaction in the near future. It is likely that the reconstituted myoglobins such as rMb(l6) or rMb(l9) are good probes to investigate the binding sites of partner protein and gating mechanism in ET processes.* Another recent model of ET In nature is the construction of multicomponent systems in artificial bilayers to mimic the mitochondria1 membrane. Recently, Groves and co-workers have reported an elegant model where electron moves from the membrane spanning Mn" porphynn to cytochrome c via trianionic zinc porphyrin which is located on the surface of the membrane and designed to match the binding site of cytoch- rome c 2x Spectral titrations of the cytochrome c to the 20-21 vesicular complex show the formation of a 1 1 adduct with a binding constant of 5 X 106 M-1. Electron transfer rate from Mn" porphyrin in the 20-DPPC/DMPC vesicle to cytochrome c was found to be first order and independent of the length of the imidazole tail of 21 with a rate constant of 103s-1, however, the rate obtained in the shorter DLPC lipid was ten times faster than that observed in DPPC/DMPC vesicles These results support the stable 20-2l-cytochrome c ternary complexation and electron transfer Ilia multiple pathway in the membrane ensembles This research is one of the new approaches to the construction of nanoscale charge-separation systems to eluci- date the ET mechanism of cytochrome c as a positive-charge membrane protein.\cytochrome c OH ~ En 0 a ~ Lipid = dipalmitoylphosphatidylcholine(DPPC) and dimyristoylphospha- tidylcholine (DMPC) or dilaurylphosphatidylcholine(DLPC) 6 Conclusion It is well-known that molecular recognition processes are important in biological systems, such as enzymatic reactions, transportation, regulation etc , as protein interfaces are fre- quently composed of intermolecular salts bridges, hydrogen bonds as well as extensive hydrophobic surface contacts ET is also controlled by molecular recognition between the oxidor- eductase and the electron carrier, which gives rise to efficient binding in the respiratory systems, although the mechanism is not clear This has stimulated a variety of studies, especially on intermolecular ET process by use of noncovalently linked donor-acceptor models In this review, we indicated new approaches to systems based on noncovalently linked donor- acceptor pairs as an ET model, focused on the molecular recognition of electron carriers, one is the specific receptor for quinones by multifunctional porphyrin and another is the reconstituted zinc myoglobins having the anionic interface for viologen dication and cytochrome c It is reasonable to think that both of them could be widely useful for the elucidation of ET mechanisms, thanks to the following advantages (I) they link recognition behaviour with ET process and (11) it is possible to monitor the stable charge-separation species in each model by spectroscopic measurements.The methodology of the construction of the noncovalently linked donor-acceptor pairs as simple models might provide a new insight into the complicated biological ET process. 7 Acknowledgements We thank our collaborators, Professor Keitaro Yoshihara dnd Dr Shigeichi Kumazaki at the Institute for Molecular Science, and Professor Hideki Masuda at Nagoya Institute of Technol- ogy We are deeply indebted to our colleagues for their excellent experiments In particular, we appreciate the significant con- tributions made by several graduate students, Tetsuo Takimurd, Takashi Miyahara, Tomohito Asai and Yutaka Hitomi 8 References I J Deisenhofer, 0 Epp, K Miki, R Huber and H Michel, Natui e, 1985 318,618 2 M R Wasielewski, Chem Rex .1992, 92, 435 3 H Kurreck and M Huber, AngeM Chem Int Ed Enql 1995 34 849 4 J L Sessler, B Wang, S L Springs and C T Brown, in Cowipiehenni c Supramolecului Chemistry Vol 4, ed Y Murdkami, Pergdmon, Oxford, 1996.p 31 1 5 K Sakamoto, H Miyoshi, K Takegami, T Mogi, Y Anraku dnd H lwamura, J Biol Chem , 1996, 271,29897 6 H Pelletier and J Kraut, Scrence, 1992, 258, 1748 7 T Tsukihara, H Aoyama, E Yamashita, T Tomizaki, H Yamaguchi, K Shinzawa-Itoh, R Nakashima, R Yaono dnd S Yoshikdwa, Sc /en(( 1995, 269, 1069, S Iwata C Ostermeier, B Ludwig dnd H Michel Natui e, 1995, 376, 660 8 Y Aoyama,M Asakawa,Y MatsuiandH Ogoshi J Am Chem Soc 1991, 113,6233 9 (a)A Hmiman, Y Kubo and J L Sessler, I Anz Cheni Soc 1992 114,388,(h)J L Sessler, B Wang and A Harriman, J Am Chem So( 1993, 115, 10418, (0 A Berman, E S Izraeli, H Levanon dnd B Wang, f Am Chem Soc 1995. 117, 8252 0 Other models of non-covalently linked porphyrinquinone systems M Gonzalez, A McIntosh, J Bolton and A Weedon, J Chem Soc Chem Conimun , 1984,1138.Y Kuroda, M Ito, T Sera and H Ogoshi J An? Chem Soc , 1993, 115, 7003, H Imahori E YoShiZdWd K Ydmada, K Hagiwara, T Okada and Y Sdkata, f Chem So( Cheni Commrrn 1995. 1133, F D'Souza, .I Am Cheni Soc , 1996,118 923. C A Hunter and R J Shannon, Chem Commun . 1996, 1361 1 (a) T Hdyashi, T Miyahara, N Hashizume and H Ogo\hi, J Am Chem Soc 1993,115,2049,(h)T Haydshi, T Miyahara, Y Aoydmd M Kobayashi and H Ogoshi, Pur e Appf Chem , 1994, 66, 797, (( ) T Hayashi, T Miyahara, N Koide, Y Kato, H Masuda and H Ogoshi, J Am Chem Soc , 1997, 119, 7281 12 H Ogoshi, Y Kuroda, T Mizutani and T Hayashi, Puic Appi Chem 1996, 68, I41 1 13 ((0T Haya5hi.T Miyahara, Y Aoyama, M Nonoguchi and H Ogoshi Chem Lett, 1994, 1749, (h)T Hayashi, T Miyahara, S Kumazaki, H Ogoshi and K Yoshihard, Anpe~ Cheni In! Ed Enql , 1996, 35, 1964 14 H Heitele, F Pollinger, T Haberle, M E Michel Beyerle dnd H A Staab, J Phyr Chem , 1984,98,7402,H A Staab J Weiser and E Baumann, Chem Be1 , 1992.125. 2275 1.5 A Harriman, D Magda and J L Sessler, J Phys Cheni , 1991, 95. 1530 16 C Turro,C K Chang,G E Leroi,R I CukierandD G Nocera 1 Am Chem Soc, 1992,114,4013 17 P J F deRege, S A Williams and M J Therien, Screnr e, 1995, 269, 1409 18 P Bertrand, B E Bowler, J Chang, B M Hoffmm, H B Grdy, A Kuki, A G Mauk. G McLendon, M J Natan, J M Nocek, A L Raphael, A G Sykes, M J Therien and S A Wdllin, Structuie and Bonding, vol 25, Springer-Verlag, Berlin, 1991 19 G McLendon and R Hake.Chem Rev, 1992, 92, 481, J M Nocek J S Zhou, S D Forest, S Priyadarshy, D N Beratan, J N Onuchic and B M Hoffman, Chem Re\ , 1996,96,2459,M M Ivkovic Jensen and N M Kostic, Biothemrstiy, 1996, 35. 15095, K Wang, H Mei L Geren, M A Miller, A Saunders, X Wang J L Wdldner Chemical Society Reviews, 1997, volume 26 363 G. J. Pielak, B. Durham and F. Millett, Biochemistry, 1996, 35, 15107. 20 J. S. Zhou and M. A. J. Rodgers, J. Am. Chem. Soc., 1991, 113, 7728. 21 T. Hayashi, Y. Hitomi, A. Suzuki, T. Takimura and H. Ogoshi, Chem. Lett., 1995, 911. 22 T. Hayashi, T. Takimura, Y. Hitomi, T. Ohara and H. Ogoshi, J. Chem. Soc., Chem. Commun., 1995, 545; 2503. 23 T. Hayashi, T. Takimura and H. Ogoshi, J. Am. Chem. Soc., 1995,117, 1 1606. 24 N. Mataga, A. Karen, T. Okada, S. Nishitani, N. Kurata, Y. Sakata and S. Misumi, J. Phys. Chem., 1984, 88, 5138. 25 Recent review on ET in zinc myoglobin and methyl viologen 18: K. Tsukahara, M. Okada, S. Asami, Y. Nishikawa, N. Sawai and T. Sakurai, Coord. Chem. Rev., 1994, 132,223. 26 T. Hayashi, A. Tomokuni, T. Takimura and H. Ogoshi, unpublished results. 27 T. Hayash;, Y.Hitomi and H.Ogoshi, submitted for publication. 28 J. Lahiri, F. D. Fate, S. B. Ungashe and J. T. Groves, J. Am. Chem. Soc., 1996,118, 2347. Received, 28th January 1997 Accepted, 27th May 1997 364 Chemical Society Reviews, 1997, volume 26

