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: El30 (USA $234). Customers in Canada will be charged the sterling price plus a surcharge to cover GST. Individuals can subscribe for &45 (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 1HN. (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 Zdeniik 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 11003 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/CS99726FX021

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

_. Volume 26cHEMICALSOCI ETY Pages 407-476 December 1997 ISSN 0306-0012REVIEWS Issue 6 CSRVBR 26(6)407476 Enzymes in organic synthesis: recent developments in aldol reactions and glycosylations Shuichi Takayama, Glenn J. McGarvey and Chi-Huey Wong 407-4 16 Polymer-supported organic reactions: what takes place in the beads? Philip Hodge 417424 Molecular and chemical basis of prion-related diseases Sheila B. L. Ng and Andrew Doig 425432 Sandwich-type heteroleptic phthalocyaninato and porphyrinato metal complexes Dennis K. P. Ng and Jianzhuang Jiang 433-442 Ultrasound in synthetic organic chemistry Timothy J. Mason 443-452 Preparation of seven and larger membered heterocycles by electrophilic heteroatom cyclization Gerard Rousseau and Fadi Homsi 453-462 CMP.NBu5Ac ~su5~ GIycosylation employing bio-systems: from enzymes to whole Ud3 $,alyl,rarde,*~ cells Vladimir KPen and Joachirn Thiem 463474 (ps lh"B/) ATPA:ZY$le ADP 1997 Indexes 475-476 Articles that will appear in forthcoming issues Lanthanide(1zr) chelates for NMR biomedical applications Silvio Aime, Mauro Botta, Mauro Fasano and Enzo Terreno Self-assembly of single electron transistors and related devices Daniel L.Feldheim and Christine D. Keating 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. Davies, Ping Zhou and Bang-Chi Chen Aspects of weak interactions Dudley H.Williams and Martin S. Westwell Covalency in semiconductor quantum dots James R. Heath and Joseph J. Shiang Equilibrium, frozen, excess and volumetric properties of dilute solutions Michael J. Blandamer The biomedical chemistry of technetium and rhenium Jonathan R. Dilworth and Suzanne J. Parrott Nonplanar porphyrins and their significance in proteins John A. Shelnutt, Xing-Zhi Song, Jiang-Guo Ma, Song-Ling Jia, Walter Jentzen and Craig J. Medforth Non-conventional hydrogen bonds Ibon Alkorta, Jose Elguero and Isabel Rozas NMR studies of carbohydrate-protein interactions in solution Ana Poveda and Jesus Jimenez-Barber0 Dibenzotetraaza[ 14lannulenes: versatile ligands for transition and main group metals Philip Mountford Carbocycles from carbohydrates via free-radical carbocyclizations: new synthetic approaches to glycomimetics Angeles Martinez-Grau and Jose Marco-Contelles Liquid-liquid equilibria in polymer solutions at negative pressure Attila Imre and W. Alexander van Hook Seven-coordinate halocarbonyl complexes of the type [MXY(C0)3(NCMe)2] (M = Mo, W; X, Y = halide, pseudohalide) as highly versatile starting materials Paul K. Baker The aza-Payne rearrangement: a synthetically valuable equilibration Toshiro Ibuka Mass spectrometric approaches to the reactivity of transient neutrals Detlef Schroder and Helmut Schwarz Christoph A. Schalley, Georg Hornung, Chemical Society Reviews, 1997, volume 26

ISSN:0306-0012    

DOI:10.1039/CS99726FP023

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Hts HIS Zn HIEEnzymes in organic synthesis: recent developments in aldol 0 reactions and glycosylations Shuichi Takayama, Glenn J. McGarvey and Chi-Huey Wong* Department of Chemistry and the Skaggs Institute of Chemical Biology, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037, USA Carbohydrates have not been as accessible as other biomole- cules such as proteins and nucleic acids and are the least exploited.As a result of their highly asymmetric and densely functionalized nature, carbohydrates are difficult to synthe- size using conventional chemistry. Enzymatic synthesis, however, with its high selectivity and mild reaction condi- tions is very useful for the preparation of carbohydrates. This review gives a brief overview of recent developments in the application of enzymatic aldol reactions and glycosyla- tions to carbohydrate synthesis.1 Introduction Advances in efficient production of biomolecules have an impact on numerous fields of science. Among the available technologies, PCR (polymerase chain reaction), solid phase synthesis, over expression of proteins in microorganisms, and protein display have allowed scientists to routinely access nucleic acids and proteins. This has led not only to a deeper understanding of these classes of molecules, but also to new developments in various fields from chemistry and biology to medicine, materials science and even computing. A third class of biomolecules, the carbohydrates have proved less accessible and are far less explored.Though there has been much progress in the area of chemical carbohydrate synthesis,l-3 it is still Professor Chi-Huey Wong received his BS and MS degree from National Taiwan University, and his PhD in Chemistry with George M. Whitesides from Massachusetts Institute of Technol-ogy. He then moved along with Professor Whitesides to Harvard University as a postdoctoral fellow for another year. He taught at Texas A&M Universityfor six years and since I989 has held the Ernst W. Hahn Chair in Chemistry at The Scripps Research Institute. His current interests lie in the area of bioorganic and synthetic chemistry, especially the development of new synthetic chemistry based on enzymatic and chemoenzy- matic reactions and the rational development of mechanism-based inhibitors of enzymes and receptors. Chi-Huey Wong Glenn J.McGarvey difficult to synthesize these highly asymmetric and densely functionalized compounds on a routine or large-scale basis using conventional chemistry. There are also no PCR equivalent replication systems or template-based expression systems available for carbohydrates. Enzymatic synthesis, however, with its high selectivity and mild reaction conditions is very useful for the preparation of carbohydrates. Advances in recombinant DNA technology has made it possible to express almost any enzyme including enzymes useful for carbohydrate synthesis and many of these enzymes are now commercially available for routine use in any laboratory. These developments are making carbohydrate structures more accessible, contribut- ing to a deeper understanding of carbohydrate recognition and opening new opportunities in various fields of science.This review gives a brief overview of recent developments in the use of these enzymes in aldol reactions and glycosylations. Also included is the application to synthesis of bioactive com-pounds. 2 Aldolase-catalyzed synthesis of novel sugars Over 30 aldolases have been identified and isolated so far, the majority of which catalyze the reversible stereospecific addition of a ketone donor to an aldehyde acceptor.47 Mechanistically, two distinct classes can be recognized (Fig. 1). Type I aldolases form a Schiff-base intermediate in the active site with the donor Professor Glenn J.McGarvey received his BS in Chemistry from the University of California at Santa Barbara then completed his PhD in Chemistry at the University of California at Davis under the direction of R. Bryan Miller. This was followed by twoyears of postdoctoral research at the California Institute of Technology in the laboratories of Professor Robert E. Ireland. He then took his present position on the faculty of the Department of Chemistry at the University of Virginia where his research interests to date have been broadly based in synthetic organic chemistry. Professor McGarvey is presently on leave from UVa and is a visiting scientist at The Scripps Research Institute. Shuichi Takayama received his BS and MS degree from Tokyo University then entered the Graduate Program in Chemistry at the Scripps Re- search Institute for his PhD studies.He has been a graduate student in Professor Chi-Huey Wong's group since 1994. His research interest is in under- standing the phenomena of lfe on a molecular level using chemical and enzymatic or-ganic synthesis. Shuichi Takayama Chemical Society Reviews, 1997, volume 26 407 0 Fructose-diphosphatealdolase (FDP A) 2-0~~02-03PO&OH + H GOPO32-* OPO32-t--OH OH OH Donor Acceptor Type I Aldolase Type II Aldolase Fig. 1 Mechanism of type I and type I1 aldolases substrate, which subsequently adds stereospecifically to the Type I1 aldolases use a Zn2+ cofactor, which acts as a Lewis acid in the active site.lO,ll Although aldolases are quite specific for the nucleophilic donor component, a high degree of flexibility is allowed for the acceptor aldehyde component permitting the synthesis of various unnatural sugars. Aldolases can also be classified according to the donor component utilized [dihydroxyacetone phosphate (DHAP), pyruvate, acetaldehyde, or glycine] .Four dihydroxyacetone phosphate-dependent aldolases having com- plementary stereoselectivities have been cloned and overex- pressed (Fig. 2), and three of them are commercially available (Table 1). 0 OH 0 OH PO+,, OH OH OH OH D-FDP D-TDP Fructose-diphosphate aldolase Tagatose-diphosphate aldolase 2 0 /\ ,TopII /G3P POLOH DHAP/ \HLCH3 OH J L-lactalde hyde 0 OH POGCH3 PO+CH3 s: OH OH OH OH L-FUC l-P L-Rha l-P Fuculose 1-phosphate aldolase Rhamnulose 1-phosphate aldolase Fig.2 Product stereochemistries generated by the four complementary DHAP aldolases Table 1 Some commercial sources of aldolases, glycosyltransferases, and gl ycosidases" Aldolase type enzymes Fructose- 1,6-diphosphate aldolase Type I Fructose- 1,6-diphosphate aldolase Type I1 Fuculose- l-phosphate aldolase Rhamnulose-1-phosphate aldolase N-Acetylneuraminic acid aldolase Transaldolase Transketolase G1 ycosyltransferases p 1,4-Galactosyltransferase (x2,6-Sialyltransferase (x2,3-Sialyltransferase (x 1,3-Fucosyltransferase (x 1,2-Mannosyltransferase Glycosidases Supplier Boehringer, Calbiochem, Fluka, Serva, Sigma, Worthington Boehringer, Fluka, Sigma Boehringer Boehringer Toyobo, Sigma, Boehringer Sigma Sigma, Fluka Supplier Boehringer, Calbiochem, Fluka, Sigma, Oxford GlycoSystems, Worthington Boehringer, Calbiochem, Sigma Calbiochem Calbiochem, Oxford Gly cosy stems Calbiochem a Many variety of glycosidases from various sources (too many to list here) are available from the suppliers listed above as well as some other suppliers.Refer to the references for suppliers of specific glycosidases. More than 100 aldehydes have been used as acceptor substrates for DHAP dependent aldolases to prepare mono- saccharides.4-7 DHAP dependent aldolases also catalyze the condensation of pentose and hexose phosphates with DHAP, consequently extending the sugar chain by three carbons while introducing two new stereogenic centres.This provides a route to novel high-carbon sugars which are difficult to obtain from either chemical synthesis or natural sources. A number of these compounds have been synthesized, including analogs of 2-keto- 3-deoxyoctulosonate (KDO) [Fig. 3(a)]. When an appropriate dialdehyde is the substrate, C-disaccharide mimetics where two sugars are linked by carbon instead of oxygen, can be prepared by enzymatic tandem aldol reactions [Fig. 3(b)] .I2 Pyruvate-dependent aldolases, such as neuraminic acid aldolase (NeuAc aldolase), have also been used extensively in synthesi~.~~ An example is illustrated in the synthesis of analogs of N-acetyl neuraminic acid, a biologically important sugar involved in cell adhesion processes and viral infection [Fig.3(c)].13,14 These 408 Chemical Society Reviews, 1997, volume 26 PO/\/OH OH FDPA * HO IHO DHAP HO FDP A FDP A DHAP HO DHAP HO HO OP HO-HO nu UI1 F IYPUM~ aiuuiaac Dyruvate.- R HO OH + - R -WCOOH HO OH substratefor sialyltransferases Fig. 3 Aldolase-catalyzed synthesis of novel high-carbon sugars OH 2. Pase I 1 HO HO OH 0 0 1. DERA t- HO co2- 2. NeuAc aldolase I OH Fig. 4 Sequential aldol reactions catalyzed by DERA unnatural neuraminic acids can be converted to their cytidine- 5’-monophospho-derivatives (CMP-derivatives) that sialyl-transferases accept as a substrate, providing a route to modified sialyloligosaccharides [Fig.3(c)].14J5 Further study of enzymatic aldol reactions has led to the development of new sequential aldol reactions resulting in the combination of three or four substrates in one pot (Fig. 4).16 The key to this reaction is the use of deoxyribose-5-phosphate aldolase (DERA) which gives aldehyde products that can serve as a substrate for further enzymatic aldol reactions. With the increasing understanding of the specificity of various aldolases, these sequential aldolase reaction processes are expected to find use in the synthesis of various uncommon monosaccharides. The enzymatic aldol reaction of azido-aldehydes have been used to prepare imino-sugars [Fig.5(~1)].~7,~7 Azido-sugars prepared by aldolase-catalyzed reactions can be converted to cyclic-imine sugars by reduction under strongly acidic condi- tions followed by neutralization, or to iminocyclitols by Pd, Pt, or Rh-catalyzed hydrogenation in which the reduction of azide to amine, rearrangement to imine, and reductive amination takes place in one pot.17 Thio-aldehydes have been used to prepare deoxythiosugars [Fig. 5(h)].4 When nitroaldehydes are used as substrates, the enzymatic products undergo an intramolecular nitroaldol reaction to give nitrocyclitols [Fig. S(C)].~When phosphonate-containing aldehydes are used as substrates, the products spontaneously undergo a Horner-Wadsworth-Em-mons olefination to give another type of cyclitol [Fig.5(6>].4 Transketolases and transaldolases are commercially available enzymes that have also been used for enzymatic aldol type reactions.6-7 Highly efficient overexpression and large-scale synthesis of carbohydrates have been reported with trans-ketolase.18 There are other enzymes that are not classified as aldolases but that catalyze aldol type reactions (often classified as synthases and transferases). These enzymes are increasingly important but will not be mentioned in this article because of space limitations and because they have not been used very much in carbohydrate synthesis. Antibodies that catalyze aldol reactions have also been made.” 3 Enzymatic synthesis of oligosaccharides, glycopeptides and glycoproteins 3.1 Glycosyltransferases In vivo, oligosaccharides are synthesized sequentially, in a one- linkage one-enzyme fashion, by a variety of glycosyl-transferases that catalyze the transfer of sugars from an activated species, such as sugar nucleotides, to a growing oligosaccharide chain.This contrasts with the formation of proteins and nucleic acids which are synthesized by a single biocatalytic machinery that forms all linkages according to a template. Due to advances in recombinant DNA technology, many of the glycosyltransferases are now available in large quantities for the in vitro synthesis of various oligosaccharides. Coupled with the regeneration of sugar nucleotides, these Chemical Society Reviews, 1997, volume 26 409 1.FDPA HO N3 0 H2, PdIC, HCI,,; : II NaOH,, OH(*) M e v H 4OH + HOqcH3 HO~ OH OH H2, Pd/C POLOH 2. phosphatase Me&DHAP H OH ____)___) cDHAP HO AAcO O G 2. phosphatase HO OH OAcOH (4 1. DHAPFDP A 02N52c 02NbOAc+ p-OAcH&NO2 OH -OH 2. phosphatase HOZ' OH 'OAc bAc (1 : 1) Fig. 5 Synthesis of imino-sugars, deoxythiosugars, and cyclitols using aldolases enzymes have been developed for large-scale synthesis. The cofactor regeneration scheme not only reduces the cost of sugar nucleotides, but also lessens the problem of product inhibition caused by the resulting nucleoside phosphates.20 This enzy- matic strategy of oligosaccharide synthesis has been applied to the kilogram-scale synthesis of the oligosaccharide sialyl Lewisx (SLex) which is in clinical trials as a new anti-inflammatory agent for the treatment of reperfusion injury [Fig. 6(a)].Other regeneration schemes have also been developed.Recently, a novel regeneration system for UDP-Gal has been used in combination with a-and P-galactosyltransferases to afford a xenotransplantation antigen [Fig. 6(b)].21 Glycosyltransferases have also been used in the solid-phase and solution-phase synthesis of glyc~peptides.~~,~~ More re- cently, glycosyl transferases in combination with the protease, subtilisin, have been used for the synthesis of novel glycopro- teins. Ribonuclease B (FWase B) with a heterogeneous carbohydrate composition was remodeled to a homogeneous species via enzymatic removal of the heterogeneous saccharide units, followed by addition of new sugars, including unnatural sugars such as mercury containing sugars, with glycosyl- transferases (Fig.7).24 Combined with the enzymatic ligation of peptide fragments to proteins, this strategy provides a powerful method for the preparation of glycoproteins. The mild reaction conditions and biocompatibility of the enzymatic method for synthesis of glycopeptides and glycoproteins are complemen- tary to the solution- and solid-phase chemical approaches and may be more suitable for the synthesis of large and complex structures. Indeed, the principles of enzymatic aldol-reactions and glycosylations have been applied to systems as complex as that of engineering cell surfaces.2s Some glycosyl transferases catalyze the formation of sugar polymers. Sequential reaction of sugar nucleotide donors with a growing oligosaccharide will give a polysaccharide.This was demonstrated in the synthesis of a medicinally important biopolymer hyaluronic acid (HA) .26 HA with a molecular mass of -500 000 has been prepared from UDP-GlcNAc and UDP- glucuronic acid (UDP-GlcA) using haluronic acid synthetase coupled with regeneration of the sugar nucleotides (Fig. 8). 410 Chemical Society Reviews, 1997, volume 26 3.2 Glycosidases Glycosidases are enzymes that activate sugars towards hydroly- sis of glycosidic bonds in vivo. Under appropriate conditions, however, the activated intermediates can be intercepted by other sugars to form new glycosidic bonds.This is especially useful when a glycosyltransferase is not available or difficult to obtain. Glycosidases also have the advantage of not requiring ex- pensive sugar nucleotides as the sugar donor. An example is the synthesis of sialyl Lewisa, in which the 1,3-linked N-acetyllactosamine was prepared via a b-galactosidase reac- tion followed by enzymatic glycosylation using sialyl and fucosyl transferases [Fig. 9(a)].4By sequentially using two different galactosidases, a xenotransplantation antigen was also synthesized [Fig. 9(b)].27,28 The carbohydrate polymer cellu- lose has been prepared using cellulase.29 Catalytic antibodies with glycosidase activities have been prepared and may also become useful for synthesis in the future.30 One of the drawbacks of glycosidase mediated glycosylations has been the low yield associated with the hydrolytic nature of the enzyme. Various strategies have been used to overcome this problem.Some of the novel strategies are depicted in Fig. 10. Glycosidase-catalyzed synthesis of disaccharides, for example, can be coupled in situ with a glycosyltransferase reaction to improve the overall yield [Fig. lO(a)].4 Another interesting way to improve the yield and facilitate product isolation of glycosidase catalyzed glycosidation reactions has been demon- strated in the galactosidase-catalyzed synthesis of N-acetyl lactosamine, one of the intermediates in the synthesis of SLex.31 The key was the use of 6-0x0 p-nitrophenyl galactose, prepared by enzymatic oxidation of the corresponding galactose deriva- tive with galactose oxidase, as the glycosyl donor.The 6-0x0 derivatives are less prone to hydrolysis resulting in improved yields of the 6'-0xo disaccharide. Reduction of the aldehyde with sodium borohydride afforded the desired product and also facilitated isolation due to formation of a boron complex [Fig. 10(b)]. Chitinase, which normally works to hydrolyze the polymer, chitin, has been used to synthesize artificial chitin in quantitative yield.32 The key to this polymerization reaction is the use of a transition state analog substrate and performing the CMP [I-UDP E4 Pi +-PPi PEP UTP HO PEP El :Glycosyltransferase El : Glycosyltransferase E2: Pyruvate kinase E3: Sugarnucleotide pyrophosphorylase OH OH E3: Pyruvate kinase E4: Pyrophosphatase HO sialyl LewisX E4: Sugarnucleotide synthase I H?PH OH N HAc Galcrl,3Gal~1,4GlcNAc-OR Xenotransplantation antigen UDP-Gal UDP El : ~1,4-Galactosyltransferase E2:al,3-Galactosyltransferase E3: Sucrose synthase E4: UDP-Glc 4'-epimerase Fig.6 Multiple enzyme systems used in the large-scale synthesis of oligosaccharides endoglycosidase H * dase B Subtilisin 8397 Glycosyl glycero1:buffer (9:l) Fragments Q GlcNAc transferases pH 6.2 (42°C)GlcNAc-Protein S(25°C or 4°C) 0 7 A = MeHgS- orHO- and glycopeptides Fig. 7 Enzymatic synthesis of glycoproteins: Synthesis of a ribonuclease B glycoform containing sialyl Lewisx Chemical Society Reviews, 1997, volume 26 411 reaction at a high pH where the enzyme can activate the ,,/ UDP-GICA \/UDP-GlcNAc \ substrate but cannot hydrolyze product [Fig.1O(c)]. 4 Some applications in glycobiology 4.1 Synthesis of sialyl LewisX mimetics Though carbohydrates themselves are not always suitable as drugs-they are too unstable and orally inactive-understand- ing the mechanism of carbohydrate recognition opens the way Hyaluronic Acid Synthase for new concepts and strategies in drug development, such as r 1 designing carbohydrate mimics with better pharmacological properties to intervene with carbohydrate mediated processes. This has been the case with sialyl LewisX (SLex), a carbohydrate moiety involved in cell adhesion processes. The chemistry developed for the synthesis of SLex and related structures has led to important discoveries in structure-function relationships [Fig.1 1(a)].33,34These discoveries, together with the conforma- tion of SLex determined by NMR [Fig. l 1(a)],20.33334 have led to Fig. 8 Enzymatic synthesis of hyaluronic acid with regeneration of sugar the rational development of SLex mimetics which may be nucleotides comparable to or even better than the natural ligand as inhibitors of selectin recognition events. Several groups have been pN02Ph-P-Gal H&o&OH OH CMP-NeuAc HWS&,,P-galactosidase HG0& SialylTP El--0 HO GDP-FUC H02CXo NHAcHO nuuI1 (42%) OH HO OH NHAc FucT HQ ....\-,, HO OH pN02Ph-P-Gal OH OH 0'go&a-galactosida_se HoQHO P-galactosidase pNO2Ph-a-Gal (50%) -HO!&$i&SEt OH N HAc NHAc SEt (20%) SEt OH Ho NHAc Xenotransplantation antigen Fig. 9 Use of galactosidases in the synthesis of sialyl Lewisa and a xenotransplantation antigen OH IHod-. . C02H OH *OH &HO 80 HO OH Ho .RO OHNHAc nu NHAcCMP-CMPNeuAc uOH OH regeneration galactoseoxidase HO NHAc HOH0&0pN02Ph -OpN02P h * cataiase HO OHOH Pgalactosidase (77%) 1) NaBH4 Ho&io&.e Ho OH (60%) ,OH chitinase Fig. 10 Strategies used to improve the yields of glycosidase catalyzed glycosylations 412 Chemical Society Reviews, 1997, volume 26 actively engaged in this effort, and several SLex mimetics developed [Fig.1 l(b)]have been shown to have affinities for selectins increased from the millimolar range for SLex to the micromolar range for the mimetic^.^^ Some key reactions used in the development of these mimetic compounds are the enzymatic aldol reactions. A key component used in the synthesis of some of the potent mimetics is (2S,3R)-2-amino- 3,4-dihydroxybutanoic acid (L-hydroxythreonine), which can be easily prepared via a threonine aldolase-catalyzed addition reaction.35 The fructose-diphosphate aldolase catalyzed aldol reaction of DHAP and a mannose derivative gave a phosphory- lated compound which had especially good inhibition against P-selectin binding.34 4.2 Synthesis of glycosidase and glycosyltransferase inhibitors Inhibition of the enzymes associated with carbohydrate bio- synthesis is also biologically and medicinally important.Both glycosidases and glycosyltransferases are important enzymes involved in the processing and synthesis of oligosaccharides and are therefore obvious targets for intervention. The reactions catalyzed by these enzymes are thought to proceed through similar transition states in which substantial sp2 character and positive charge develop at the anomeric centre of the reacting sugar [Fig. 12(a)].This mechanistic rationale has led to the development of transition-state analog inhibitors of glycosi- dases and glycosyltransferases. Various 5-, 6-, and 7-membered ring imino-sugars have been synthesized using aldolases [Fig.12(b)] These nitrogen-containing heterocycles are po- tent inhibitors of glycosidases and have also been used as key components in the synthesis of glycosyltransferase inhibitors. An inhibitor of a-I ,3-fucosyltransferase, for example, has been prepared by the attachment of the N-acetyllactosamine moiety to an iminocyclitol-type a-fucosidase inhibitor [Fig. 12(c)].37 Combined with new high throughput assays, the nitrogen heterocycles can also serve as core structures for combinatorial approaches to the development of glycoprocessing enzyme inhibitors.l7 5 Conclusion and future prospects As we have briefly reviewed, enzymatic aldol reactions and glycosylations have made access to carbohydrate molecules Fig.11 (a) Sialyl LewisX (SLex) functional groups crucial for selectin binding and NMR structure of SLex. (b)Chemo-enzymatic synthesis of SLex mimetics Chemical Society Reviews, 1997, volume 26 413 + h '&OH 5 Me 0--,6+ -02c I HoOH&:-OH HO 'OHI Me 0--,6+ I Acceptor-O-H -----:B-5 Fucosidase Reaction a-l,3-Fucosyltransferase Reaction 0 K, = 1.4 prn K, = 5.5prn HO HO I I HO ' I OH : .-.., HO OH Me= H OH M e kOH I( = 5.6 nM K, = 6.9 nMHo60H H IC50 = 31 pM in the presence of 0.03 rnM GDP I(=5pM Fig. 12 (a) Transition-state of glycosidases and glycosyltransferases. (b) Transition-state analog inhibitors of fucosidase. (c) An azatrisaccharide fucosyltransferase inhibitor shown as a complex with GDP.more routine and available on a large-scale basis. Many glycoprocessing enzymes useful for synthesis are already available and new ones are becoming accessible at an ever increasing rate. Indeed, it is believed that any oligosaccharide in the mammalian system can be prepared in large quantities based on the glycosyltransferase methodology as methods for the regeneration of all relevant sugar nucleotides have been developed.6 In cases where an enzyme that catalyzes a certain reaction is difficult to obtain or non-existent, the use of alternative biocatalysts (for example, glycosidases and catalytic antibodies) or novel substrates (for example, Fig. 5) is useful. These enzymatic tools are accelerating research in glycobiology and medicine as well as provide opportunities to open new doors in other fields of science.With an increasing number of these enzymes becoming commercially available (Table l), it is anticipated that use of enzymatic aldol reactions and glycosyla-tions will become more and more routine and find many applications in various fields, as was the case with proteins and nucleic acids. 6 Acknowledgement We thank many of our coworkers, whose names are listed in the references, for their contributions to this research. We thank Dr Shang-Cheng Hung for helpful discussions. Our work was financially supported by the National Institutes of Health, the National Science Foundation, Cytel Corporation in San Diegoand Sandoz Pharma Preclinical Research.414 Chemical Society Reviews, 1997, volume 26 7 References I T. Ogawa, Chem. Soc. Rev., 1994, 23, 397. 2 S. J. Danishefsky and M. T. Bilodeau, Angew. Chem., Int. Ed. Engl., 1996,35, 1380. 3 Preparative Carbohydrate Chemistry,ed. S. Hanessian, Marcel Dekker, New York, 1997. 4 H. J. M. Gijsen, L. Qiao, W. Fitz and C.-H.Wong, Chem.Rev., 1996,96, 443. 5 C.-H. Wong, R. L. Halcomb, Y. Ichikawa and T. Kajimoto, Angew. Chem., Int. Ed. Engl., 1995, 34, 412. 6 C.-H. Wong and G. M. Whitesides, Enzymes in Synthetic Organic Chemistry, Pergamon, Oxford, 1994. 7 W.-D. Fessner and C. Walter, Top. Curr. Chem., 1996, 184, 97. 8 N. Blom and J. Sygusch, Nature Struct. Bid., 1997, 4, 36. 9 J. Sygusch, D. Beaudry and M. Allaire, Proc.Natl. Acad. Sci. USA, 1987,84,7846. 10 S. J. Cooper, G. A. Leonard, S. M. McSweeney, A. W. Thompson, J. H. Naismith, S. Qamar, A. Plater, A. Berry and W. N. Hunter, Structure, 1996, 4, 1303. 11 M. K. Dreyer and G. E. Schulz, J. Mol. Biol., 1996, 259, 458. 12 0.Eyrisch and W.-D. Fessner, Angew. Chem., Znt.Ed. Engl., 1995,34, 1639. 13 C.-C. Lin, C.-H. Lin and C.-H. Wong, Tetrahedron Lett., 1997, 38, 2649. 14 J. L.-C. Liu, G.-J. Shen, Y. Ichikawa, J. F. Rutan, G. Zapata, W. F. Vann and C.-H. Wong, J. Am. Chem. SOC.,1992, 114, 3901. 15 M. D. Chappell and R. L. Halcomb, J. Am. Chem. Soc., 1997, 119, 3393. 16 H. J. M. Gijsen and C.-H. Wong, J. Am. Chem. Soc., 1995, 117, 7585. 17 S. Takayama, R. Martin, J. Wu, K. Laslo, G. Siuzdak and C.-H.Wong, J. Am. Chem. Soc., 1997,119, 8146. 18 K. G. Morris, M. E. B. Smith, N. J. Turner, M. D. Lilly, R. K. Mitra and J. M. Woodley, Tetrahedron: Asymmetiy, 1996, 7, 2185. 19 J. Wagner, R. A. Lerner and C. F. I. Barbas, Science, 1995, 270, 1797. 20 Y. Ichikawa, Y.-C. Lin, D. P. Dumas, G.-J. Shen, E. Garcia- Junceda, M. A. Williams, R. Bayer, C. Ketcham, L. E. Walker, J. C. Paulson and C.-H. Wong, J. Am. Chem. Soc., 1992,114,9283. 21 C. H. Hokke, A. Zervosen, L. Elling, D. H. Joziasse and D. H. Van Den Eijnden, Glycoconj. J., 1996, 13, 687. 22 0. Seitz and C.-H. Wong, J. Am. Chem. Soc., 1997, 119, 8766. 23 M. Schuster, P. Wang, J. C. Paulson and C.-H. Wong, J. Am. Chem. Soc., 1994, 116, 1135. 24 K. Witte, P. Sears, R. Martin and C.-H.Wong, J. Am. Chem. Soc., 1997, 119, 21 14. 25 L. K. Mahal, K. J. Yarema and C. R. Bertozzi, Science, 1997, 276, 1125. 26 C. Deluca, M. Lansing, I. Martin, F. Crescenzi, G.-J. Shen, M. O’Regan and C.-H. Wong, J. Am. Chem. Soc., 1995,117,5869. 27 G. Vie, C. H. Tan, M. Scigelova and D. H. G. Crout, Chem. Commun., 1997, 169. 28 K. G. I. Nilson, Tetrahedron Lett., 1997, 38, 133. 29 S. Kobayashi, K. Kashiwa, K. Tatsuya and S. Shoda, J. Am. Chem. Soc., 1991,113,3079. 30 K. D. Janda, L.-C. Lo, C.-H. L. Lo, M.-M. Sim, R. Wang, C.-H. Wong and R. A. Lerner, Science, 1997,275,945. 31 T. Kimura, S. Takayama, H. Huang and C.-H. Wong, Angew. Chem., Int. Ed. Engl., 1996, 35, 2348. 32 S. Kobayashi, T. Kiyosada and S. Shoda, J. Am. Chem. Soc., 1996,118, 13113. 33 P. Sears and C.-H. Wong, Proc. Natl. Acad. Sci. USA, 1996, 93, 12086. 34 C.-H. Wong, F. Moris-Varas, S.-C. Hung, T. G. Marron, C.-H. Lin, K. W. Gong and G. Weitz-Schmidt, J. Am. Chem. SOC.,1997,119, 8152. 35 V. Vassilev, T. Uchiyama, T. Kajimoto and C.-H. Wong, Tetrahedron Lett., 1995, 36, 4081. 36 F. Moris-Varas, X.-H. Qian and C.-H. Wong, J. Am. Chem. Soc., 1996, 118,7647. 37 L. Qiao, B. W. Murray, M. Shimazaki, J. Schultz and C.-H. Wong, J. Am. Chem. Soc., 1996, 118,7653. Received, 2 7th June 1997 Accepted, 5th August I997 Chemical Society Reviews, 1997, volume 26 415

ISSN:0306-0012    

DOI:10.1039/CS9972600407

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Polymer-supported organic reactions: what takes place in the beads? Philip Hodge Department of Chemistry, University of Manchester, Oxford Road, Manchestei-, UK M13 9PL The use of polymer-supported reactants in organic synthesis is currently of considerable interest, especially in the context of combinatorial syntheses. To carry out successfully reac- tions using polymer-supported reactants it is important to be aware of what takes place inside the beads. Examples are presented in this article which show that compared to the analogous homogeneous reaction systems, polymer-sup- ported reactions can show substrate selectivity, be slower or faster, follow a different reaction course, or give a sig- nificantly different stereochemical result. 1 Introduction In 1963Merrifield first described in the literature his method of ‘solid phase’ peptide synthesis.’ In this method the first amino acid residue of the peptide to be synthesised is bound to polystyrene beads through an ester linkage formed using the carboxyl group of the amino acid.As a consequence the peptide subsequently synthesised is attached to the beads via the carboxyl terminus. The polystyrene beads are crosslinked and are, therefore, totally insoluble in all organic solvents. Thus, at each stage in the synthesis the supported peptide can be separated cleanly and easily from the other species present. At the end of the synthesis the peptide produced is detached from the polymer support by cleaving the ester linkage. Merrifield first synthesised a tetrapeptide using this approach,l but he had soon developed a machine for automated peptide synthesis,* and synthesised ribonuclease A3 an enzyme with 124 amino acid residues.This novel approach so revolutionalised peptide synthesis that Merrifield was awarded the 1984Nobel Prize for Chemistry. Also in 1963 Letsinger and Kornet described an alternative ‘solid phase’ peptide synthesis procedure in which the first amino acid was bound to the polymer support through the amino Philip Hodge graduated from the University of Manchester in 1960 and then worked for his PhD with Arthur Birch and Rod Rickards on naturally occurring pyrroles. After postdoctoral work in collabot-atron with Professor- Sir Ewart Jones at Oxford and then with Syntex in California he joined the staff of Lancaster University.Noting that most of the smarter molecules in nature are reactive macromolecules, he has spent the last 25 years carry- ing out research at the organic chemistry-polymer chemistry in- terface working on both the ap- plications ofreactive polymers to organic synthesis and the appli- cations of organic chemistry to the synthesis of novel polymers. He moved to the Chair of Poly-mer Chemistry in the Department of Chemistry at the University of Manchester in 1989. group.4 This research work was subsequently developed into a basis for ‘solid phase’ oligonucleotide synthesis,5 and this general synthetic approach has proved invaluable for synthesis- ing oligonucleotides for many applications in genetic engi- neering.The advantages of the ‘solid phase’ method are not, however, limited to the synthesis of natural polymers and there are many applications in other areas of organic synthesis. Whilst organic reactions using polymer-supported species had been carried out prior to Merrifield’s work, for example, the use of the acid forms of cation-exchange resins as supported acid catalysts for esterifications, his work stimulated other researchers to study a wide range of synthetic organic reactions using polymer- supported substrates, polymer-supported reagents, or polymer- supported catalysts. Over the last 30 years or so more than a thousand papers unconnected with peptide synthesis have been published on these other topics.It is evident from these studies that the polymer support plays an active and crucial role in supported syntheses. With the upsurge in interest in polymer- supported reactions, especially in the context of combinatorial chemistry637 and of the automation of organic syntheses, it is appropriate to briefly review in this article what has been learnt about ‘what goes on in the polymer beads’, especially as much of this fundamental work has been published other than in the standard organic chemistry journals. Due to space limitations this article cannot in any way be comprehensive. Instead, just a few selected examples are discussed which illustrate some of the complexities of polymer-supported reactions.It is important to appreciate these complexities if, for example, combinatorial syntheses and possibly the screening of the products of such syntheses are to be carried out meaningfully. ‘Solid phase’ peptide syntheses are, in effect, examples of reactions using supported substrates. More specifically they are examples of reactions involving supported protecting groups. In most ‘solid phase’ peptide syntheses it is carboxylic acid groups that are protected. Syntheses using supported substrates are also the type of reactions used in combinatorial chemistry. They involve, in sequence, the attachment (linking) of the initial substrate to the support, various polymer-supported synthesis steps, and then the detachment of the final product from the support.Because no separation of the supported species is possible, the supported synthesis reactions must be very clean and high yielding. This requires a very careful choice of reaction conditions, especially as substantial substrate loadings ( > 1.5 mmol-I g) are often required in order to obtain useful amounts of the final products. Reactions using polymer-supported reagents are much less demanding since such reagents are used in only one reaction and not every functional site need react. The requirement for high loadings usually remains, however. Polymer-supported catalysts are particularly attractive because not every site needs to react, low loadings are often acceptable, and the recovered catalyst is often available for immediate re- use.It is tempting to think that the reactions of polymer-supported species will be just the same as those of low-molecular mass analogues. For many reactions carried out under homogeneous conditions this is the case. Indeed, the successful analyses of the kinetics of, for example, free radical polymerisations and Chemical Society Reviews,1997, volume 26 417 condensation polymerisations are actually based on this as- sumption. However, the key point with reactions using species bound to insoluble polymer beads is that the reaction systems are heterogeneous. This has far-reaching effects. The polymer behaves as a separate phase. With a uniformly functionalised polymer bead of diameter ca. 100 vm, i.e. of a size that can be filtered off easily and is therefore commonly used, more than 99% of the functional groups will be within the bead.It is, therefore, clear that with beads of any significant loading reactive species in solution will have to enter into the beads to react and that inside the beads there is often a significantly high concentration of reactive sites. The major differences between reactions on polymer sup- ports and their low-molecular mass analogues can be grouped loosely into three main types of effect. These are: (i) effects resulting from the need for the soluble reactants to gain access to the supported reactants, (ii) microenvironmental effects and (iii)site-site interactions. These effects are not independent of each other and to complicate matters further the contributions and importance of the various effects can change as a reaction and/or synthesis proceeds. They do, however, provide a convenient framework for the following discussion. 2 Access of soluble reactants to supported reactants The supports which have been used most extensively for polymer-supported organic syntheses are microporous poly- styrene beads crosslinked with 1 or 2% of divinylbenzene. For reactive species in solution to gain access to the reactive sites in the beads, these beads must be swollen by the reaction solvent.To a first approximation the solvents which will swell the beads best are those that would dissolve the corresponding linear polymers. Indeed swelling represents an attempt by the polymer chains to dissolve.The extent of swelling decreases markedly as the percentage of crosslinking increases and a 1% crosslinked bead swells significantly more than a 2% crosslinked bead. If the percentage crosslinking is much less than 1%, however, polystyrene beads tend to become physically fragile and, unless they are handled very carefully, they can disintegrate. It is important to stress here that the functionalities attached to the beads can significantly affect the swelling properties, especially with highly functionalised beads, and that during chemical reactions the swelling properties may change considerably as one functionality is transformed into another. The choice of reaction solvent is therefore crucial in polymer-supported reactions and the optimum solvent may not be the same one as that commonly used in the analogous reaction using low- molecular mass reactants.It should be noted that the reactions in the swollen beads take place in a gel phase and not, as commonly described, a ‘solid phase’. Reactions in the solid state are very different from polymer-supported reactions. Reaction systems which serve to illustrate the importance of the choice of swelling solvent come from early studies of supported transition metal complex catalysts. Kagan’s research group prepared 2% crosslinked polystyrene beads containing 0.5 mmol g-1 of 2,3-0-benzylidene-2,3-dihydroxy-l,4-bis(di-pheny1phosphino)butane residues 1.8 Rhodium(1) complexes prepared from these beads could be used successfully to catalyse the reactions between dihydrosilanes and prochiral ketones in benzene at 20 “C: see Scheme 1.The chemical yields were high and the percentage enantiomeric excesses (% ee) achieved were very similar to those obtained (S58% ee) in the analogous low-molecular mass reactions. The same polymer- supported catalyst also catalysed the hydrogenations (see Scheme 2) of a-methylstyrene, 2-ethylhex- 1 -ene, a-ethylstyrene 3 and methyl a-phenylacrylate 4 in benzene at 20 “C but in the last two cases the % ee values achieved were even lower than in the corresponding low-molecular mass reactions (<2.5% ee vs. d 15% ee). a-Acetamidoacrylic acid 5 is practically insoluble in pure benzene and attempts to hydrogenate it in benzene-ethanol mixtures failed as did 418 Chemical Society Reviews, 1997, volume 26 1 CH3 2 attempts to hydrogenate a-methylstyrene and 2-ethylhex- 1 -ene under these conditions.The reactions failed because ethanol did not swell the crosslinked polystyrene matrix. The use of ethanol as a solvent is, however, highly desirable as the % ee obtained in such hydrogenations are often higher with this solvent. Stille and his group9 overcame the problem by incorporating the same catalytic groups (8% of repeat units) into a lightly crosslinked poly(2-hydroxyethyl methacrylate) 2. The hydroxyethyl moie- ties interact well with ethanol and the matrix swells well in this solvent. The derived rhodium(1) complexes successfully cata- lysed the hydrogenations of a-acetamidoacrylic acid 5, its (3-phenyl derivatives 6, and a-phenylacrylic acid 7 in benzene- ethanol (1 vol:5 vol) and gave the desired products in essentially the same ee (5240% vs.73%; 86% vs. 81%; 5844% vs. 63% respectively) as those obtained with the analogous low-molecular mass catalyst. The dominant config- urations were the same in both reaction systems. Even when reactions using polymer-supported reagents or catalysts do proceed smoothly, if diffusion of the soluble substrate into the active sites in the beads is rate limiting, and it often is, this can result in the supported reactant displaying a significant size selectivity. This is illustrated by some work of R = alkyl ; X = aryl ; Y = alkyl or aryl Scheme 1 Asymmetric hydrosilylation of prochiral ketones R2 H2 * ,R2 R3CH=C: -R’CHp-CH, ~~ 0 ~catalyst 1 C02R’ 4 R’ = CH3; R2 = C6H5; R3 = H 5 R’ = H; R2 = -NHCOCH3; R3 = H 6 R’ = H; R2 = -NHCOCH3; R3 = C6H5 7 R’ = H; R2 = CeH5; R3 = H Scheme 2 Catalytic hydrogenations carried out with polymer-supported chiral catalysts Grubbs and Kroll on the hydrogenation of olefins.lO A catalyst 8 was prepared by equilibrating 2% crosslinked polystyrene beads in which 8% of the repeat units were benzyldiphenyl- 8 SO~H S02NHNH2 9 10 phosphine residues with a twofold excess of tris(tripheny1- phosphine)-chlororhodium(1). When used as a catalyst with benzene as the solvent at 25 "C under 1 atm of hydrogen the relative rates of hydrogen uptake, with hex- 1-ene arbitrarily set to be 100, were octadecene (mixture of isomers) 19; cyclohex- ene 39; cyclooctene 15; cyclododecene (cis-and trans-mixture) 8.8; and A2-cholestene 1.2.With a similar soluble catalyst the relative rates of hydrogen uptake, again relative to hex- 1-ene arbitrarily set to be 100, were all in the range 7 1-100. Thus with the supported catalyst the rate of reduction depended greatly on the molecular size of the olefin. Going from an acyclic to a cyclic olefin or increasing the ring size of a cyclic olefin decreased the rate of reduction. The large rigid A2-cholestene showed the most dramatic decrease in reduction rate. These differences arise because the more bulky a molecule the more slowly it diffuses to the reactive sites in the beads.The effects are particularly pronounced in the present case probably because most of the catalyst sites will involve two or more polymer-supported phosphine ligands. These sites will, there- fore, serve as extra crosslinks. This makes diffusion through the matrix more difficult than it otherwise would be. It also means that many of the active sites will actually be on crosslinks. The latter are clearly regions which are particularly crowded. In most other reaction systems access to the active sites is less of a problem and bulky molecules such as steroids, for example, can react without difficulty." Reaction solvent restrictions may sometimes be overcome by using macroporous or macroreticular polymers.Usually these are 2040% crosslinked polystyrene beads prepared by suspen- sion polymerisation in the presence of a porogen. Various types of internal structure are possible depending on the amount and type of porogen used. Such beads have a rigid porous structure that scarcely swells in most solvents. The open texture allows a wide variety of solvents to enter the pores but not necessarily the highly crosslinked near-rigid framework of the bead. Some commercial polymer beads including certain anion-exchange resins are specially prepared for use in organic solvents. They can provide a convenient means of making various anions available for reaction in non-aqueous solvents. For example, the periodate form of such resins containing 1.3-2.0 mmol of oxidant per g can be used successfully to cleave 1,2-diols in ethanol, chloroform, dichloromethane, diethyl ether or benzene as well as water.12 However, in other cases a substantial fraction of the reactive groups are located in highly crosslinked regions and are often not readily available for chemical reaction.Evidence for such effects was found by Emerson et aZ.13They converted a commercial macroreticular cation-exchange resin containing more than 5.2 mmol 8-1 of residues 9 into various polymer-supported benzenesulfonylhydrazides 10. These were then reacted with a range of aldehydes and ketones, presumably to give hydrazones. Starting with a polymer containing 2.8 mmol g-1 of residues 10, acetone in benzene reacted with virtually every site (2.7 mmol g-I), but pentan-2-one and cyclopentanone only reacted respectively with 1.9 and 1.8 mmol g-l of the groups even when a substantial excess of ketone was used or when more highly loaded starting polymers were used.Starting with polymers containing 4.2 and 4.8 mmol g-1 of residues 10, 1.2 and 1.7 mmol g-1 reacted respectively with cyclohexanone and cyclohex-2-enone. When these poly- mers were first treated with an excess of glucose in water to remove the more accessible residues 10, only 0.4 mmol g-l reacted with these six-ring ketones. These results suggest that as far as the present reactions are concerned only ca. I .O mmol g-1 of the residues 10 are in the more accessible parts of the macroporous beads: the rest are in the highly crosslinked regions.An effect relating to access arises if the polarity of the microenvironment within the beads differs significantly from that of the solvent outside the beads. The difference can either encourage or discourage low-molecular mass reactants from diffusing into the beads. If the diffusion barriers are not too high equilibria may be set up between the soluble reactants in the beads and the soluble reactants outside the beads. Takagi converted crosslinked poly(acry1ic acid)s into supported per- oxyacids 11 and carried out detailed studies of their reactivity.14 -7yH-CH2j-0'/c '0-OH 11 In this connection he studied the distribution of benzene and cyclohexane, as models for cyclohexene, between various peroxyacid resins and solvents.With several resins changing the solvent from dioxane to tert-butyl alcohol, for example, approximately doubled the concentrations of the model com- pounds in the resins. He observed that in epoxidation reactions conversion to epoxides was very poor when the solvent was less polar than the resin since then the olefin tended to stay in the solvent outside the beads. Sometimes it is desirable to be able to react non-polar functional groups in polystyrene beads with salts and it might be expected that this could cause some difficulties. However, often an excellent practical solution is to use phase transfer catalysis. Well known examples are the chemical modification of chloromethylated polystyrenes by reaction with cyanide, car- boxylates, phenoxides, or thiolates.l"l6 Such reactions are a convenient means to introduce desired functionalities into polystyrenes.It is evident from the above discussion that the polymer supports used for other than peptide or nucleotide synthesis are at present far from ideal in several respects and that there is considerable scope to prepare improved supports. This is likely to be a very active area of research in the future. Already numerous attempts have been made.17-lY It is, however, not a trivial matter to identify new supports that are easy to prepare, that have a satisfactory physical form that permits agitation during reactions and filtrations without problems, that have a reasonably high capacity (> 1.0 mmol g-I), and that contain repeat units that, unless required to do so, will not react with the diverse range of reagents involved in, say, a multistage combinatorial synthesis.Many of the supports used for peptide or nucleotide synthesis are simply lightly crosslinked micro- porous polystyrenes with novel linker groups which allow the easy detachment of the products when required. They generally Chemical Society Reviews, 1997, volume 26 419 have relatively low capacities, and some of them, for example glass beads, have exceedingly low capacities. A significantly different type of support are those (TentaGels) where the hydrophobic properties of crosslinked polystyrenes have been substantially offset by carrying out a graft polymerisations of ethylene oxide inside the beads.20 In other cases poly(N,N- dimethylacry1amide)s are used and the poor physical properties of these polymers overcome by depositing them inside rigid supports.21,22 One way forward is to investigate beads prepared using longer more flexible crosslinking agents than di-vinylbenzene.Such networks are likely to be superior to the common polystyrene beads.23 3 Microenvironmental effects It has already been noted that the microenvironment in a bead may alter the concentrations of low-molecular mass reactants present relative to those in the surrounding solution. The present section is concerned with cases where the microenvironment in a polymer bead can result in a change in the direction or the rate of a particular polymer-supported reaction.One of the earliest reported examples of a significant microenvironmental effect involves the reactions of several alkylbenzenes with a crosslinked poly(ma1eimide) 12 in which 70% of the repeat units were N-brominated to give residues 13.24 Although N-bromosuccinimide reacts with ethylbenzene R 12 R=H 13 R=Br in carbon tetrachloride in the presence of benzoyl peroxide to give a high yield of the a-bromo derivative 14, with the polymer-supported reagent 13 a significant amount of the dibromide 15 was also produced. Cumene reacts with N-bro- mosuccinimide under similar conditions to give high yields of bromide 16 or dibromide 17, depending on the amount of reagent used.In contrast cumene gave a mixture of the tribromide 18 (48%), monobromide 19 (15%), and monobro- Br Br 14 15 18 20 19 mide 20 (13%) when 2.3 equiv. of the supported reagent 13 were used and an 85% yield of tribromide 18 when 3.7 equiv. were used. p-Cymene and p-bromocumene behaved similarly. The authors suggest that the differences between the results obtained with N-bromosuccinimide and supported reagent 13 arise because the microenvironment within the polymer beads is relatively polar and this favours dehydrobromination reactions. Thus, dehydrobromination of the initial product 16 from 420 Chemical Society Reviews, 1997, volume 26 cumene would afford a-methylstyrene, and this would then react with bromine, formed by reaction of the supported reagent 13 with the hydrogen bromide liberated by dehydro-bromination, to give dibromide 17.Repetition of these dehydro- bromination and bromination reactions leads to the other products. Several further results support this interpretation. For example, treatment of dibromide 17 with the unbrominated crosslinked polymaleimide 12 resulted in dehydrobromination to give bromides 19 and 20, and N-bromosuccinimide reacted with cumene in acetonitrile, a more polar solvent, to give tribromide 18. A closely related system to that just discussed has recently been studied by Kondo et al.z5 They prepared a polymer by copolymerising 2,4-diamino-6-vinyl- 1,3,5-triazine with mol% of divinylbenzene and then reacted the product with tevt-butyl hypochlorite in methylene dichloride to give polymer beads containing 4.6 mmol g-l of residues 21.This supported CI HN NH CI 21 reagent reacted smoothly with cyclohexanol in methylene dichloride at room temperature to give cyclohexanone in 98% yield and with butane-1,4-diol to give y-butyrolactone in 76% yield. However, the analogous low-molecular mass reagent failed to react with the same substrates under similar conditions. The authors suggested that this was a microenvironmental effect. The polarity of the microenvironment in the beads may be significantly higher than that in the analogous reaction in solution and this may favour the oxidation reaction. Another example of a microenvironmental effect, in this case one which leads to an increase in reaction rate, concerns the conversion of n-alcohols into n-alkyl chlorides.' Reaction of octan- 1-01 with carbon tetrachloride and 4-diphenylphos-phinylisopropylbenzene22 or polymers containing residues 23 produces 1-chlorooctane.When a linear polymer containing residues 23 is used the rate of the reaction is about two times 22 23 24 faster than when phosphine 22 is used. With a lightly crosslinked polymer- Containing residues 23 the same effect 1s observed except that now the reaction is about 5 times faster. The rationalisation proposed depends on the fact that as the reactions proceed phosphonium salts are formed. In the reactions using the phosphine 22 these salts are either dispersed throughout the reaction medium or they precipitate out and thus play no further part in the reaction.With the polymers, however, the salt residues remain in the vicinity of the unreacted phosphines and they provide a favourable polar micro-environment for reaction of the latter with carbon tetrachloride. This last reaction (see Reaction 1) is the rate-limiting step in the whole process and is favoured by a polar environment. The CI-CC13Ar3P.Qn -Ar3P-CI + -CCI3 (1) Ar = aryl residues crosslinked polymer probably produces a greater effect than the linear polymer because with the latter the polar interactions can to some extent be reduced by the chains uncoiling. The crosslinks restrict the extent to which this can happen. It should be noted in this reaction system that as the polarity in the crosslinked beads increases they would be expected to swell less thus making the microenvironment within them even more polar, and that as the polarity in the beads increases the alcohol substrate will tend to accumulate there rather than remain in the non-polar carbon tetrachloride solution outside.Also it should be noted that the reaction involving the phosphine 22 takes place more rapidly than that involving simply triphenylphos- phine. This reminds us of the need to use good low-molecular mass models when studying polymer effects. Alexandratos and Miller have taken the study of microenvi- ronmental effects a step further and have actually sought to tailor the microenvironment in 2% crosslinked polymers containing phosphine residues 24 for carrying out the Mitsu- nobu reaction of benzoic acid with benzyl alcohol (Reaction 2).26 The rate-determining step in the Mitsunobu reaction is the reaction shown in Reaction 3.In proceeding to the transition (1) C6H5C02H + C6H5CH20H + C6H5C02CH2C6H5 (2) (I) phosphine and diethyl azodicarboxylate state this SN2 reaction involves charge dispersal and is, therefore, expected to be favoured by a non-polar environment. Alexandratos and Miller found that for a range of polymers suspended in tetrahydrofuran at 25 “C the percentage conver- sion of benzyl alcohol after 6 min was greatest (94%) with a crosslinked polystyrene in which 18% of the repeat units were residues 24. As the loadings of residues 24 increased, i.e.as fewer unsubstituted phenyl residues were present, the percent- age conversion under standard conditions dropped. For exam- ple, with a polymer in which 100% of the repeat units not involved in crosslinking were residues 24 the percentage conversion was only 42% after 6 min. When phenyl residues were replaced by the more polar residues derived from either methyl methacry late or from methacrylic acid percentage conversions dropped drastically (S2%). In most of these reaction systems high yields of ester were obtained given sufficiently long reaction times. The microenvironment in the vicinity of the polymer backbone can be expected to be sterically crowded. This is, for example, the reason why reactions at the phenyl residues of polystyrene occur mainly at the meta- and para-positions and not the ortho-positions.It is also the reason why chlorination of polymers containing residues 25 occurs mainly at the side chain position rather than the backbone position. This reaction forms +CH-CH~ + IQ 25 a convenient route for the preparation of chloromethylated polystyrenes.27 In general steric effects will be greatest when a reactive functional group is directly attached to the polymer backbone, but as the functional groups are separated from the backbone by ‘spacer groups’, steric effects would be expected to disappear rapidly and functional group accessibility and mobility to increase. 19 Most polymer-supported reactants are prepared from polystyrenes and here the benzene ring itself will act as a small rigid spacer group. The support might be expected to influence the course of reactions greatly if polarities are such that a substrate moiety bound to a polymer support prefers to interact with the support itself rather than with the solvent. Extreme examples of this situation occur when a hydrophobic substrate is simply adsorbed to a support and then undergoes reaction with aqueous reagents. When prochiral ketones adsorbed onto cellulose triacetate are reduced with aqueous potassium borohydride or prochiral enones adsorbed onto cellulose triacetate are epoxi- dized with alkaline hydrogen peroxide the chiral support clearly influences the course of the reactions as modest levels of asymmetric synthesis occur.28 Other examples of such effects are given in the next section.4 Site-site interactions The ease with which polymer-supported reactive groups can react together has long been an intriguing topic. In the 1960s and early 1970sorganic chemists tended to assume that the fact species were attached to a polymer automatically resulted in substantial site isolation, but detailed studies carried out since then have clearly shown that this is not the case.29 In this connection it should be remembered that in most of the polymer-supported reaction systems studied the overall con- centration of the reactive groups in the beads is reasonably high. For example, with a loading of reactive functional groups of 1 .O mmol g-l on a lightly crosslinked gel that swells in the reaction solvent by a factor of 3, the concentration of reactive groups is 0.33 mol dm-3.With a loading of reactive groups of 0.5 mmol g-’ on a highly crosslinked macroporous support which swells only modestly in the reaction solvent, the overall concentration of reactive groups will be ca. 0.5 mol dm-3, but if the functional groups were introduced by chemical modification of preformed beads they will be located mainly in the pores and there the local concentration will probably be in excess of 1.0 mol dm-3. An example of a reaction system which clearly demonstrates that in many circumstances a high proportion of polymer- supported reactive groups can reach each other easily comes from studies of Wittig reactions involving one of the less common ways of generating ylide~.~O Consider first the reactions that occur when triphenylphosphine reacts with carbon tetrabromide in methylene dichloride or with carbon tetrachloride.These are summarised in Scheme 3. Here X-Scheme 3 Reactions occurring when carbon tetrabromide or carbon tetrachloride are treated with triphenylphosphine. X = Br or C1 Reaction ‘C’ involves two phosphorus-containing species reacting together and it generates the dihalomethylene ylide. When, starting with polymers containing triphenylphosphine moieties 23, the analogous reactions occur, the analog of Reaction ‘C’ involves two supported species reacting together: see Scheme 4. Moreover, this particular reaction has to compete with that of the phosphine residues 23 directly with the carbon Chemical Society Reviews, 1997, volume 26 421 X-IC’ I (C6H5)2 p-x\ X Scheme 4 Site-site reactions occumng when polymer-supported phos- phines (23) are treated with carbon tetrabromide or carbon tetrachloride.The reaction is the polymer-supported analogue of reaction ‘C’ in Scheme 3. X = Br or C1. tetrahalide, i.e. the analogue of Reaction ‘A’ in Scheme 3. When the polymer-supported reactions are carried out in the presence of suitable aldehydes or ketones Wittig reactions will occur and dihalo-olefins with result. The yields of these olefins will be an indication of the extent of site-site interactions in these systems.In practice when a 1% crosslinked polystyrene containing 2.5 mmol g-’ of phosphine residues 23 was reacted with 0.6 mol dm-3 equivalents of carbon tetrabromide and 0.4 equiv. of benzophenone in chloroform, the 1,l -dibromo-olefin 26 could be isolated as crystals in 89% yield. In the analogous reaction with carbon tetrachloride, as both reactant and solvent, the 1 ,l-dichloro-olefin 27 was isolated in 87% yield. These figures 26 Ha = Br 27 Ha=CI indicate that at least 70% of residues 23 reacted together. Clearly increasing the percentage of crosslinking would be expected to reduce chain mobility and thus be a major factor in decreasing the ease of site-site interactions. This proved to be the case in the present reaction system. Thus as the percentage crosslinking increased successively from 2 to 4 to 8 to 15 and to 37% the minimum number of sites reacting together decreased respectively from 47 to 41 to 14 to 2 and to 0%.Changes in the loading of reactive sites are expected to have a significant but less dramatic effect. The above reaction system is relatively simple in that the supported residues come together, react and then separate. No new crosslinks are formed. Some systems are more complex and site-site interactions introduce new crosslinks which would be expected to reduce the ability of the beads to swell in the reaction solvent. An example of such a system involves acetal/ ketal formation with the polymer-supported diol28.16 When 1% crosslinked polystyrene beads containing 3.15 mmol 8-1 residues 28 were reacted with terephthaldehyde, some mole- cules of the latter bound at just one aldehyde group, some bound at both.To determine the proportions, the free aldehyde groups in the ‘singly bound’ molecules were fully reduced (as monitored by FT-IR spectroscopy) with sodium borohydride, then the bound molecules were released and the mixture of recovered aldehydes analysed. Molecules which had been ‘singly bound’ were now present as 4-hydroxymethylbenzalde-422 Chemical Society Reviews, 1997, volume 26 hyde. Those that had been ‘doubly bound’ were recovered unchanged, i.e. as terephthaldehyde. It was found that approx- mately equal amounts of terephthaldehyde were ‘singly bound’ and ‘doubly bound’.More extensive studies of this same type were carried out using 5a-androstane-3,17-dione29. This I S-CH2 I H CH OH I CH20H 29 28 substrate is almost rigid and the 3-keto group is significantly more reactive to ketal formation than the 17-keto group. These features should significantly favour ‘single binding’ (via the 3-position). Indeed, when the dione 29 was bound to various resins the only ketone band in the FT-IR spectrum was that at 1745 cm-l due to a ketone group in a five-membered ring, i.e. the 17-ketone. After reduction of these keto groups, detachment of the steroids from the support, and analysis the proportions of ‘single binding’ and ‘double binding’ could be estimated. Several factors were investigated which were expected to favour greater ‘single binding’ than was achieved with terephthaldehyde. These were (i) the use of larger excesses of diketone; (ii) the use of polymer with lower loadings of diol groups per g; (iii) the use of a 20% crosslinked macroporous polymer; and (iv) the use of a shorter attachment time in the expectation that the 3-keto groups would react more rapidly. However, in no case was useful ‘single binding’ achieved; the ‘double binding’ in all cases was in the range 2348%.It is evident from the two examples discussed above, and from numerous other reaction systems studied in recent years,29 that site-site interactions are possible between a large fraction of the reactive sites on both lightly crosslinked polystyrene beads and macroporous polystyrene beads.This is not surpris- ing as with a linear polymer in solution there is no reason why, given time, all sites should not encounter others. Whilst crosslinking will reduce mobility it will need to be very extensive to achieve permanent site isolation. Although macro- porous beads have high crosslinking, the ability to reduce site- site interactions is significantly offset because, as noted above, if prepared by chemical modifications the more accessible reactive sites tend to be concentrated in the pores. It should also be noted that in many cases the supported substrate molecule will itself serve as a ‘spacer group’ and facilitate site-site interactions. This could completely negate any reduction in site-site interactions the polymer itself may achieve.In the first of the two examples discussed above, in a competitive situation almost all the phosphine residues 23 in the 15 and 37% crosslinked supports reacted with the carbon tetrahalides in solution (the analogue of Reaction ‘A’ in Scheme 3) rather than with the other phosphorus-containing residues (Reaction ‘C’). It is very likely that with highly crosslinked supports a degree of permanent site isolation is achieved. Support for this view comes from some early work on polymer- supported catalysts by Grubbs et al.31J2 They found that a catalyst prepared from 20% crosslinked macroporous polysty- rene beads and containing 1.0 mmol g-1 of PS ‘titanocene’ residues 30, prepared as outlined in Scheme 5, was 25-120 times more active for the hydrogenation of the hex-l-ene in hexane than the corresponding soluble catalyst.The higher activity of the supported catalyst was attributed to permanent site isolation which allowed a catalytically significant amount of an active monomeric titanium species, presumably residue 30, to survive. Only a dimeric titanium species was found in CI J I ,,GI CTi i,, (iii) 30 Scheme 5 Synthesis of a polymer-supported titanocene catalyst. (i) Na+CsHs-; (ii) CpTiC13; (iii) BunLi. solution. The variation in catalytic activity as a function of the loading of the beads reached a maximum at 0.14 mmol of Ti per gram.33 These results suggest that permanent site isolation is only achieved at very low loadings on highly crosslinked supports and will, therefore, only be useful in the context of polymer-supported catalysts.If it is desired to allow low-molecular mass species in solution reacting with a polymer-supported site to compete effectively with reactions between supported sites then one way to achieve this would be to use modestly loaded lightly crosslinked polymer supports which swell extensively in the reaction solvent, so reducing the concentration of the supported sites and making it relatively easy for the species in solution to diffuse into the beads, and where the polarities of the reactive species in solution, the reaction solvent and the micro-environment in the beads encourage the low-molecular mass reactive species to concentrate in the beads.A simple novel polymer-supported system where significant site isolation is probably achieved relatively easily and which results in a major change in the stereochemical course of a reaction comes from work of the author’s gro~p.3~This involves the reductions of 6-keto 5a-steroids, such as 3~-hydroxy-5a-cholestan-6-one31, whilst adsorbed to the inner surfaces of Amberlite XAD-4, a type of macroporous crosslinked polystyrene beads which have internal surfaces of ca. 750 m2 g-l. The reduction was achieved by aqueous potassium borohydride with the assistance of phase transfer catalysts. Under these reaction conditions, in the presence of the aqueous medium, the lipophilic steroid remained adsorbed to the polystyrene support.At the end of the reaction period the steroidal products were easily recovered by washing the beads with an appropriate organic solvent. Reductions of 6-ketoste- roids with borohydride in solution or with the steroid as a suspension in aqueous borohydride generally afford ca. 80% of the 6P-alcoho1, which is axial, and ca. 20% of the 6a-alcohol, which is equatorial. This is in contrast to most other borohy- dride reductions of cyclic ketones which give mainly the equatorial alcohol (‘product development control’). The differ- ence arises because to give the 6a-alcohol the incoming reagent has to pass the 19-methyl group and for steric reasons this is difficult. The reagent, therefore, approaches from the less hindered a-face and the product is now mainly the axial 6fl-alcohol (‘steric approach control’).In the reductions of the 6-ketosteroids adsorbed on the Amberlite XAD-4 the 6a-alcohol was found to be the main product and it formed up to 90% of the alcohol fraction. This reversal of the normal ratio was attributed to the difficulty of one adsorbed species reacting with another. Thus, in solution the initial reduction is brought about by -BH4; subsequent reductions are brought about by the various alkoxyborohydrides produced: see Scheme 3. The latter are both more reactive and more sterically demanding. In the supported system the reductions with alkoxyborohydrides involve two adsorbed steroid molecules reacting together and they can only do so with difficulty, if at all.As a consequence, in the supported system most of the reduction is brought about by -BH4 itself. This reductant is not, apparently, subject to ‘steric approach control’ and the product formed is that expected from ‘product development control’. It should be noted that this type of reaction system involves the simplest possible way of attaching and detaching substrates to a polymer support and it could find applications with other reactions. Finally in this section and still on the subject of the stereochemical course of polymer-supported reactions, Daunis et al. have achieved substantial asymmetric synthesis in a supported reaction system that depends totally on helpful site- site interactions.35 Recognising that many enolate anion reactions systems involve aggregates including solvent, Daunis et al.prepared a series of crosslinked polyacrylates in which ca. 1.0 mmol g-I of residues 32 were surrounded on average by three or four chiral pendant groups. Best results were obtained when the latter were residues 33 derived from (S)-prolinol. The residues 32 were condensed with glycine tert-butyl ester then the enolate 34 was generated by treatment with lithium di- +CH-CH+ I c OHH/ +0 32 33 +CH-CH+ I 34 isopropylamide in tetrahydrofuran. Due to the chiral residues 33 the prochiral lithium enolate 34 was present in a chiral environment. The enolate was alkylated at 20 “C with methyl Chemical Society Reviews, 1997, volume 26 423 iodide and the product cleaved off the support by treatment with hydrochlonc acid.Conversion of the alanine hydrochlonde into free alanine gave the latter in 85% overall yield with an 82% ee of the (S)-enantiomer A similar experiment where the alkylat- ing agent was isopropyl iodide gave valine in 84% yield with an 84% ee of the (S)-enantiomer. The recovered polymer could be re-used successfully A minor but useful point to note in these experiments is that since polyacrylates have a more flexible backbone than polystyrenes, to achieve a supported reactant with good physical properties 10% of the crosslinking agent N,N’-dimethylethylenebisacrylamidewas needed 5 Conclusions In this article some of the various effects that can occur when reactions are carried out using polymer-supported species have been considered Some of the effects result from the need for the low-molecular mass reactants to gain access to the supported reactive sites, other effects can result if the microenvironment in the beads differs from that in solution, and others can result from the presence or absence of site-site interactions Examples have been quoted which show that compared to the analogous reaction systems in solution, polymer-supported reactions can show substrate selectivity, be slower or faster, follow a different course or give a different stereochemical result What is absolutely clear is that it is unwise to aqsume that a polymer- supported reaction proceeds in just the same way as the analogous reaction in solution 6 Acknowledgement The author thanks Monbusho for a Visiting Professorship at Tokyo Institute of Technology during which time this article was written 7 References 1 R B Merrifield, J Am Chem Soc, 1963, 85, 2149 2 R B Merrifield, J M Stewart and N Jemberg, Anal Chem , 1966,38, 3 B Gutte and R B Memfield, J Bid Chem , 1971, 246, 1922 4 R L Letsinger and M J Kornet, J Am Chem SOC , 1963,85, 3045 5 R L Letsinger and V Mahadevan, J Am Chem SOC , 1965,87,3526 and 1966,88,53 19 6 See.for example, N K Terrett, M Gardner, D W Gordon, R J Kobylecki and J Steele, Tetrahedron, 1995, 51, 8135 7 G Lowe, Chem Soc Rev, 1996, 25, 309 8 W Dumont, J C Poulin, T P Dang and H B Kagan, J Am Chem SOC, 1973,95, 8295 9 N Takaishi, H Imai, C A Bertello and J K Stille, J Am Chem Soc, 1976,98 5400 and 1978,100,264 10 R H Grubbs and L C Kroll, J Am Chem Soc , 1971.93, 3062 11 C R Hamson, P Hodge, B J Hunt, E Khoshdel and G Richardson, J Orp Chem , 1983, 48, 3721 12 C R Hamson and P Hodge, J Chem Soc Perkrn Trans I, 1982, 509 13 D W Emerson, R R Emerson, S C Joshi, E M Sorensen and J E Turek, J Org Chem, 1979,44,4634 14 T Takagi, J Appl Pol Sci , 1975,19, 1649 15 J M J Frechet, M D de Smet and M J Farrall, J Org Chem , 1979, 44, 1774 16 P Hodge and J Waterhouse, J Chem Soc Perkin Tians I, 1983, 2319 17 A Guyot, P Hodge, D C Sherrington and H Widdecke, React Polym , 1992,16, 233 18 M Kempe and G Barany, J Am Chem Soc , 1996,118, 7083 19 J M Brown and J A Ramsden, Chem Commun , 1996, 2117 20 E Bayer, Angew Chem , 1991,103, 117 21 E Atherton, R C Sheppard and A J Rosevar, J Chem SOL Chem Commun , 1981, 1151 22 P W Small and D C Sherrington, J Chem SOC Chem Commun, 1989, 1589 23 S Itsuno, Y Sakurai, K Ito, T Maruyama, S Makahama and J M J Frechet, J Org Chem , 1990, 55, 304 24 C Yaroslavsky, A Patchornik and E Katchalski, Tetrahedron Lett, 1970, 3629 25 S Kondo, S Kawasoe, M Ohira, T Atarashi, K Ikeda, H Kunrsada and Y Yuki, Macromol Rapid Commun , 1995, 16, 291 26 S D Alexandratos and D H J Miller, Maciomolecules, 1996, 29, 8025 27 S Mohanraj and W T Ford, Macromolecules, 1986, 19, 2470 28 J C Bnggs, P Hodge and Z -P Zhang, React Polym, 1993, 19, 73 29 For an excellent review of the early work see W T Ford, in Polymeric Reagents and Catalysts, ed W T Ford, ACS Symp Ser No 308, Washington DC, 1986, ch 11 30 P Hodge and E Khoshdel, React Polym , 1985,3, 143 31 R H Grubbs, G Gibbons, L C Kroll, W D Bonds and C H Brubaker, J Am Chem Soc, 1973,95, 2373 32 W D Bonds, C H Brubaker, E S Chandrasekaran, C Gibbons, R H Grubbs and L C Kroll, J Am Chem Soc , 1975, 97, 2128 33 R H Grubbs, C P Lau,R CukierandC H Brubaker, J Am Chem Soc , 1977,99,4517 34 J C Briggs, P Hodge and Z -P Zhang, Tetrahedron, 1997, 53, 3943 35 M Calmes, J Daunis, H Ismaih, R Jacquier, J Koudou, G Nkusi and A Zouanate, Tetrahedron, 1990, 46, 6021 Received, 27th June 1997 Accepted, 5th August 1997 424 Chemical Society Reviews, 1997, volume 26