ISSN:0306-0012    

DOI:10.1039/CS9972600355

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

I \-Photo-induced electron and energy transfer in non-covalently bonded supramolecular assemblies Michael D. Ward School of Chemistry, University of Bristol, Cuntock’s Close, Bristol, UK BS8 ITS Covalently linked chromophore-quencher complexes are widespread in the area of transition-metal photochemistry and as models for photosynthesis. This review surveys recent examples of supramolecular complexes in which interacting chromophore and quencher fragments are instead held together by non-covalent interactions such as hydrogen bonding, aromatic n-stacking, hydrophobic interactions and labile metal-ligand coordinate bonds. The use of these methods to assemble multi-component photo-active com- plexes has led to the preparation of many highly sophisti- cated systems for energy transfer or charge separation which would not be accessible by ‘conventional’ synthetic meth- odology.1 Introduction The synthesis of polynuclear complexes containing photo- chemically active groups, such as metal polypyridyl complexes or metalloporphyrins, is one of the most widely studied topics in contemporary inorganic chemistry.’ The impetus behind it is the possibility of harvesting solar energy. In nature, the energy of sunlight is harnessed during photosynthesis and used to drive the endothermic reaction sequence by which water and C02are converted to sugars, and a great deal of effort is directed at understanding the primary events of photosynthesis in which absorption of light by chlorophyll is followed by a long-distance electron-transfer to a quinone group, generating a long-lived charge-separated state.There is also the possibility of preparing ‘unnatural’ systems by use of light-absorbing metal complexes which are not naturally occurring and therefore were not at nature’s disposal during evolution. The ultimate idea is however the same: to use the light energy absorbed by the molecule to drive endothermic reactions, such as splitting of water to H2 and 02 which could be used as a fuel source. ~~ ~~~ Mike Ward took his BA at Robinson College in the University of Cambridge, and stayed put to do his PhD with Ed Constable on the coordination chemistry of multidentate pyridine-bused ligands. He then went to Strasbourg for a year us a Royal Society post-doctoral fellow, to make catenates with Jean-Pierre Sauvage, before taking up his current position as a lecturer in the School of Chemistry at the University of Bristol.His research interests cover all aspects of the coordination chemistry of multidentate ligands, in particular the various forms of interaction (magnetic, electronic and photochemical) between metal centres in multinuclear com-plexes, and the design and syn- thesis of new ligands. Michael 1). Ward The basis of these photochemically active compounds (chromophores) is that following absorption of a photon of light they enter a long-lived electronically excited state. Of course all chemical compounds are in principle capable of electronic transitions in which an electron is promoted from a ground state to a higher-energy state-usually a HOMO --+ LUMO transi- tion-but in the vast majority of cases the excited state collapses very quickly back to the ground state, with evolution of heat as the electronic energy is converted to increased vibrational motion in the molecule.If the electronic excited state of the molecule survives for long enough however, there is the possibility that it can interact with another molecule before it is deactivated, and the reactions of molecules in their electron- ically excited state are completely different from those which they undergo in their ground If no such interaction occurs then the excited state will be deactivated either thermally, or sometimes by emission of a photon (lumines- cence).Luminescent complexes are particularly useful in this area, as they tend to have long-lived excited states, and the loss of luminescence (quenching) is an obvious sign that the excited state complex is reacting with another group rather than undergoing radiative decay. The two main mechanisms by which an electronically excited molecule can pass its energy on to another molecule (a ‘quencher’) are electron transfer and energy transfer. These are illustrated schematically in Fig. 1. It will be seen that promotion of an electron from a filled orbital to a higher-energy empty orbital means that in the excited state the molecule is simultaneously a stronger oxidising agent and a stronger reducing agent than it was in its ground state.The promoted electron is in a high-energy orbital and can transfer out to an electron-poor quencher; i.e. it acts as a reducing agent. Alternatively, the low-energy hole left by the promoted electron can accept an electron from an electron-rich quencher; i.e. the excited-state chromophore acts as an oxidising agent. The type of electron transfer that occurs depends on the nature of the species that is interacting with the excited-state chromophore. Energy transfer, in contrast, involves no net electron transfer; instead the excited-state energy of the chromophore is trans- ferred to the quencher, which itself enters an electronically excited state. This is more likely to occur if the quencher has a low-energy excited state available and is not amenable to oxidation or reduction.These basic principles have been thoroughly and clearly described elsewhere, and the reader interested in the photophysical and photochemical principles which underlie the work described in this review is referred to these articles. 1,2 In order to study the interactions between chromophores and quenchers under controlled conditions, very many molecules have been prepared in which these components are linked by a covalent bond. This allows reasonably precise knowledge of the distance between the two groups, their relative spatial orienta- tion, and the nature of the pathway linking the two components which can act as a conduit for the electron or energy transfer.It has therefore been possible to relate the rate and efficiency of the interaction-such as electron transfer from chromophore to quencher after light absorption-to the distance between the components, the conformation of the bridge, the presence or Chemical Society Reviews, 1997, volume 26 365 preparations and properties of such sy~tems.33~ A couple of illustrative recent examples of covalently linked chromophore- quencher systems are in Fig. 2.576 The porphyrin (chromo- EXCITATION phore)-quinone (quencher) system is one of the most com- monly studied because of its relevance to naturally occurring Ground-state Excited-state photosynthesis, in which the excited-state chlorophyll (a chromophore chromophore porphyrin complex) transfers an electron to a nearby quinone If + Excited-state Electron-rich Reduced Oxidised chromophore quencher chromophore quencher OXIDATIVE ELECTRONt+ -+ TRANSFER Excited-state Electron-poor Oxidised Reduced chromophore quencher chromophore quencher + ENERGY====3 TRANSFERI+ + ++ Exciied-state QUw~cher with Ground-state Excited-state chromophore a~essible chromophore quencher excited state Fig.1 Electron transfer and energy transfer quenching of a chromophore excited state absence of direct orbital overlap between the components, and so on. Such studies have contributed a great deal to understanding of the fundamental photophysical processes which it is desired to exploit further in solar energy harvesting, and there are many excellent reviews dealing with TJ? CHROMOPHORE CHROMOPHORE Fig.2 Examples of covalently linked chromophore-quencher assemblies, based on (a)a polypyridyl-Ru" 5). and (h)a porphyrin chromophore and a quinone quencher (ref 6). 366 Chemical Society Reviews, 1997, volume 26 (an electron acceptor) following excitation by a photon of light.3 The [Ru(bipy)3I2+ chromophore (bipy = 2,2'-bipyridine) is also a very popular system because of its particularly suitable photophysical characteristics, its high stability, and the ease with which it can by synthetically modified. Most of the examples in this article will involve one or other of these systems .2,4 The purpose of this review is to look at some recent examples of chromophore-quencher systems in which the two compo- nents are associated by non-covalent interactions.Whilst conventional chemical synthesis relies on manipulation of covalent bonds, the recently emerging area of supramolecular chemistry relies on weaker non-covalent interactions such as metal-ligand coordinate bonds, hydrogen-bonding, aromatic n-stacking, hydrophobic interactions and so on to control the assembly of large, structurally sophisticated species whose preparation would be way beyond the scope of more conven- tional synthetic methods.7.8 There are two principal reasons for the use of weaker interactions of this type to control the assembly of chromophore-quencher systems. The first is ease and diversity of synthesis. The covalently linked systems can be very difficult to prepare and are limited in that the assembly of the molecule is irreversible; it is not usually possible to undo the components and reassemble them in a different way.Each synthesis is therefore a difficult, one-off process. Being able to prepare the component parts, mixing them together, and letting a collection of non-covalent interactions assemble the compo- nents for you has obvious appeal. The second reason is that nature relies on supramolecular methods of assembly. In the photosynthetic reaction centre for example the complex array of our components is not held together by covalent bonds between them; rather, the components are held in a spatially well-defined arrangement by the surrounding protein, with a collection of the weak non-covalent interactions between them which suffice to QUENCHER QUENCHER 0 chromophore and a viologen-type quencher (ref.mediate the energy or electron-transfer processes as well as holding the components in place.9 Accordingly, the last few years has seen a surge of interest in use of supramolecular methods to synthesise chromophore- quencher complexes, both as models for the naturally occurring photosynthetic reaction centres and to explore new ‘unnatural’ systems for light-harvesting. A selection of these form the subject of this review, and are subdivided according to the type of interaction responsible for linking the components:8 coor- dinate bonds; hydrogen bonds (by far the largest category); hydrophobic interactions; aromatic n-stacking; and, finally, combinations of two or more types of interaction.2 Components linked by coordinate bonds Coordinate bonds (dative bonds between ligand lone-pairs and metals) may be crudely subdivided into two types; those involving a kinetically inert metal centre and those involving a kinetically labile metal centre. The former are essentially like covalent bonds in that, once formed, they are more or less permanent and the ligands are not involved in equilibria between bound and unbound states. With kinetically labile metals however monodentate ligands may exchange rapidly, so association by such an interaction is easily reversible and more akin to e.g. supramolecular association via hydrogen-bonding, and these are the examples of interest for our purposes. Most examples of chromophore-quencher assembly iia labile coordinate bonds involve axial ligation of pyridine ligands to zinc-porphyrin units.The association constant K for such complexes is ca. 103-104 dm3 mol-1. The principle is illustrated by complex 1, in which a pyromellitimide group with a pendant pyridyl group is axially ligated to a zinc(I1)-porphyrin fragment. 10 This is a ‘donor-spacer-acceptor’ assembly, so-called because in this case the excited state of the porphyrin unit acts as an electron donor, the pyridyl group is the spacer, and the pyromellitimide acts as an electron-accepting quencher. Simply adding an excess of the free pyromellitimide-pyridyl ligand to a solution of the zinc-porphyrin complex results in complete loss of the characteristic luminescence of the porphyrin fragment, because assembly of the five-coordinate complex 1 allows electron-transfer to occur from the excited-state por- phyrin to the pyromellitimide, with a rate constant kET of 2.1 X 1010 s-1.The same principle has been exploited in complex 2, in which a quinone electron-acceptor is tethered between two Zn”-porphyrins by two axial pyridyl ligands. I I In this case the ‘two-point’ binding of the electron-acceptor means that the association constant between the two components is much higher ( lo7 dm3 mol-I), so simply mixing the two components results in more or less complete association even at the low concentrations typically used for spectroscopic studies.The electron-transfer rate constant kET was found to be 1.6 X 1010 s-1, a similar value to that for 1 which reflects the similar nature of the electron-donor and the comparable separation between chromophore and quencher. In 1 and 2, the shortest through-space route between chromophore and quencher is also the through-bond pathway, since the chain of atoms linking the chromophore and quencher lies more or less on the straight line between them. It is therefore not possible to tell whether the electron-transfer in these systems is occurring through space or whether it requires the chain of bonds to act as a ‘wire’. In 3 however this ambiguity is removed. Axial coordination of the two pyridyl groups of one component (upper in the diagram) to the Zn**-porphyrin centres of the other (lower) results in a close spatial approach of the free-base porphyrin chromophore and naphthalene-diimide quencher fragments.* Tbe through-space separation of the components is at most 10A,owhereas the through-bond distance around the ring is about 35 A. As with 2, the two-point binding which holds the complex together results in a high association constant, of about 3 X lo* dm3 mol-I. The assembly of the complex results in a loss of about 70% of the fluorescence .C02Me MeO2C L 1 2 intensity from the free-base porphyrin of the upper component. In other words, following excitation, 70% of the excited-state free-base porphyrin units are quenched by electron-transfer to the naphthalene-diimide electron acceptor, and this electron transfer occurs through space (or, more accurately, through the solvent molecules occupying the cavity between the two components).3 Components linked by hydrogen bonds 3.1 Hydrogen bonds as an interface for electron or energy transfer Hydrogen bonding has been by far the most popular method for supramolecular assembly of chromophore-quencher groups over the last few years. This is because of its two principal characteristics; its directionality and its selectivity.x The directional features of hydrogen bonds mean that when two components are associated via hydrogen-bonding, it is possible to know the separation between the components and, in some cases, their relative orientation.In this respect hydrogen bonds behave like the coordinate bonds described above, and unlike some of the weaker non-covalent interactions described below. The selectivity, particularly in multiply hydrogen-bonded systems, means that considerable control can be exerted over the association process by careful use of exactly complementary components, and this is apparent in the examples described below. Although the strength of an individual hydrogen bond is relatively weak (of the order of 10kJ mol-I), multiple hydrogen Chemical Society Reviews, 1997, volume 26 367 R' ,/hv I H electron-transfer A A R R 3 (R, R' = substitutedaryl groups) bonds can be much stronger, typically of the order of 50 kJ therefore, the double hydrogen-bond interface between compo- mol-1 or more, which is beginning to approach the strength of nents is more effective at mediating electron-transfer than an the weakest covalent or coordinate bonds.interface of comparable length composed of carbon-carbon o For hydrogen bonding to be of any use in this area it is bonds. necessary that the bond can act as an effective conduit for Complex 6 is a nice example of a carefully designed multiple electron transfer or energy-transfer in addition to its structural hydrogen bond between two complementary components. l4 properties. Complexes 4 and 5 were compared to evaluate The pendant barbiturate group attached to the porphyrin forms this.13 In 4, the Fe"1-porphyrin centre acts as an electron six hydrogen bonds with the exactly complementary bis(diami- acceptor following excitation of the ZnILporphyrin centre, and dopyridine) receptor to which is attached a fluorescent dansyl kET is 8.1 x 10' s-1 across the double hydrogen bond formed by (dimethylaminonaphthalene-sulfonyl) group.The multi-point association of carboxylic acids. In 5, with a saturated carbon hydrogen-bonding ensures a very strong association (K = 106 covalent bridge, kET is 4.3 X 109 s-1. Somewhat surprisingly, dm3 mol-I), and in the associated complex the fluorescence of R R IR 4 [R = 3,4,5-C6H2(OMe)3] k Ph Ph Ph 5 (R = OMe) Ph 368 Chemical Society Reviews, 1997, volume 26 the dansyl group is almost completely quenched by efficient energy-transfer to the porphyrin fragment. Complexes 7-9 are all examples of complexes in which the electron-donor [a ZnI1-porphyrin] is associated with an electron acceptor via a triple hydrogen bond.For 7 and 8, the electron acceptor is a quinone and the interface is a cytosine-guanine hydrogen bond of the type which occurs in double-stranded nucleic acids. Complex 8 is however more conformationally rigid than 7, which gives a better-defined inter-component separation, and therefore allows a better understanding of the intercomponent electron-transfer pathway which occurs in both cases.I-5 Whereas the electron-transfer in 7 (kET = 4 X loxs-1) could occur in part by diffusional encounter between the components which transiently brings them spatially close together, in 8 the electron-transfer (kET = 8 x 108 s-1) most likely occurs through the hydrogen bonded bridge.Complex 9 uses a different mode of hydrogen bonding to link the components; here the quinone associates with the hydroxy groups of a calixarene group pendant from the Zn"-p?rphyrin, but in such a way that it is held spatially close (9 A) to the porphyrin core. The inter-component electron-transfer (LET = 8 X loxs-I) is therefore more likely to be through space or through the intervening solvent, rather than taking the much longer through-bond route (c5 complex 3). The assembly 10 is an example of a multi-chromophore (as opposed to a chromo- phore-quencher) assembly of porphyrin units i9ia triple cyto- sine-guanine hydrogen-bonds, and rapid inter-component en- ? ergy-transfer between components was observed [from the Zn"-porphyrin unit which has the higher-energy excited state, to the free-base porphyrin unit which has the lower-energy excited statel.17 This is of particular relevance to attempts to model the behaviour of the primary light-absorption process of the photosynthetic reaction centre, which contains an array of several light-harvesting porphyrin units.9 The use of metal-polypyridyl complexes with pendant hydrogen-bonding substituents has lagged behind the use of porphyrin complexes but is beginning to be developed.The author's group in Bristol have prepared complexes 11-14, in which luminescent RulI or 0s" tris-bipyridyl cores are function- alised with the nucleobases adenine, thymine, cytosine or guanine respectively.It is well-known that excitation of a [R~(bipy)~]2+-typechromophore can result in energy-transfer to an [O~(bipy)~]2+chromophore if the two components are directly attached.4 Complexes 11-14 were prepared to see if a similar result could be obtained across a hydrogen bond. Thus. mixing components 11 and 12 results in association but with a rather small stability constant (K = 102 dm3 mol-I), which means that only a fraction of one percent of the components associate at the concentrations normally used for spectroscopic studies; the properties of the associated pair are difficult to detect in the presence of large excesses of the free component parts.However, use of the pair 13 and 14, with the triple cytosine-guanine hydrogen-bond replacing the weaker double adenine-thymine hydrogen bond, overcomes this: the associa- H I H 7 [R = CH~CGH~O(CH~)~O(CH~)~OCH~] ,OR w RO OR 8 (R= SiMe2But) R' R R 9 R=Bu' Chemical Society Reviews, 1997, volume 26 369 H 12+ ‘RL (4,4’-But2blpy)2 (4,4’-But2blpy)2 11 12 1 2+ \, .. ‘RL ’R; (4,4’-But2blpy)2 (4,4’-But2blpy)2 13 14 H2NYNYo12+ (R = SiMe2Bu) RO OR tion constant K is ca. 6000 dm3 mol-l in CH2C12, and in the associated pair the luminescence of the Ru“ energy-donor is quenched. Along similar lines, Sessler has prepared complex 15 370 Chemical Society Reviews, 1997, volume 26 (R = SiMe2Bu)wR‘= RO OR for possible attachment by hydrogen-bonding to porphyrin chromophores.l9 An associative interaction related to hydrogen bonding is the salt bridge between a deprotonated carboxylate and a protonated ‘sapphyrin’ pentapyrrolic macrocycle in assembly 16.20 Here, the anion is effectively ‘chelated’ by the four protons, and the geometry of this interaction serves to ensure that the two components must be mutually perpendicular.The pathway linking them is therefore clearly defined; in particular face-to- face association, which could provide an alternative energy- transfer pathway and thereby complicate interpretation of the results, is not possible. Excitation of the porphyrin fragment resulted in nearly complete (96%) energy-transfer across the salt bridge to the sapphyrin, with