ISSN:0306-0012    

DOI:10.1039/CS9972600417

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Molecular and chemical basis of prion-related diseases P.1’5 Sheila B. L. Ng and Andrew J. Doig* Department oj Biomolecular- Sciences, UMIST, PO Box 88, Munchester, UK M60 I QD E-muI: Andrew, Doig@ um1st. ac .uk Prion-related diseases include scrapie in sheep, bovine spongiform encephalopathy in cattle and Creutzfeldt-Jakob disease in humans. The infectious agent for these diseases surprisingly contains no nucleic acid, but is a protein (PrP) which exists in two conformations, PrPC and PrPSC. The infectious PrPSC form has a higher P-sheet and lower a-helix content than PrF. The structures of PrP and models for how PrPSC is able to replicate by converting PrPC to PrPSc are discussed. 1 Introduction The prion-related diseases affect both humans and animals.They are also known as transmissible spongiform encephalo- pathies because they frequently cause the brain to become riddled with holes. Human prion diseases that have, to date, gained considerable recognition and concern include Creutz- feldt-Jakob disease (CJD), Kuru, Gerstmann-Straussler-Scheinker syndrome (GSS) and fatal familial insomnia (FFI). Scrapie in sheep, bovine spongiform encephalopathy (BSE) in cattle, transmissible mink encephalopathy, feline spongiform encephalopathy and chronic wasting disease of captive mule deer and elk are common prion diseases found in animals.’ All these diseases are characterised by vacuolation in the brain and usually result in ataxia (a condition in which the reflexes are diminished, and the absence of perfect coordination is shown through tremors and jerky movements), motor disturbance, dementia, and progression to a fatal outcome.The presence of abnormal fibrils (linear polymers) or insoluble plaques in the infected brain of all these diseases is another important characteristic and the probable cause (see below).2 While they may vary in their course and rapidity of development, they all cause brain neuron degeneration and ultimately, death. In the early 1980s, the name ‘prions’, an abbreviation for proteinaceous infectious particles,3-5 was proposed for the infectious agents that cause these devastating diseases. The suggestion that these causative agents consist solely of protein breaks the Central Dogma of Molecular Biology which implies that the conveyers of transmissible diseases require genetic Sheila Ng studied Biotechnology at Ngee Ann Polytechnic (Singa- pore) for three years before join- ing Genelabs Diagnostics Pte Ltd as a Laboratory Technologist in 1992.She is currently a final year undergraduate reading biochem- istry at UMIST. Sheila Ng material that is composed of nucleic acid (DNA or RNA) for the establishment of an infection in a host. Prions are defined as ‘novel infectious pathogens distinct from bacteria, fungi, parasites, viroids, and viruses, with respect to both their structure and the diseases they cause’.’ Table 1 gives the properties of the major prion-related diseases of which the best known is bovine spongiform encephalopathy (BSE), also known as ‘Mad Cow Disease’.This is a progressive, lethal central nervous system disease of cattle, characterised by the appearance in neurons in the brain of affected cattle of vacuoles (clear holes), that give the brain the appearance of a sponge (hence the term, spongiform). The condition was first identified in cattle in the UK in 1986 and epidemiology led to the conclusion that the bovine agent had originated from the scrapie agent. The source of the epidemic was soon traced to a food supplement that included sheep and cattle offal and carcasses. The BSE epidemic in cattle in Britain reached a peak in 1993 when more than 1000 cases per week were being reported. More than 160000 infected cows have been identified to date, involving more than 50% of the dairy herds in the UK, costing about 2 billion pounds sterling in compensation to farmers.The British government banned the use of animal-derived feed supplements in 1988, but it was not until 1991-1992 that the ban was strictly enforced, thus delaying the eradication of the disease. BSE has now been shown to have been transmitted to a number of other species, including mice, domestic cats, pigs and macaques. The epidemic of BSE in the UK and other countries raises the possibility of a significant threat to public health through the consumption of BSE-infected tissues. 2 Identification and properties of prion protein 2.1 Experiments carried out to determine the composition of the disease-causing agent In search for the cause of these novel illnesses, purification of the infectious material from scrapie-infected brains was an essential first step of investigation.The scrapie agent was first purified from infected Syrian golden hamster brains.6 The brain ~~~~~ ~~~~ indrew Doig read Natural Sci- we and obtained his Chemistry ghD with Professor Dudley Wil- iams at the University of Cum- >ridge.He was a NATO Post- l‘octoral Fellow with Professor tobel-t Baldwin in the Depart- nent of Biochemistry, Stanford Yniversity, before becoming a decturer in the Department of 3iomolecular Sciences, UMIST, n 1994. Andrew Doig Chemical Society Reviews, 1997, volume 26 425 Table 1The pnon diseases Disease Organism Typical symptoms Route of acquisition Distribution Span of overt illness Scrapie Sheep Loss of coordination and weight, followed (1) Infection of pasture with placental tissue Found in many parts of the world, possibly Three to four months by irritability, some develop an intense itch, leading them to scrape off their wool scrapie agent followed by ingestion (2) Genetic disorder started in Spain Incidence is related to the breed of sheep encephalopath y Bovine spongiform cow and weight, followed by apprehensiveness, Loss of coordination Meat and bone meal containing sheep and cow carcasses Seen in adult cattle of either sex More than 160000 cases have 20 months to 18 years agitation and and offals been identified muscle contractions Kuru Human Loss of coordination, Infection (probably Known only in highlands Three months to one Creutzfeldt-Jakob disease Human often followed by dementia Dementia, followed by loss of coordination, although sometimes through cannibalism, stopped by 1956) Spontaneous conversion of PrP (in ‘sporadic’ disease) Sometimes of Papua New Guinea Sporadic form 1 person per million worldwide Inhented form some Year Typically about one year, range is one month to more than the sequence is reversed inheritance of a mutation in the gene 100 extended families have been identified 10 years coding for the pnon Infectious form about protein (PrP) Rarely, 80 cases have been via an inadvertent identified New consequence of a variant CJD Around medical procedure 20 individuals in UK (iatrogenic) have been identified Possibly (ipon vCJD), (2 cases in France, through dietary exposure to bovine 1 case in Italy, 1 case in Germany) offals Gerstmann-S traussler- Human Loss of coordination, Inhentance of a Some 50 extended Typically two to six Scheinker disease often followed by mutation in the PrP families have been years dementia gene identified Fatal familial insomnia Human Trouble sleeping and disturbance of nervous mutation in the PrP Inhentance of a Nine extended families have been identified Typically about one year system, followed by gene insomnia and dementia extracts of the infected hamsters were subjected to a range of reticulum while 23 amino acids are removed from the tests, designed to reveal the nature of the disease-causing C-terminus on addition of glycosylphosphatidylinositol (GPI).component. Surprisingly, procedures which modify nucleic PrP is attached to cell surface membranes by its GPI. Most of acids did not lead to a decrease in scrapie infectivity.7 In the PrP molecules carry bi-, tri-, and quater-nary neutral and contrast, the scrapie agent was inactivated by protein dena- asparagine-linked oligosaccharides at two sites. A disulfide turants, indicating that an essential component of the infectious bond links the only two cysteine residues in the mature prion agent was protein because experiments involved procedures protein. that denature (unfold) or degrade proteins reduced scrapie infectivity.Nevertheless, it is still very resistant to degradation 2.4 Detection of plaques in some prion diseases for a protein. The protease resistant fragments of PrPSc, PrP 27-30, accumu-late as insoluble deposits of protein, or amyloid plaques, in the 2.2 Identification of prion protein isoforms brains of some patients, implying that amyloid formation is Researchers identified a protease-resistant protein of about 27 to essential for the formation of PrPSc.lO,ll Similar plaques are also 30 kDa in fractions purified from Syrian hamster brains, seen in Alzheimer’s disease, although Alzheimer’s amyloid designated PrP 27-30. N-Terminal amino acid sequencing of consists of a different protein. PrP 27-30 allowed the selection of a clone from a hamster brain chromosomal DNA library and the identification of the PrP 2.5 Normal function of PrP gene.8 The PrP gene was found to be encoded within a single Mice devoid of PrPC can develop and reproduce normally, exon as a protein of 254 amino acids.9 The same PrP gene was though abnormalities in the action of the neurotransmitter also found in other mammals, including humans.Knowledge of GABA (y-ammobutyric acid) suggest a role for PrP in nerve the gene sequence led to the identification of the normal PrP signalling.12 PrP may also be involved in the regulation of sleep, gene product, a protein of 33 to 35 kDa, designated PrPC. PrF as well as circadian activity rhythms.13 is protease-sensitive and the cellular isoform of the prion protein PrP (hence the superscnpt C), whereas the abnormal, 2.6 The prion hypotheses, virino vs.protein infectious, scrapie-form, designated PrPSc, is protease-resistant. We have come to expect that all life forms (from viruses to PrPSc is now used to refer to the protein molecules that bacteria to plants to humans) hand down the blueprints to all constitute the prions causing all scrapie-like diseases (hence the their progeny via their DNA (or RNA for some viruses). This superscript Sc) of animals and humans. concept, however, is not adopted by the various scrapie-related prion diseases discussed here. Two different views have 2.3 Primary structure of prion protein dominated this field of research. The first view argues that the A signal peptide of 22 amino acids at the NH2-terminus is infectious agent consists of, or contains, an informational cleaved from full-length PrPC in the rough endoplasmic molecule (independent of the affected animal) that is capable of 426 Chemical Society Reviews, 1997, volume 26 Table 2 Properties of the prion protein, PrP Properties PrPC PrPSc Primary sequence Same Same Molar mass/kDa 33-35 33-35 Position where Chromosomal gene Chromosomal gene instructions specifying amino acid sequence reside Proteolytic resistance Destroyed Reduced to a completely by protease-resistant proteolytic enzymes core of about 27-30 kDa, termed PrP 27-30 Solubility Soluble in presence Forms insoluble of detergent aggregates Turnover rate Rapid Very slow Presence Found in all cells, Found only in the especially on brain and to a lesser suface of neurons extent, other tissues of animalshumans with prion disease Content levels Constant throughout Levels in brain rise disease during disease progression in cases of prion disease Aggregate Does not aggregate Aggregates have a characteristics in the form of fibrillar structure insoluble polymers with similar properties as amyloid plaques Infectivity Non-infectious Infectious mutation.The agent is given the name, ‘virino’, which is conceived of as a small, informational nucleic acid molecule susceptible to mutation which, with the host protein, PrPC, combines after infection. The other view proposes that the causative agent contains only protein and derives its informa- tion from the structure of that protein.It is suggested that the progression of disease is accompanied by the conversion of normal host-encoded PrPC to the abnormal form, PrPSC. According to this hypothesis, when this abnormal protein is transferred to a new host, it initiates a cascade and catalyses the modification of the new host’s other PrPC so that they in turn, become abnormal. Hence PrPSc is infectious. 3 Isoforms Despite the uncertainty in the relationship between PrPSc and the causative agent(s) of the transmissible spongiform ence- phalopathies, one aspect appears quite clear: the main differ- ence between normal PrPC and scrapie PrPSc is confor- mational. 3.1 Absence of a chemical modification that might distinguish PrPSC from PrPC To date, PrPSc is the only component of the infectious prion protein identified.A post-translational process was suggested to mediate the conversion of the normal, cellular PrP isoform PrPC or a precursor into abnormal PrPSc form by two findings. Throughout the course of scrapie infection, investigators found only one PrP mRNA in normal and scrapie-infected brains, and its level remained ~0nstant.l~ Furthermore, the entire PrP gene was found within a single exon, hence eliminating the possibility of alternatively spliced species of PrP during mRNA processing. PrPSc often has the same amino acid sequence as PrPC, as shown by mass spectrometry and Edman degradation sequencing.Mass spectrometry showed no post-translational chemical modifications other than the GPI attached to the C-terminus and two Asn-linked oligosaccharides already known to occur on both PrPSc and PrPC, indicating that conformation alone distinguishes the PrP isoforms. Indeed, a substantial conformational difference was observed using FTIR spectroscopy and circular dichroism spectroscopy. PrPC con- sists primarily of a-helices, while PrPSc contains 6-sheet.15 3.2 Molecular modelling of PrP suggested a four-helix bundle protein Molecular modelling was used to predict the structure of the normal protein based solely on its amino acid sequence. Secondary structure prediction algorithms identified four poten- tial helical regions, suggesting that PrPC might be a four-helix- bundle protein (Fig.The four putative a-helical regions, Fig. 1 Proposed models of PrPc (a) and PrPsc (b).I6 Residues implicated in the species barrier3’ (Asnl08, Metll2, Met129 and Ala133) are shown as balls and sticks. (Reproduced with permission from ref. 16). Chemical Society Reviews, 1997, volume 26 427 H1-H4, were found to reside within residues 90-231 of the protease-resistant core of PrP 27-30. 3.3 Synthetic peptides corresponding to three of the four putative a-helical regions of PrP can fold into P-sheets The four putative helices were synthesised as peptides of 13-1 7 amino acids and their secondary structures found by FTIR spectroscopy. When the peptides were dried from a helix- promoting solvent such as 1,l,1,3,3,3-hexafluoropropan-2-o1, they had spectra indicative of a-helices with amide I bands showing maxima at 1650-1660 cm-I.However, on addition of aqueous buffers, only the spectrum of H2 remained unchanged while the other three a-helical regions immediately adopted CJ-sheet conformations. 17 Furthermore, the three peptides that folded into CJ-sheet conformation were individually observed to precipitate slowly from aqueous solution as amyloid fibrils. 3.4 P-Sheet structures aggregate into amyloid fibrils enhanced by mutations Some synthetic PrP peptides were found to have a high (3-sheet content and be capable of amyloid formation. Mutations in these regions enhanced their ability to aggregate. PrP molecules that arise from mutated genes probably do not adopt the abnormal, scrapie conformation as soon as they are synthesised, or people with mutant genes would become ill in their early childhood or before birth.It was therefore suspected that mutations in the PrP gene made the proteins produced from these mutations more susceptible to flipping from an a-helical to a (3-sheet shape. 3.5 Peptides in their P-sheet form can induce a-helix to @-sheet conformational transitions Further evidence supporting the proposition that PrPSc can induce the switching of an a-helical PrP molecule to a (3-sheet form came from studying the conformational transition with synthetic peptides.18 It was found that the highly conserved hydrophobic peptide H 1 (residues 109-1 22) was capable of inducing a CJ-sheet structure in other helical PrP peptides.These include peptide H2 (residues 129-141) and the longer, more hydrophilic peptides containing the H 1 sequence such as 104H 1 (residues 104-122), both of which have a coil or a-helical structure in solution. 3.6 In vitro conversion of PrPC into PrPSc By identifying conditions which allowed the reversible denatu- ration of PrPSc, a cell-free system was set up to test if de nuvu conversion of PrPC to PrPSc could be observed in vitro.19 Treatment of PrPSc preparations with a high concentration (6 M) of the denaturant GuHCl irreversibly destroyed its ability to induce the conformational conversion and its resistance to proteinase K. On the other hand, the converting activity and protease-resistance of PrPSC could be restored by diluting the denaturant to concentrations less than 3 M.This incomplete denaturation suggested that the converting activity and recovery of the characteristic protease-resistance of PrPSc requires the maintenance of some native PrPSc structure and that irreversible denaturation can only be achieved by exceeding the threshold denaturant concentration at which the native structure was completely destroyed. 35s-Labelled PrPC from an uninfected source was mixed with PrPSc in order to evaluate whether the conversion of PrPC to a proteinase K-resistant form could occur during PrPsc renatura- tion. After two days incubation, proteinase K digestion was shown to eliminate the original full-length 35S-PrPC and produce 35s-labelled proteinase K-resistant PrP.This observa- tion established that PrPC can be converted selectively to proteinase K-resistant forms similar to PrPSc in a cell-free system. 3.7 Presence of a P-sheet in an NMR structure of mouse PrP While the model predictions of the structure PrPc suggested that it was comprised entirely of a-helices, l6 another group of investigators found a CJ-sheet in an NMR structure of an autonomously folding mouse prion protein domain comprising residues 121-231, PrP( 121-231) (Fig. 2).20Attempts to express PrP( 108-23 1) in the periplasm of E. coli resulted in proteolytic cleavage after residues 112, 118 and 120. PrP( 121-231) was found to contain three a-helices and a two-stranded antiparallel (3-sheet.It is possible that the short (3-sheet in PrP(121-231) serves as a nucleation site for the conformational transition from a-helix to CJ-sheet seen in the PrPC to PrPSC conversion. No folds similar to PrP(121-231) could be identified in the Brookhaven Protein Data Bank. The orientation of the three helices in PrP( 121-231) (Fig. 2) was distinctly different from the proposed four helix bundle model of PrP( 109-2 17) [Fig. 1 (a)]. Invariant residues in PrP sequences are found within the hydrophobic core and the two glycosylation sites are located on the protein surface. All six residues for which mutation is believed to be associated with inherited prion disease are located in, or near, regular structural elements. The residues that vary with familial prion diseases are all solvent accessible. Fig.2 NMR structure of PrP( 121-231) (Reproduced with permission from ref. 20.) 4 Diversity A difficulty exists in the ‘protein-only’ hypothesis for the infectious agent of prion diseases: the existence of multiple prion strains with markedly different, and apparently inherit- able, characteristics. These various strains can be distinguished on the basis of species response, incubation period, clinical disease, neuropathological manifestations and PrPSC distribu- tion in brain tissue.2’ About twenty phenotypically different strains of scrapie and BSE have been isolated in mice. This discovery of ‘strain variation’ therefore poses an interesting challenge to the ‘protein-only’ hypothesis as to date only pathogens containing nucleic acids are known to occur in multiple strains.4.1 Each strain could be associated with a different PrP conformer A possible explanation for prion diversity is that prions can adopt multiple conformations. A prion might be able to convert normal PrP into the infectious form highly efficiently when folded in a particular conformation, but when folded another way, its efficiency of conversion might be decreased. Similarly, one ‘conformer’ could be attracted to the neurons in one specific 428 Chemical Society Reviews, 1997, volume 26 part of the brain, whereas another conformer resides in a different site, thus exhibiting different symptoms.PrPc and PrPSc have been shown to adopt at least two different structures but the exact number of conformations a PrP can adopt is unknown. 4.2 New variant CJD+vidence for BSE transmission to humans New variant Creutzfeldt-Jakob disease is a form of human prion disease recently reported in the United Kingdom, affecting, unusually, young people and having a highly consistent and unique pathological pattern.22 These patients were homozygotes (for methionine) at residue 129 of PrP. This important finding could be an indication of the arrival of a new risk factor for CJD in the United Kingdom, possibly through dietary exposure to bovine offals. In order to differentiate new variant CJD from the other three forms (iatrogenic, sporadic and acquired, see Table l), molecular analyses involving mainly the use of Western blots, proteinase K denaturation and transgenic technology were performed.22 Sporadic and iatrogenic CJD were demonstrated to be associated with three distinct patterns of human PrPSc on Western blots after proteolytic cleavage by their differing band sizes.Types 1 and 2 were related to different phenotypes of sporadic CJD. A third type was seen in some CJD cases, in which the route of exposure to prions was through intra- muscular injection of human hormones. On the other hand, iatrogenic CJD resulting from CNS exposure was found to typically resemble classical sporadic CJD. Despite showing similarity to the PrPSc band sizes of Type 3 CJD, the new variant CJD could be distinguished from the other three types of CJD by a characteristic pattern of band intensities.This unique molecular marker has clearly separated new variant CJD from sporadic CJD and confirmed the proposal that new variant CJD exists as a distinct and new subtype of prion disease. A glycoform pattern similar to that of new variant CJD was observed in BSE itself, experimental mice, naturally transmitted BSE in domestic cats and experimental BSE in macaques. These results strongly support the hypothesis that new variant CJD has resulted from the transmission of BSE to humans. The cause of sporadic CJD remains unclear but may involve spontaneous conversion of PrPC to PrPSc as a rare random event. Prion strain variation therefore involves post-translational modifications of PrP, which persist or can be converted between isolated strains (in the case when PrP genotypes are mis-matched).This is consistent with the model of prion propaga- tion whereby strain variation results in post-translational modification of PrP (such as a conformational change) with no nucleic acid involved. 4.3 Barriers to prion infection between species The phenomenon known as the ‘species barrier’ has been observed in prion research since the mid-1980s. It is the observation that it is difficult for prions made by one species to cause disease in animals of another species. For instance, difficulties were observed in the transmission of scrapie from one animal species to another as well as in transmission of human diseases, such as Kuru and CJD to various animal species, including non-human primates.This could be partly due to the reduced efficiency of interactions between endoge- nous PrPC and exogenous PrPSc which differ in their amino acid sequence. One notable example of the ‘species barrier’ is the BSE epidemic or ‘Mad Cow Disease’ in Great Britain in early 1986, an indication of the first transmission of a prion to a new species showing lower infectivity and longer incubation time than subsequent passages in that species. Furthermore, the transmission of BSE to the marmoset and macaques has since raised the possibility of BSE transmission to humans. Worry- ingly, bovine PrP has recently been noted to be more homologous with human PrP than is sheep PrP in the region between codons 96 and 167.However, it is still unknown whether the differences in the amino acid sequence between bovine and sheep PrP in this central domain could account for the apparent different susceptibility of humans to bovine and sheep prions. 5 Replication The structural findings that demonstrated the conformational difference between PrPC and PrPSc have also led to suggestions of two plausible models for prion replication which could account for the pathogenesis of infectious, inherited, and sporadic forms of prion disease. Any such conformational model requires that one amino acid sequence can code for at least two conformations, depending on its state of complex and surrounding environment.The proposed models raised an important issue in the PrPC to PrPSc conversion as to whether the presence of PrPSc polymers/multimers is required or whether the formation of PrPSc aggregate is solely a conse- quence of the overproduction of PrPSc. 5.1 Model 1: a catalytic or template-assisted model (Fig. 3) This conformational model23 proposes that random fluctuations in the structure of PrPC result in the reversible generation of a partially unfolded monomer, an intermediate designated PrP*, involved in the formation of PrPSc. The concentration of PrP* would normally be low and PrPSc formed in insignificant amounts. Catalysis PrPC-P~P*-PrP*-PrPsc -PrPSC Fig. 3 Catalytic model for prion replication Exogenous prions containing PrPSc act as templates to induce the conversion of PrP* into PrPSc in the infectious form of prion disease.However, the insolubility of PrPSc renders this process irreversible and so the formation of PrP* followed by PrPSc promotes increasing concentration of PrPSc. Mutations in PrP could cause destabilisation of PrPC, inducing its conversion into the intermediate PrP* state resulting in formation of PrPSc. In the case of sporadic prion diseases, the production of PrPSc could result from sufficient accumulation of PrP* under rare situations. Indeed, the development of spongiform degeneration is observed in transgenic mice overexpressing the wild-type PrP gene. Somatic mutations could also play a part in this form of disease by destabilising PrPC and promoting its conversion into PrPSc through PrP*. For instance, both Met and Val are commonly found at amino acid 129 in human PrP.Homo- zygosity (Met-Met) at codon 129 was found to predispose to sporadic CJD in these studies. Further experiments showed that those patients carrying the Asp I78 to Asn mutation would have insomnia if the codon 129 specifies Met in the mutant allele; if the codon 129 specifies Val, they would have dementia. 5.2 Evidence for structural differences between the PrPSc structure involved in prion infectivity and amyloid formation Organic solvents that perturb protein conformation were employed in the study of the structure of PrP amy10id.~~ 1,1,1,3,3,3-Hexafluoropropan-2-o1(HFIP), a solvent known to promote a-helix formation, was observed to modify the structure of PrP amyloids from their original rod shape into flattened ribbons with a more regular substructure.Increasing the concentration of HFIP resulted in a steady reduction of (S-sheet content, proteinase K resistance of PrP27-30 and prion infectivity. Congo red dye green gold birefringence under polarised light is characteristic of amyloid. HFIP was found to have reversibly decreased the binding of Congo red to the rods while inactivating pnon infectivity in an irreversible manner. A Chemical Society Reviews, 1997, volume 26 429 structurally related solvent, 1 ,1, 1 -trifluoropropan-2-01, how- ever, did not inactivate prion infectivity, but was shown to exhibit similar effects as HFIP by altering the morphology of the rods and abolishing Congo red binding.These results obtained from the use of electron microscopy and differential Congo red binding have distinguished the specific (3-sheet-rich structures required for prion infectivity from those needed for amyloid formation hence indicating that amyloid formation is not essential for prion infectivity, in conflict with the nuclea- tion-dependent polymerisation model described below. 5.3 Model 2: a nucleation-dependent polymerisation model (Fig. 4) The slow onset of neurodegeneration is a characteristic of both the human prion diseases and Alzheimer's disease." The brain pathology of these diseases shows similarity by the appearance of aggregated peptides, often in the form of amyloid plaques.Amyloid exists as ordered non-crystalline polymers and can be defined as a one-dimensional crystal in which the inter-molecular packing in the plane perpendicular to the direction of fibril growth is non-uniform. Amyloid is capable of forming different types of insoluble, amorphous aggregates which can form rapidly when the protein concentration exceeds the solubility. There is a kinetic barrier to amyloid formation caused by a lag in forming an amyloid nucleus which can subsequently propagate. This rate-determining step was proposed to be mechanistically relevant to that accelerated by the infective agent of scrapie during amyloid formation. This led to the hypothesis of a possible replication mechanism for the conver- sion of PrPC into PrPSc which assumes that PrPSc is an aggregate in which an alternative conformer of PrP is stabilised by intermolecular interactions.I In this model, the initial slow reversible formation of a nucleating PrPsc multimer leads to a seeding process, during which PrP monomers are added to the growing polymer in an abnormal PrP conformation.*S The growing polymers can break apart, generating new seeds. In the sporadic form of disease, the reversible nucleation process is normally slow; on the other hand, PrPSc acts as a ready-made nucleation seed in the infectious form of disease and inherited mutations might contribute to the increase in affinity of the abnormal conformation of PrP for the polymer seed.Both replication and infection involve the nucleation of poly-merisation according to this mechanism. Nucleation-dependent polymerisation is common among many well-characterised processes including protein cry stallisation, flagellum assembly, microtubule assembly, sickle-cell haemoglobin fibril formation and actin polymerisation. The process is similar to crystallisa- tion from a metastable supersaturated solution. PrPC- slow nucleus formation-- / fast aggregationL- nucleus , ,I I I, l 0 , l PrPSC Fig. 4Nucleation-dependent polymerisation model for prion replication' Three distinctive features are characteristic of a nucleation- dependent polymerisation mechanism. First, there is a time lag before the aggregates become detectable.Secondly, there is a critical concentration. After completion of polymerisation, the solution contains mainly monomers and high polymers at equilibrium. The monomer concentration at this point is referred to as the critical concentration, below which polymer- isation will not occur. Thirdly, a supersaturated solution can be seeded by a preformed nucleus, resulting in immediate polymerisation. In contrast, the growth of a linear polymer neither requires nucleation nor can be seeded. Linear polymer- isation is characterised by the accumulation of intermediates in a sequential manner: no time lag is observed, and supersaturated solutions are unstable and rapidly aggregate. 5.4 Evidence for nucleation-dependent in vitro amyloidformation PrP and the P-amyloid protein of Alzheimer's disease share a similar sequence that may be responsible for the initiation of protein aggregation and amyloid formation in vivo in both cases.26 Part of the PrP sequence resembles the amyloidogenic C-terminal portion of the (3-amyloid protein of Alzheimer's disease including amino acids 96-111 and is also highly conserved across species.Two synthetic peptides correspond- ing to PrP residues 96-111 (PrP96-lllM and PrP96-111V) were synthesised. These peptides were observed to be sparingly soluble and formed rigid, unbranched amyloid fibrils, as shown by electron microscopy (EM), FTIR, Congo Red staining, and X-ray diffraction.'* The existence of a kinetic barrier to amyloid formation was demonstrated by both peptides in thermody- namic solubility measurements, aggregation kinetics and seed- ing experiments.This suggested that the rate-determining step for aggregation is the formation of an ordered nucleus. Furthermore, seeding was shown to occur only with the amyloid fibrils of PrP96-1llM and PrP96-111V but not with a related peptide with a shuffled sequence. This effect was consistent with the proposal outlined above that the aggregation of PrP involves a nucleation-polymerisation event analogous to the seeding of a cry stallisation experiment. 5.5 Nucleation-dependent polymerisation mechanism accounts for predisposition of PrP homozygotes to CJD A nonpathogenic polymorphism is known to occur at amino acid 107 (codon 129) within PrP.This polymorphism involves the conservative substitution of valine for methionine and both homozygous genotypes are predisposed to sporadic and in- fectious CJD. For instance, 21 in 22 individuals with sporadic CJD were homozygous for methionine or valine at position 107, while the population in general is approximately 50% homo-zygou~.*~The chemical basis for this genetic effect has been investigated using peptide models of PrP. Peptides derived from the PrP 118-133 sequence (containing methionine or valine at position 129) were observed to form amyloid via a nucleation- dependent mechanism, lo suggesting that homogeneous peptide amyloid (Met 129 or Val1 29) is more stable than heterogeneous amyloid (Met 129 and Val 129).lo Further studies were carried out with the objective of modelling possible mechanistic differences in prion formation between position 129 homozygotes and heterozygotes.** This was achieved by studying amyloid fibril formation from supersaturated peptide solutions and comparing homogeneous solutions with heterogeneous mixtures (1 : 1 Met 129 :Val 129) at the same total peptide concentration. These experiments demonstrated that heterogeneous supersaturated solutions showed longer nucleation times than the homogeneous solu- tions, indicative of a less stable heterogeneous nucleus. In addition, this phenomenon suggested the formation of non- productive heterogeneous oligomers (hence the slower nuclea- tion rate) and that Met129 inhibits growth of Val129 fibrils (and vice versa).Seeding was observed when Met129 fibrils were added to supersaturated homogeneous solutions of Met 129 or Vall29. This was followed by a measurement of a comparable growth rate10 which indicated that seeding, unlike nucleus formation, is insensitive to the polymorphism. Seeding of the heterogeneous solution with Met129 fibrils also proved suc- cessful, but with a slower observed growth rate, thus providing further evidence for the proposal of mutual growth inhibition. In an attempt to compare their thermodynamic stability, the solubilisation rates of fibrils formed from homo- and hetero- geneous solutions were measured.28 The results showed that the fibrils from the heterogeneous solution dissolved at approx- imately twice the rate of the homogeneous fibrils and reached a higher solubility than the homogeneous solutions.This sug- 430 Chemical Society Reviews, 1997, volume 26 gested that the two peptides formed separate homogeneous fibrils in the heterogeneous solution. By influencing nucleation time and critical concentration, a conservative sequence polymorphism (Met vs. Val) can therefore exert a dramatic effect on amyloid formation. In summary, because the slower nucleation and higher solubility of the peptide mixtures are analogous to the higher PrP critical concentration for nucleus formation in a heterozygote, the PrP concentration in vivo would therefore be more likely to be below the critical concentration. In this way, heterozygotes would be protected against amyloid formation. 5.6 Nucleation-dependent polymerisation could account for virus-like features of prion The existence of prion strains has been difficult to explain by the protein-only hypothesis as strain variation has hitherto been explained by sequence variation in DNA or RNA.According to the nucleation-dependent polymerisation model, strain varia- tion could results from different growth faces in prion fibrils. The strain dependence of scrapie incubation time could be a result of imperfect complementarity between the interacting surface of the foreign seed and host PrP, leading to slower initial growth and delayed disease onset. After growth using host PrP, the face of the seed assumes the characteristics of the host, hence accounting for the observation that scrapie infectivity takes on the strain of the host.In this way, scrapie strains act as alternative conformations or packing arrangements of PrPsc polymers, analogous to the alternative crystal forms observed for many proteins. 6 Conclusions Intensive research on prion biology and diseases both currently and over the past three decades has involved a wide spectrum of disciplines including virology, neurology, neuropathology, molecular biology, cell biology and protein chemistry. These studies have taken on new urgency since the accumulation of evidence that a new prion strain has crossed the ‘species barrier’ to humans from BSE infected cattle.6.1 Nature While a large accumulation of evidence supports the involve- ment of prion or PrP in the pathogenesis of these transmissible spongiform encephalopathies, it is still not completely certain that there is no nucleic acid within the transmissible agent of prion-related diseases, though it appears unlikely. Of particular significance is the demonstration that PrPC could be converted into PrPSc under defined conditions in vim, a useful piece of evidence in proposition of the ‘protein-only’ hypothesis. Nevertheless, the cell-free production of amyloid fibrils from either precursor proteins or peptides observed in most amyloid diseases has failed to transmit disease by the amyloid fibrils themselves.2’ On the other hand, the existence of a nucleic acid component might be convincingly ruled out if the generation of a new form of prion-related disease infectivity could be produced in a cell-free system.6.2 Misfolding The prediction of protein folds from their amino acid sequences is an impressively long-standing challenge in molecular biology and biophysics, also commonly known as the Protein Folding Problem. Protein aggregation is a widespread phenomenon that can occur during protein folding in vivo as well as in vitro. The likelihood of intermediates within a folding pathway generating incorrectly folded species is dependent on several factors, including protein concentration, pH, temperature, ionic strength and redox environment. Kinetic competition which may result in formation of aggregates also exists between the correct and misfolded forms.Prion-related diseases can be characterised as ‘protein folding diseases’ involving a misfolded, abnormal form of the infectious agent, that is covalently indistinguishable but conformationally distinct from the normal form. The patho- genic mechanisms of some of these diseases are also known to involve aggregation. In prion diseases, considerable evidence has accumulated in support of the importance of PrP(90-145), and in particular PrP(109-122), in the ato P conversion during PrPSc formation. These regions have been implicated as the key regions in mediating the interaction of PrPC and PrPSc during the conformational transition as well as producing effects that are indicative of the prion disease species barrier.Furthermore, elucidation of the PrP( 121-23 1)NMR-derived tertiary structure provides additional structural details of the conversion process. While the exact molecular mechanism of PrPSc replication is still unclear, an existing challenge remains to relate the misfolding of the prion protein to the resulting cellular pathology such that the three different (inherited, infectious and sporadic) forms of prion diseases can be accounted for. 6.3 Analogues Some striking similarities with prion diseases exist in neu- rodegenerative disorders such as Alzheimer’s disease and Parkinson’s disease. In all these diseases, the more widespread ills mostly occur sporadically but also sometimes in a familial manner.All these disorders only appear during the middle to later stages of life and are marked by similar pathology: degeneration of neurons, aggregation of protein deposits as plaques, the growth of larger glial cells (which serve to support and nourish nerve cells) in response to neuronal damage in the brain and an absence of white blood cells in the brain. Therefore, ongoing research on prion diseases might provide a crucial relevance to other more common neurodegenerative disorders. 6.4 Therapies A thorough understanding of the molecular processes under- lying the conformational transition from PrPC to PrPSc is a precondition to devising rational approaches or therapeutics to the prevention and treatment of prion disease.These approaches could involve interfering with the PrPC to PrPSc structural change, stabilisation of the PrPC form or reduction of the PrP concentration. In a recent study, reagents which stabilise proteins in their native state have been found to interfere with the conversion process of PrP by reducing the rate and extent of PrPSc formation.30 The reagents appeared to have no effect on the existing population of PrPSc molecules in scrapie-infected mouse cells, but instead interfered with the formation of PrPSc from newly synthesised PrF. This observation therefore led to the suggestion that the action of the reagents was to stabilise the a-helical conformation of PrPC, hence inhibiting the conforma- tional conversion into the P-sheet form of PrPSc.The further observations that long-term glycerol treatment appeared to slow down or even prevent PrPSc formation could possibly reveal potential strategies for treatment of prion-related diseases. Identification of more potent stabilising reagents and develop- ment of novel ways of introducing them into the central nervous system are likely to result in effective therapies and treatments for prion disease in mammals. 7 Acknowledgements We thank the Medical Research Council (G9625094) for financial support and Simon Hubbard for critical review of the manuscript. 8 References 1 S. B. Prusiner, Science, 1991, 252, 1515. 2 P. A. Merz, R. A. Somerville, H. M. Wisniewski and K. Iqbal, Am Neuropathol. (Berlin), 198 1, 54, 63.3 T. Alper, W. A. Cramp, D. A. Haig and M. C. Clarke, Narure, 1967,214, 764. 4 J. S. Griffith, Nature, 1967, 215, 1043. 5 S. B. Prusiner, Science, 1982, 216, 136. Chemical Society Reviews, 1997, volume 26 431 6 S. B. Prusiner, D. F. Groth, S. P. Cochran, F. R. Masiarz, M. P. McKin- ley and H. M. Martinez, Biochemistry, 1980, 19, 4883. 7 S. B. Prusiner, Biochemistry, 1982, 21, 6942. 8 S. B. Prusiner, D. F. Groth, D. C. Bolton, S. B. Kent and L. E. Hood, Cell, 1984, 38, 127. 9 K. Basler, B. Oesch, M. Scott, D. Westaway, M. Walchli, D. F. Groth, M. P. McKinley, S. B. Prusiner and C. Weissmann, Cell, 1986, 46, 417. 10 J. H. Come, P. E. Fraser and P. T. Lansbury,Proc. Natl. Acad. Sci.USA, 1993,90,5959. 11 J.T. Jarrett and P. T. Lansbury, Cell, 1993, 73, 1055. 12 J. Collinge, M. A. Whittington, K. C. L. Sidle, C. J. Smith, M. S. Palmer, A. R. Clarke and J. G. R. Jeffreys, Nature, 1994, 370, 295. 13 I. Tobler, S. E. Gaus, T. Deboer, P. Acherrnann, M. Fischer, T. Rulicke, M. Moser, B. Oesch, P. A. McBride and J. C. Manson, Nature, 1996, 380, 639. 14 B. Oesch, D. Westaway, M. Walchli, M. P. McKinley, S. B. H. Kent, R. Aebersold, R. A. Barry, P. Tempst, D. B. Teplow, L. E. Hood, S. B. Prusiner and C. Weissmann, Cell, 1985, 40, 735. 15 K.-M. Pan, M. Baldwin, J. Nguyen, M. Gasset, A. Serban, D. Groth, I. Mehlhorn, Z. Huang, R. J. Fletterick, F. E. Cohen and S. B. Prusiner, Proc. Natl. Acad. Sci. USA, 1993, 90, 10962. 16 Z. Huang, J.-M. Gabriel, M.A. Baldwin, R. J. Fletterick, S. B. Prusiner and F. E. Cohen, Proc. Natl. Acad. Sci. USA, 1994, 91, 7139. 17 M. Gasset, M. A. Baldwin, D. Lloyd, J.-M. Gabriel, D. M. Holtzman, F. E. Cohen, R. Fletterick and S.B. Prusiner, Proc. Natl. Acad. Sci. USA, 1992,89, 10940. 18 J. Nguyen, M. A. Baldwin, F. E. Cohen and S. B. Prusiner, Biochemistry, 1995, 34, 4186. 19 D. A. Kocisko, J. H. Come, S. A. Priola, B. Chesebro, G. J. Raymond, P. T. Lansbury and B. Caughey, Nature, 1994,370, 471. 20 R. Riek, S. Hornemann, G. Wider, M. Billeter, R. Glockshuber and K. Wuthrich, Nuture, 1996, 382, 180. 2 I M. E. Bruce, in Methods In Molecular Medicine: Prim Diseases, ed. H. Baker and R. M. Ridley, Humana Press, 1996, 13, 223. 22 J. Collinge, K. C. L. Sidle, J. Meads, J. Ironside and A. F. Hill, Nature, 1996,383, 685. 23 F. E. Cohen, K-M. Pan, Z. Huang, M. Baldwin, R. J. Fletterick and S. B. Prusiner, Science, 1994, 264, 530. 24 H. Wille, G. F. Zhang, M. A. Baldwin. F. E. Cohen and S. B. Prusiner, .I. Mul. Bid., 1996, 259, 608. 25 P. T. Lansbury and B. Caughey, Chem. Biol., 1995,2, 1. 26 K. Halverson, P. E. Fraser, D. Kirschner and P. T. Lansbury, Biochemistry, 1990, 29, 2639. 27 M. S. Palmer, A. J. Dryden, J. T. Hughes and J. Collinge, Nutwe, 1991, 352, 340. 28 J. H. Come and P. T. Lansbury, .I. Am. Chem. Soc., 1994, 116,4109. 29 B. Caughey and B. Chesebro, Trends Cell Biol, 1997, 7, 56. 30 J. Tatzelt, S. B. Prusiner and W. J. Welch, EMBO .I., 1996, 15, 6363. 31 H. M. Schiitzl, M. Da Costa, L. Taylor, F. E. Cohen and S. B. Prusiner, J. Mol. Biol. 1995, 245, 362. Received, 21st Muy 1997 Accepted, 17th July 1997 432 Chemical Society Reviews, 1997, volume 26