a rate constant of 1.8 X 109 s-I.The particular appeal of this system is that simple anions (sulfate, phosphate etc.) could be used to direct the assembly of large numbers of luminescent chromophores by formation of multiple salt-bridge interactions of this type. As an alternative to the molecular components described in all of the above examples, a small ‘nanocrystallite’ of the semiconductor Ti02 (diameter about 22 A)has been used as an electron-donor. Irradiation of the Ti02 fragment at 355 nm results in promotion of an electron across the band gap, from the valence-band to the conduction band. Charge-separation has therefore resulted, giving a high-energy electron and a low- energy hole (cf. Fig.1 which depicts the analogous situation for an individual molecule). In complex 17, a particle of Ti02 is encapsulated by a diamidopyridine derivative containing long alkyl chains which attach to the Ti02 surface by adsorption.21 The diamidopyridine forms a triple hydrogen-bond with the complementary uracil-based component to which a doubly alkylated 4,4’-bipyridyl fragment (a viologen) is attached. Viologen groups are good electron acceptors, and following irradiation of the Ti02 particle, electron-transfer to the viologen occurs across the triple hydrogen-bond. Several control experi- ments showed that the hydrogen-bond is essential for the electron-transfer to occur; no electron-transfer either through space or by diffusional encounter of the donor and acceptor components was detected.3.2 Proton-coupled electron-transfer Because hydrogen bonds involve (by definition) protic func- tional groups, they are also appropriate for studying proton- coupled electron-transfer in which inter-component proton transfer accompanies electron transfer, with the electron energytransfer Et' 'Et 16 (bandgapphoton) transfer and the proton transfer both mediated by the same interface. Coupling of proton transfer and electron transfer is common in many simple redox processes, such as the 2e/2H+ reduction of quinone (Q) to hydroquinone (benzene- 1,4-diol, H2Q), or the 2e/2H+ reduction of metal-oxo complexes [L,M=O] to aqua complexes [L,M-OH2].In biological systems it is of fundamental importance during photosynthesis. Absorp- tion of light results in a charge-separation process, i.e. generation of an electron on one side of the membrane and a positive hole on the other side. The electron is used to reduce NADP+ to NADPH on one side of the membrane, a process which also consumes protons. The hole is used ultimately to oxidise water to 02 on the other side of the membrane, a process which also liberates protons. The resulting proton imbalance generates a thermodynamic gradient which is the driving force for production of adenosine triphosphate (ATP). In the particular case of photoinduced electron transfer, where the transfer is initiated by electronic excitation of one component, the coupling of proton motion to electron transfer has recently been demonstrated in a few cases.In complex 18, there is a symmetrical double hydrogen-bond between the carboxylic acid substituents attached to the RuI1 fragment (electron donor in its excited state) and the dinitrobenzene (electron acceptor). In complex 19 in contrast the bridge is asymmetric, between a protonated amidinium group attached to the Ru" complex and a deprotonated carboxylate on the electron acceptor.2* Two pieces of evidence suggest that the photo- induced electron-transfer is indeed coupled to proton transfer. Firstly, although the thermodynamic driving force for electron transfer is significantly less for system 18 than for 19, the intramolecular electron transfer rate is faster (kET = 8.0 x lo6 s-l and 4.3 X 106 s-l, respectively) which is the opposite of what would be expected. This is because for 18 the double proton transfer is overall symmetrical [Fig.3(a)]and therefore does not result in any charge redistribution in the bridge. Consequently there is no need for the solvent interactions around the bridge to change, and the activation energy barrier is electron-transfer I,NO2 H 19 \ I -f-JJo;:o~Noz /\\N' Et Et NO2 20 Chemical Society Reviews, 1997, volume 26 371 (b) electron transfer H nmY+-H NIIIIIIIIHIIIIIIIIO NIIIIIIII~IH =O11111111110 -’;)-@ -@++ &’;)-@-H\ ,m~ ~ IIIIIIIIII~N-H l ~ ~ ~ ~ ~ ~ N-H~ lll~l~~~l~oo N-H / /H H H/ Fig.3 Examples of proton-coupled electron transfer, in which (a)double proton exchange within a carboxylic acid dimer results in no charge redistribution within the bridge (cf.complex IS), and (h)single-proton transfer from amidinium to carboxylate results in a redistribution of charge within the bridge (cfi complex 19). The bonds indicated in bold are those directly involved in the proton transfer. low. In contrast, the proton transfer in 19 results in a charge redistribution in the bridge, [Fig. 3(b)],which in turn requires changes in solvation. This imposes an activation energy barrier on the proton transfer and therefore, since the two processes are coupled, the electron transfer is also slowed. The second piece of evidence comes from the kinetic isotope effect; replacement of the H atoms in the hydrogen-bonded bridges of 18 and 19 by D atoms slows down the kET values by a factor of about 1.4 in each case.This means that cleavage of the 0-D bond (for 18) or the N-D bond (for 19) is involved in the rate-determining step; if no movement of the protons occurred with the electron transfer then substitution of H for D would not affect the rates. Systems such as 20 and 21 with porphyrins as the electron- donors have also been studied and show similar behaviour;l7 in 20, deuteration of the hydrogen-bonded bridge slowed down the photo-induced electron transfer by a factor of about 1.7. Why should proton motion be coupled to electron transfer at all? For example, in 19 it is easy enough to see why the proton transfer should be hindered (because of the change in solvation of the components required to stabilise the charge redistribu- tion), but why should that also slow the electron-transfer rate? The answer to this is that the strength of the electronic interaction between the electron donor and acceptor groups is strongly dependent on the position of the H atoms in the hydrogen bond.The equilibrium O-H-..O arrangement for a hydrogen-bond is not optimal for acting as a conduit for electron-transfer; it has been estimated that the electron-transfer rate is about four times faster if the hydrogen bond is in the symmetrical O.-.H-..O arrangement, which corresponds to the transition state between the O-H...O and O..-H-O extremes.The electron transfer will therefore occur just at the instant that the hydrogen bond is converting from one extreme to the other, which accounts for the experimental observations described above for complexes 18 to 21.24 4 Components linked by hydrophobic interactions Non-polar compounds tend to aggregate in polar solvents, particularly water, to minimise unfavourable solute-solvent interactions; this is responsible in part for the formation of micelles when detergents (with long hydrophobic tails) are added to water. This is a weak and directionally non-specific process but can be sufficient to permit components to associate to a sufficient extent to allow an electronic interaction between them which would not occur otherwise.An elegant example is provided by complex 22, which contains a cyclodextrin ‘bowl’ attached to each face of a porphyrin chromophore.25 Cyclodex- trins are basically hollow cylindrical molecules with a conical taper, whose hydrophobic interiors provide a suitable refuge for small non-polar molecules to hide from a polar solvent.* In the 372 Chemical Society Reviews, 1997, volume 26 presence of various hydrophobic quinone derivatives in water, quenching of the porphyrin luminescence was observed by electron transfer (kET = 10’ s-l) to the quinone held in the cyclodextrin cavity. Quinones without hydrophobic substi-tuents, which did not enter the cavity, did not cause any quenching of the porphyrin excited state. The excited state of the fluorescent pyrene derivative 23 is very efficiently quenched by the nucleosides 2’-deoxythymi- dine (dT), 2’-deoxycytidine (dC) and 2’-deoxyguanosine (dG) in aqueous solution, by an electron-transfer process in each case.2h The extent of quenching is much greater than would be expected to arise from random collisional encounter of the chromophore and quencher, which is ascribed to association of the components to give a non-covalently bonded [chromo- phore...quencher] complex. Although hydrogen-bonding and aromatic n-stacking between the components is feasible, the principal reason for this association is thought to be a hydrophobic interaction, since both chromophore and quencher components contain hydrophobic domains.Interestingly, quenching by dC and dT shows a substantial kinetic isotope effect when H20 as the solvent is replaced by D20, indicating the presence of a proton-coupled electron-transfer as the quenching step.Note that hydrogen bonding between these components is not thought to be significant; the proton transfer in this instance comes from the solvent (cf. complexes 18-21 in which proton coupling arose from shifts in the positions of the protons within the hydrogen bond). Non-protonated dC and dT are poor electron acceptors, and without assistance would not be able to accept an electron from the excited state of 23 as their reduction potentials are too negative; they are thermodynam- ically incapable of quenching 23 in non-polar organic solvents. However when protonated the positive charge makes them much easier to reduce, so the electron-transfer quenching that is observed can only occur if it is coupled to simultaneous protonation by the solvent to stabilise the reduced nucleoside.The solvent therefore plays two roles in the quenching; a kinetic one (driving the components together by a hydrophobic interaction) and a thermodynamic one (making the electron- transfer process energetically favourable by permitting simulta- neous proton transfer). 5 Components linked by aromatic stacking interactions The tendency of planar aromatic systems to ‘stack’ in an approximately parallel face-to-face arrangement (cf. the struc- ture of double-stranded DNA) has been known for a long time. Although its origins are still not completely understood, this type of interaction offers a well-established way of promoting self-assembly in supramolecular complexes.8 The interaction is known to be considerably strengthened when one of the / qu'none guest HO, @ cyclodextnnhost 22 aromatic systems is electron-poor and the other electron-rich, such thdt there is an electrostatic donor-acceptor component to the interaction and therefore some degree of charge transfer between the components Complex 24 contains d tris(bipyridy1)-ruthenium(rr) chromo-phore to which is appended dialkoxyphenyl substituents 77 2x A cyclic electron-acceptor (BXV4+) containing two viologen-based groups can associate with these by a n-stacking interaction between the electron-rich dialkoxyphenyl and the electron-poor methylviologen aromatic rings The association constant K is 1200 dm3 mol-' in water This association provides an intramolecular pathway for photoinduced electron transfer from the excited-state of the ruthenium chromophore to the methylviologen [cf Fig 2(a),displaying the same compo- nents but covalently linked] Following excitation, electron transfer to the BXV4+ quencher affords the photo-product Ru3+ (BXV )?+, which is particularly long-lived ((a 1 ps) because electrostatic repulsion between the oxidised Ru?+and the reduced (BXV )?+ prevents the back-transfer of an electron which would regenerate the Ru~+(BXV)4+ ground-state Viologen-based acceptors have been incorporated into supra- molecular assemblies with porphyrin-based electron-donors in the same way (eg complex 25) 29 Assembly of chromophore and quencher components using n-stacking has also been achieved vzu intercalation of the component parts into DNA strands Compounds with externally directed planar aromatic groups may bind to DNA by insertion of the aromatic group between the parallel base-pairs Com-plexes 26 and 27 can intercalate into DNA by virtue of their dipyridophenazine (26) or phenanthrenequinone-diimine (27) groups 30 The Ru" complex 26 is the chromophore, the RhIli complex 27 is capable of quenching the luminescence of ruthenium-based chromophore by accepting an electron from the photo-excited stdte In the presence of DNA, both components bind strongly by intercalation with association constants of K > lo6 dm3 mol-i, and on excitation of the chromophore, rapid electron transfe! to the quencher is observed over distances as long as 40 A This electron transfer occurs through the stacked array of 25 or more aromatic HO OH 23 \ transfer / heterocycles in the DNA \trdnd and I$ very fcM (X, r > 10') S-1) 6 Components linked by a combination of interactions The examples above were chosen to illustrctte the individudl non-covalent interactions between chromophore and quencher specie\ which can promote energy transfer or electron transter In many cases however a combination of interdction\ is u\ed to achieve the desired associdtion, and the exampla in thi\ \ection illustrate some examples of the careful we of severd type\ of interaction in concert to xhieve assembly between compo nent\ In complex 28 the two hydroxy groups on one \ide of the porphyrin plane provide a mean\ of selectively binding pu~N-quinones in such a fashion that face to face aromdtic n-\tacking of the quinone with the porphyrin dlso occurs, although mo\t of the strength of the interaction is thought to arise from the hydrogen bonds The two-point hydrogen bonding result\ in association constants of the order of 100 dm3 mol-I (in chloroform solution) for d variety of pal a-quinones Complex 29 is an extension of this principle in which four-point hydrogen bonding between the convergent hydroxynaphthyl substituents on the zinc-porphyrin and 2,3,5,6-tetrdmethoxy-p-benzoquinone is strong (K = 2 5 X lo5 dm mol in toluene) and highly specific, other pu~a-quinones are bound much less strongly v In both cases very fast (pico5econd timescale) and efficient electron transfer occur\ from the photo- excited porphyrin to the quinone, resulting in complete quenching In 30 the quinone quencher binds to the chromo- phore IZU a double hydrogen-bond with a peripheral hydro- quinone group, which again also results in face-to-fdce x-stacking of the two aromatic fragment5 33 A different type of approdch to the assembly of components is exhibited by 31, which contains covalently-linked chromo- phore (zinc-porphyrin) and quencher (quinone) unit\ linked by a covalent bridge 34 The molecule has a U-shdped \tructure, in which the chcomophore and quencher component\ dre \epdrated by about 9 A (centre-to-centre) on either side of the central Chemical Society Reviews, 1997, volume 26 373 (bipy) Rh'.25 27 28 (ethyl substituents on octaethylporphyrincore omitted for clarity) cavity. The size and shape of the cavity are ideal for the binding of an aromatic guest molecule. Dihydroxybenzene derivatives bind within the cavity of 31 by a combination of aromatic n-stacking with the porphyrin and quinone fragments on either side of the cleft, and hydrogen bonding to the amide carbonyl groups at the bottom of the cleft. The presence of an aromatic guest in the cleft greatly increases the efficiency of electron- transfer between chromophore and quencher.In CC14, very little electron-transfer across the cleft from chromophore to quencher could occur. The much longer pathway around the covalent bond system of the molecule could be discounted. On addition of hexyl-3,5-dihydroxybenzoatehowever, 75% of the lumines- cence intensity from the porphyrin unit was quenched, because incorporation of the aromatic guest into the cleft provided a facile pathway for electron-transfer, across the n-stacked system of aromatic groups (cf. electron-transfer through the stacked aromatic nucleobases of DNA).3O As a final example, complex 32 is particularly elegant because formation of the chromophore-spacer-quencher as-sembly relies on three types of non-covalent interaction, all mediated by the macrocyclic receptor unit.35 The quinone quencher is held in this macrocyclic receptor by a combination 374 Chemical Society Reviews, 1997, volume 26 of hydrogen bonding to the amide protons and n-stacking with the phenyl rings in the cyclic framework (K =: 3000 dm3 mol-I).The whole quencher-receptor assembly was then attached to the axial position of a zinc-porphyrin chromophore via coordination of the peripheral pyridyl group of the macrocyclic receptor. The complex, therefore, relies on a combination of metal-ligand coordinate bonds, n-stacking, and hydrogen bonding to assemble the components in a very specific way. Efficient quenching of the porphyrin lumines- cence by the quinone occurs within this assembly, by photo- induced electron transfer from porphyrin to quinone; this quenching is much more efficient than that which occurs between the porphyrin unit and free quinone by diffusion in solution.7 Conclusion Supramolecular methods clearly offer great promise for the assembly of high-nuclearity, structurally sophisticated com-plexes; nature uses such methods all the time, and synthetic chemists have started to use them more and more in the last few years. The application of such methods to the assembly of Ph A Ph 30 0 31 porphyrin plane \aromatic guest molecule ,Ar 32 photochemically active molecules, both for understanding natural photosynthetic processes and for the preparation of ‘unnatural’ systems for light-harnessing, is catching up rapidly and offers immense scope for further development 8 References 1 V Balzani and F Scandola, Sup1amoleculai Photoc hemistry, Ellis Horwood, Chichester, 1991 2 A Juris, V Balzani, F Barigelletti, S Campagna, P Belser and A von Zelewsky, Coord Chem Rev , 1988, 84 85, T J Meyer, Acc Chem Res , 1989, 22, 163 3 H Kurreck and M Huber, Angew Chem ,Int Ed Engl , 1995,34,849, M R Wasielewski, Chem Rev, 1992, 92, 435 4 J -P Sauvage, J -P Collin, J -C Chambron, S Guillerez, C Coudret, V Bdlzani, F Bangelletti, L De Cola and L Flamignl, Chem Rev, 1994, 94, 993, V Balzani, A Juns, M Venturi, S Campagna and S Serroni, Chem Rev, 1996,96,759 5 L F Cooley, C E L Headford, C M Elliott and D F Kelley, J Am Chem Soc , 1988, 110,6673 6 M R Wasielewski, D G Johnson, W A Svec, K M Kersey, D E Cragg and D W Minsek, in Photochemical Enet gy Con\ ersion, ed J R Norris and D Meisel, Elsevier, 1989, p 135 7 J -M Lehn, Supramolecular Chemistry, VCH, Weinheim, 1995 8 D Philp and J F Stoddart, Angew Chem Int Ed Engl, 1996, 35, 1154 9 W Kuhlbrandt, Nutwe, 1995, 374, 497, and refs therein, J Barber, Nature, 1988,333, 114 and refs therein 10 C A Hunter, J K M Sanders, G S Beddard and S Evans. J Chem SOC Chem Commun , 1989, 1765 11 H Imahon, E Yoshizawa, K Yamada, K Hagiwara, T Okada and Y Sakata, J Chem Sue Chem Commun , 1995, 1133 12 C A Hunter and R K Hyde, Angen Chem Int Ed Enql , 1996,35, 1936 13 P J F de Rege, S A Williams and M J Therien, Science, 1995,269.1409 14 P Tecilla, R P Dixon, G Slobodkin, D S Alavi, D H Waldeck and A D Hamilton, J Am Chem Soc . 1990, 112,9408 15 A Harriman, Y Kubo and J L Sessler, .I Am Chem Soc , 1992,114, 388, A Berman, E S Izraeli, H Levanon, B Wang and J L Sessler J Am Chern Soc, 1995,117, 8252 16 T Arimura, C T Brown, S L Springs and J L Sessler, Chem Commun , 1996, 2293 17 J L Sessler, B Wang and A Harriman, J Am Chem So( , 1995,117, 704 18 C M White, M F Gonzalez, D A Bardwell, L H Rees, J C Jeffery, M D Ward, N Armaroli, G Calogero and F Barigelletti, J Chern Soc Dalton Trans, 1997, 727, N Armaroli, F Bangelletti, G Calo-gero, L Flamtgni, C M White and M D Ward, submitted for publication 19 J L Sessler, C T Brown, R Wang and T Hirose, Inorg Chim Acta 1996,251, 135 20 V Kral, S L Springs dnd J L Sessler, J Am Chem Sor , 1995, 117, 8881 21 L Cusack, S Nagaraja Rao, J Wenger and D Fitzmaurice, Chem Muter, 1997, 9, 624, L Cusack, S Nagaraja Rao and D Fitzmaurice, Chem Eui J, 1997,3,202 22 J A Roberts, J P Kirby and D G Nocera, J Ant Chem Soc , 1995, 117, 8051 23 C Turro, C K Chang, G E Leroi, R I Cukier and D G Nocera, J Am Chem Soc , 1992, 114, 4013, J P Kirby, N A van Dantzig, C K Chang and D G Nocera, Tetrahedron Lett , 1995,36, 3477 24 R I Cukier, .I Phys Chem, 1994, 98, 2377, R I Cukier, .I Phyc.Chem , 1996, 100, 15428 and refs therein 25 Y Kuroda, M Ito, T Sera and H Ogoshi, J Am Chem Soc , 1993,115. 7003 26 V Y Shafirovich, S H Courtney, N Ya and N E Geactinov, J Am Chem Soc , 1995,117,4920 27 M Seller, H Durr, I Willner, E Joselevich, A Doron and J F Stoddart, J Am Chem Soc, 1994,116, 3399 28 M Kropf, E Joselevich, H Durr and I Willner, J Am Chem SOL , 1996, 118,655 29 M J Gunter and M R Johnston, J Chem Soc Perkin Trans I, 1994, 995 30 C J Murphy, M R Arkin, Y Jenkins, N D Ghatlia, S H Bossmann, N J Turro and J K Barton, Science, 1993, 262, 1025, M R Arkin, E D A Stemp, C Turro, N J Turro and J K Barton, J Am Chem Soc , 1996,118,2267 31 Y Aoyama, M Asakawa, Y Matsui and H Ogoshi, J Am Chem SOL , 1991,113,6233 32 T Hayashi, T Miyahara, S Kumazaki, H Ogoshi and K Yoshihara, Angew Chem Int Ed Engl, 1996,35, 1964 33 F D’Souza, J Am Chem Soc, 1996, 118,923 34 J N H Reek, A E Rowan, R de Gelder, P T Buerskens, M J Crossley, S De Feyter, F de Schryver and R J M Nolte, Anqefi Chem Int Ed Engl, 1997, 36, 361 35 C A Hunter and R J Shannon, Chem Commun , 1996, 1361 Received, 22nd May I997 Accepted, 27th June I997 Chemical Society Reviews, 1997, volume 26 375