ISSN:0306-0012    

DOI:10.1039/CS9972600425

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Sandwich-type heteroleptic phthalocyaninato and porphyrinato metal complexes Dennis K. P. Ng“ and Jianzhuang Jiangb Department of Chemistry, The Chinese University of Hong Kong, Shatrn, NT, Hong Koiig, PR China h Department of Chemistry, Shandong University, Jinan 250100, PR China Over the past decade, a substantial number of sandwich-type metal complexes with mixed phthalocyaninato and/or porphyrinato ligands have been synthesised. Owing to the intriguing n-n interactions associated with these complexes, this family of sandwich compounds displays unique phys- ical, spectroscopic and electrochemical properties which have aroused great interest from researchers in various disciplines. This article seeks to summarise recent develop- ments in this field of research and explain the growing interest in this fascinating and important class of com-plexes.1 Introduction Tetrapyrrole derivatives such as phthalocyanines (e.g. 1) and porphyrins (e.g. 2) are well-studied macrocyclic compounds. Owing to their highly delocalised n electron systems, these compounds exhibit a wide range of intriguing physical and chemical properties which render them useful in various disciplines ranging from catalysis, materials science to medi- cine.’ When these macrocycles are held in close proximity by metal ions in sandwich-type complexes, the n-systems interact strongly with each other, which results in characteristic features in this class of compounds. For example, the bis(phtha1ocyani- nato)lutetium(IIr) [LU(PC)~] is the first intrinsic molecular semiconductor, with a conductivity of 5 X 10-5 Q-1 cm-1 in thin films.2 This sandwich compound is also one of the most promising electrochromic display materials.la,3 It offers not only high contrast, but virtually a full spectrum of colours available by tuning the electric potential applied. The bis(por- Dennis K. P. Ng was born in Hong Kong in 1965. He studied chemistry at The Chinese University of Hong Kong, where he received his BSc with first class honours (1988) and, under the supervision of Tien-Yau Luh, his MPhil(l990). He then joined the research group of Malcolm L. H. Green at the University of Oxford and obtained his DPhil in 1993. He was the recipient of u Croucher Foundation Studentship (I988-1 990) and Scholar- ship (I 990-1 993).He became a Research Fellou, in Chemistry at the California Institute of Technology working with Seth R. Murder (I993-1 994) before returning to his alma mater. He is presently an Assistant Professor at The Chi- nese University of Hong Kong. His current research interests focus on the chemistry of tetra-pyrrole compounds and syn-thetic models for the active sites of tungsto-and molyhdo-enzymes.Dennis K. P. Ng phyrinato) metal complexes M(Po~)~, first reported by Buchler et al. in 1983,4 have also been put forward as structural and spectroscopic models for the ‘special pair’ found in the reaction centre protein of photosynthetic bacteria.5 This special pair is actually a dimer of closely spaced bacteriochlorophyll b [(B~hl)~]where the initial photoinduced electron-transfer step of photosynthesis occurs.Because of these unique properties, an increasing number of studies of the sandwich-type phthalocya- ninato and porphyrinato metal complexes have been stimulated in recent years. / 29H,31 H-phthalocyanine 21H,23Kporphine 1 2 Homoleptic sandwich compounds containing the same phthalocyaninato or porphyrinato ligands have been extensively studied.6 A vast number of double-decker and triple-decker complexes [M(P)2 and M2(P)3, where P = Pc or Por; see Section 7 for abbreviations] of lanthanides, actinides and group 4 transition metals have already been described. Recently, heteroleptic analogues with mixed macrocyclic ligands have Jianzhuang Jiang was born in Heilongjiang, China in 1963.He received his BSc (I985), MSc (1988) and PhD (1992, with Qinglian Zhang) .from Peking University. During his doctoral study (1990-1992), he obtained a fellowship from the Jupunesti Menbusho and carried out pi-t of his PhD uiork at O~aku University under the guidance of Ginya Adachi. After spending tM’oyeai-s as a postdoctora1,fellou. at Peking Uni\w-sity (again with Qinglian Zhang:), he becume a tiisiting scholar ut The Chinese Uni\w-sity o~HoF~~ Kong, Mihere he c*ollabot-ated with Dennis K. P. Ng on thc topic of this article. He joined the Shandong Unityei-siij in April 1996 and is nm’ u Pu~essorof Inorganic Chem-istry. HiJ research interests lie in the lanthanide cmnplexes of porphyrins and phthalocya-nines.Jianzhuang Jiang Chemical Society Reviews, 1997, volume 26 433 received increased attention. The individual chromophores in these complexes may display very different optical and redox properties that allow an investigation of the n-n interactions and the extent of hole delocalisation. To date, there are about 40 papers describing these heteroleptic complexes and virtually all of them have been published over the past decade. The purpose of this article is to provide an overview of the current status in this field of research. Only heteroleptic double- and triple- decker complexes-i.e. with different phthalocyaninato and/or porphyrinato ligands-will be reviewed with emphasis on their syntheses, structures, spectroscopic and electrochemical prop- erties.2 Synthetic methods 2.1 Synthesis of heteroleptic phthalocyaninato metal complexes A straightforward but not efficient method for the preparation of heteroleptic double-deckers M(Pcl)(Pc2) involves the mixed condensation of two different phthalonitriles in the presence of a metal salt. As expected, the reactions give statistical mixtures of products containing a different number, type and position of substituents. For example, treatment of an equimolar mixture of 4-propoxyphthalonitrile (3) and 4-tert-butylphthalonitrile (4) with lutetium acetate at 290 “C gives a mixture of double- deckers from which the mi~-(PrO)~(Bu~)~Pc’~Lu can be isolated by repeated column and thin-layer chromatographic procedures in 4.6% yield (Scheme l).7 This purified product still exists as a mixture of many constitutional isomers.An improved method to M(Pcl)(Pc2) has been reported by L’Her et aZ.8 Reaction of Lu(OAc)3 with LiZ(Pc) and Li2[Pc’- (Bu‘)~], L~~[PC”(R)~] (R = CH3,OCH3) or Li2(Nc) produces the unsymmetrical double-deckers Lu(Pc)(Ring) in 15-22% yield, along with the symmetrical double-deckers Lu(Pc)2 and Lu(Ring)2 [Ring = Pc’(But)4, Pc”(CH&, Pc”(OCH&, Nc], which can be separated by column chromatography. Reaction temperature appears to be the most critical parameter in these reactions. The mixed double-deckers M(Pc1)(Pc2) can also be prepared in a stepwise manner.Treatment of phthalonitrile with Lu(OAC)~.~H~Oand DBU gives Lu(Pc)2 and the half-sandwich compound Lu(Pc)(OAc)(H20)2. The latter complex reacts with Na2(Nc) in 1-chloronaphthalene leading to the formation of LU(OAC)~*~H~O DBU acN CN Pro CN CN 3 4 Lu(OAc)3 t R R R R R R‘ R = Pro or But Scheme 1 Lu(Pc)(Nc) (5).9 Alternatively, reaction of Lu(Pc)(OAc) with 2,3-dicyanonaphthalene, DBU and ammonium molybdate in hexan-1-01 also produces Lu(Pc)(Nc) in 25% yield (Scheme 2).’* The half-sandwich compound Lu(Pc)(OAc)(H20)2 also reacts with the metal-free phthalocyanine Hz(Pc-crown) 6, which possesses four fused benzo-15-crown-5 moieties.” An 5 Scheme 2 434 Chemical Society Reviews, 1997, volume 26 equimolar mixture of these compounds in refluxing 1 -chlorona- phthalene gives mainly the double decker Lu(Pc)(Pc-crown) together with a small amount of the triple-decker (Pc)Lu(Pc- crown)Lu(Pc).However, by using an excess amount of Lu(Pc)(OAC)(H~O)~(3 : 1 mol ratio), the latter complex is obtained as the main product. It is worth noting that the symmetrical double-decker LU(PC)~ is also produced in all these reactions indicating that the lutetium ion can be released from the half-sandwich compound under these reaction conditions. n 6 The heteroleptic sandwich compounds Eu(Pc)[Pc"(R)~] (R = n-C7HI5, OC5Hll) have recently been synthesised with an analogous stepwise procedure. 12 Treatment of Eu(acac)3.nH20 with Li2(Pc) yields the half-sandwich compound Eu(Pc)(acac), which acts as a template to induce cyclic tetramerisation of the corresponding 4,5-disubstituted phthalonitriles (Scheme 3).2.2 Synthesis of heteroleptic porphyrinato metal complexes Symmetrical bis(porphyrinat0) lanthanide complexes Ln(Por)2 are generally prepared by treating Ln(a~ac)~-nH~O with metal- free porphyrins.4 A plausible mechanism is shown in Scheme 4. It is believed that the protonated species LnH(p~r)~ are initially formed which undergo deprotonation and oxidation to give Ln(p~r)~.The triple-deckers Ln2(Por)3 are also produced as side products. This method can be extended to unsymmetrical Ln(Por')(Por2). Treatment of Ln(acac)3-nH20 (Ln = Ce,13314 Eu15) with H2(OEP) and H2(TPP) in 1,2,4-trichlorobenzene gives a mixture of products from which Ln(OEP)(TPP) can be separated by column chromatography.But for Ln = La, the protonated LaH(OEP)(TPP) has been isolated instead, along with a small amount of La2(0EP)2(TPP).16 The half-sandwich compounds Ln(TPP)(acac) (Ln = Sm,17 Eu,IS Dy,lg Yb,I8 Lu19) also react with Li2(0EP) leading to the isolation of Eu(OEP)(TPP) or LnH(OEP)(TPP) (Ln = Sm, Dy, Yb, Lu). The latter species (for Ln = Sm, Lu) undergo deprotonation in basic solvents such as DMF and pyridine to form the corresponding anions [Ln(OEP)(TPP)]-.17,19 Similarly, the actinide analogues M(OEP)(TPP) (M = Th, U) can also be prepared by treating LiZ(0EP) with M(TPP)C12 (M = Th, U), which can be generated from MCl4 and H2(TPP).20 An alternative route involves the dimethylamide Th(NMe2)4 as the starting material. Upon treatment with an equimolar mixture of H2(0EP) and H,(TPP>, it converts into Th(OEP)(TPP) in 50% yield.21 Functionalisation of homoleptic bis(porphyrinat0) metal complexes provides an alternative pathway to heteroleptic counterparts.Treatment of Zr(TTP)2 (7) with I2 and AgN02 gives Zr(TTP)(TTP-N02) in which the nitro group is attached to a (3-pyrrole site on one of the TTP rings. Reduction of this complex with NaBH4 in the presence of 5% Pd/C leads to the Lb(PC)Eu(acac)3*nH20 -Eu(Pc)(acac) R R = n-C7H15, OC5Hii Scheme 3 HZ(WLn( aca~)~*nH~O * Ln(Por)(acac) -2 H(acac) HZ(W Ln2(Por)3 41-H(acac) + Hi, + e-LnH(Por)pLn(Por)2 --H', -e-Scheme 4 formation of Zr(TTP)(TTP-NH2) (8).*2 This amine undergoes condensation with anthraquinone 9 in excess pyridine or pyromellitic dianhydride 10 and hexylamine giving the coupled products 11 and 12 respectively (Scheme 5). These compounds containing an electron acceptor attached to a cofacial porphyrin pair, serve as excellent structural models of the photosynthetic reaction centre.23 2.3 Synthesis of mixed phthalocyaninato and porphyrinato metal complexes The preparation of mixed phthalocyaninato and porphyrinato lanthanide complexes usually involves the treatment of Li2(Pc) with Ln(Por)(acac) generated in situ from Ln(acac)3.nH20 and H2(Por) or the reaction of Li*(Pc) with Ln(acac)3*nH20 followed by treatment with H2(Por).These reactions normally give mixtures of the mixed double-deckers Ln(Pc)(Por) and the triple-deckers Ln2(Pc)(Por)2 and Ln2(Pc)2(Por) the yields of which depend on the nature of the lanthanide metals and the macrocycles, and the ratio of these reactants.For instance, reaction of Ln(acac)3*nH20 (Ln = Sm, Eu, Gd) with H2(TPP) and Li2(Pc) affords mainly L~*(Pc)(TPP)~ (13) (62-68%) and L~~(Pc)~(TPP)(14) (15-17%), together with a small amount of Ln(Pc)(TPP) (15) (Scheme 6).24 Similar reactions using Ln(acac)3.nH20 (Ln = Er, Lu, Y), however, lead to the Chemical Society Reviews, 1997, volume 26 435 formation of the heteroleptic double-deckers as the sole product.25 By employing H2(TPyP) instead of H2(TPP), the lithiated species Li[Ln(Pc)(TPyP)] (Ln = Eu, Gd) have been isolated in moderate yields which are believed to be the primary products.Upon exposure to air, solutions of these compounds convert slowly to the corresponding neutral complexes Ln( Pc)(TPyP) .26 Weiss et al. have employed the latter method to synthesise the gadolinium ~andwiches.~7 Treatment of Li2(Pc) with 2 equiv. of Gd(a~ac)~-nH~Oproduces Gd(Pc)(acac), which fur- ther reacts with l equiv. of H2(TPP) giving Gd(Pc)(TPP) in 65% yield and a small amount of Gd2(Pc)(TPP)2. However, by using 4 equiv. of the metal salt and 2.5 equiv. of the metal-free porphyrin, the triple-decker Gd,(Pc)(TPP)2 has been isolated in 78% yield. Recently, Weiss et al. have also reported a series of novel triple-deckers in which two different metal ions are sandwiched by three macrocycles.28 Reaction of Ml(Por)(acac) (MI = Gd, Lu, Y; Por = OEP, TPP), generated in situ, with the double- deckers M2(Pc)(TPP) (M2 = La, Ce) results in the formation of the respective mixed triple-deckers (Por)M1 (Pc)M2(TPP) in good yields.The actinide double-deckers M(Pc)(Por) (M = Th, U; Por = OEP, TPP, TTP) can be prepared by treating M(Por)C12 with an excess of Na2(Pc).2* Similarly, treatment of M(Por)C12(M = Zr, Hf; Por = OEP, TPP), prepared from II 0 11 Zr + Zr (11) NaBH4 0 0 5% PdIC 7 a C6H13NH2 12 Scheme 5 Ln Ln(acac)ynH20 Ln = Sm, Eu, Gd 13 14 15 Scheme 6 436 Chemical Society Reviews, 1997, volume 26 MC14(THF)2 and Li2(Por)(DME),, with Na2(Pc) produces the corresponding M(Pc)(Por) .6 3 Structure None of the known heteroleptic bis(phthalocyaninat0) metal complexes have yet been structurally characterised, whereas the molecular structures of two mixed porphyrinato metal com- plexes, namely Ce(OEP)(TPP)13 and SmH(OEP)(TPP),17 have been determined. The coordination geometry of Ce(OEP)(TPP) is a square antiprism in which the porphyrin N4 planes are separated by 2.768 A and the mean Ce-N (OEP) bond distance (2.471 A) is slightly shorter than the Ce-N (TPP) distance (2.480 A).A perspective view of the complex (Fig. 1) shows that the two porphyrin rings are domed. The distance dc~of the C2oN4 mean plane and the N4 mean plane which is a measure of the doming is 0.32 for the OEP and 0.295 A for the TPP.The average dihedral angle 4of the individual pyrrole rings with respect to the corresponding C20N4 mean plane is 14.8" for the OEP and 13.0" for the TPP. These data show that the OEP ring is somewhat more deformed than the TPP ring which may arise from a stronger covalent interaction between the OEP ligand and the metal centre.13 Fig. 1ORTEP drawing of Ce(OEP)(TPP). (Reproduced with permission from ref. 13). The molecular structure of SmH(OEP)(TPP) (Fig. 2) as determined by X-ray diffraction analysis also exhibits a nearly perfect square antiprismatic geometry where the samariuy centre is displaced 1.470 8, from the TPP N4 plane and 1.5 16 A from the OEP N4 plane; the porphyrin Nqplanes, which are almost parallel with a dihedral angle of 0.858", are therefore separated by 2.986 A.The average Sm-N (TPP) bond distance (2.538 A) is slightly shorter than the Sm-N (OEP) distance (2.563 8,) which is just the reverse of that observed in the structure of Ce(OEP)(TPP). This observation suggests that the proton may reside on the OEP ring and the complex may be formulated as SmIII(HOEP)(TPP). The OEP ring again is more deformed than the TPP ring in the stpcture as shown by the distance dcN[0.316 (OEP) and 0.233 A (TPP)] and the mean dihedral angle @ [14.69" (OEP) and 10.68" (TPP)].I7 The structures of C~(PC)(TMPP)~~ and La(Pc)(TPP)25 have also been determined which are similar to those of the mixed porphyrin analogues.The metal centre is bonded to four isoindole nitrogens of a Pc ring and four pyrrole nitrogens of a Por ring forming a slightly distorted square antiprism. It lies closer to the N4 mean plane of Por than the N4 plane of Pc by 0.152 [for Ce(Pc)(TMPP)] or 0.231 8, [for La(Pc)(TPP)], probably due to the larger cavity size of Por. Both macrocycles are also domed displaying a saucer shape. The neodymium complex Nd2(Pc)2(TMPP) is the first heteroleptic phthalocyaninato and porphyrinato complex that has been structurally characterised.30 The two Nd ions are sandwiched between two Pc rings and one TMPP ring in a Fig. 2Molecular structure of SmH(OEP)(TPP). (Reproduced with permis- sion from ref. 17). triple-dec ker-like structure, i. e. (Pc)Nd(Pc)Nd(TMPP).The two Pc ligands exhibit a staggered orientation creating a distorted square antiprismatic environment around the metal centre. In contrast, the other Nd ion adopts a distorted cubic geometry which can minimise the nonbonding interactions between the four 4-methoxyphenyl groups of TMPP and the inner Pc ring. The inner Pc ring is almost planar while the outer Pc ring and TMPP ring are slightly domed. The structures of the cerium triple-deckers Ce2(Pc)2(TMPP) and C~~[PC"(CH~)~](TPP)~ have also been reported.31 Re- markably, the former complex possesses a symmetrical struc- ture, (Pc)Ce(TMPP)Ce(Pc), different from the neodymium analogue.30 The latter complex also adopts a similar structure in which the substituted Pc ring lies between two TPP rings.In both structures, the coordination geometry of both cerium centres is a distorted cube. Fig. 3 shows the molecular structure of (TPP)Ce(Pc)-Gd(0EP)-an elegant triple-decker which has three different macrocyclic ligands and two different metal centres.28 Both metal centres are octacoordinated and the coordination poly- hedrons of the Ce and Gd ions are a slightly distorted cube and a square antiprism, respectively. The outer TPP and OEP rings are domed towards the metal centres, while the inner Pc ring is Fig. 3 ORTEP plot of the structure of (TPP)Ce(Pc)Gd(OEP). (Reproduced with permission from ref. 28). Chemical Society Reviews, 1997, volume 26 437 also domed with its N4 mean plane lying closer to the Gd ion than the C2oN4 mean plane.This has been attributed to the smaller ionic radius of GdIII compared with CeIII and the stronger n-n interactions between the OEP and Pc rings with respect to those between the TPP and Pc rings. 4 Spectroscopic properties 4.1 lH NMR spectra As observed previously for homoleptic phthalocyaninato com- plexes of trivalent metals, all of the mixed Pc analogues M111(Pc1)(P~2)do not give satisfactory NMR spectra because of their paramagnetic nature associated with the unpaired electron in one of the Pc rings, and their limited solubility in common organic solvents. However, by adding a small amount of hydrazine hydrate into [2H7]DMF or [2H~]DMS0 solutions of these complexes, satisfactory 'H NMR spectra can be obtained because the paramagnetic neutral species are reduced to the more soluble monoanions [MIII(Pc1)(Pc2)J- in which both of the macrocycles are diamagnetic dianions.83l2 There has been some controversy in the location of the proton in M"IH(Porl)(Por2).Coutsolelos et al. have recently studied the 1H NMR spectra of LuIIIH(OEP)(TPP), which contains a diamagnetic lutetium ion.19 The spectrum in CDCI3 shows two upfield signals at 6 -8.57 and -9.54 in 1 :4.3 ratio, which can be assigned to the proton attached to the TPP and OEP rings, respectively. The observed upfield shift for these protons may arise from the ring current generated by the macrocycles. These data indicate that the proton in question resides preferentially on the more basic OEP nitrogen and this complex can be described as a 1 :4.3 mixture of the tautomers LuIII(OEP)(HTPP) and LulI1(HOEP)(TPP) in solutions.These upfield signals disappear in the spectrum recorded in [*H7]DMF suggesting that the proton is removed by the basic solvent. The 'H NMR data of Zr(TTP)(TTP-N02) have been reported by Girolami et a1.22 The spectrum recorded at 20 "C shows eight singlets of equal intensity for the methyl groups along with 14 doublets and one singlet due to the (3-pyrrole protons. These data indicate clearly that the two porphyrin rings are not rotating with respect to one another on the NMR time-scale. No evidence of rotation has been observed even at temperatures up to 150 "C. However, the variable temperature 1H NMR spectra of Zr(TTP)(TTP-N02) and 8 have revealed that the tolyl groups can rotate along the respective C(ipso)-C(meso) axis.Due to the steric hindrance arising from the (3-pyrrole substituents, the rotation barriers of two of the tolyl groups are ca. 4 kcal mol-I (1 cal = 4.184 J) higher than those of the remaining six tolyl groups. This kind of dynamic motion is well-known for monomeric and sandwich-type porphyrinic compounds.'*32 324 Through a series of sophisticated NMR experiments, Bertini et al. have explored the structure and dynamics of LnH(TPP)2 and LnH(OEP)(TPP) (Ln = Dy, Yb).l* The phenyl rings have been found to rotate at a rate of ca. 30 s-1 as measured for YbH(TPP)2. The rotational barrier has also been determined for Th(TPP)2.21 The value (14.5 kcal mol- l) is comparable to that obtained for the same process in monoporphyrin Th(TPP)C12 (13.8 kcal mol-1).The similarity of these activation energies suggests that the rotation is not significantly hindered in the sandwich compound, despite the close proximity of the two porphyrin macrocycles. Heteroleptic double-deckers containing a diamagnetic ion such as CeIV,13,29 Th1V,20,21 23-1" and HFV6 usually give normal IH NMR spectra with signals falling in the 0-10 ppm region. For double- and triple-decker complexes with paramagnetic metal centres such as CeIII,31 SmIII,24 EuI11'57243 and UIV,20 'H NMR spectra may still be obtained but the signals are generally broad and spread over a wide region. 4.2 Mass spectra Mass spectrometry is a versatile technique in the characterisa- tion of these sandwich-type complexes. It requires only a tiny 438 Chemical Society Reviews, 1997, volume 26 amount of sample (< 1 mg) and does not depend on the nature of the metal centre(s). Apart from the conventional desorption- ionisation methods such as fast atom b0mbardment,~4.26 liquid secondary ion12 and field desorptioqx the newly developed electrospray ionisation (ESI) can also be used in studying these complexes.The heteroleptic double-and triple-deckers Ln(Pc)(TPP) (15),Ln2(Pc)(TPP)2 (13) and L~~(Pc)~(TPP) (14) (Ln = Sm, Eu, Gd) have recently been examined by using positive-ion ESI-FTICR mass spectrometry.32 The spectra of all these compounds show intense signals corresponding to the singly charged molecular radical cations.The protonated species [M + HI.+ may also be present in the case of double- deckers as revealed by comparing the isotopic patterns. Multiply charged molecular ions up to +5 have also been observed which may be formed from successive oxidation of the ligands. Interestingly, even-charged species have not been detected with significant abundances in all cases. Apart from the molecular mass information, tandem mass spectrometry offers additional structural information on these complexes. This is one of the advantages of FTICR that allows an isolation of a particular species and a subsequent study of its gas-phase reactions. The mass-selected collision-induced dis- sociation (CID) experiments have been performed on the europium triple-de~kers.