ISSN:0306-0012    

DOI:10.1039/CS9972600365

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Asymmetric synthesis of building-blocks for peptides and peptidomimetics by means of the p-lactam synthon method Recent advances in the syntheses of enantiopure synthetic building blocks useful €or non-protein amino acids, dipep-tides, oligopeptides using solid state peptide synthesis,isoserines (norstatines), dipeptide isosteres [hydroxy(ke-to)et hylene, hydroxyethylamine, hydroxyethylene and dihy-droxyethylene isosteresl, taxoids, polyamines, poly(amino alcohol)^ and poly(amino ether)s, and other biologicallyactive compounds through applications of the P-lactarn synthon method is reviewed. 1 Introduction The fi-lactam skeleton has attracted significant interest among synthetic and medicinal chemists over the years mainly because it is the core structure of natural and synthetic D-lactam antibiotics.' The importance of fi-lactams as synthetic inter- mediates, however, had not been widely recognized when we siarted the development of the P-luctani spthon method in the early 1980s.Through the 1980sand 1990s,we and others have successfully demonstrated the usefuhess of enantiopure fi-iactams as versatile intermediates for the asymmetric synthe- sis of a variety of protein and non-protein amino acids, peptides, peptide turn rnirnetics, peptidornimetics, taxoid antitumour agents, heterocycles and other types of compounds of biological and medicinal interest. Thus, the /?-lacram synthon nieihod has now been fully established as a powerful synthetic rnethod.2.l We will describe here recent advances in the asymmetric synthesis of useful building blocks for peptides and peptidomi- metics by means of the fi-lactoni syntliori method.2 Preparation of enantiopure P-lactams as key intermediates Enantlopure 3-amino- and 3-hydroxy-P-lactarns are the key intermediates for the synthesis of peptides and peptidomimet-ics. Thus, we describe first the currently available methods for the synthesis of 3-amino- and 3-hydroxy-P-Iactams with high enantiopurity. Asymmetric ketene-imine [2 i-21 cycloaddition and ester enolate-imine cyclocondensation are the two major methods that have been successfully used for this purpose. 2.1 Through [2 + 21 cycloaddition of athiral ketenes to chiral imines The asymmetric Staudinger reaction, i.c.,ketene-irnine [2 + 21 cycloaddition, using enantiopure imines has been studied for some time, but it is rather difficult to achieve extremely high diastereoselectivity in general, mainly because of the unfavour-able arrangement of chiral moieties in the transition sta1e.l Nevertheless, enantiopure diastereorners can be isolated with-out difficulty.Several recent exampfes are shown below. 3-Phthalimido-P-Iactam 3a with 80% de (de = diastereo-rneric excess) was obtained by using imine 1 derived from (R)-phenylethylamine (Scheme 1 ).5 Chemical Society Reviews, 1997, volume 26 377 1 2 74% Et3N1 1 %Gc,0 + 0+N>Ph MeMe 3a (90 : 10) 3b Scheme I (R)-1 -Naphthylethylamine was used as the chiral auxiliary for the synthesis of 3-phenoxy-P-lactam 6a with 66% de.6 The separation of the two diastereomers by column chromatography on silica gel is reported to be facile (Scheme 2).When (S)-1-(2-~hlorophenyl)ethylaminewas used as the chiral auxil- iary, the reaction proceeded with only 42% de (Scheme 2).7 PhO?"' 0 + Ar 4 5 Ar/ 6a 6b Ar = 1-Np, 83:17 Ar = 1 -CI-C6H4, 71 :29 Scheme 2 0-TPS-(R,R)-Threonine PNB ester (TPS = triphenylsilyl; PNB = 4-nitrobenzyl) was employed as the chiral auxiliary in the synthesis of 3-azido-fl-lactam 9a to achieve 90% de (Scheme 3).8Interestingly, when the chiral auxiliary has a free 13-hydroxy group, no selectivity is observed. The [2 + 21 cycloaddition of azidoketenes to benzy-lideneamines bearing a chiral 13-lactam backbone was found to be extremely stereoselective (>99% de), leading to the formation of enantiopure bis-(3-lactams (Scheme 4).9 3-Benzy- lidenearnino-(34actams, 12a (3R,4S) and 12b (3S,4R), were obtained through the [2 + 21 cycloaddition of azidoketene to the enantiopure imino ester 10 followed by chromatographic separation of two diastereomers, lla and llb (80%, lla/llb = 51/49), followed by reduction of the azido moiety and then imine formation.Each 3-benzylideneamino-~-lactam12 was converted into the corresponding bis-(3-lactam 13a or 13b through [2 + 21 cycloaddition with azidoketene to give 13a or 13b. In these cycloadditions, only one of the two possible 378 Chemical Society Reviews, 1997, volume 26 C02PNB 7 8iEt3N + C02PNB C02PNB 9a R = H, 5050 9b R = TPS, 955 Scheme 3 oflyMe yMePhCH=N xCO~BU' + 0pN I I CO~BU' CO~BU' 10 lla llb 0hNYMe 0&Ay"" 0hhyMe CO~BU' CO~BU' / CO~BU' 12a 14a: X = H 15a: X =Ac Me 1 A~NH*COHN ~COHN*co2~Ut (S,R,S)-1 6 0kNyMe &Ay".CO~BU' CO~BU'12b 14b: X = H 15b:X =Ac AcNH-COHN '%OHN AC02But (R,S,S)-l6 Scheme 4 Reagents: (a) N3CH2COCl, Et3N, CH2C12, -78 "C to room temp. (h)H2 (1 atm), 5% Pd-C, MeOH, 0-5 "C. (c) PhCHO, Na2S04, CH2C12. (6)Ac20, N-methylmorpholine, CHC13. (e) H2 (1 atm), 10% Pd-C, EtOH, 50 OC. stereoisomers was formed in each case. The relative stereo- chemistry was determined by 1H NMR and the absolute configuration was determined by converting 13a and 13b to known tripeptides (S,R,S)-and (R,S,S)-16. When a chiral auxiliary is attached to the C-terminus of an imine such as 18 derived from (lS,2S)-2-arnino-l -phenylpro- pane- 1,3-diol, excellent diastereoselectivity is achieved and the predominant isomers are isolated in > 90% yields (Scheme 5).'0 BocN cis-4-benzyl-P-lactam 31a with >96% de in 55% yield (Scheme 9).14 For the [2 + 21 cycloaddition of chiral ketenes 32 generated from 25 (a: S, b: R) with chiral imino esters 33 derived from alanine, valine, phenylalanine and methionine, it has been shown that no appreciable double induction is observed and Rh -CI + "CH2Ph 25a 1Et3N 26 27a 80-90 Yoyield 27b 95:5-97.3 Scheme 7 2825a J-AH &AH 0 0 29a 29b R = TIPS, 59 Yoyield, 80:20 R = TBDPS, 50 Yo yield, 9O:lO Scheme 8 PhOCO 22b (86:14) Ph",'..,,\22/ n Et3N, CH2C12 "/,, ...+Go --I-23 "C room temp OpN\CHr;h 17 18 RX = phthalimido, 91% 19 RX = PhO, 9.4% Scheme 5 2.2 Through [2 + 21 cycloaddition of chiral ketenes to imines Another approach to the asymmetric ketene-imine [2 + 21 cycloaddition is to attach a chiral auxiliary to a ketene. This strategy has been shown to be very successful for the asymmetric synthesis of 3-amino-6-lactams. With the use of (+)-tartaric acid derivative 20 as the ketene precursor, (S-lactam 22 (PMP = 4-methoxyphenyl) was ob- tained with 72% de (Scheme 6).l13l2 When ephedrine derived chiral ketene precursor 23 was used, >90% de was achieved for the synthesis of P-lactam 24 (Scheme 6).Oxazolidinones derived from (S)-and (R)-phenylglycine are excellent chiral auxiliaries for the asymmetric ketene-imine [2 + 21 cyc10addition.l~ For example, the reaction of (S)-oxazolidinon-3-ylacetyl chloride 25a with N-benzylaldi- mines 26 gave 13-lactam 27a with 90-94% de in 80-90% yield (Scheme 7).'3 The same oxazolidinone chiral auxiliary was used for the asymmetric synthesis of trans-p-lactam 29. By the introduction of a bulky a-siloxy group (TIPS = triisopropylsilyl; TBDPS = tert-butyldiphenylsilyl) to an imine, the relative stereochemistry of the resulting P-lactam 29a was switched to completely trans with 60-80% de (Scheme 8). Usually, enolizable imines cannot be used for the ketene- imine [2 + 21 cycloaddition because of its facile isomerization to enamines.However, N-bis(trimethylsilyl)methylaldimines such as 30 were found to circumvent this problem. Accordingly, the reaction of 25a with 30 in the presence of triethylamine gave PhOCO 20 21 22a base P 0)-kPMP PMP 23 21 24a (>955) 24b Scheme 6 Chemical Society Reviews, 1997, volume 26 379 2.3 Through asymmetric cyclocondensation of chiral ester enolates with achiral imines The ester enolate-imine cyclocondensation provides another efficient route to (3-lactams, which can be applied to asymmetric synthesis of 3-amino- and 3-hydroxy-(3-lactams. 25a 30 55% Et3N1 Me3Si Me3Si 31a (>98.2) 31b Scheme 9 only the chiral centre in the ketene played a key role in the asymmetric synthesis (Scheme lo).*The reaction gave only one of the two possible diastereomers in all cases examined.The 6-lactams 34 thus obtained were converted to the corresponding N-protected dipeptides through hydrogenolysis over Pd/C in MeOH and then saponified. The N-protected dipeptides 35 can be used for fragment condensation with other N-terminus-free peptide units. Removal of the chiral auxiliary was carried out by a modified Birch reduction with lithium in liquid NH3/THF/ ButOH to afford enantiopure dipeptides 36 in excellent yields (Scheme 1O).* 25a:(S) I1 25b:(R) 32 35 34 Id The cyclocondensation of chiral lithium enolates generated in situ from N,N-bis(sily1)glycinates 37 with N-PMP-imines 38 gave trans-3-amino-P-lactams 39 exclusively or as the predo- minant product depending on the chiral auxiliary R* used (Scheme 11).Among the chiral auxiliaries examined, the best results were obtained with (-)-menthy1 and (+)-or (-)-trans- 2-phenylcyclohexyl, which gave 39 with >99% ee.3 M\" ,Me [?-CH2COOR* Me/ Me 37 39 R1 = Ph, pF-c~H4, pCF3-C6H4,4-Me0-C6H4,3,4-(Me0)2C6H3 Scheme 11 3-Hydroxy-(3-lactams 42 with high enantiomeric purity (90-98% ee) have been obtained through highly efficient chiral ester enolate-imine cyclocondensation of enantiopure O-TIPS- hydroxyacetate 40 with N-TMS-imines 41a or N-PMP-imines 41b (Scheme 12). A variety of aryl, alkenyl and alkyl substituents can be introduced to the C-4 position of 3-hydroxy- P-lactam~.~>~~--'7Among the chiral auxiliaries examined, (+)-and (-)-trans-2-phenylcyclohexyl gave the best results.N-TMS-imines of aromatic aldehydes work very well in this reaction, giving 6-lactam 42 in one step. N-PMP-imines are suitable for alkenyl-, alkyl-, and fluoroalkyl aldimines that are rather labile. The resulting N-PMP-p-lactams 43 are readily converted to 42 by oxidative cleavage using CAN [cerium(Iv) ammonium nitrate]. The N-acylated derivatives of 6-lactam 42 have been successfully used for the practical syntheses of Taxol (paclitaxel) and Taxotere (docetaxel) as well as for the development of highly potent new taxoid antitumour agents.*,1*-20 As described above, a variety of 3-amino- and 3-hydroxy- (3-lactams with excellent enantiopurities are now available based on asymmetric ketene-imine [2 + 21 cycloaddition and chiral ester enolate-imine cyclocondensation. This forms the foundation for the applications of these p-lactams to the asymmetric syntheses of non-protein amino acids, peptides, dipeptide isosteres and peptidomimetics that are not easily prepared by conventional methods.The next section reviews recent applications of the p-lactam synthon method along these lines. TIPSO,, .,,R1 TIPSOCH2COOR* 1 LDA * 40 2 R1CH=N-TMS(41a) 3 H20 96-98%ee I R* = (+)-trans-2-phenylcyclohexyl I TIPSO,,,,TIPSOCH&OOR* LDA pi TIPSO',,>p1 40 2 R1CH=N-PMP(41b) +$&MPE PNH 36 0 0 R = Me, Pr', PhCH2, MeS(CH2)2; Ar =Ph 72-95% 43 42 96-98Y0ee Scheme 10 Reagents: (a)NEt,, CH2Cl2, -78 "C. (b)CH2C12, -78 to 0 "C, 2 h.(c)(i) H2, Pd-C, MeOH, SO "C, S h, (ii) 1 M NaOHlTHF, room temp., 1 h, H?O+.(d)Li/NH?/Bu'OH. -78 "C, 15 min. R' = Aryl, Alkenyl, Alkyl, etc. Scheme 12 380 ChemicaE Society Reviews, 1997, volume 26 3 Synthesis of enantiopure non-protein amino acids, their derivatives, and dipeptide isosteres from P-lactam intermediates Non-protein amino acids are defined as the amino acids which are not found in protein main chains because they do not have a specific transfer-RNA and codon triplet or they do not arise from protein amino acids by post-translational modification Their origins and functions are so diverse that many things remain to be explored The majority of naturally occurring non- protein amino acids originate in plants and micro-organisms Many of them are important components of therapeutic drugs and compounds of medicinal interest such as antibiotics, decarboxylate inhibitors, aminotransferase inhibitors, protease inhibitors and anticancer agents Thus, the development of efficient methods for the synthesis of enantiopure non-protein amino acids is significant in organic synthesis and medicinal chemistry 3.1 cx,P-Diamino acids and their derivatives Enantiopure a,P-diamino acids are often components of peptidic antibiotics such as lavendomycin or glumamycin 21 It has been shown that the hydrolysis of enantiopure 3-amino- p-lactams affords the corresponding a,fi-diarnino acids 2 As Scheme 13 illustrates, the acid hydrolysis of (S,R)-44 gives (S,R)-4S in quantitative yield as hydrochlorides, that is reduced to (S,R)-diamino alcohol 46 in high yield with LiAIH4 (LAH) The ciy-fi-lactam (S,R)-44 is readily converted to the ttans-f3-lactam (R,R)-44through imine formation, deprotonation and protonation Then, (R,R)-44 is transformed to (R,R)-4S and (R,R)-46in the same manner as (S,R)-44 Since the other set of diastereomeric P-lactams, (R,S)-44can be obtained by using the other enantiomer of the chiral auxiliary in the chiral ketene- imine [2 + 21 cycloaddition process (see Scheme 7), four diastereomers of a,fi-diamino acids 45 and diamino alcohols 46 are accessible by this method The h ans-P-lactams (R,R)-44 and (S,S)-44can also be obtained through chiral ester enolate- imine cyclocondensation (see Scheme 11) This protocol has been combined with the alkylationc at C-3 position (I ide sup a) and applied to the syntheses of enantiopure a-alkyl-a$-amino acids (SR) 44 (R R) 44 .p H 2COOH H NH2 (SR)45 yH H NH2 H NH2 R+OH ,+OH H 'NH2 H NH2 (SR)46 (R R) 46 Scheme 13 3.2 P-Hydroxy-P-amino acids (isoserines), their dipeptides and dipeptide isosteres a-Hydroxy-f3-amino acids (isoserines) with correct relative and absolute configurations are key components of a large number of therapeutically important compounds For example, (2R,3S)-3-amino-2-hydroxy-5-methylhexanoic acid (norstatine), (3R,4S)-4-amino-3-hydroxy-5-methylheptanoicacid (statine), and their analogues have been used extensively as crucial amino acid residues in peptide-based inhibitors of enzymes such a\ renin2* and HIV-I protease 23 N-Benzoyl-(2R,3S)-3-phenyl-isoserine and N-tert-butoxycarbony1-(2R,3S)-3-phenylisoserine moieties are at the C-13 positions of the exciting anticancer agents paclitaxel and docetaxel 73 25 Accordingly, it is very important to develop efficient methods for the synthe\es of isoserines with excellent enantiopurity As Scheme 14 shows, norstatine (47, R = But) and cyclohexylnorstatine (47, R = cyclohexylmethyl), key com- ponents of renin inhibitors, have been prepared quantitatively through ring-opening hydrolysis of the corresponding (IJ-lac- tams 42 obtained through highly efficient chiral ester enolate- imine cyclocondensation (see Scheme 12) 7-Other nor\tdtine analogues are obtained in the same manner 2 42 47 R = Bu' CyCH2 Ph PhCHICH Ph(CH2)2 Cy(CH2)2 Scheme 14 Cyclohexylnorstatine isopropyl ester SO ha\ a150 been prepared through ring opening hydrolysis of P-lactam 48 obtained by chromatographic separation of the desired dia5ter- eomer from ketene-imine [2+ 21 cycloddducts, followed by hydrogenolysis on Pd/C (Scheme 15) n Me PCOOPr Phfi N H 0A'LPh OCH2Ph 49 48 H2 10% PdiC 99'& Me2CHOH room tempI OH 50 Scheme 15 N-Acylation of (3-lactam4 with, L' q benzoyl and mt-butoxycarbonyl (Boc), increases the electrophilic nature of (IJ-lactams Thus, 1-acyl-P-lactams readily react with various nucleophrles such as amines, alcohols, metal alkoxides and metal enolate5 2 27 For instance, the ring-opening coupling of I -acyl-fi-lactams 51 with a-amino acid methyl esters proceed4 smoothly at ambient temperature under neutral conditions to give N-Boc-dipeptides containing isocerine residue5 52 in excellent yields (Scheme 16) This novel peptide coupling method has been successfully applied to a solid-pha\e peptide synthesis using the Wang resin-bound amino acid residues 53 (Scheme 16) 2 It is worth mentioning that (3R,4S)-1 -benzoyl-3-siloxy-4-phenylazetidin-2-one and (3R,4S)- 1-Boc-3-siloxy-4-phenyl-azetidin-2-one have successfully been applied to the practical syntheses of paclitaxel and docetaxol.through highly efficient ring-opening coupling with metalated baccatin I11 deriva-tives 2 20 The reaction of 1-Boc-3-(protected hydroxy)-P-lactam 55 with Grignard reagents affords the corresponding a-hydroxy- Chemical Society Reviews, 1997, volume 26 381 HO,,,. R'pN': R2 + H2N 'kOOMe 0 Boc I CH2C12,room temp. c BocHN uN'COOMe Hz~ OH 52 R' = Bu', Ph, CyCH2, PhCH = CH; R2 = PhCH2, Bu', Pr', Indolemethyl; R3 = Me, But 54 Scheme 16 @-amino ketones 56 in high yields (Scheme 17). Cuprates can be used as highly chemoselective alkylating agents for this reaction, but yields are moderate.2* This reaction has been successfully applied to a practical synthesis of a key component 58 of a potent renin inhibitor (Scheme 17).29 BocNH 0 i II or R2CuLi, Et20,-1 0 "C to room temp.56 55 n OTES 57 58 Scheme 17 N-Boc-13-lactams 59 react with ketone and ester enolates to yield hydroxy(ket0)ethylene dipeptide isostere~,~~ Two exam- ples are shown in Scheme 18: 4-isobutyl-P-lactam 59a and 4-cyclohexylmethyl-(-lactam 59b give ring-opening coupling products 60 and 61, respectively, in good to excellent yields through reactions with lithium enolates.30 Lithium enolates of ketones and esters such as acetone, acetophenone, phenyl ethyl ketone, ethyl acetate, ethyl propionate and methyl 2-methylpro- panoate have been employed. The reaction is faster with ester enolates than ketone enolates. 