~~ The spectrum of Eu2(Pc)2(TPP) [Fig.4(a)] shows the parent ion [Eu2(Pc)2(TPP)]+ and several fragment ions including [Eu(Pc)(TPP)J+ and [Eu(Pc)zJ+. The formation of these two fragments suggests that this triple-decker adopts an asymmetrical structure, i.e. (Pc)Eu(Pc)Eu(TPP). In contrast, the mass spectrum of Eu*(Pc)(TPP)2 obtained under CID conditions [Fig. 4(b)] shows the existence of [Eu(Pc)(TPP)]+ with no indication of the formation of [Eu(TPP)~]+, which supports the symmetrical structure (TPP)Eu(Pc)Eu(TPP). By closely examining the fragmentation patterns of these triple-deckers (Scheme 7), it has been postulated that a collision-induced intramolecular charge trans- fer process occurs prior to fragmentation in these complexes.L: 0.5 [Eu(Pc)]+1 500 1000 1500 2000 m/z Fig. 4 CID mass spectra of (a) Eu~(Pc)~(TPP)and (b)Eu*(Pc)(TPP)z. 4.3 IR and resonance Raman spectra Phthalocyanine and porphyrin n-radical anions (Pc*-and Pore-) display characteristic IR marker bands that are not present in the dianions Pc2-and Po$-. This allows an investigation of the hole or unpaired electron (de)localisation in these mixed sandwich complexes and their cations. For instance, the IR spectra of the double-deckers EuIII-(OEP)(TPP)15 and [MIV(OEP)(TPP)J+ (M = Ce,13 Th21) show a strong band around 1530 cm-1 that falls in the region typical for OEP radical. No (or a very weak) marker band for TPP radical at ca.1270-1300 cm-1 is observed. This therefore suggests that the hole is preferentially localised on the OEP ring which can be rationalised by the fact that the OEP macrocycle is significantly easier to oxidise than TPP. [Eu( Pc)(TPP)]+ [Eu(TPP)]+ [Eu(TPP)]+ [Eu(Pc)(TPP)]+-Scheme 7 This has been corroborated by the resonance Raman spectroscopic studies of [CeIV(OEP)(TPP)In+(n = 0, 1) reported by Bocian et al.33 The frequencies of certain skeletal modes of both the OEP and TPP macrocycles are shifted upon oxidation. These shifts are large for skeletal vibrations of the OEP ring and small for those of the TPP ring. In addition, the shifts observed for the OEP and TPP modes in CeIV-(OEP)(TPP) are larger and smaller, respectively, than those of these vibrations in the homoleptic complexes Ce1V(OEP)2 and CeIV(TPP)2.These results indicate that the electron is removed primarily from a molecular orbital comprised mainly of an OEP orbital. It has been estimated that this molecular orbital is ca. 80% OEP-like in character. The heteroleptic complexes MII1(Pc)(TPP) (M = La, Pr, Nd, Eu, Gd, Er, Lu, Y) give a strong IR band at 1317 cm-I, which is a diagnostic band for Pc-, while the marker band due to TPPa- is not seen.25 This is consistent with the localisation of a hole mainly on the Pc ligand. However, a band at 1258 cm-1 attributable to TPP.-, in addition to the band at 1317 cm-1 due to Pea-, appears in the IR spectra of the singly oxidised species [MIII(Pc)(TPP)]+ suggesting that these monocations can be described as di-n-radicals [MIII(Pc.-)(TPP*-)]+.For the OEP analogues [MJv(Pc)(OEP)]+ (M = Zr,6 Hf,6 Th,20 U20), both diagnostic bands belonging to the Pc*- (1310-1321 cm-1) and 0EP.-(1515-1547 cm-I) are observed which suggests a delocalisation of hole over both macrocycles. Tran-Thi et al. have studied the resonance Raman spectra of a series of double- and triple-decker complexes of cerium and gadolinium.27 With reference to the Raman spectra of some monophthalocyaninato complexes, they have proposed that the strong band at ca. 1500 cm-1 is a good marker band to characterise neutral metalated phthalocyanines and this band experiences an upshift of ca. 15 cm-I upon removal of one electron from the Pc2- ring.The shift in frequency of this band can thus be used to probe the extent of hole delocalisation in these complexes. For example, the Raman band at 1515 cm-1 for the anion [GdlI1(Pc)(TPP)]- shifts to 1529 cm-l in the corresponding neutral compound showing that the hole is localised predominantly on the Pc ring in Gdrrr(Pc)(TPP). In contrast, the marker band for the cerium analogue CeIV(Pc)(TPP) upshifts by only 7 cm-1 upon one-electron oxidation which is consistent with a hole delocalisation over both macrocycles in [CeTV(Pc)(TPP)]+. This delocalisation may originate from the strong n-n interactions between the macrocycles which are very close to each other due to the ion contraction in the CeIV ion. Resonance Raman studies of the triple-deckers [GdIIIZ- (Pc)(TPP)2]+" (n = 0, 1) have also revealed that the ligand preferentially oxidised is the Pc moiety.But for the cerium analogues [CeII12(Pc)2(TPP)]+-" (n = 0, l), in which the metal centre can be oxidised, the marker band upshifts by only 3 cm-upon oxidation which is comparable with that observed in [CeIII(Pc)(TPP)]- and CeIV(Pc)(TPP). This small shift is in agreement with a metal-based oxidation and the oxidised product can thus be formulated as [(Pc2-)CeIV(TPP2-)CerI1-(Pc2-)] +.27 4.4 Electronic spectra The mixed bis(phtha1ocyaninato) metal complexes display characteristic electronic spectra which are very similar to those of the homoleptic analogues. Fig. 5 shows a typical UV-VIS spectrum of Lu'I'(Pc)(Nc), along with the spectra of LuIII(Pc)2 and LuIII(Nc)2 for comparison.8 The spectrum shows the B band (or Soret band by analogy to porphyrins) at 326 nm, the Q band at 706 nm and two n-radical anion bands at 450 and 987 (not shown) nm.All of these can be attributed to n-n* electronic transitions. It is worth noting that the Q band lies roughly in the middle of the corresponding bands of the symmetrical counter- parts indicating that the hole is delocalised over the two n-conjugated systems on the electronic timescale. This class of compounds also displays a near-IR band at ca. 1500 nm which can be ascribed to an electronic transition from a filled Pc-Nc n-bonding orbital to the HOMO, which is a Pc-Nc n-anti bonding orbital, 991 2 300 400 500 600 700 800 Alnm Fig.5 UV-VIS spectra of Lu(Pc)* (-), Lu(Pc)(Nc) (-) and Lu(Nc)* (---). (Reproduced with permission from ref. 8). The absorption spectra of MIV(OEP)(TPP) (M = Ce,14 Th,20,21,34U20) exhibit monoporphyrin-like B and Q absorption bands. For example, the spectrum of the thorium double-decker (Fig. 6) shows the B(0,O) band at 393 nm and the Q(0,O) band at 584 nm, which is much weaker than the Q(1,O) band at 544 nm. In addition, there are two absorptions labelled Q' and Q" in Fig. 6 which are not seen in the corresponding monomeric 400 150 100 5010, L 350 450 550 650 7i 0 Ahm Fig. 6 Absorption spectrum of Th(OEP)(TPP)in CHC13. (Reproduced with permission from ref. 21). Chemical Society Reviews, 1997, volume 26 439 metalloporphyrins and bis(porphyrinat0) metal complexes with large macrocycle spacing.The features of these bands vary systematically with the ionic radius of the metal ion and thus the distance between the two rings. Based on these observations, it is believed that these Q' and Q" absorptions are highly characteristic of strongly coupled porphyrins. Oxidation of this compound results in a blue-shift of the B(0,O) band. It appears at 370 nm in [Th(OEP)(TPP)]+ and 364 nm in [Th(OEP)(TPP)]2+. Similar to bis(phthalocyaninat0) complexes of trivalent metals, the monocation also exhibits a near-IR transition at 1318 nm, which shifts to 956 nm in the dication.21 This spectroscopic feature is of particular interest since a broad transition at 1300 nm is also found for the oxidised special pair [(BC hl)2] +.5 The fluorescence spectrum of Th(OEP)(TPP) has also been recorded by Holten et al.34A broad fluorescence emission band appears at ca.800 nm, which is lower in energy than the Q' absorption maximum and is also red-shifted relative to typical emission bands from monomeric porphyrins. This has been attributed to states that acquire an increased amount of charge resonance character as the electronic interactions between the porphyrins become stronger. Excitation of quinone-substituted double-decker 11 also gives a fluorescence signal of which the intensity decreases considerably with increasing solvent polar- ity.23 This is in contrast to that observed for Zr(TTP)2 (7) in which the fluorescence intensity remains relatively unchanged in a range of solvent polarities. This indicates that the singlet excited state of 11 undergoes electron transfer to form a charge- separated state (porphyrin+-quinone-).The heteroleptic double-deckers MIV(Pc)(Por) (M = Ce, Th, U, Zr, Hf)h3)927 and [MIII(Pc)(Por)]- (M = La, Pr, Nd, Eu, Gd, Er, Lu, Y),25-27 in which both of the ligands are dianionic, exhibit common spectral features. Fig. 7 shows the absorption spectrum of [Eu(Pc)(TPyP)]-, which is representative of the spectra of these complexes.26 The spectrum displays strong Pc and TPyP B bands at 334 and 412 nm, respectively. The Pc Q bands appear at 584, 635 and 794 nm while the former two visible bands should also contain a contribution of the TPyP Q bands.By analogy with the bis(porphyrinat0) complexes, the remaining absorption at 477 nm can be ascribed to a n-n* transition arising from molecular orbitals delocalised over both macroc ycles. I ": 800 1200 1600 A Inm 300 400 500 600 700 800 900 Alnm Fig. 7 UV-VIS spectra of Li[Eu(Pc)(TPyP)] in CH30H (-) and Eu(Pc)(TPyP) in CHCl? (---); the inset shows the near-IR spectrum of Eu(Pc)(TPyP). (Reproduced with permission from ref. 26). The absorption spectra of the related [MIV(Pc)(Por)]+ and MIII(Pc)(Por) are also very ~imilar.6,209~5--27 Both series of complexes contain a one-electron-oxidised n-radical tetrapy- rrole ligand which imparts unique spectral features on them.As shown in Fig. 7, the spectrum of Eu(Pc)(TPyP) displays the Pc B band (324 nm), TPyP B band (403 nm) and a strongly attenuated Q band at 729 nm, which suggests the presence of Pc*-.~~This accompanied with the results from IR and resonance Raman studies of MI1*(Pc)(TPP) (see Section 4.3) seem to indicate that only weak JI-n interactions are present in such complexes and it can therefore be assumed that the hole is mainly localised on the Pc ligand.25~27 The spectrum of Eu(Pc)(TPyP) also shows two bands at 467 and 978 nm, which have counterparts in the spectra of L~III(Pc)~ and may arise from the Pc--. The near-IR band appears at 1218 nm that, based on the above assumption, can be ascribed to an intramolecular ring- to-ring charge transfer transition in which the TPyP2- acts as an electron donor and the Pc.- as an acceptor.However, as shown by Weiss et al.,25 the absorption maximum of this near-IR absorption is red-shifted when the ionic radius of the central metal increases. There is a linear correlation between the wavenumbers of these bands and the metal ionic radii indicating that the energies of the near-IR transitions depend on the ring to ring separations. This reveals that interactions between the two tetrapyrrole ligands cannot be completely ignored. In the heteroleptic triple-deckers LnI[12(P~)(Por)~and Ln11r2(Pc)2(Por) (Ln = Ce, Nd, Sm, Eu, Gd),24,273031 all of the macrocyclic ligands are dianionic. It can be envisaged that their electronic spectra resemble those of MIV(Pc)(Por) and [MfII(Pc)(Por)]- to a certain extent. Fig.8 shows the spectra of Sm1112( Pc) (TPP)2 and Sm11I2( Pc)2( TPP) , which are typical .24 The absorptions at ca. 350 and 420 nm can be attributed to the Pc and TPP B bands, respectively. The remaining absorptions in the visible region may be due to the Q bands of the macrocycles in which the lower-energy Q bands are contributed mainly by the Pc. It is noteworthy that the relative intensities of the Pc and TPP B bands are related to the Pc:TPP ratio in the com-plexes. 300 400 500 600 700 800 Alnm Fig. 8 UV-VTS spectra of Sm2(Pc)(TPP)2 (---) and Sm2(Pc)2(TPP) (-) in CH2C12.(Reproduced with permission from ref. 24). 4.5 EPR Spectra The presence of an unpaired electron in M"I(Pc)(TPP) (M = La, Pr, Nd, Eu, Gd, Er, Lu, Y),25 [Ce*v(Pc)(TPP)]+29 and [ThlV(OEP)(TPP)]+21 has been confirmed by EPR spec-trometry.All solid state spectra show a signal at g = 2.002-2.01 1 which is typical of organic radicals. The EPR spectra of [MIV(Pc)(Por)]+ (M = Zr, Hf; Por = OEP, TPP) show the presence of four lines in addition to the central one at 110 K in frozen CH2C12 solutions.6 A representative spectrum for [ZrJV(Pc)(OEP)]+ is given in Fig. 9(a) (top). The overall appearance of the spectrum is characteristic of a randomly orientated solid in the triplet state which is supported by the weak half-field absorption at g = 4.004 observed in the solid state spectrum [Fig. 9(b)].This has been attributed to the formation of n-radical anion dimers denoted as [(PO~--)M~~(PC~-)...(PC~-)M~V(PO~--)]~+.The av- erage distance (I-)between the two unpaited electrons in the dimer has been estimated to be ca.9 A by the equation: 440 Chemical Society Reviews, 1997, volume 26 r = (0.65g2/D)1/3, where D is the zero-field splitting parameter. As shown in Fig. 9(a),the relative intensity of the triplet (S = 1) vs.the doublet (S = 1/2) component of the signal increases with the concentration of the compound in solution. Since the intensity of the signal is proportional to the concentration of the radicals, the equilibrium constant for the monomer-dimer equilibrium can be determined. The large values lO7-lO* dm3 mol-l calculated for these compounds in 10-5-10-3 M frozen solutions suggest that mainly dimeric species exist in frozen solutions over this concentration range.(a)frozen solution (b)solid state g = 4.004 x 500 Fig. 9 EPR spectra of [Zr(Pc)(OEP)][SbC16] (a)at different concentrations in frozen CH2C12at 110 K and (b)in the solid state at 298 K. (Reproduced with permission from ref. 6). M = mol dm-3. The EPR spectra of [ThIv(Pc)(Por)]+ (Por = OEP, TPP) also show the presence of a concentration-dependent mixture of monomeric (S = 1/2) and dimeric (S = 1) species at 110 K in frozen CH2C12 or toluene solutions, but not in the solid state.20 The calculated distance betweeno the two unpaired electrons in the dimer varies from 8.8 to 9.0 A, which is consistent with two interacting Pc planes separated by ca. 3.3 A.Such intermole- cular x-n interactions are not observed for the symmetrical analogues whose EPR spectra show only a single absorption in both the solid state and frozen solutions.5 Electrochemical and spectroelectrochemical properties Similar to homoleptic bis(phtha1ocyaninato) metal complexes, the mixed MIII(PcI)(Pc2) usually undergo a reversible ligand- based oxidation and a reversible ligand-based reduction, along with some other less well-defined redox processes as revealed by cyclic voltammetry.*~I0.l2 The potentials of the first oxidation and reduction of these compounds lie between the corresponding values for the symmetrical analogues indicating that there are strong n-n interactions in these complexes.The cyclic voltammogram of CeIV(OEP)(TPP) in CH2C12 shows two oxidation waves at +0.402 and +0.968 V vs. SCE assignable to the two successive ligand-based oxidations, and one reduction wave at -0.436 V attributable to a metal-based reduction.13 The first oxidation lies slightly closer in potential to that of CeIV(OEP)2 than that of CeIV(TPP)2. It appears that the mixed sandwich is oxidised first at the OEP ring which is in accord with the IR and resonance Raman data (see Section 4.3). By using a toluene-acetonitrile mixture as solvent which has a larger negative potential window (ca. -2.7 V), Kadish et al. have found that the Th and U analogues undergo three oxidations and four reductions.20 All of them, except for those located at the potential limits of the solvent, are diffusion- controlled one-electron transfer processes.Chemical oxidation of these compounds and related sandwich-type complexes is usually achieved by using phenoxathiinylium hexachloroan- timonate as oxidant.6,13,20.25.27 Kadish et al. have also studied the spectroelectrochemistry of MIV(Pc)(Por) (M = Zr, Hf; Por = OEP, TPP).h After the first two oxidations of MIV(Pc)(OEP), the OEP B band diminishes in intensity and shifts to the blue while the Pc B band remains almost unchanged. For the TPP analogues MIV(Pc)(TPP), the oxidations, however, lead to a decrease in intensity for both the Pc and TPP B bands, and a hypsochromic shift for the latter absorption. These observations indicate that the HOMO of MIV(Pc)(OEP) (M = Zr, Hf) has more OEP than Pc character, while the HOMO of MIV(Pc)(TPP) (M = Zr, H9 may comprise an equal contribution from Pc and TPP macrocycles.The first two reductions of all these compounds affect mainly the Pc B band which attenuates as the Por B band remains essentially unchanged. The LUMO of these double-deckers can thus be represented mainly by the LUMO of the Pc ring. They have also found that the first two oxidation potentials of these sandwich compounds strongly depend on both the ionic radius of the central metal and the type of porphyrin, but the first two reduction potentials are almost independent of these factors.6 The oxidations become easier when the ionic radius of the metal decreases. Thus the HOMO-LUMO gap (i.e.the potential difference between the first oxidation and the first reduction) decreases with a decrease in the ring-to-ring separation. A similar linear correlation between the half-wave potentials of the first oxidation and the first two reductions and the radii of the central metals has also been found for MIII(Pc)(TPP).25 The electrochemistry of two series of heteroleptic triple- deckers EuT1I2(Pc)(Por)2 and Eu"$(Pc)2(Por) (Por = TPP, TMPP, TBPP, TClPP) has recently been examined by cyclic voltammetry.35 All these compounds undergo four reversible one-electron oxidations and up to three reversible one-electron reductions. The redox potentials of the first series of compounds depend slightly on the nature of the substituents attached to Por.The shifts are expected, based on the electron-donating property of the tert-butyl and methoxy groups, and the electron-withdrawing property of the chloro moiety. Although optical and resonance Raman studies of Gd1112(Pc)(TPP)2 suggest that the ligand preferentially oxidised is the Pc ligand,27 the dependence of the first oxidation potentials of these triple- deckers on the Por substituents may argue that the HOMO should also contain a substantial Por character. In contrast, the first oxidation potentials of Eu%(Pc)Z(Por) are almost inde- pendent of the Por substituents which may be rationalised by a Pc-based oxidation. 6 Concluding remarks Sandwich-type complexes containing different phthalocyani- nato and/or porphyrinato ligands have been known for about a decade.Synthetic routes to heteroleptic double- and triple- decker complexes of various metals have been established and the electronic structures of these compounds have also been probed by a wide range of spectroscopic techniques. Under- standing their electronic and photophysical properties is not only of much general interest but may also shed light on the factors governing the highly efficient charge separation during photosynthesis. Although considerable progress has been made Chemical Society Reviews, 1997, volume 26 441 in this field, there is clearly much room for further investigation. Functionalisation of sandwich-type compounds is still in its infancy By attaching functional groups or special moieties to these compounds, it may be possible to tune their electronic properties without altenng the nng-to-nng separation. The properties associated with these units may also be imparted to the parent sandwich compounds.Applications of these com- plexes in matenals science are also another promising, but not yet developed area We hope the grounding provided in this article will stimulate further research into this novel class of compounds 7 Abbreviations acac acetylacetonato CID collision-induced dissociation DBU 1,8-diazabic yclo [5.4.01undec -7-ene DME 1,2-dimethoxyethane DMF NJ-dimethylformamide DMSO dimethyl sulfoxide ESI electrospray ionisation FTICR Founer transform ion cyclotron resonance HOMO highest occupied molecular orbital LUMO lowest unoccupied molecular orbital Nc dianion of 2,3-naphthalocyanine OEP dianion of octaethylporphynn Pc dianion of phthalocyanine Pc' dianion of tetrasubstituted phthalocyanine PC" dianion of octasubstituted phthalocyanine Por general porphynnato dianion SCE saturated calomel electrode THF tetrahydrofuran TBPP dianion of meso-tetra(4-tert-butylpheny1)porphynn TClPP dianion of rneso-tetra(4-chloropheny1)porphynn TMPP dianion of meso-tetra(4-methoxypheny1)porphyrin TPP dianion of rneso-tetraphenylporphyrin TTP dianion of meso-tetra@-toly1)porphynn TPyP dianion of meso-tetra(4-pyndy1)porphynn 8 References (a) Phthalocyanines-Properties and Applications, ed C C Leznoff and A B P Lever, VCH, New York, 1989, vol 1, 1993, vol 2, 1993, vol 3, 1996, vol 4, (b) Metalloporphyrrns in Catalytic Oxidation, ed R A Sheldon, Dekker, New York, 1994, (c) Porphyric Pesticides- Chemistry Toxicology and Pharmaceutical Applications, ed S 0 Duke and C A Reberz, Amencan Chemical Society, Washington DC, 1994 M Passard, J P Blanc and C Maleysson, Thin Solid Films, 1995,271, 8 R B Daniels, J Peterson, W C Porter and Q D Wilson, J Coord Chem , 1993,30, 357 J W Buchler, H -G Kapellmann, M Knoff, K -L Lay and S Pfeifer, Z Naturforsch Tell B, 1983, 38, 1339 J W Buchler and G Heinz, Chem Ber ,1996,129,1073 and references therein 6 R Guilard, J -M Barbe, A Ibnlfassi, A Znneh, V A Adamian and K M Kadish, Inorg Chem , 1995,34, 1472 and references thereln 7 Y Liu, K Shigehara, M Hara and A Yamada,J Am Chem Soc ,1991, 113,440 8 F Guyon, A Pondaven, P Guenot and M L'Her, Znorg Chem , 1994, 33, 4787 and references therein 9 N Ishkawa, 0 Ohno and Y Kaizu, Chem Phys Letr , 1991, 180, 51 10 M Bouvet, P Bassoul and J Simon, Mu1 Cryst Liq Cryst , 1994,252, 31 11 N Ishikawa and Y Kaizu, Chem Phys Lett, 1994, 228, 625 and references therein 12 J Jiang, W Liu, W -F Law, J LinandD K P Ng,Znorg Chim Acta, in press 13 J W Buchler, A De Clan, J Fischer, P Hammerschmitt, J Loffler, B Scharbert and R Weiss, Chem Ber , 1989, 122, 2219 14 0 Bilsel, J Rodnguez and D Holten, J Phys Chem, 1990, 94, 3508 15 J W Buchler and J Loffler, Z Naturforsch Teil B, 1990,45, 531 16 J W Buchler, M Kihn-Botulinski, J Loffler and B Scharbert, New J Chem , 1992,16,545 17 G A Spyroulias, A G Coutsolelos, C P Raptopoulou and A Terzis, Inorg Chem , 1995,34,2476 18 I Bertini, A Coutsolelos, A Dikiy, C Luchinat, G A Spyroulias and A Troganis, Inorg Chem , 1996, 35, 6308 19 G A Spyroulias and A G Coutsolelos, Inorg Chem, 1996, 35, 1382 20 K M Kadish, G Moninot, Y Hu, D Dubois, A Ibnlfassi, J -M Barbe and R Guilard, J Am Chem Soc, 1993,115, 8153 21 G S Girolami, P A Gorlin, S N Milam, K S Suslick and S R Wilson, J Coord Chem , 1994,32, 173 22 G S Girolami, P A Gorlin and K S Suslick, Znorg Chem , 1994,33, 626 23 G S Girolami, C L Hein and K S Suslick, Angew Chem Int Ed Engl , 1996,35, 1223 24 J Jiang, R L C Lau, T W D Chan, T C W MakandD K P Ng, Inorg Chim Acta, 1997,255, 59 25 D Chabach, M Tahiri, A De Clan, J Fischer, R Weiss and M El Malouli Bibout, J Am Chem SOL , 1995,117, 8548 26 J Jiang, T C W Mak and D K P Ng, Chem Ber , 1996,129,933 27 T -H Tran-Thi, T A Mattioli, D Chabach, A De Clan and R Weiss, J Phys Chem , 1994, 98, 8279 28 D Chabach, A De Clan, J Fischer, R Weiss and M El Malouli Bibout, Angew Chem Int Ed Engl, 1996,35, 898 29 M Lachkar, A De Clan, J Fischer and R Weiss, New J Chem , 1988, 12,729 30 M Moussavi, A De Clan, J Fischer and R Weiss, Znorg Chem , 1986, 25,2107 31 D Chabach, M Lachkar, A De Clan, J Fischer and R Weiss, New J Chem , 1992,16,431 32 R L C Lau, J Jiang, D K P Ng and T -W D Chan, J Am Soc Mass Spectrom , 1997,8, 161 33 J K Duchowski and D F Bocian, Znorg Chem , 1990,29,4158 34 0 Bilsel, J Rodnguez, S N Milam, P A Gorlin, G S Girolami, K S Suslick and D Holten, J Am Chem Soc , 1992,114,6528 35 J Jiang, W Liu, W -F Law and D K P Ng, Inorg Chim Acta, in press Received, 9th May 1997 Accepted, 2nd July 1997 442 Chemical Society Reviews, 1997, volume 26