382 Chemical Society Reviews, 1997, volume 26 THF, -78 "C BocHN ~ 0' C 02B ut 96% OTIPS 59a 60 THF, -10 "c ~EtO LBocHN HPh 86% OTIPS 59b 61 Scheme 18 The N-Boc-epoxides 63 are important intermediates for the synthesis of various non-protein amino acids.The epoxides 63 can be readily prepared from the corresponding 13-lactams 59 in two steps (Scheme 19).3l TIPSO,,, .,,R I. MeOH, DMAP, ! TEA, reflux 2. LAH, diethyl ether) 'OcNH b O H f ref lux OH80-86~~ 6259 R U6575% 63 R = Bu', CyCH2, 2-methylprop-I -enyl Scheme 19 Hydroxyethylamine isosteres are readily synthesized from epoxides 63. For example, epoxide 63c undergoes facile ring- opening reactions with methyl esters of (S)-phenylalanine and (S)-proline to give the corresponding dipeptide isosteres 64 and 65, respectively in good isolated yields (Scheme 20).31 HCI.Phe-OMe BocHN ,,,o TE:e:zoHt BocL : HN '~~Otvle 74% OH 63c 64 A HCI.Pro-OMe C02Me 63c 65 Scheme 20 Hydroxyethylene isosteres are also easily obtained from epoxides 63.An example is shown in Scheme 21.3' The reaction of an allyl-Grignard reagent with epoxide 63a gives alcohol 66 in excellent yields. The hydroxy and the amino groups are protected as an oxazolidine and the subsequent ozonolysis and oxidative work-up afford the corresponding N,O-protected hydroxyethylene isostere 68 in good isolated ~ield.3~ X Y Bc +N\ 66 0 Me63a 69 70 dirnethoxypropane Li, NH3 THF. Bu'OHPPTS,90-95 "C BocN t c--%-0 H30+ 85% '3,67 Y ,,,Me H2N /ICOOH H2N *CONHMe 72 71 1)MeOH-CH2C12 Scheme 2203,-78 "C BocN .Lc02H2) H202, NaOh 3) 0.1 M HCI 80% \ 68 Scheme 21 3.3 a-Alkyl-a-amino acids, their derivatives, and their dipeptides Among the non-protein amino acids, a-alkyl-a-amino acids have been attracting medicinal and biochemical interest because many of them serve as powerful substrate-based inhibitors of enzymes such as decarboxylases and aminotransferases.21 a-Alkyl-a-amino acid residues also serve as conformational modifiers of physiologically active peptides, bringing in conformational restraints. a-Alkyl amino acids also provide a challenging synthetic problem since conventional enzymatic resolution in combination with racemization cannot be applied effectively.Accordingly, the asymmetric synthesis of a-alkyl- a-amino acids with excellent enantiopurity has been ex-tensively investigated, and the p-lactam synthon method provides one of the most efficient routes to these amino acids.2.3 Two types of extremely stereoselective alkylations of (3-lactams have been studied: (i) the alkylation of the C-3 carbon (Type I), and (ii) the alkylation of the side chain ester enolate (Type 2). For Type 1 alkylation, an electrophile should attack the C-3 position of the p-lactam enolate from the opposite side of the C-4-aryl group to avoid steric conflict. In the Type 2 alkylation, the lithium enolate forms a chelate with the (3-lactam oxygen and then the electrophile attacks from the back side of the aryl group at C-4.E' X The Type 1 alkylation has been applied to the asymmetric synthesis of (S)-a-methylphenylalanine72a (X = Y = OMe) and (S)-a-methyldopa 72b (X = Y = OH) with > 99.5% ee as shown in Scheme 22.2 The methylation at C-3 of cis-@-lactam 69 gave 70 with >99.5% de. The Type 1 alkylation is applicable to trans-(3-lactam 73 (Scheme 23) since the reaction goes through the same Type 1 enolate as that from the corresponding cis-(3-lactam. The alkylation proceeded smoothly for methylation or allylation to give 74 with >99.5% de in >90% yield.2 This protocol has been applied to bicyclic (3-lactam 75 to give 76 with 297% de in 5149% yield (Scheme 23).32 2.1. LiHMDSRX > &N\0 PMP THF, -78 "C 73 R = Me, ally1 1.KHMDS, -78 "C 2. RX,-78 "C * 75 R = Me, allyl, Bu Scheme 23 PhvN$-fPh PMP 74 MPh O' I okNfiJ 0 76 The Type 1 alkylation has also been applied to 3-DMPSO- (3-lactam 77 (DMPSO = dimethylphenylsiloxy) to afford 78, which is transformed to a-methyl- and a-allylphenylisoserines 79 as well as 1 -Boc-fi-lactams 80 with >99% ee (Scheme 24).33 l-Boc-3-methyl-(3-lactam 80a has been used for the synthesis of the 2'-methyl analogue of d0cetaxel.3~ DMPSo+,,. .\Ph I) LDA, -78 "C RDMPSO/,,,, h,.,,ph THFn" ____) 2) RXp\ 0 PMP >99% ee PMP 77 / 78 1) CAN 2) Bog0TEA, DMAP 2)6M HCI R J Ph 0 DMPSOI,,,, i I1 80 R = Me, ally1 79 Scheme 24 The Type 2 alkylation was applied to the asymmetric synthesis of (R)-a-alkylalanines 84 (X = PhO) as well as (S)-phenylalanyl-(R)-a-alkylalanines 83 (X = CbzNH) (Cbz = benzyloxycarbonyl) (Scheme 25).2 The Type 2 alkylation has been applied to the sequential asymmetric double alkylation of chiral (3-lactam ester 85 that is a chiral glycinate as well as a phenylglycinate equivalent (Scheme 26).2 The (3-lactam ester 85 ( > 99% ee) was prepared Chemical Society Reviews, 1997, volume 26 383 0pt Ph through asymmetric [2 + 21 cycloaddition of a chiral ketene to 1, LDAlTHF fey?-butyl-N-benzylideneglycinate(see Scheme 10).As shown* Me 2. RBr in Scheme 26, the salient feature of this method is that the quaternary chiral centre of the desired configuration can be COOBu' COOBu' achieved just by changing the order of the addition of two alkyl 81 82 R = PhCH2, Et X = CbzNH, PhO H2/PdC R Me 1 a4 t 83 Scheme 25 I COOBu' 1.LiHMDS 2. R'X / 86-R 1. LiHMDS \ 2. R2X COOBu' 86a' 1. LiHMDS 2.Mel * Ph -78 "C, THF 95% CO'BU' 85 1. TFA, 20 'C 2. Li/NH3/THF'zph?Me ButOH,-78"C H2N C0NH2c02H -3. Dowex 5OX-2 (S,R)-90 62% 384 Chemical Society Reviews, 1997, volume 26 1.LiHMDS 2.R2X 86a-R halides used (Rl # RZ). The reactions were carried out using the combinations of methyl iodide, allyl bromide, and benzyl bromide, and doubly alkylated (3-lactam esters 87 with > 99% de were obtained in high yields. The 0-lactam esters 87 were converted to the corresponding (S,S)-and (S,R)-phenylalanyl- a-alkylalanines 88.2 The sequential asymmetric triple alkylation of 85 was performed through the combination of Type 2 and Type 1 alkylations (Scheme 27).2 After the completion of the sequen- tial Type 2 asymmetric double alkylations of the glycinate moiety with methyl iodide and allyl bromide, the side chain of the resulting @-lactarn ester 87 had no acidic protons.Thus, the Type I alkylation with methyl iodide took place at the C-3 of (3-lactam87. It was found that the first Type 2 double alkylation as well as the subsequent Tlpe 1 alkylation proceeded with virtually complete stereoselectivity to give 89, which is converted to (S)-a-methylphenylalanyl-(R)-a-allylalanine90 in good yield.QvphR' 0hN*Ri 1 1.TFA 87-s Ph R,' R' Hr,, l-,..I (S,/?)-a8 ATFA I COOBu' 87-R Scheme 26 1. LiHMDS 2. CH2=CHCH2Br CO~BU' 87a-R 89 Scheme 27 3.4 Polyamines, poly(amino alcohols), and poly(amino ethers) via bis-6-lactams Polyamines, poly(amino alcohols) and poly(amino ethers) are very useful compounds which can be directly synthesized from azetidines with various substitution patterns. However, the azetidine skeleton has been one of the most difficult amines to synthesize because of its ring strain. Only a few effective synthetic methods have been developed to date, and the selective reduction of (3-lactams serves as one of the most straightforward and efficient routes to enantiopure azeti-dines.35 It has been shown that monochloroalane (AIH2Cl) and dichloroalane (AIHC12) are the best reducing agents that can convert enantiopure 6-lactams 91 to the corresponding azeti- dines 92 in high yields without racemization (Scheme 28).35 The reductive cleavage of N-C2 bond of the resulting azetidines 92 affords the corresponding amino alcohols or diamines 93 in virtually quantitative yields (Scheme 28).35 This protocol has been successfully applied to bis-(3-lactams.Two types of bis-(3-lactams, i.e., bis-P-lactams in which two (3-lactam rings are directly connected and bis-13-lactams that have an alkanamide linker between two (3-lactam rings, have been examined. As Scheme 29 shows, the chloroalane reduction of these two types of bis-(3-lactams exemplified as 94 and 96 gives the corresponding bisazetidines 95 and 97, respec-tively.35 The reductive N-C2 bond fission and removal of benzyl group through hydrogenolysis of azetidines and bisazetidines on Pd-C or Raney nickel affords the corresponding hydroxy- lamines, diamines, polyamines, poly(amino alcohols) and poly(amino ethers) in high to quantitative yields as exemplified in Scheme 30.35 It should be noted that the reductive cleavage of azetidines and bisazetidines is much faster than that of benzyl-oxygen bond, and thus the poly(amino alcohols) can be obtained as their benzyl ether~.3~ These hydroxylamines, diamines, polyamines, poly(amino alcohols) and poly(amino ethers) serve as chiral chelating agents or catalysts as well as versatile enantiopure building blocks for peptidomimetics and chiral macrocycles.X Ar x, Ar A~TNHR .OMe )Me HZ/PdC MeOH, HCI,, 81% 'CH2Ph 98 99 MeOH, room temp ' iOOo/o Phj 100 101 /bH HO <OH 102 103 H2IPdC NMeOH, room temp""^ qh100% I -NH ,Ph f"'HL H"n? 2 105Ph Scheme 30 R R 90 91 X = PhCH20, N3 Scheme 28 Et20 0hNy.Me 85%I CO~BU' 94 061?,,e H2IPd-C AIHzCI- hNHMPh 0 Et20 0 3370 \I\f 'Od2Ph 96 97 Scheme 29 Y 92 Y = OH, NH2, AcNH .NyMe 1 'OH 95 4 Conclusion This short review has described recent advances in the asymmetric synthesis of enantiopure building blocks for peptides and peptidomimetics by means of the p-lactarn synthon method.3-Amino- and 3-hydroxyl-(3-lactams with excellent enantiopurity can be obtained via asymmetric ketene-imine [2+2] cycloaddition or chiral ester enolate-imine cyclo-condensation. The reductive cleavage of the N-C4 bonds of 3-amino-(3-lactams provides non-protein amino acids, their derivatives, and their dipeptides. Extremely stereoselective Type 1 and Type 2 asymmetric alkylations as well as sequential double and triple asymmetric alkylations of 3-amino-(3-lactams, followed by selective hydrogenolysis afford a variety of a-alkyl-a-amino acids, their derivatives, and their dipeptides, that cannot easily be achieved by other methods. The Type 1 asymmetric alkylation has also been successfully applied to 3-hydroxy-P-lactams which give a-alkyl-a-hydroxy-13-amino acids and their derivatives.Facile hydrolysis of 3-hydroxy-P-lactams as well as N-acyl-3-hydroxy-(3-lactamsaffords a-hydroxy-(3-amino acids (isoserines or norstatins). The ring- opening coupling of N-acyl-3-hydroxy-(3-lactamswith a-amino esters and enolates and further manipulations provide isoserine- containing dipeptides and various dipeptide isosteres. The selective reduction of enantiopure (3-lactams and bis-(3-lactams with chlorohydroalanes, followed by reductive ring cleavage of the resulting azetidines and bisazetidines furnishes a unique and efficient route to hydroxylamines, diamines, polyamines, poly(amino alcohols) and poly(amino ethers) which are also Chemical Society Reviews, 1997, volume 26 385 useful and versatile building blocks for peptidomimetics and chiral macrocycles.5 Acknowledgments This research was supported by grants from the National Institutes of Health (NIGMS) and the Center for Biotechnology at Stony Brook which is sponsored by the New York State Science and Technology Foundation. General support from Ajinomoto Co., Inc. is also gratefully acknowledged. 6 References 1 G. I. Georg, The Organic Chemistry of PLactams, VCH, New York, 1992, and references cited therein. 2 I. Ojima, Acc. Chem Res. 1995, 28, 383. 3 I. Ojima, in Advances in asymmetric synthesis, J. P. Inc., New York, 1995, VOI. 1, pp. 95-146. 4 G. I. Georg and V. T. Ravikumar, in The Organic Chemistry of PLactams, ed.G. 1. Georg, VCH Publishers, New York, 1992; p. 295. 5 J. Aszodi, A. Bonnet and G. Teusch, Tetrahedron, 1990,46, 1579. 6 G. 1. Georg and Z. Wu, Tetrahedron Lett., 1994, 35, 381. 7 Y. Hashimoto, A. Kai, and K. Saigo, Tetrahedron Lett., 1995, 36, 8821. 8 A. K. Bose, M. S. Manhas, J. M. V. D. Veen, S. S. Bari and D. R. Wagle, Tetrahedron 1992, 48,483 1. 9 I. Ojima, K. Nakahashi, S. M. Brandstadter and N. Hatanaka, J. Am. Chem. Soc., 1987,109, 1798. 10 M. Jayaraman, A. R. Deshmukh and B. M. Bhawal, Tetrahedron, 1996, 52, 8989. 11 N. Ikota and A. Hanaki, Heterocycles, 1984, 22, 2227. 12 R. D. G. Cooper, B. W. Daugherty and D. B. Boyd, Pure Appl. Chem., 1987, 59, 485. 13 D. A. Evans and E. B. Sjogren, Tetrahedron Lett., 1985, 26, 3783. 14 C.Palomo, J. M. Aizpurua, M. Legido, R. Galarza, P. M. Deya, J. Dunoguks, J. P. Picard, A. Ricci and G. Seconi, Angew. Chem., Znt. Ed. Engl., 1996, 35, 1239. 15 I. Ojima, S. D. Kuduk, J. C. Slater, R. H. Gimi and C. M. Sun, Tetrahedron, 1996, 52, 209. 16 I. Ojima, S. D. Kuduk, P. P. Pera, J. M. Veith and R. J. Bernacki J. Med. Chem., 1997,40,279. 17 I. Ojima, J. C. Slater, P. Pera, J. M. Veith, A. Abouabdellah, J.-P. BCguC and R. J. Bemacki, Bioorg. Med. Chem. Lett., 1997, 7,133. 18 I. Ojima, J. C. Slater, E. Michaud, S. D. Kuduk, P.-Y. Bounaud, P. Vrignaud, M.-C. Bissery, J. Veith, P. Pera and R. J. Bernacki, J. Med. Chem., 1996,39, 3889. 19 I. Ojima, J. C. Slater, S. D. Kuduk, C. S. Takeuchi, R.H. Gimi, C.-M. Sun, Y. H. Park, P. Pera, J. M. Veith and R. J. Bernacki, J. Med. Chem., 1997,40,267. 20 R. A. Holton, R. J. Biediger and P. D. Boatman, in Taxol: Science and Applications, ed. M. Suffness, CRC Press, New York, 1995, pp. 97-121. 21 M. J. Jung, in Chemistry and Biochemistry ofAmino Acids, ed. G. C. Barrett, Chapman and Hall, New York, 1985, p. 227. 22 S. Thaisrivongs, D. T. Pals, L. T. Kroll, S. R. Turner and F.-S. Han, J. Med. Chem., 1987, 30, 976. 23 J. R. Huff, J. Med. Chem., 1991,34, 2305. 24 Tuxane Anticancer Agents: Basic Science and Current Status, ed. G. I. Georg, T. T. Chen, 1. Ojima, and D. M. Vyas, ACS, Washington DC, 1995. 25 M. Suffness Taxol: Science and Applications, CRC Press, New York, 1995.26 Y. Kobayashi, Y. Takemoto, T. Kamijo, H. Harada, Y. Ito and S. Terashima, Tetrahedron, 1992, 48, 1853. 27 E. W. Ng, Dissertation, State University of New York at Stony Brook, 1996. 28 C. Palomo, J. M. Aizpurua, J. M. Garcia, M. Iturburu and J. M. Odrio- zola, J. Org. Chem., 1994, 59, 5 184. 29 D. M. Spero, S. Kapadia and V. Farina, Tetrahedron Lett., 1995, 36, 4543. 30 I. Ojima, E. W. Ng and C. M. Sun, Tetrahedron Lett., 1995, 36, 4547. 31 I. Ojima, H. Wang, T. Wang and E. Ng, 217th American Chemical Society National Meeting, San Francisco, April 13-1 7, 1997; Abstr. ORGN 52. 32 P.-J. Colson and L. S. Hegedus, J. Urg. Chem., 1993, 58, 5918. 33 T. Wang, F. Delaloge and I. Ojima, 218th American Chemical Society National Meeting, Las Vegas, September 7-11, 1997; Abstr. ORGN 210. 34 J. Kant, W. S. Schwartz, C. Fairchild, Q. Gao, S. Huang, B. H. Long, J. F. Kadow, V. Farina and D. Vyas, Tetrahedron Lett., 1996, 37, 6495. 35 I. Ojima, M. Zhao, T. Yamato, K. Nakahashi, M. Yamashita and R. Abe, J. Org. Chem., 1991, 56, 5263. Received, 24th April 1997 Accepted, 16th June 1997 386 Chemical Society Reviews, 1997, volume 26