ISSN:0306-0012    

DOI:10.1039/CS9972600433

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Ultrasound in synthetic organic chemistry I-RfACHES VIOLENT UNSTABLESIZE COLLAPSE Timothy J. Mason Sonochernistiy Centre, School of Natural and Environmental Sciences, Coventry University, Coventry, UK CVl5FB High-power ultrasound can generate cavitation within a liquid and through cavitation provide a source of energywhich can be used to enhance a wide range of chemical processes. Such uses of ultrasound have been grouped under the general name sonochemistry. This review will concen- trate on applications in organic synthesis where ultrasound seems to provide a distinct alternative to other, more traditional, techniques of improving reaction rates and product yields. In some cases it has also provided new synthetic pathways. 1 Introduction The use of ultrasound in chemistry (sonochemistry) offers the synthetic chemist a method of chemical activation which has broad applications and uses equipment which is relatively inexpensive.The driving force for sonochemistry is cavitation and so a general requirement is that at least one of the phases of the reaction mixture should be a liquid. When laboratory research in sonochemistry began it seemed to be mainly a method of initiating intransigent reactions especially those which depended upon the activation of metallic or solid reagents. Its development in the past 15 years however has revealed that it has far wider applicability than this and also that it presents a significant scientific challenge to understanding its underlying physical phenomenon-acoustic cavitation.The ever expanding number of applications of sonochemistry in synthesis has made the subject attractive to many experimen- talists and interest has spread beyond academic laboratories into industry and chemical engineering. ‘-5 It was in 1986 that the first ever International Symposium on Sonochemistry was held at Warwick University UK as part of the Autumn Meeting of the Royal Society of Chemistry.6 This meeting was significant in that it was the beginning of serious interest in the uses of ultrasound in chemistry as a study in itself. Of course sonochemistry dates back much further than this. Its origins can be traced to the early part of this century with the discoveries of echo sounding and the mechanical use of power ultrasound for emulsification.The formation of the Royal Society of Chemistry Sonochemistry Group in 1987 followed Professor Mason obtained a BSc (1967) in chemistry and PhD (1970) in organic chern- istiy jrom Southampton Uni-versity. After periods at Am- herst College, USA, York Uni- versity and Bradford University he joined Coventry Polytechnic (now University) in 1975.He is currently chairman of the RSC Sonochemistry group and Pres- ident of the European Society of Sonochemistiy and was awarded a DSc in 1996. by a European Society in 1990 and then other national groups has meant that the subject has expanded greatly over the last few years. There are a range of applications for the uses of ultrasound in chemistry which include synthesis, environmental protection (the destruction of both biological and chemical contaminants) and process engineering (improved extraction, crystallisation, electroplating and new methods in polymer technology). 2 Fundamental aspects Ultrasound is defined as sound of a frequency beyond that to which the human ear can respond.The normal range of hearing is between 16 Hz and about 18 kHz and ultrasound is generally considered to lie between 20 kHz to beyond 100 MHz. Sonochemistry generally uses frequencies between 20 and 40 kHz because this is the range employed in common laboratory equipment. However since acoustic cavitation can be generated well above these frequencies, recent researches into sonochem- istry use a much broader range (Fig.1). High frequency ultrasound from around 5 MHz and above does not produce cavitation and this is the frequency range used in medical imaging. Fig. 1 Sound frequency ranges Like all sound energy, ultrasound is propagated via a series of compression and rarefaction waves induced in the molecules of the medium through which it passes. At sufficiently high power the rarefaction cycle may exceed the attractive forces of the molecules of the liquid and cavitation bubbles will form. These bubbles will grow over a few cycles taking in some vapour or gas from the medium (rectified diffusion) to an equilibrium size which matches the frequency of bubble resonance to that of the sound frequency applied.The acoustic field experienced by the bubble is not stable because of the interference of other bubbles forming and resonating around it. As a result some bubbles suffer sudden expansion to an unstable size and collapse violently. It is the fate of these cavities when they collapse which generates the energy for chemical and mechanical effects (Fig. 2). There are several theories which have been advanced to explain the energy release involved with cavitation of which the most understandable in a qualitative sense is the ‘hot spot’ approach. Each cavitation bubble acts as a localised micro- reactor which, in aqueous systems, generates temperatures of Chemical Society Reviews, 1997, volume 26 443 Fig. 2 Sound propagation in a liquid showing cavitation bubble formation and collapse several thousand degrees and pressures in excess of one thousand atmospheres.In addition to the generation of extreme conditions within the bubble there are also major mechanical effects produced as a result of its rapid collapse. These are also of significance in synthesis and include very rapid degassing of the cavitating liquid (in the rarefaction cycle the newly formed bubbles will fill with gas and be expelled from the liquid) and rapid crystallisation (brought about through seed crystal generation on implosion). 3 Laboratory equipment The first requirement for sonochemistry is a source of ultrasound and whatever type of commercial instrument is used the energy will be generated via an ultrasonic transducer-a device by which mechanical or electrical energy can be converted to sound energy.There are three main types of ultrasonic transducer used in sonochemistry: liquid-driven (effectively liquid whistles), magnetostrictive (based on the reduction in size of certain metals, e.g. nickel, when placed in a magnetic field) and piezoelectric. Most of the current equip- ment used for sonochemistry utilises transducers constructed of piezoelectric ceramics. These are brittle and so it is normal practise to clamp them between metal blocks for protection, The overall structure is known as a piezoelectric ‘sandwich’. Usually two ceramic elements are combined so that their overall mechanical motion is additive (Fig. 3).Piezoelectric trans- ducers are very efficient and, depending on their dimensions, can be made to operate over the whole ultrasonic range. Fig. 3 Construction of a piezoelectric sandwich transducer The two most common sources of ultrasound for laboratory sonochemistry are the ultrasonic cleaning bath and the ultra- sonic horn & probe system.7 These generally operate at frequencies of around 40 and 20 kHz, respectively. 444 Chemical Society Reviews, 1997, volume 26 3.1 The ultrasonic cleaning bath The simple ultrasonic cleaning bath is by far the most widely available and cheapest source of ultrasonic irradiation for the chemical laboratory. Although it is possible to use the bath itself as a reaction vessel this is seldom done because of problems associated with corrosion of the bath walls and containment of any evolved vapours and gases.The normal usage therefore involves the immersion of standard glass reaction vessels into the bath which provides a fairly even distribution of energy into the reaction medium (Fig. 4). The reaction vessel does not need any special adaptation, it can be placed into the bath, thus an inert atmosphere or pressure can be readily maintained throughout a sonochemical reaction. The amount of energy which reaches the reaction through the vessel walls is low- normally between 1 and 5 W cm-2. Temperature control in commercial cleaning baths is generally poor and so the system may require additional thermostatic control. Fig. 4 The ultrasonic cleaning bath in sonochemistry 3.2 The ultrasonic probe This apparatus allows acoustic energy to be introduced directly into the system rather than rely on its transfer through the water of a tank and the reaction vessel walls (Fig.5). The power of such systems is controllable and the maximum can be several hundred W cm-2. The probe system is more expensive than the bath and it is slightly less convenient in use because special seals will be needed if the horn is to be used in reactions which involve reflux, inert atmospheres or pressures above (or below) ambient. Fig. 5 The ultrasonic probe system in sonochemistry 4 An attempt to formulate some rules governing sonochemical activity One of the earliest tenets of sonochemistry was that it is particularly good at assisting reactions involving solid reagents. This is generally but not exclusively correct.A number of groups are attempting to gain an understanding of the underlying principles of sonochemistry in order to be able to predict which type of reaction would be most susceptible to sonication. As a result of these efforts some guidelines have been identified. An empirical classification of sonochemical reactions into three types was proposed by J.-L. Luche and was based upon the purely chemical effects induced by cavitation8 Other (mechanical) effects of cavitation bubble collapse (e.g. emulsification) were considered to be physical rather than chemical and judged to be ‘false’ sonochemistry. These so-called ‘false’ effects are often important and have been included in the following interpretation of the three original types of reaction susceptible to sonochemical enhancement.Type I Homogeneous systems which proceed via radical or radical-ion intermediates. This implies that sonication is able to effect reactions proceeding through radicals and further that it is unlikely to effect ionic reactions. Type 2 Heterogeneous systems proceeding via ionic interme- diates. Here the reaction is influenced primarily through the mechanical effects of cavitation such as surface cleaning, particle size reduction and improved mass transfer. This is what has sometimes been referred to as ‘false sonochemistry’. Type 3 Heterogeneous reactions which include a radical pathway or a mixed mechanism i.e.radical and ionic. Radical reactions will be chemically enhanced by sonication but the general mechanical effect referred to above may well still apply. If the radical and ionic mechanisms lead to different products ultrasound should favour the radical pathway and this could lead to a switch in the nature of the reaction products. In this article the term sonochemistry will be used to encompass any beneficial effect on synthesis induced by cavitation whether it is chemical or physical. 4.1 Reactions which exemplify the ‘rules’ of sonochemistry4.1.I Homogeneous liquid-phase reactions Any system involving a homogeneous liquid in which bubbles are produced is not strictly homogeneous, however, in sono- chemistry it is normal to consider the state of the system to which the ultrasound is applied.Sonochemical syntheses in homogeneous conditions are not extensively reported which suggests that cavitation is less effective in promoting reactions under these conditions. The few studies which have appeared indicate that sonochemical effects generally occur either inside the collapsing bubble where extreme conditions are produced, at the interface between the cavity and the bulk liquid where the conditions are far less extreme or in the bulk liquid immediately surrounding the bubble where the predominant effects will be mechanical (Fin. 6). In order for a chemical to experience the extreme conditions generated inside the cavitation bubble during collapse it must enter the bubble and so should be volatile.The ‘concentration’ of cavitation bubbles produced by sonication using conven- tional laboratory equipment is very small and so overall yields in this type of reaction are low. Thus in the sonication of water small quantities of OH. and H. radicals are generated in the bubble and these undergo a range of subsequent reactions including the generation of H202, The highly oxidising HO- species can react with other moieties in the bubble or migrate to the bulk solution where they have only transient existence. Such radicals can have a significant effect on both biological and chemical species in aqueous solution and can be detected chemically.9 Organic solvents will also slowly decompose on sonication but solvent decomposition is normally only a minor contribution to any sonochemical reaction taking place in the medium.A synthetically useful reaction which takes place in the collapsing bubble is the production of amorphous iron from the sonolysis of Fe(CO)5 (0.4 M) in decane under argon. lo Volatile iron pentacarbonyl enters the bubble and is decomposed during collapse. The fact that an amorphous (rather than crystalline) material is produced confirms that very high temperatures are generated in the bubble and that extreme cooling rates are involved. Conventional production of amorphous iron requires rapid cooling from the vapour to solid state of the order of loh K s-l. Sonolytic decomposition of iron pentacarbonyl in pentane (a more volatile solvent) yields Fe3(CO)12 rather than the metal indicating that the cavitation collapse is not so extreme in this solvent.Since this original report the study of cavitation induced decomposition of iron and other metal carbonyls has continued and expanded. In the case of molybde- num hexacarbonyl the product is nanostructured molybdenum carbide which has proved to be an excellent dehydrogenation catalyst.11 Sucrose has a negligible vapour pressure and so cannot enter the bubble during sonication. A study of the effect of sonication on the rate of acid catalysed inversion of this material revealed no appreciable effect. It is tempting to conclude from this that sonochemistry has no effect on involatile materials in solution.This is not entirely correct because bubble collapse produces very large shear forces in the surrounding liquid capable of breaking the chemical bonding in polymeric materials dissolved in the fluid.’ Over the last few years, increasing interest has been shown in this procedure since the net result of polymer- chain rupture is a pair of macroradicals, which may recombine randomly (resulting in a reduction in molar mass and possibly leading to a monodispersed system) or act as a radical site on which to polymerise another monomer added to the solution (resulting in block copolymerization). Small accelerations, in the range 4-15%, have been found for the rate of acid catalysed hydrolysis of a number of esters of carboxylic acids.’ In the case of methyl ethanoate the effects (at 23 kHz) were attributed to the increased molecular motion induced by the pressure gradients associated with bubble collapse. Similarly, the hydrolysis of the 4-nitrophenyl esters of a number of aliphatic carboxylic acids at 35 “C showed ultrasonically (20 kHz) induced rate enhancements which were all in the range of 14-15% (Scheme 1).The activation energy for the hydrolysis of each of the substrates varied considerably with the alkyl substituent (R = Me, Et, Prl, Bur) on the carboxylic acid and so the uniform increase in rate could not be associated with any cavitational heating effect. Here again, the modest sonochemical effect was considered to be the result of mechanical effects.Scheme 1 The effect of ultrasonic irradiation on the hydrolysis of 2-chloro-2-methylpropane in mixed aqueous ethanolic solvents Chemical Society Reviews, 1997, volume 26 445 of different compositions revealed more evidence for the influence of mechanical effects. 1 The rate enhancement induced by ultrasound (at 20kHz) was found to increase with increase in the alcohol content and to decrease as the reaction temperature was raised. A maximum rate increase of 20-fold was observed at 10 "C in 50% (m/m) solvent composition. This composition is closely coincident with the structural maximum for the binary ethanol-water solvent system. It is logical to suppose that if the sonochemical enhancement is associated with solvent disrup- tion then the maximum effect would be observed at this composition.4.1.2 Heterogeneous systems In any heterogeneous system cavitation which occurs in the liquid phase will be subject to the same conditions as have been described above for homogeneous reactions. There will be a difference however when bubbles collapse at or near any interface and this will depend upon the phases involved. If cavitation bubbles are formed at or near to any large solid surface the bubble collapse will no longer be symmetrical. The large solid surface hinders liquid movement from that side and so the major liquid flow into the collapsing bubble will be from the other side of the bubble. As a result of this a liquid jet will be formed which is targeted at the surface with speeds in excess of 100 m s-* (Fig.7). The mechanical effect of this is equivalent to high pressure jetting and is the reason why ultrasound is so effective in cleaning. Depending upon the conditions used this powerful jet can activate surface catalysis, force the impregna- tion of catalytic material into porous supports and generally increase mass and heat transfer to the surface by disruption of interfacial boundary layers. Fig. 7 Cavitation near to a solid surface For this reason the use of ultrasound in conjunction with almost any electrochemical process will be beneficial and has been the subject of extensive study. The subject has become known as sonoelectrochemistry. 12 The particular advantages which accrue include (a)degassing at the electrode surface, (b) disruption of the diffusion layer which reduces depletion of electroactive species, (c) improved mass transport of ions across the double layer and (4continuous cleaning and activation of the electrode surfaces.All of these effects combine to provide enhanced yield and improved electrical efficiency. When the solid is particulate in nature, cavitation can produce a variety of effects depending on the size and type of the material (Fig. 8). These include mechanical deaggregation and dispersion of loosely held clusters, the removal of surface coatings by abrasion and improved mass transfer to the surface. Mechanical deagglomeration is a useful processing aid and is illustrated in the effect of sonication (in a bath) of titanium dioxide pigment in water.A powder sample made up in water consisting initially of agglomerates (volume mean diameter ca. 19 pm) was rapidly broken apart ( < 30 s) to provide a limiting size of 1.6 pm particles. Furthermore, the sonicated sample 446 Chemical Society Reviews, 1997, volume 26 showed no sign of re-agglomeration even after being allowed to stand for a period of 24h. Fig. 8 Cavitation in a particulate medium The 0-alkylation of 5-hydroxychromones is a difficult process probably as a result of hindrance to ionisation caused by hydrogen bonding between the carbonyl and OH group coupled with some dispersion of the resulting phenoxide 0-charge. Thus, using 5-hydroxy-4-oxo-4H-l-benzopyran-2-carboxplic acid ethyl ester as substrate in N-methylpyrrolidinone (NMP) a low yield (28%) of the 0-propyl product is obtained after 1.5 h at 65 "C using 1-iodopropane and potassium carbonate as base (Scheme 2).7 The yield was greatly increased under sonication OH 0 OR 0 PrnI/K2C03 NMP COOEt COOEt Scheme 2 (probe 20kHz) and the scope of the reaction was expanded by using a range of different haloalkanes (Table 1).Power ultrasound would be expected to be effective in enhancing this reaction via the reduction of the particle size of KzC03 powder. This factor was investigated by first sonicating NMP containing K&03 at 65 "C. The appropriate proportions of 1-iodopropane and chromone were then added to the resulting very fine suspension and the reaction was run under conventional conditions.This resulted in around 90%product formation in 90 min at 65 OC, a reactivity similar to that obtained under continuous ultrasonic irradiation except that the reaction was approaching a definite limit at 90% yield. The fall-off suggests that the surface of the remaining K2C03 had become deacti- vated, and this was confirmed when sonication of the residual mixture rapidly completed the reaction. Table 1 0-Alkylation of a hydroxychromone Alkyl group Yield (stirred) (%) Yield (sonicated) (%) PrnI 28 100 Bun1 57 97 BnBr 59 97 GLC yields, 90 min in NMP at 65 OC, sonication with 20 kHz probe system. A study of the Ullmann coupling reaction has provided evidence that the mechanical effects of surface cleaning coupled with an increase in surface area cannot fully explain the extent of the sonochemically enhanced reactivity.The reaction of 2-iodonitrobenzene to give a dinitrobiphenyl using conven- tional methodology requires heating for 48 h and the use of a tenfold excess of copper powder (Scheme 3). The use of power ultrasound affords a similar (80%) yield in a much shorter time (1.5 h) using only a fourfold excess of copper.7 During these Scheme 3 studies it was observed that the average particle size of the copper feI1 from 87 to 25 pm but this increase in surface area was shown to be insufficient to explain the large (50-fold) enhancement in reactivity produced by ultrasonic irradiation.The studies suggested that sonication assisted in either the breaking down of intermediates and/or the desorption of products from the surface. An additional practical advantage was that sonication prevented the adsorption of copper on the walls of reaction vessels, a common problem when using conventiona1 methodology. The collapse of cavitation bubbIes at or near the interface of immiscible liquids will cause disruption and mixing, resulting in the formation of very fine emulsions (Fig. 9). This is essentially a mechanical effect but it has been utilised in the hydrolysis of benzoate esters where the emulsion was produced by a probe system1' Using 10% NaOH the conventional hydrolysis (Scheme 4) under refhx, gave a very low yield after 90 min; however sonication at room temperature afforded near complete hydrolysis in 1 h.Fig. 9 Cavitation in a two phase liquid medium Me+ooMe -10% NaOH Me Me Scheme 4 Other applications of sonochemically induced emulsification are in phase transfer catalysis, emuIsion polymerisation and two phase enzymatic syntheses. 4.1.3 Reuctions 'switched' by ultrasound An extremely good way of demonstrating that sonochemistry is different from other methods of enhancing chemicaI reactions is to find specific reactions for which ultrasound has changed the product distribution. The first report of a reaction exhibiting 'sonochemical switching' came from Ando et aE.14 The system consisted of a suspension of benzyI bromide and alurnina- supported potassium cyanide in toluene as solvent (Scheme 5).The aim was to produce benzyl cyanide by nucleophilic displacement of the bromine by supported cyanide. Under stirring alone the reaction provided diphenylmethane products ilia a Friedel-Crafts reaction between the brorno compound and the solvent, catalysed by Lewis acid sites on the surface of the solid phase reagent. In contrast, sonication of the same constituents produced only the substitution product, benzyI cyanide. The explanation for this was based upon cavitation producing a structural change to the catalytic sites of the solid support, possibly by masking them through cavitationaIly induced cyanide absorption. stir 24 h 75% \ A sonicate 24 h 77% Scheme 5 The same group have reported an example of sonochemical switching in a homogeneous reaction.The decomposition of lead tetraacetate in acetic acid the presence of styrene at 50 "C generates a small quantity of diacetate via an ionic mechanism. Under otherwise identical conditions sonication of the mixt.pre gives 1-phenyfpropyl acetate predominantly through an inter- mediate methyl radical which adds to the double bond (Scheme 6).15 These results are in accord with the proposition that radical processes are favoured by sonication. I OAc OAc O+b+\ \3 radical mixed ionic pathway pathway pathway Scheme 6 Another example of sonochemical switching and is found in the Kornblum-Russell reaction (Scheme 7). 4-Nitrobenzyl bromide reacts with 2-lithio-2-nitro-propane via a predomi- nantIy polar mechanism to give, as a final product, 4-nitrobenz- aldehyde.16 An alternative SET pathway exists in this reaction leading to the formation of a dinitro compound. Sonication changes the normal course of the reaction and gives preferen- tially the latter compound, in amounts depending on the irradiation conditions and the acoustic intensity. Scheme 7 A sonochemical switch has also been observed in the formation of the indanone nucleus from o-ally1 benzamides (Scheme 8).17 The ketyl radical anion cyclizes to 2-methylin-danone and liberates an amide ion which deprotonates the ally1 moiety. The resulting carbanion then undergoes cyclization to a-naphthol. Under sonication the first step of the process is accelerated and the ketyl is generated much more rapidly so that only the cyclization to 2-methylindanone occurs.The Kolbe electroIysis of cyclohexanecarboxylate in aqueous methanol generates a mixture of products in which bicyclohexyl Chemical Society Reviews, 1997, volume 26 447 90%0 stirredIIq:2 <+ Li/THF ultrasound 90% Scheme 8 predominates (49%). In the presence of ultrasound (38 kHz) however the product distribution was changed quite sig- nificantly reducing the yield of bicyclohexyl to only 7.7% (Scheme 9).12 The major products were the result of two electron processes through a cyclohexane carbocation which gave cyclohexene (34%) by elimination and cyclohexyl methyl ether (32%) by solvent attack.A characteristic of many sonoelectrochemical processes is that the average cell potential under sonication is less than that required conventionally. In this case a current density of 200 mA cm-2 could be maintained at a potential of 7.3 V with ultrasound compared with 8.3 V under silent conditions. /-GO 2 J single-electron product I I -€-t two-electron products Scheme 9 5 Some applications of ultrasound in synthesis 5.1 The activation of metals Ultrasound can be used to accelerate reactions involving metals through surface activation which can be achieved in three ways (a) by sonication during the reaction, (b) as a pre-treatment before the metal is used in a conventional reaction or (c) to generate metals in a different and more reactive form.A classic use of ultrasound is in the initiation and enhance- ment of synthetic reactions involving metals as a reagent or catalyst. One such example is the preparation of a Grignard reagent. A long-standing problem associated with Grignard reagent synthesis is that in order to facilitate reaction between the organic halide and the metal in an ether solvent all of the reagents must be dry and the surface of the magnesium must be clean and oxide free. Such conditions are difficult to achieve and so many methods of initiating the reaction have been developed most of which rely on adding activating chemicals to the reaction mixture. A very simple method of initiating the reaction is by sonication of the reaction mixture in an ultrasonic bath which avoids the need for the addition of chemical activators. The quantitative effects of ultrasound on the induction times for the formation of a Grignard reagent using magnesium turnings in various grades of ether have been examined (Scheme lO).l8 Using damp, technical grade ether 448 Chemical Society Reviews, 1997, volume 26 Br I CH3-CH2-CH-CH3 MgBr Mg turnings I F CH3-CH2-CH-CH3 ether Scheme 10 ultrasonic irradiation is able to initiate Grignard formation in under 4 min compared with several hours using the traditional method involving periodic crushing of the metal.The formation of cyclopropanes through the SimmonsSmith reaction involving zinc dust and CH212 and an alkene suffers from several experimental drawbacks some of the major ones being the sudden exotherm which occurs after an unpredictable induction period, foaming and the difficulties in removing finely divided metal from the reaction products.The conven- tional method for enhancing this reaction relies upon activation of the zinc metal by using it in the form of a zinc-silver or zinc- copper couple and/or using iodine or lithium in conjunction with the metal. The experimental difficulties have been eliminated using a sonochemical procedure where no special activation of the zinc was required and good and reproducible yields were obtained using zinc metal in the form of mossy rods or foil (Scheme 1l).'9 CH3(CH2)7CH=CH(CH&COOCH3 Zn/CH2I21CH3( CH&CH -CH( CH2)7COOCH3 V Scheme 11 The dehydrogenation of tetrahydronaphthalene to naph-thalene using 3% Pd/C in digol under the influence of sonication is accelerated by ultrasonic irradiation (Scheme 12).20 The conventional thermal reaction in digol at 200 "C reached 55% conversion in 6 h (but thereafter reaction ceased) and only 17% reaction was obtained in the same time at the lower temperature of 180 "C.Under sonication at 180 "C the reaction reached completion in 6 h. Pulsed ultrasound (at 50% cycle) was as effective as continuous sonication and even a 10% cycle gave over 80% yield. These results offer considerable energy savings, particularly on processes carried out on a large scale. Scheme 12 Surface activation is of great use in catalysis where metal powders such as nickel, which are generally poor catalysts, can be activated by sonication before use.Normally, simple nickel powder is a reluctant catalyst for the hydrogenation of alkenes yet ultrasonic irradiation offered a reactivity comparable with Raney nickel.21 In this case, ultrasound gave an unexpected decrease in surface area due to aggregation of particles, with electron micrographs indicating a smoothing of the nickel surface. Auger electron spectroscopy revealed an increase in the nickel/oxygen ratio at the surface. The explanation suggested was that abrasion from interparticle collisions removes the oxide layer of the nickel giving the observed increased reactivity. A simple pre-sonication of 3 pm nickel in ethanol prior to use is quite capable of converting this powder from an extremely poor into an acceptable catalyst for the conventional hydrogenation of oct- 1-ene.The reduction of metal salts to a finely divided very reactive free metal generally involves refluxing the metal salt in THF with a very active metal like potassium. The conditions for the production of these so-called Rieke powders can be ameliorated using ultrasound such that equally reactive metal powders can be produced using lithium in THF at room temperature. An example of the use of sonochemically generated Rieke powders is in the preparation of organosilicon compounds (Scheme 13).?2 C13SiH (94%) 'Rieke' Ni powder Scheme 13 A novel method of generated finely divided zinc metal is by the use of pulsed sonoelectrochemistry using an ultrasonic horn as the cathode.23 Normal electrolysis of ZnC12 in aqueous NH4Cl affords a zinc deposit on the cathode.When the electrolysis is pulsed at 300 ms on/off and the cathode is pulsed ultrasonically at a 100:200 ms on/off ratio the zinc is produced as a fine powder. This powder is considerably more active than commercial zinc powder, e.g. in the addition of ally1 bromide to benzaldehyde (Scheme 14). H\ /OH Scheme 14 5.2 Enzymatic syntheses An area of sonochemistry which is deserving of far greater attention is the use of ultrasound to modify enzyme or whole cell reactivity. High power ultrasound will break biological cell walls releasing the contents but it can also denature enzymes.It is therefore very important that when ultrasound is used in conjunction with biological material the conditions of sonica- tion must be carefully regulated. Controlled sonication has been used to 'stimulate' a suspen- sion of baker's yeast to provide an inexpensive source of sterol cyclase (Scheme 15, Table 2).'4 This technique provides an Scheme 15 enantioselective enzymatic synthesis of a sterol in gram quantities. Significantly, sonication has no effect on the activity of the isolated cell-free cyclase system, a tesult which demonstrates how cell membrane disruption can occur without damage to the contents. Table 2 Conversion of squalene oxide to sterol with baker's yeast Enzyme source Conversion (%) Enantiomer conversion (96) Whole yeast 9.5 19 Pre-treated yeast" 41.9 83.9 Presonication at 0 "C using a probe system (20 kHz) for 2 h.Enzymatic reaction at 37 "C, 12 h. When an enzyme is used in a two phase synthesis one of the important requirements is an efficient emulsification/mixing system. Sonication provides such a method which has been used in the synthesis of peptides (Scheme 16).25 The methodology is effective using different solvent combinations (Table 3). BOC-Gly + Phe-N2H2Ph Papain Aqueous emulsion I BOC -Gly- Phe-NzHzPh Scheme 16 Table 3 Dipeptide synthesis in an aqueous emulsion" Organic phase Stir Sonicate Diethyl ether 71 89 Light petroleum 12 62 u Water (75%) with organic solvent (25%) at 37 "C 12 h, 38 kHz ultrasonic bath.Another and probably the most spectacular example of the correct choice of optimized sonicating conditions has been reported for the microbial conversion of cholesterol to chol- estenone (Scheme 17).26 Optimum conditions involved irradia- tion pulses of 2.8 W power applied for 5 s each 10 min and this gave a 40% yield increase. Scheme 17 5.3 Phase transfer and related reactions The effect of cavitation on a suspended solid has been described above (section 4.1.2). Such effects become very important in the case of reactions involving solid-liquid phase transfer catalysis. The N-alkylation of indole with RBr [R = CH3(CH2)11] in toluene at 25 "C in the presence of solid KOH produces a 19% yield in 3 h using rut-butylammonium nitrate (Scheme 18).This yield is substantially improved by sonication to around 90% after only 80 rni11.~7 RBr/KOH (solid) H R Scheme 18 In some cases sonochemistry can completely remove the need for PTC as is the case in the generation of dichlorocarbene by the direct reaction between powdered sodium hydroxide and chloroform at 40 "C using an ultrasonic bath.28 Under these conditions styrene can be cyclopropanated in 96% yield in 1 h when a combination of both sonication and mechanical stirring is used. Significantly the yield is much reduced to 38% in 20 h with sonication alone because the power of the bath is not sufficient to disperse the solid reagent into the dense chloroform (Scheme 19).One route to amino acids is via the synthesis of aminonitriles. The direct reaction between an aldehyde, KCN and NH4C1 in acetonitrile leads to a mixture of products but in the presence of alumina and sonication the reaction can be made more specific (Scheme 20).29 In the case of benzaldehyde the yield of the Chemical Society Reviews, 1997, volume 26 449 0 0NaOH (solid)/CHC13 Scheme 19 CN O+ H I I I H-C-OH H-C-OH H-C-NH2C' Scheme 20 target aminonitrile is poor under normal stirred conditions with benzoin and hydroxynitrile predominating (Table 4). The presence of alumina suspended in acetonitrile increases the proportion of aminonitrile but the overall results make it clear that the optimum reaction conditions require the presence of suspended alumina together with sonication and then the yield of target aminonitrile reaches 90%.Table 4 Strecker synthesis of an aminonitrile0 Conditions Cyanohydrin Benzoin Aminonitrile Stir 38 21 6 Stir + A1203 19 9 64 Sonicate 45 22 23 Sonicate + A1203 3 7 90 0 25 "C. 38 kHz ultrasonic bath. 5.4 Miscellaneous synthesesSynthetic applications of the sonolysis of iron carbonyl which lead to useful ferrilactones synthons have been described (Scheme 21). These are prepared easily and in good yields from vinyl epoxides and either iron pentacarbonyl or, for conven- ience and safety, diiron nonacarbonyl. The use of ferrilactones together with ultrasonically assisted reactions of samarium diodide and sodium phenylcyanide in natural product syntheses have been reviewed.30 + ,CO),Fe4 RO/-v (I) R'NH2 /Lewis acid I/ co R' Scheme 21 A rather difficult double Wittig reaction (Scheme 22) has been effected with enhanced efficiency under sonication.3 The process constitutes a novel type of annelation of an aromatic ring when applied to o-quinones.It is possible to considerably simplify experimental procedures with ultrasound which allows the use of bases which are insensitive to moisture or air. Scheme 22 Trialkylboranes are generally obtained through the stepwise reaction of borane with an alkene. With hindered alkenes however the reaction is very slow. Sonication promotes rapid reaction even with highly hindered substrates (Scheme 23).32 Synthetic applications of this technique include the hydrobor- ation/oxidation of vinyl groups.Neat 99% yield 1 h ultrasonic bath (5 h normal) * HO*/ -OTBDMS OTBDMS (i) 9-BBN ,THF , ultrasound 89 '/o(ii) NaOH , H202 Scheme 23 Sonochemistry has been used to improve a Friedel-Crafts alkylation reaction used for the synthesis of the anti-inflamma- tory agent ibuprofen (Scheme 24).33 When performed under classical conditions (2 h at 25 "C) the reaction afforded only 17% yield and for this reason the normal synthesis is via a less direct route. Under the influence of ultrasound, using a cleaning bath, but under otherwise identical conditions the yield was improved to 50%. YOOH A A Scheme 24 6 Conclusions Sonochemistry is an expanding field of study which continues to thrive on outstanding laboratory results.34 Applications can be found over a range of chemical systems, however it is in 450 Chemical Society Reviews, 1997, volume 26 heterogeneous reactions that sonochemical syntheses are most widely developed.The potential improvements afforded by sonication suggest that all chemical laboratories nowadays should be equipped with at least one small cleaning bath for simple trials While an empirical understanding of the subject has taken sonochemists a long way towards predicting possible applica- tions considerable attention is currently being paid to gaining an understanding of what actually goes on in the collapsing bubble and in its immediate environment.In this area the chemists are finding very fruitful cooperation with engineers, physicists and mathematicians making sonochemistry a truly interdisciplnary study Recent laboratory studies have revealed that for a few heterogeneous reactions high speed stirring has a similar effect to sonication 35 Thus in the cyclopropanation of styrene (Scheme 19)the yield can be improved from 3% with magnetic stirring through 20% at 8000 rpm to 70% at 24000 rpm Such results are intriguing in that they confirm the importance of mass transfer in sonochemistry and could suggest that high speed stirring involves hydrodynamic cavitation Unlike sonica- tion however stirring at such very high speeds is unlikely to become a viable prospect in industry Whatever the real laws of sonochemistry might be it is clear that sonochemistry has arrived, that sonochemistry is expanding and that chemists from all disciplines will find within the subject plenty that will be of interest to them.7 References T J Mason and J P Lorimer, Sonochemisti y, Theoi y, Applications and User of Ultiasound in Chenzistiy, Ellis Horwood Publishers, Chichester, 1988 Ultiasound rty physical hioloqicul and chemical effects, ed K S Suslick, VCH, Mannheim, 1988 C Einhorn, J Einhorn and J -L Luche, Synthesis--Stuttgait, 1989, 787 Sonochemisti y The uses oj ulti asound in chemist!y, ed T J Mason, Royal Society of Chemistry, Cambridge, 1990 Current tiends in sonochemistry, ed G J Price, Royal Society of Chemistry, Cambridge, 1993 Special edition of the journal Ultrasonics covering the RSC Sonochem istry Symposium, Warwick 1986, Ultrasonics, 1987, 25, January iFsue T J Mason, Piac tical Sonochemistry A useis guide to applications in c henustry and c hemical enqineei inq, Ellis Horwood Publishers, Chich- ester, 1991 J -L Luche, Sonochemistry, from experiment to theoretical con siderations, Adtances in Sonothemisrig, ed T J Mason, JAI Press, London, 1993, vol 3, p 85 9 P Riesz, Free radical generation by ultrasound in aqueous solutions of volatile and non-volatile solutes, Ad1 ances in Sonochemistr y, ed T J Mason, JAI Press, London, 1991, vol 2, p 23 10 K S Suslick, S -B Choe, A A Chichowlas and M W Grimstaff, Nature, 1991, 353, 414 11 T H Hyeon, M M Fang and K S Suslick, J Am Chem Soc , 1996, 118,5492 12 D J Walton and S S Phull, Sonoelectiochemistiy, AdIanceJ in Sonochemistiy, ed T J Mason, JAI Press, London, 1996, vol 4, p 205 13 S Moon, L Duchin and J V Cooney, Tetiahedion Lett, 1979, 20, 3917 14 T Ando and T Kimura, Ultrasonic organic synthesis involving non- metal solids, Adlances in Sonochemistiy, ed T J Mason, JAI Press, London, 1991, vol 2, p 21 1 15 T Ando, P Bauchat, A Foucaud, M FUjita, T Kimura and H Sohmiyd, Tetr uhedi on Lett , 1991, 32, 6379 16 M J Dickens and J L Luche, Tetr ahedi on Lett , 199I, 32,4709 17 J Einhorn, C Einhorn and J -L Luche, Tetiahedioii Lett, 1988, 29, 2183 18 J D Sprich and G S Lewandos, Inor-g Chim Acts, 1982, 76, 1241 19 0 Repic and S Vogt, Tetr ahedi on Lett , 1982, 23, 2729 20 T J Macon,J P Lonmer,L Paniwnyk,P W WrightandA R Harris, .I Cutal , 1994, 147, 1 21 K S Suslick and D J Casadonte, .I Am Chem SOL , 1987, 109, 3459 22 W L Parker.P Boudjouk and A B Rajkumar, I Am Chenz Soc . 199I. 113,2785 23 A Durant, J L Delplancke, R Winand and J Reisse, Teti ahedr 017 Lett , 1995,36,4257 24 J BUjOns, R Guajardo and K S Kyler, J Anr Chem Soc . 1988, 110, 604 25 K Tadasa, Y Yamamoto, I Shimoda and H Kayahara, J Fac Agrrc Shinshu UniL , 1990, 26, 21 26 R Bar, Biotechnol Bioeng , 1988, 332, 655 27 R S Davidson, Ultiasonics, 1987, 25, 35 28 S L Regen and A Smgh, J 01y Cheni , 1982,47, 1587 29 T Hanafusa, J Ichihara and T Ashida, Chem Lett, 1987, 687 30 C M R Low, Ultrasonics Sonochemistiq, 1995, 2, 153 31 C Yang, D T C Yang and R G Harvey, Syn Lett, 1992,799 32 G E Keck, A Palani and S F McHardy, .I 01q Chem , 1994, 59, 31 13 33 C Garot, T Javed, T J Mason, J L Turner and J W Cooper, Bulletin des Societes Chimiques Belqes, 1996, 105, 755 34 T J Mason and J -L Luche, Ultrasound as a new tool for synthetic chemists, Chemist, I’ undei E.iti eme oi Non classic ul Conditions, ed R van Eldick and C D Hubbard, John Wiley, New York, 1997,p 317 35 J Reisse, presented at NATO Advanced Study Institute on Sonochem istry and Sonoluminescence, Leavenworth, Washington, USA, August 1997 Received, 9th May I997 AI cepted, 30th June I997 Chemical Society Reviews, 1997, volume 26 451