ISSN:0306-0012    

DOI:10.1039/CS9972600377

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Approaches to the synthesis of ingenol Sanghee Kim? and Jeffrey D. Winkler” Department of Chemistry, The University of Pennsylvania, Philadelphia, Pennsylvania 19104, USA Ingenol is a highly oxygenated tetracyclic diterpene which mimics the function of diacylglycerol, the endogenous activator of protein kinase C. One of the most imposing challenges in the synthesis of ingenol is the establishment of the highly strained ‘inside-outside’ or trans-intrabridgehead stereochemical relationship of the carbocyclic ring system of the ingenanes. The approaches that have been examined in several laboratories towards the synthesis of the ingenanes are summarized. Our own work, using the intramolecular dioxenone photocycloaddition to establish this unique ster- eochemical feature, is described.1 Introduction Ingenol 1 (Fig. 1) is a highly oxygenated tetracyclic diterpene isolated initially from the Euphorbia ingens species of the Euphorhiaceue plant family. 1 Diverse ingenane types with different oxidation states at C-3, C-4, C-5, C- 12, C- 13, C- 16 or C-20 have also been isolated.* Various esters of ingenol are able to substitute for diacylglycerol 2, the endogenous activator of protein kinase C (PKC), and thereby exhibit antitumour or tumour-promoting activity. Protein kinase C is the phosphor- ylating enzyme which mediates cellular signal transduction for a large class of hormones and cellular effectors that activate phosphatidylinositol 4,5-bis(phosphate) t~rnover.~ In addition to ingenol 1, several other natural products including teleocidin 3, asplysiatoxin 4, bryostatin 6 and esters of phorbol5 mimic the function of diacylglycerol 2.4 Although several proposals for a pharmacophore common to these structurally dissimilar activa- tors of PKC have been described,s the requisite structure- activity relationships (SAR), that could lead to new therapeutic agents for the treatment of inflammatory and proliferative diseases, remain to be definitively established.6-7 Sanghee Kim was born in 1966 in Mokpo, Korea. He received his BS degree in pharmacy from Seoul National University in 1988, where he also obtained his MS degree in medicinal chemistry under the supervision of Professor Deukjoon Kim in 1990.After finishing military service as a lieutenant, he worked for a year as an assis- tant researcher at the Korea Institute of Science and Tech- nology.In 1992, he joined Pro- fessor Jeffrey Winkler’s group at the University of Pennsyl-vania and earned his PhD in organic chemistry in 1997. He is now working as a post-doctoral associate at The Scripps Research Institute in the laboratories of Professor K. C. Nicolaou. Sanghee Kim The structurally and biologically related diterpene phorbol 5 was recently synthesized by Wender.8 However, despite the efforts of many groups, ingenol 1 has not yet yielded to total synthesis.9 While the high degree of oxygenation, notably the cis-trio1 (from C-3 to C-5 located on the p face of A and B rings), represents an important challenge to synthesis, the most imposing obstacle to the synthesis of ingenol is the establish- ment of the highly strained ‘inside-outside’ or trans intra-bridgehead stereochemistry of the B,C ring system.This unique stereochemical feature would appear to play a very important role in the biological properties of the ingenanes, as Paquette” has reported that a highly functionalized ingenane analogue 7 (Scheme l), which has cis rather than trans intrabridgehead stereochemistry (the C-8 epimer of ingenol), is completely devoid of biological activity. In this review, the approaches that have been examined in several laboratories towards the synthesis of the ingenanes will be summarized. Also, our own work, in which intramolecular dioxenone photocycloaddition is used to establish the critical inside-outside stereochemical relationship, will be described.2 Inside-outside stereochemistry of ingenol Bridged bicyclic systems can exist as three different stereo- isomers: an out-out isomer 8, an in-in isomer 9 and an in-out isomer 10 as shown in Fig. 2.10 Usually the in-in isomer 9 is most unstable because of the severe repulsive interaction between the inside atoms. However, the energy difference between in-out and out-out isomers varies depending on the system.IO For example, the out-out isomer of bicyclo[4.4.4]te- tradecanes is less stable than the in-out isomer by 12 kcal mol-1 (1 cal = 4.184 J), presumably as a consequence of severe eclipsing interactions along each of the three chains of the bridged bicyclic ring system.In the ingenane ring system, the in-out isomer is generally more strained than the out-out Jeffrey D. Winkler graduated cum laude in Chemistry from Harvard College in 1977, and received his PhD degree with Professor Gilbert Stork. After an American Cancer Society Postdoctoral Fellowship in the laboratories of Professor Ronald C. D. Breslow, he joined the Chemistry Depart- ment at the University of Chi-cago as an Assistant Professor in 1983. He moved to the Uni- versity of Pennsylvania in 1990, where he is currently Professor of Chemistry and a Member of the University of Pennsylvania Cancer Center. Jeffrey D. Winkler Chemical Society Reviews, 1997, volume 26 387 16 OAO 'OH OH 0 Diacylglycerol (DAG) 2 HO CHpOR;! Teleocidin B-1 20 3 lngenol 1 I OH OH CH20R1 Debrornoaplysiatoxin (DAT) 4 Phorbol 5 0 Bryostatins 6 Fig. 1 KNH2, NH3 AL'2.H30+ o,&o,,H Me0 Me0 32% from 15 15 16 17 hv 85% I, ___) ORR = TBDMS 18 19 7 R = palrnitate Scheme 1 corresponding in-out and out-out bicyclo[4.4.l]undecan-7-one conformers differ in strain energy by 3.3 kcal mol-I. Ingenol itself is more strained than its C-8 epimer (isoingenol) by 5.9 kcal mol-1.11 Only a few other natural products exhibit the phenomenon of 'inside-outside' stereochemistry, such as 3-a-acetoxy-15~-hydroxy-7,6-secotrinervita-7,11-diene and secotrinerviten- out -out 8 in -in 9 in -out 10 2P,3a-diol, macrocycles 13 and 14 in Fig. 3 respectively.10 The inside-outside intrabridge stereochemical relationship in inge- no1 1 is depicted in the MM2 minimized three dimensional picture (Fig.3). 3 Synthetic approaches toward ingenol H out The biological activity of ingenol and its unusual structure have stimulated the interest of many chemists. A number of 11 12 approaches to this complex ring system have been published Fig. 2 since the mid 1980s, but the total synthesis of ingenol has not 388 Chemical Society Reviews, 1997, volume 26 yet been accomplished, due to the structural complexity and the instability of these natural products. The published approaches for the synthesis of ingenol include: (1) both inter- and intra-molecular [6 + 41 cycloaddi-tions of tropones to dienes,12-14 (2) intramolecular [4 + 31 cycloaddition of cyclic oxyallyls to a tethered furan,l5 (3)Lewis acid catalysed aldol cyclization, l6 (4) photo induced ring expansion-ring contraction,'7 (5) a ring contraction strategy based on the Ireland-Claisen rearrangement,ll,l* (6) trans-formation of an out-out system to an in-out isomer by [1,5]-hydrogen sigmatropy 19 and (7) intramolecular dioxenone photocycloaddition and fragmentation.2O Only the latter three strategies provide the requisite inside-outside intrabridgehead stereochemical relationship of ingenol.3.1 Paquette's approach to cis-ingenane Paquette and co-workers have synthesized the less strained isoingenol7, which is epimeric with the natural product at C-8, @AcO~'.' 13 14 l:iK.3 Me0 C02Me -NaH0 58% Me0 20 1. TMSl 2. PCC C02Me3.LHMDS 43ally1 bromide 60% overall yield 23 possessing the fully functionalized AB ring of ingenol.17 The B,C rings of the ingenol skeleton were generated by sequential inter- and intra-molecular alkylation of the partially reduced tetralone substrate 15 with dichlorobutene yielding the cis intrabridgehead stereochemistry instead of the requisite trans relationship. The photo-induced isomerization of 18 led to the formation of perhydroazulene 19, containing the ABC ring system of ingenol. The stereocontrolled elaboration of the cis-trio1 and the unsaturations in the A and B rings of ingenol were successfully accomplished.However, this highly functionalized isoingenol 7, which has out-out intrabridgehead stereochem- istry, was devoid of the biological activity associated with the naturally occurring ingenane esters. These results underscore the importance of the trans-intrabridgehead stereochemistry for the biological activity of the ingenanes. 3.2 Mehta's approach to the isoingenane framework Mehta has reported a construction of the isoingenane ABC framework.l6 He employed a sequential titanium-catalysed intramolecular variant of the Mukaiyama reaction for the synthesis of seven-membered B ring (21 to 22), followed by base promoted aldol cyclization for the formation of the 5-membered A ring (23 to 24). This approach, however, did not yield the inside-outside chemistry.3.3 Harmata's approach to the isoingenane framework Cyclic oxyallyl cations 26, derived from cycloheptanone 25 underwent intramolecular [4 + 31 cycloaddition with a tethered furan to give two polycyclic products (27 and 28) in a 7.3 to 1 ratio.'5 The major cycloadduct 27 is the result of an ex0 approach of the dienophile to the diene. Both cycloadducts possess the ABC ring structure of isoingenol with cis intra-bridgehead stereochemistry. 3.4 Rigby's approach to cis-and trans-ingenane Rigby has employed both inter- and intra-molecular [6n + 4n] cycloadditions of tropones to dienes to construct the cis fused bicyclo[4.4.1] system. These thermally allowed cycloadditions produce a functionally rich BC ring building block for the construction of C-8-isoingenol intermediates. In his intermole- 1 LHMDS TMSCl 66% overall 21 22 1.PdC12 cuc12. 02 2. NaH 59% C02Me 0 24 Scheme 2 D 1. LDA 2. TfSO2CI 3. LiC104,TEA Et20, 56% J 7.3 25 26 27 28 Scheme 3 Chemical Society Reviews, 1997, volume 26 389 29 30 31 32 83% COpEt 1 HgS04 1. oso4 * ___) 2. Me2CuLi 2. MsCl 3. NaH 3 CH2(C02Et)2 33 55% overall weld 34 38% overall yield NaH 35 u Scheme 4 cular cycloaddition approach,l3 bridgehead enolate alkylation followed by aldol condensation was used to install the A ring of isoingenol (32 to 34). Regioselective dihydroxylation of the C ring followed by activation and treatment with sodiomalonate afforded a tetracyclic isoingenane 35 as shown in Scheme 4.The use of thermal, metal-free, and metal-promoted intra- molecular higher-order cycloadditions has been reported re-cently by Rigby.12 He found that tropone with various diene tethers underwent em-selective [6n + 4n] cycloaddition under delivery approach to the establishment of the C-8(3 bridghead hydrogen stereochemistry.19 The out-out bicyclo[4.4. llunde- cane 41 was produced as a single (endo) isomer by chro- mium(0)-mediated higher-order cycloaddition between com-plex 40 and 2,4-hexadiene. Selective dihydroxylation of 41 and protection of the resulting diol as an acetonide gave 42. E,uo-face selective epoxidation of 42, followed by epoxide opening with a lithium base led to the exclusive formation of dienol44.This regioselective elimination process can be rationalized by promoted reactions proceeded via an exclusive endo-selective to achieve ring opening due to the conformation of the bicyclic thermal conditions (36 to 37). He also reported that the metal- the fact that the C-8 hydrogen is the only one properly aligned undecane system. An alkoxide accelerated [1,5]-hydrogen sigmatropic rearrangement was employed for delivering the 13-H on the alkoxide carbon (C-1 1) to the bridgehead position BH(C-8) with retention of stereochemistry to yield the in-out bicyclo[4.4. llundecane 45. 3.5 Funk’s approach to trans-ingenane37 Funk has reported a clever solution to the problem of the inside- outside stereochemistry of ingenol, in which an intramolecular version of the Ireland-Claisen rearrangement leads to the transformation of 12-membered lactone 50 to a more strained trans fused bicyclo[4.4.llundecane ring system 52.’ It should be noted that the crucial trans relationship of C-8 and C-10 (ingenol numbering) was set early in the sequence by ster- H eoselective alkylation of a chiral homocarene enone-ester 4739 obtained from (+)-3-carene 46. Both conjugate addition of methyl cuprate to 47 at C-11 and alkylation of 48 with the butenol chain at C-8 occurred opposite to the cyclopropane ring. pathway (38 to 39). 80 OC yield 84% * \ / 36 1) hv(Pyrex) * 2) (Me0)3P 52% overall yield Scheme 5 More recently, he reported an elegant protocol for the It is noteworthy that the resulting macrobicyclic lactone 50 does conversion of the out-out bicyclo[4.4.llundecane system into not suffer from the ring strain present in ring-contracted trans- the more strained in-out stereoisomer by using an internal bridged bicyclo[4.4.1] system, i.e. 52, which is obtained on LINEt2 KH, 18-crown-6 dioxane, reflux .. Y 43 Scheme 6 390 Chemical Society Reviews, 1997, volume 26 rearrangement of the ketene acetal51 derived from lactone 50, via a boatlike transition state. Unfortunately, the first approach, shown above, afforded an ingenane nucleus in which the functionality is not optimally situated for the construction of the A-ring. A subsequent approach using the same strategy delivered a more readily elaborated Claisen product.The new version makes use of an n to n -2 ring contraction (53 to 54) instead of an n to n -4 rearrangement (50 to 52) as in Scheme 7, and leads to the first synthesis of the tetracyclic ring system with the inside-outside stereochemis try. 4 Construction of the tricyclic nucleus of the ingenane diterpenes by dioxenone photocycloaddition and fragmentation methodology Winkler and co-workers reported the first synthesis of a tricyclic ingenane ring system having the correct trans-intrabridgehead stereochemical relationship.21 In order to construct the inside-outside ring system, they employed the LiMeCuCN -78 OC, 86% 346 intramolecular version of the modified de Mayo reaction, the dioxenone photoaddition-retroaldol fragrnentation.22-23 4.1 The de Mayo reactions de Mayo found that irradiation of the enol form 59 of a P-diketone 58 in the presence of an alkene produced a four- membered ring aldol product 150.~~This photoadduct 60 underwent a facile retro-aldol reaction in situ to produce a 1,5-diketone 61, thereby alleviating the strain energy of the cyclobutane intermediate. Attempted extension of these results to 6-keto esters gave a different [2 + 21 photocycloaddition, yielding oxetane products.34 A solution to this problem was reported by Baldwin,25 who used dioxenone heterocycles 62as covalently locked enol tautomers of the corresponding S-keto ester.However, it still proved too difficult to predict the regiochemical outcome of the photocycloaddition with un-symmetrical alkenes.Winkler and co-workers found that regiochemical control of the dioxenone photocycloaddition can be greatly improved using the intramolecuIar version of this reaction.26 Irradiation of /,,, /A LDAy TBSo-/d-C'-2 CH2CHC02Me Triton B 54% overall yield 47 48 1 TBAF 2 K2C03. MeOH, H20 TlPSOTf _____)reflux CI OTBS 'I 49 52% overall yield 50 51 Scheme 7 1 LHMDS CIP(OEt)2 2 toluene, 95 OC 3 HF -88% over 3 steps 2 EtnCuMgBr Me02C'"& 82% overall yield EtTr' nU 53 54 55 1 NaOMe 2 LHMDS 81Yo overall yield " 56 57 Scheme 8 Chemical Society Reviews, 1997, volume 26 391 r-1 58 59 60 61 0 0 0 62 63 64 65 Scheme 9 a tethered dioxenone 66 produced photoadduct 67 which on bridged bicyclic ring systems such as those of paclitaxol and retro-aldol fragmentation provided six-, seven-and eight- ingenol which are not accessible using standard de Mayo membered ring esters 68 in excellent yield with exceedingly reaction conditions.Irradiation of dioxenone 69 led to the high ( >50 :1) levels of regiochemical control. exclusive formation of photoadduct 70, which on fragmentation produced trclns-bicycl0[5.3.l]undecan-l1-one-3-carboxylic acid 71, the smallest bridged bicycloalkanone with 'inside- outside' stereochemistry at that time. We proposed that the unusual trans intrabridgehead stereochemical relationship could be a consequence of the chairlike folding of the nascent six-membered ring, as shown in conformation 69A.Photocy-cloaddition of conformer 69A should lead to the intermediate 66 67 cyclohexane diyl 72 @-bond formation) instead of diyl 73, based on recent mechanistic studies from our lab~ratory.~~ The alternative formation of the eight-membered ring diyl 73 (a-bond formation) would be expected to produce the less strained cis-ring-fusion stereochemistry in the cycloaddition, since the rate of intersystem crossing for the triple diyl is slower than ring flipping.27 This methodology has been extended to the construction of the trans-bicyclo[4.4. llundecane moiety which constitutes the BC ring system of the ingenanes. The dioxenone 76 was derived 68 from retrosynthetic analysis of the ingenane 74 which lacks Scheme 10 most of the functionality of ingenol except for the C-9 carbonyl, the C-20 oxygen functionality and the critical C-8p hydrogen.A model system was first examined to establish the feasibility 4.2 Construction of bridged bicyclic ring systems by of this methodology, shown in Scheme 12.28 In contrast to intramolecular dioxenone photocycloaddition and previous results which proceeded to give a single photoadduct fragmentation methodology in high yield (Scheme ll), irradiation of 77 under the usual This intramolecular dioxenone photocycloaddition-fragmen- conditions (4.8 mM in 1 :9 acetone-acetonitrile, Pyrex immer- tation methodology can be used for the construction of various sion well, 0 OC, 450 W, medium-pressure Hg lamp) led to the 1 p-TsOH, MeOH 0 2 LiOH,MeOH THF 0 * aco2HH 69 70 71 0YO 0YO w +&J& H H 69A 72 73 Scheme 11 392 Chemical Society Reviews, 1997, volume 26 were subjected to fragmentation and Barton decarboxylation to give a single bicycloundecanone 81 with the highly strained trans intrabridgehead stereochemistry. The diastereomeric ketoacids 80 become enantiomeric ketones 81 on decarbox- ylation.We anticipated that annulation of the ingenol A ring onto the photosubstrate 77 would reduce the degrees of freedom of the 74 20 alkyl chain and could lead to a higher facial selectivity as well as yield. The requisite photosubstrate 76 was prepared as outlined in Scheme 13. Reductive alkylation of bicyclo-[3.3.0]octenone 82 with hexenyl iodide followed by carboxyla- tion and dioxenone formation provided photosubstrate 76. Irradiation of 76 through a Pyrex filter afforded a single photoadduct 75 in high yield.The adduct was then fragmented to keto acid 74 with potassium hydroxide.28 The exclusive formation of the inside-outside isomer 74 can be explained by approach of the alkene to the dioxenone in the pseudochair conformation 76A. The other possible conformation is a pseudoboat, 76B, which suffers from transannular eclipsing 76 non-bonded interactions which are not present in 76A. Fig. 4 Two different approaches for the synthesis of C-3 oxygenated analogues of ingenol have been studied by Winkler and co- formation of two photoadducts, 78 and 79, in a 4.3 : 1 ratio in workers.20 In the first approach? the angular functionalization of 30% yield.The diastereomeric photoadducts 78 and 79, which C-3 (ingenol numbering) oxygenated enone 87 was achieved by result from the addition of alkene from either face of dioxenone, [2 + 21 photocycloaddition with allene. Ozonolysis of the U 77 78 79 A 80 81 Scheme 12 TFA, TFAAu"/P 58% H H 82 83 84 KOH aa% ____) eHC02HH 76 75 74 0 0 76A 760 Scheme 13 Chemical Society Reviews, 1997, volume 26 393 photoadduct 88 in methanol gave the keto-ester 89, which was transformed to the desired photosubstrate 90 through a nine step sequence. However the length of this sequence and low overall yield demanded a more efficient route.The second approach is outlined in Scheme 15.The reductive alkylation of 86 with hex-5-enyl iodide produced the desired angularly alkylated product 93 in 45-52% yield. This result reduced the length of the sequence by six steps and delivered C-3p oxygen stereochemistry found in the natural product. Carboxylation of 93 followed by ester exchange with anisyl alcohol, and condensation with acetone under acidic conditions led to the formation of dioxenone 94. Irradiation of 94 gave the desired photoadduct 95 in 61 % yield. Fragmentation of 95 with p-toluenesulfonic acid in refluxing methanol gave the ingenane tricycle 96 with the correct C-Sp hydrogen stereochemistry as a 03, MeOH Me2S 82% 3:2 mixture of C-6 a:p ester epimers 96a and 96b, re-spectively.Basic fragmentation condition30 (1% KzC03, MeOH) of 95 provided the 3p-hydroxy ingenane tricycle 97 as a mixture of C-6 p and a ester epimers in a ratio of 6 to 1, respectively. 4.3 Advanced ingenol analogues With an improved method for the preparation of C-3 oxy- genated ingenane tricycles in hand, we began to investigate the elaboration of the functionality in the A/B ring system and prepared a series of functionalized analogues, some of which were found to be biologically active in vitr0.31 In an effort to prepare the first C-3, C-4 dihydroxylated ingenane analogues via dihydroxylation of the A3g4 alkene, the elimination reaction of C-3 hydroxy group of various ingenanes 88 89 90 tN 82% D Me0 91 92 Scheme 14 1 LDA, MeOCOCN HO@ LI, NH3 5-hexenyl-1 -iodide 45-54Yo * 3HO 2 anisyl alcohol 3 TFA,TFAA 55% overall yield acetone 86 93 94 61%tN ~ @1 96a 3 : 2 ratio 96b MeOZCO 95 K2C03, MeOH 95% * 97a 6 : 1 ratio 97b Scheme 15 394 Chemical Society Reviews, 1997, volume 26 was examined.5 The Mitsunobu bromination of 97b (C-6p The regioselectivity of selenation for ketone 107 was also ester), 97a (C-6a ester), 100 and 104 proceeded to give the examined.30 The selenation of C-6a ester ketone 107a with inverted C-3a bromides 98b, 98a, 101 and 105, respectively.different bases such as LDA, LHMDS, KHMDS also occurred These compounds were submitted to standard elimination at C-2 to give exclusively A1,* enone 109 in good yield after conditions (LiCl, DMF, reflux) to produce the corresponding oxidative elimination.However, treatment of ketone 107a with alkenes. Bromide 98b led to the exclusive formation of the A394 excess phenylselenyl chloride in acetic acid, followed by alkene product 99. The compound 101, which lacks the ester oxidation with hydrogen peroxide resulted in the 2-chloro A172 substituent at C-6, also produced only the A394 alkene 102. In enone 111. Different regioselectivity was observed when the contrast, it was found that the C-6a ester 98a yield only the A233 C-6P ester ketone 107b was treated with phenylselenyl halide. alkene 103 in the same reaction conditions (unlike the C-6p Selenation, in both acetic acid and basic conditions (KHMDS in isomers, elimination of the C-6a isomers needed long periods of THF), occurred at the more substituted carbon (C-4).Surpris-refluxing). The elimination reaction of a silyl ether 105 was ingly, after oxidative elimination of the selenide 112 with examined to probe the involvement of the C-6a ester function- hydrogen peroxide, the A5,6 unsaturated ester 114, which ality during the reaction.Under the same conditions, the A2,3 resulted from enone 113 in the reaction conditions, was product 106 was the only alkene product observed, albeit in low obtained. yield, indicating that the ester functionality was not involved These results suggest the importance of the stereochemistry during the reaction.of the C-6 substituent in the regioselective formation of alkenes LICI, DMF * HO H R$fjH 98b R = C02CH3 (84%) 99 R = C02CH3 (70%) 101 R = H (62%) 102 R = H (70%) Ph3P, CBrd LICI,DMF * DBHCH2C12HO H *k ti R 97a R=C02CH3 98a R = C02CH3 (85%) 103 R = C02CH3 (70%) 104 R= CHzOTBDPS 105 R = CH20TBDPS (68%) '06 R = CHZOTBDPS (10%) Scheme 16 PhSeBr H202 53% overall -PhSeBH OH 5ICO~CH~ k02~~3 107a 108 109 PhSeCl AcOH * 107a 110 111 -PhSeBr KHMDS 107b 112 113 114 Scheme 17 Chemical Society Reviews, 1997, volume 26 395 or enones. The C-6a substituent induces the formation of the less substituted alkene products, whereas the C-613 substituent has the opposite effect. We reasoned that a C-6 substituent induced the different conformation of the ingenane tricycle nucleus as shown in Fig.5. 202~~3 107a 107b Fig. 5 After extensive investigation, we could introduce the in- genane B ring A6,7 alkene into 96 as outlined in Scheme 18. Selenation (KHMDS, PhSeC1) of a mixture of C-6 ester isomers 96 gave a single selenide; oxidative elimination of the resulting selenide using H202-CH2C12 produced a mixture of products consisting of a 5 : I ratio of the A6-7 115 and A5,6 unsaturated ester 116 respectively.2OJO While selenation-oxidation of C-6 esters led to the formation of the A637 unsaturated ester 115, it was found that reaction of 96 with NBS-AIBN in refluxing carbon tetrachloride, followed by treatment of the derived mixture of cw-bromoesters with excess lithium chloride in refluxing DMF led to the exclusive formation of the As,6 unsaturated ester 116a, the first example of the selective functionalization of C-5 as the As-6 alkene. Compound 116a is the key intermediate for the completion of the B ring functionalization, which will be discussed in the following sections.The synthesis of advanced ingenol analogue 122 is outlined in Scheme 19.5The final compound 122 has all the functionality of the AB rings of ingenol except the C-2 methyl and C-5 hydroxy groups. It is noteworthy in that stereoselective dihydroxylation (99 to 117) at C-3 and C-4, and regioselective introduction of A6-7 unsaturation (117 to 118) have been 1 PhSeCl KHMDS (71%)* 2 30% H202 (93%) Me02C0 Me02CO C02CH3 C02CH3 Me02C0 C02CH3w w w 96 115116 1 NBS AIBN, CC14 * 2 LiCI, DMF Ho C02CH3 Me02C0 60% overall yield w C02CH3 achieved.However, this key intermediate could be obtained only from the C-6(3 ester of 96 or 97, which are the minor isomers from retro-aldol fragmentation. The isomerization of the C-6a ester 96a to the 13 ester 96b has not been achieved under a variety of conditions. We next turned our attention to the introduction of the A637 alkene at early stage of the synthetic scheme.30 Unsaturated ester 115, obtained as mentioned above (Scheme 18), was reduced with DIBAL-H in THF to give the desired allylic alcohol 123. Protection of the resulting primary allylic alcohol with tert-butyldiphenylsilyl chloride, followed by oxidation of C-313 alcohol with PCC produced ketone 124.In order to introduce the A l72 alkene, a variety of reaction conditions were investigated. Even though high regioselectivity had been realized with the saturated analogues (Scheme 17), regioselec-tive functionalization of the A677 alkene ketone 124 (i.e. selenation or enol silyl ether formation) could not be achieved. Efforts to introduce the 4-hydroxy group before the unsaturation only led to decomposition of starting material under various reaction conditions. However, we could functionalize C-2 selectively through activation as a (3-ketoacid. Carboxylation of 124 followed by treatment with Eschenmoser’s salt led to the exclusive formation of a-methylene ketone 125.Initial attempts to prepare 125 directly from 124 using Eschenmoser’s salt have not been successful. Isomerization of 125 to the endocyclic enone 126 was achieved by using rhodium trichloride as a catalyst. The enone 126 was reduced under Luche conditions and resulting a-alcohol was converted to the (3-benzoate under Mitsunobu conditions to provide 127, an analogue of ingenol containing both A and B ring unsaturations. Having achieved the synthesis of compound 127 which has all the functionality of the A and B rings except the C-4 and C-5 hydroxy groups, we began to investigate possible routes for the introduction of these remaining functi~nalities.~~ In order to introduce the C-5 hydroxy functionality, allylic oxidation and allylic bromination of 115 were investigated.Neither allylic oxidation with Se02 nor allylic bromination with NBS-AIBN produced the desired corresponding C-5 alcohol or bromide. The A6j7 unsaturated ester route was abandoned and we began working with the As,6 unsaturated ester 116a. Unfortu-nately, it was found that the lack of reactivity of unsaturated esters (116 and 116a) under a variety of reaction conditions (such as ene reaction with singlet oxygen, dihy- droxylation with Os04 or epoxidation) prevented the intro- duction of the C-5 hydroxy group. Presumably, the sterically congested environment around C-5 causes the unsaturated esters (116 and 116a)to be unreactive. Reaction of C-3 hydroxy As-6 unsaturated ester 116a with methanesulfonyl chloride followed by treatment with base led to the formation of the diene ester 128 in excellent yield.Epoxidation of 128 with mCPBA occurred with high degree of regio- and stereo-chemical control to give 129. Treatment of the epoxy un-+ 96 116a Scheme 18 396 Chemical Society Reviews, 1997, volume 26 oSo4. 1. dimethoypropane * NMO 2. KHMDS, PhSeBr 72% 3. H202 C02Me 79% over 3 steps 99 117 118 1. DIBAL-H @ @1.TBSOTf 2. H202'.:;::* @ 2. HC104 2. NCS, Me2S 71yo HO HO TEA 0 HO 46% overall 0 HO OH 67% OTBS Yield OTBS 119 120 121 1. DIBAL-H (70%) * 2.BzCI (63%) 3.HC104 (59%) 122 Scheme 19 2. PCC 2. Eschenmoser's salt 115 123 124 RhC13 1. CeCI3, NaBH4 79% (100%)L 2.DEAD, TPP BzOH (65%) LOP 125 126 127 Scheme 20 saturated ester 129 with osmium tetroxide led to the exclusive formation of epoxy diol 130. The dihydroxylations of the 3,4-dihydroxy A5-6 unsaturated ester and the 3-hydroxy un- saturated esters 116 were not successful under the same reaction conditions. We reasoned that the epoxide changed the con- formation of the ingenane nucleus to reduce the steric environment around C-5. Epoxy diol 130 was treated with camphorsulfonic acid in wet acetone to yield the tetraol, which on exposure to excess dimethoxypropane produced the acet- onide 131 as a single isomer. The structure and relative stereochemistry of 131 have been verified by X-ray crystallo- graphic analysis. The selective elimination of C-6 hydroxy group of 131 to introduce the A6,7 unsaturation was next investigated.Various standard elimination reaction conditions were examined, such as Burgess reagent, Ph3BiBr2/12, PPh3/CC1&H3CN, Furukawa reagent, Martin sulfurane, and base treatment of the C-6 triflate or mesylate. None of the above conditions produced the desired unsaturated ester. We decided to reduce the C-6 ester to the corresponding alcohol, because of the instability of the A6,7 unsaturated esters we experienced in many cases. The selective reduction of a-hydroxy ester in the presence of the hindered C-9 ketone was conducted with LiAlH4 in THF. The resulting primary alcohol was selectively protected with TBDPSCI or TBDMSC1. The elimination of the C-6 hydroxy group of 132 was also investigated for the introduction of unsaturation of the B ring.Various standard conditions only resulted in decomposi- tion of the starting compound without formation of the desired allylic alcohol 134. The lack of literature procedures for the efficient trans-formation of a diol or a trio1 to an allylic alcohol led us to investigate new methodology for achieving this transformation. We turned our attention to cyclic sulfates,32 the chemistry of which is analogous to that of an epoxide, only more reactive. The elimination of an unactivated cyclic sulfate had not been reported in the literature, whereas some examples of elimination reactions with activated cyclic sulfites were known.33 The C-4/C-6 cyclic sulfate 133 was prepared by treatment of diol 132 with thionyl chloride in pyridine, followed by oxidation with RuC13 and NaI04.32 Reaction of the cyclic sultate 133 with DBU in refluxing toluene, with subsequent hydrolysis in mild acidic conditions afforded the desired allylic alcohol 134 in 75% yield.To the best of our knowledge, this is the first example of an elimination reaction of a non-activated cyclic sulfate.29 The eliminated product 134 has all the functionality of the B ring of ingenol, including the C-4, C-5 hydroxy groups Chemical Society Reviews, 1997, volume 26 397 \? I \ HoYC02CH3 * H mCPBA 80-90% \ C02CH3 b oso4 73%-\ CO~CHB 116a 128 129 130 2'propane dimethoxy-83% &H 80% 1 SOCla 2 RuC13, Na104 90% \ 131 \ 132 / 133 a) DBU b) HZS04 aq THF 75% yield 0 HO BzO HO HO OH 1 34 135 Scheme 21 with the correct stereochemistry, as well as the C-3 hydroxy group, which might be utilized for full functionalization of the A ring.5 Biological evaluation of ingenol analogues The process of PKC activation is complex, requiring multiple components. At this time, neither the basis for the binding of a wide variety of structural types (Fig. 1) to the PKC receptor nor the pharmacophore of PKC activators has been unambiguously established.4,5.*0 Even though the crystal structure of phorbol- bound PKC confirmed some predictions of tumour promoter models,6 the results for phorbol esters do not necessarily translate to in vivo and to other type of tumour promoters.Several groups have studied the pharmacophore of the seemingly structurally unrelated activators of PKC, using synthetic structure-activity and molecular modelling studies.5 The molecular modelling approach is used to find similarities in the three-dimensional array of homologous functional groups.5 A synthetic structure-activity study is used to find the structural requirements for activation of PKC, focused on the synthesis of specifically modified derivatives of natural products. Recently, we evaluated the biological potency of synthetic ingenane analogues which contain some oxygen functionalities along with correct inside-outside intrabridgeheaded stereochemistry present in the natural product.5,20,31 The C-3 monobenzoate of ingenoll36 and its analogues were evaluated for their ability to interact with the regulatory site on PKC, as quantitated by the inhibition of [3H]PDBU binding to the enzyme reconstituted in the presence of phosphatidylserine and assayed for 5 min at 37 "C.Ingenol3-monobenzoate 136 yielded an apparent K, of 0.15 f 0.03 nM (mean k SEM, n = 4) for protein kinase C-a, the C-4 deoxy ingenane analogue 137 had a K, of 165 k 21 nM (mean kSEM, n = 3) and the C-4 hydroxylated analogue 122, assayed in parallel, had a K, of 561 k 94 nM (mean +_ SEM, n = 3). Ingenol analogue 7 which has cis rather than trans intrabridgehead stereochemistry (C-8 epimer of ingenol, iso- ingenol) is completely inactive as mentioned before.These data have important implications for the understanding of structure- activity relations for the ligand binding site on the protein kinase C. Further testing of these and related compounds, as well as the 398 Chemical Society Reviews, 1997, volume 26 BzOaHOH LOH 136 137 aHBzO BzO@OHHO LOH 122 138 *OHBzO HO 135 Fig. 6 synthesis and biological evaluation of more highly function- alized ingenol congeners such as 138(which can be obtained by deprotection of 127, Scheme 20), 122 (prepared in Scheme 19) and 135, are currently underway in our laboratory and our results will be reported in due course. 6 References 1 H. Opferkuch and E. Hecker, Tetrahedron Lett, 1974, 15, 261 2 R. R. Upadhyay and G.Mohaddes, Curr Sci , 1987,56, (20),1058. 3 For recent reviews, see P. J. Blackshear, J Biol Chem, 1993, 268, 1501,L. V. Decker and P. J. Parker, Trends in Biochem Sci , 1994,19, 73. 4 R. Schmidt and A. Aitken, in Naturally Occuring Phorbol Esters, ed. F. Evans, CRC Press, Inc., Boca Raton, 1986, pp. 245; E. Hecker, in Carcinogenesis A Comprehensive Survey, Mechanisms of Tumor Promotion and Cocarcinogenesis, vol. 2; ed. T. J. Slaga, A. Sivak and R K. Boutwell, Raven, New York, 1987, p 11. 5 J. D Winkler, B.-C. Hong and S Kim, Synlett, 1995, 533, and references cited therein. 6 Recently X-ray crystal structure of PKC-phorbol ester complex have been advanced to explain the pharmacophore of the PKC activators, see Chem Eng News, 1995, Oct. 23, p.21, and references cited therein. 7 B. Shearer, B. Sullivan, J. Carter, R. Mathew, P. Waid, J. Connor, R Patch and R Buch, J Med Chem , 1991,34,2931; and for a review of PKC regulation; P. C Gordge and W. J. Ryves, Cellular Signallzng, 1994, 6, 871. 8 P Wender and F. McDonald, J Am Chem Soc , 1990,112,4956. 9 For recent review of the synthesis of tumour-promoting diterpenes, see: J H Rigby, in Studies in Natural Products Chemistry, ed. A.-u. Rahman; Elsevier Science Publishers, Amsterdam, 1993, 12, p. 233. 10 For a recent review of inside+utside chemistry, see- R. W. Alder and S. P East, Chem Rev 1996, 96, 2097 11 R. Funk, T A. Olmstead and M. Parvez, J Am Chem Soc , 1988,110, 3298. 12 J. H. Rigby, S. D Rege, V. P Sandanayaka and M.Kirova, J Org Chem , 1996, 61, 842. 13 J H. Rigby and S. V Cuisiat, J Org Chem , 1993,58, 6286. 14 R. Funk and G. Bolton, J Am Chem Soc , 1986, 108,4655 15 M Harmata, S. Elahmad and C. L. Barnes, Tetrahedron Lett, 1995,36, 1397. 16 G Mehta and V Pathak, J Chem Soc ,Chem Commun , 1987, 876 17 L. Paquette, R. Ross and J. Springer,J Am Chem Soc , 1988,110,6192 and references cited therein. 18 R Funk, T A. Olmstead, M Parvez and J. B. Stallman, J 011:Chew , 1993,58,5873. 19 J. H. Rigby, V. S Claire, S V. Cuisiat and M J Heeg, J Org Chem , 1996,61,7992 20 J. D. Winkler, B.-C. Hong, A Bahador, M G Kazanietz and P. M. Blumberg, J Org Chem , 1995, 60, 1381 and references cited therein. 21 J. D Winkler, K. Henegar and P.Williard, J Am Chem Soc , 1987, 109,2850. 22 For a review describing this general strategy, see J. D. Winkler, C. M. Bowen and F. Liotta, Chem Rev, 1995, 95, 2003. 23 For other reviews of [2 + 21 photocycloaddition, see M. T Crimmins, Chem Rev, 1988, 88, 1453 and D I Schuster, G Lem and N. A Kaprinidis, Chem Rev, 1993, 93, 3. 24 P. de Mayo, Pure Appl Chem , 1964,9, 597 25 S. W Baldwin and J. M. Wilkinson, J Am Chem Soc , 1980, 102, 3634. 26 J. D Winkler, J P. Hey, F J Hannon and P G. Williard, HeterocycfeT, 1987,25, 55 27 J. D. Winkler, B.-C. Hong, J P Hey and P. G. Williard, J Am Chem Soc, 1991, 113,8839 28 J. D. Winkler, K. E Henegar, B.-C Hong and P G Williard, .I Am Chem Soc , 1994,116,4183. 29 J. D Winkler, unpublished results. 30 J. D Winkler, S Kim and B.-C. Hong, 212th American Chemical Society National Meeting, 1996, Aug., ORGN, No 161 31 For the first synthesis of a biologically active analogue of ingenol, see J D. Winkler, B.-C Hong, A. Bahador, M G. Kazanietz and P M. Blumberg, Bioorg Med Chem Lett, 1993, 3, 577, and ref 20. 32 B B. Lohray, Synthesis, 1992, 1035. 33 (a)S K. Kang, D. H. Lee, Y S Kim and S. C. Kang, Synfh Commun , 1992, 22, 1109; (h)Y. L Bennani and K. B Sharpless, Tetiahedron Lett, 1993, 34, 2083. 34 M. Tada, T Kokubo and T. Sato, Bull Chem SOL Jpn, 1970, 43, 2162 35 J D Winkler and B. Shao, Tetrahedron Lett, 1993,34, 3355 Received, 24th April 1997 Accepted, 16th June 1997 Chemical Society Reviews, 1997, volume 26 399

ISSN:0306-0012    

DOI:10.1039/CS9972600387

出版商:RSC    

年代:1997

数据来源: RSC