ISSN:0306-0012    

DOI:10.1039/CS9972600443

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Preparation of seven and larger membered heterocycles by electrophilic heteroatom cyclization Paris-Sud, 91405 Orsay, France The preparation of seven and larger membered ringcompounds from linear substrates is known to be a difficult process. We report results for the preparation of heterocycles with these ring sizes by electrophilic heteroatom cycliza- tions. Examination of the literature shows that the obtention of seven-membered compounds is possible in good yields by a proper choice of the electrophilic reagents, while for eight- and larger membered compounds these cyclisations are possible only if steric constraints are introduced into the chain. 1 Introduction Since the discovery of the iodolactonization reaction of P,y-and y,S-unsaturated carboxylic acids by Bougault,l electrophilic heteroatom cyclization has been the subject of much research.2 However, if the formation of three to six membered ring heterocyclic compounds has been reported (Scheme l), few examples could be found in the literature for larger rings, until recently. This absence of results was indeed the result of unsuccessful attempts.3 exo-cyclisation endmyclisationG = 0,N, S E+ = H+,X+ (CI, Br, I),RSe+,RS+, M"+ (Ag, Hg,TI, Pd, Pb, Te...), ...Scheme 1 The difficulty in preparing seven and medium ring com- pounds by cyclization methods has been known for some time.4 In an interesting study involving the formation of lactones (ring sizes 3 to 23) from o-bromoalkanoates, Illuminati and Man- dolini5 have illustrated this problem well (Fig.1). A maximum rate of cyclization was observed for the formation of Dr Ge'rard Rousseau obtained his PhD at Orsay under the super- vision of Professor J. M. Conia in 1976, and then carried out his post-doctoral research at Hal--vard university with Professor R. B. Woodward. After his retui-n to Orsay, he ~~or-ked successively on the chemistry of ketene acetals, the utilisation of enzymes in or- ganic chernistiy, and more re-cently the chemistry of medium ring compounds. He is director of research at the CNRS. Dr Gerard Rousseau 106 .::... 1051..... " 0 5 10 15 20 25 Ring size Figure 1 y-butyrolactone and then the rate decreased dramatically for the 7-(by a factor of 104) and 8-membered ring lactones (by a factor of 106).A slow increase in the cyclization rate was then observed. From a synthetic point of view, these constant rate values mean that good yields should be expected for the formation of 4-6 membered ring lactones, while low to very low yields should be obtained in the other cases, To circumvent this intrinsic factor, different solutions have been sought. The most obvious, i.e. the use of high-dilution conditions is not applicable in this case, due to the fact that the kinetics for electrophilic cyclization are second order. The solution to this problem must be found in the modification of the substrates and/or the electrophiles. Nucleophilic atoms which can be used in these electrophilic heteroatom cyclizations are After completing his Freshman and Sophomore years at Santa Monica College (CA), Fadi Homsi transferred to the Uni-versity of Colorado at Boulder where he graduated in 1992.While in Boulder. he workedjor-three years both as un under-graduate and postgraduate on the synthesis oj liquid crystals with Professor David M. Walha. Then he retuiwed to France, and is currently ,finishing his PhD working with Dr Ge'rurd Rousseau.Fadi Homsi Chemical Society Reviews, 1997, volume 26 453 mainly 0, N and S. No example with S is known for the formation of rings larger than six. In Table 1 we report the common names of medium ring heterocycles which are not familiar to the majority of organic chemists. The mechanism of electrophilic cyclizations has not yet been completely elucidated.Formation of an onium ion or a concerted attack of the nucleophile on a n halogen double bond complex have been proposed.6 It is probable that both mechanisms can exist depending on the nature of the electro- phile. 2 Oxygen as nucleophile 2.1 Formation of cyclic ethers by exo-mode cyclization It has been reported that hept-6-en01 derivatives react with iodine in acetonitrile to give oxepanes and oxocanes in low yields (15-30%).7 In all cases the oxocanes (8-membered ethers) were the minor products (Scheme 2). 1 Scheme 2 These yields were subsequently improved by modifying the substrates and the electrophiles. It was reported that an isoxazoline, substituted in the three position by a trityl group, reacted with iodine in methylene chloride to give an oxepane in 80% yield.8 (Scheme 3).The driving force of this reaction seems to be the formation of the thermodynamically favoured nitrile function. However, this strategy did not allow the preparation of the corresponding oxocane. Nyy-----TGrI2 Ph3C 80% (as: trans, 1:3) Scheme 3 We found that instead of modifying the substrate, it was possible to modify the iodide reagent. Indeed, using bis(co1li- dine)iodine(I) hexafluorophosphate in methylene chloride we observed the iodoetherification of hept-6-en- l-ols in good yields (Scheme 4).9 The success observed in these cyclizations compared to the iodine reagent is due to the absence of a reactive counter anion.No diastereoselectivity was observed. In previous work, 1,4-dioxepane was obtained in only 12% yield using bis(collidine)iodine(I) perchlorate. lo Different substrates were used to study the competition between 3-exo-/7-exo cyclization. Only the 3-ex0 cyclization Table 1 Chemical names of medium ring heterocycles Ring size Ether Lactone 7 Oxepane Oxepan-2-one 8 Oxocane Oxocan-2-one 9 Oxonane Oxonan-2-one 10 Oxecane Oxecan-2-one 11 Oxac ycloundecane Oxacycloundecan-2-one 454 Chemical Society Reviews, 1997, volume 26 HO Fi R = H, alkyl, aryl t 6595% (CIS: trans, 50:50) Scheme 4 was observed which reflects the importance of the entropy factor in these cyclizations. With the same reagent, oxocanes were obtained in moderate yields.This last result illustrates well the difficulty of obtaining 8-membered ring compounds compared to 7-membered ring compounds.’ (Scheme 5). 0 RR as above OH RDR R = H 27% R=Me 24% Scheme 5 We have also examined the preparation of oxepenes by iodoetherification, starting from 3,6-heptadien- l-ols. This kind of substrate is interesting since 4-ex0, 5-end0,7-exo and 8-end0 cyclizations are in competition. Depending on the substitution at the terminal double bond all these modes of cyclization could be observed, albeit in low yields (2444%) (Scheme 6). The unsubstitued substrate (R1 = R2 = H) led mainly to the oxepene (36% yield). Introduction of an E-methyl (Rl = H; R2 = Me) favoured the 5-end0 cyclization (29% yield), while its 2isomer (R1 = Me; R2 = H) led mainly to the oxocene (18% yield).Introduction of two methyls (Rl = R* = Me) had a negative effect and only the 5-endo and 4-ex0 cyclizations products could be detected in low yields (14 and 9% respectively). In fact, contrary to what was expected, substitu- tion at the terminal double bond had a negative effect due to the increase of steric hindrance which disfavoured the formation of 7-or 8-membered rings. This phenomenon appears to be general for these ring sizes (vide infra). The presence of a substituent in the 6-position allowed the exclusive formation of oxepenes in good yields. The corre- Amine Lactam Azepane Azepan-2-one Azocane Azocan-2-one Azonane Azonan-2-one Azecane Azecan-2-one Azacycloundecane Azacycloundecan-2-one I Scheme 6 sponding oxocene was formed, again in low yield.9 (Scheme 7).R = Me 70% R = OMe 60% asabove -py 18% Scheme 7 It has been reported that this cyclization could also be carried out with silanols, but apparently with lower yields than with alcohols. The presence of substituents on the carbon-carbon double bond allowed the competitive formation of endo cyclization products1 1 (Scheme 8). u i : P h as above L Ph Other reagents were tested for the formation of oxepanes. Biomimetic cyclizations of different dienols were attempted for a short synthesis of aplysistatin. Low yields (25-30%) were observed using mercury(I1) trifluoroacetate or tetrabro-mobenzoquinone as electrophiles.12 The cyclization was im- proved using conformationally more rigid molecules (66-92% yields)I2" (Scheme 9). 18-32'/0 Scheme 10 While NBS or NIS are unable to induce the formation of oxepanes from hept-6-enols,Y it was recently reported that 1,2,4-trioxepanes could be obtained in moderate yields by cyclization of the corresponding hydroperoxides. This partial success is probably due to the higher nucleophilicity of the terminal oxygen of the hydroperoxide compared with that of the alcohol (Scheme 10). 13 Phenylsulfenoetherification was reported to be a possible route for the formation of oxepanes, using phenylthiomor- pholine in the presence of trifluoromethanesulfonic acid to generate the episulfonium ion (Scheme 1 l).14 In this case also, Ph, ,Ph Ph, ,Ph 60% 14% Ph, ,Ph Ph ph I 8YO 27% Scheme 8 aplysistatin Hg(CF3C00)2 HgOCOCF3 * Br Br CO~BU' OPNBH OPNB Scheme 9 Chemical Society Reviews, 1997, volume 26 455 /\ PhS-N 0 U HOP * CF3SO3H / CHzCI2 W 80% as above * @Sph + PhCH: 74% ""OC H2P h 3 2 Scheme 11 CH2SePh PhSeOS02C F3 &OH toluene 23 % (cstrans,85:15) 0 &N -SePh 0 I2 cat./CH2CI2 56'/o Scheme 12 compounds were isolated as side or main products.It appears that for this cyclization, thallium triacetate was more efficient,20 since the oxepane was obtained as a unique product (Scheme 15). Oxocanes could not be obtained with these conditions. \ While with thallium triacetate no 6-exo-/7-endo competition SePh P 71'/o Scheme 13 the presence of the substituents on the terminal carbon of the double bond allowed, surprisingly, the competitive formation of Z = CH2 R = Ph 79% 4% oxocanes.Z=O R=Me 70% 20% Phenylselenoetherification also appears to be an efficient method to prepare oxepanes, though few examples are availa- ble. Cyclization was observed in the reaction of unsaturated alcohols with PhSeOS02CF3 or N-phenylselenophtha1imideI6 (Scheme 12). Phenylselenobromide was found to be efficient for the formation of 1 ,3-dioxocane.17 The success of this cyclization 72% seems due in part to the presence of an electronically enriched Scheme 15 carbon-carbon double bond (Scheme 13).(Compare, however, with the example shown in Scheme 7). This is the best example was observed during the cyclization, with benzene tellurenyl reported to date concerning the preparation of an oxocane by acetate (PhTeOAc) such a competition did take electrophilic cyclization. Palladium was also found to be useful for the cyclization of Phenylselenochloride was equally efficient for the cycliza- heptenols and led to the formation of 1,6-disubstitued oxepanes tion of 1,7-octadiene in the presence of water18 (Scheme 14). (Scheme 16).22 Alkenols are in general inert in the presence of TsOH. It was reported that if a silicon atom was fixed on the carbon-carbon 2 PhSeCl * double bond, then the cyclization could take place and an CH3CNIH20 PhSen S e P h oxepane was isolated in low yield.2-3 Such a reaction could probably be improved by using more efficient electrophiles 34% (Scheme 17).Scheme 14 Electrophilic cyclizations are not limited to double bonds, and examples starting with acetylenic and allenic derivatives Treatment of hept-6-enols with lead tetraacetate was found to which are generally more reactive24 have been reported for the lead in some cases to oxepanes.19 However, very often these formation of 5-and 6-membered heterocycles.* Study of such electrophilic cyclizations were not specific and rearrangement substrates for the obtention of larger ring sizes should give 456 Chemical Society Reviews, 1997, volume 26 HOh 6kO-CH3 0 I (dba)3Pd2CHC13 PPh3/THF 48% Scheme 16 17% Scheme 17 interesting results (see for example the results reported in Section 3). 2.2 Formation of cyclic ethers by endo-mode cyclization As we have seen in the Introduction, the presence of a highly carbocationic stabilising group on the carbon-carbon double bond can favour an endo-mode cyclization.For example, we found that the oxepanes were favoured over the more stable tetrahydropyrans if a methoxy group was fixed to the double bond. This endu-cyclization was also exclusively observed for the formation of oxocanes, however, in this case the ring closure was less efficient (Scheme 18).9 We attribute this last result to HO OMe CH2C12 -V' n = 1 84% n =2 23% Scheme 18 the fact that introduction of a methoxy group (or another substituent) on the terminal carbon atom of the double bond is unfavourable, due to the increase in steric hindrance.This effect seems more pronounced for the formation of 8-membered ring heterocycles (and probably for the larger ring sizes) compared to smaller ring sizes. This substituent effect also seems to be responsible for the 7-end0 cyclization of allenic derivatives, using silver tetrafluo- roborate, instead of the more obvious 6-ex0 cyclization.25 In the same way a 7-end0 cyclization was observed for the reaction of a highly sensitive phenol with phenylselenochloride26 (Scheme 19). Palladium also appears to favour the endo-cyclization mode, since the oxepane was found as the unique product in the cyclization of hex-5-en- 1-01.27 Unfortunately, the scope of this reaction was not examined (Scheme 20).2.3 Formation of lactones When we started our work in this field nothing was known about the possibility of preparing medium ring lactones (mediolides)? t We suggest the term mediolide for medium ring lactones, in comparison with macrolides for large ring lactones. 70% PhSeCl -EtOAct-78 "C 32% Scheme 19 Bun3SnAr/CuCI2 [PdC12*(PhCN)2] * OAr 73% Scheme 20 by an electrophilic cyclization. Only rare examples were reported concerning E-caprolactones and large ring lactones. After several unsuccessful attempts,3 the first E-caprolactones were obtained from substrates possessing a certain rigidity brought by proline or benzo rings28 (Scheme 21).NBS DMF * 0 89% N$cN COOH NC > 60% Scheme 21 Reaction of phenylselenochloride with hept-6-enoic acid was reported to lead to E-caprolactone. This reagent was inefficient for larger membered ring lactones. However, it was found that N-phenylselenophthalimidecould be used to prepare large ring lactones (14 to 16) (Scheme 22).29 In these conditions small amounts of the endu-cyclization products were isolated. Important advances in this field were made using bis(col1i- dine)iodine(I) hexafluorophosphate as the electrophile. In a first study we checked the reactivity of 4-oxahept-6-enoic acid with different iodine reagents. These results are summarised in Scheme 23. The superiority of I+(collidine)z PF6- in this reaction (due to the very low nucleophilicity of the anion) led us to study the reactivity of this reagent with a wide range of unsaturated acids.While hept-6-enoic acids led in high yields to the corresponding E-caprolactones, the results were much more disappointing for larger ring sizes. Indeed the 8-and 1 1-membered ring lactones were obtained in low yields (Scheme 24).30 We decided to examine which kind of structural modifica- tions of the carbon chain would favour the cyclization. Three factors were examined : the oxygen effect, the gem-dimethyl effect and the influence of a carbon-carbon double bond. The introduction of an oxygen atom in the carbon chain is known to favour the cyclization for numerous ring sizes.37 However, an exception was noticed concerning the seven-membered heterocycles. In fact, this negative effect could be suppressed either by adding substituents on the carbon-carbon Chemical Society Reviews, 1997, volume 26 457 C02H PhSeCl CH&12/- 78 "C * O r s e p h 70% 0 @-SePh 0 c C 0 2 H CH2CI2/roomtemp.SePh n =l 50% 14% n =3 54% 15% Scheme 22 double bond or changing the position of the oxygen in the chain (Scheme 25).30 A positive oxygen effect was observed for the formation of larger ring sizes, since lactones in the range 8-13 were obtained in moderate yields. For these ring sizes a competition between the em-and endo-cyclization was often observed (Scheme 26). Comparison of these results with those reported for the selenolactonization reaction (Scheme 22) shows that iodine favours the endo-cyclization mode.We found that all these cyclizations occurred under kinetic control. This means that the decrease of CH-HC intramolecular non-bonding interactions used to explain the oxygen effect3' is not a satisfactory one for these results. We suggested that the oxygen atom induces a stabilisation effect by formation of an intermediate. This stabilisation decreases the activation entropy of the reaction and consequently favours the cyclization.30 The gem-dialkyl effect is well known in the formation of five- and six-membered ring compounds.32 We decided to carry out a study of this effect for the formation of E-caprolactones and mediolides.For caprolactones almost no effect could be detected. However with mediolides a positive effect was observed, since 8-12 membered lactones were formed in moderate yields (Scheme 27). Here, a competition between the em and endo cyclization was also e~tablished.~~ It seems obvious that the presence of an unsaturation on a chain should favour the formation of cyclic compounds. In our case we decided to study the reactivity of a$-unsaturated acids. As reported in Scheme 28, the presence of a cis double bond favours the cyclization. For example, 8-membered lactones were obtained in exceptional ~ields.3~ As we have previously noted, the proportion of endo cyclization increases appreciably with the lactone size. The beneficial effect of a cis double bond was also observed in the k1 R' R2 59% 70% 45% 45 Yo 5% 39% 75% 75% Scheme 25 I2/CH3CN 0% ICYCH2C12 0% NISIAgOTWCH2CIz 0% I+(~ollidine)~C1O4-/CH2CI2 I+(~yridine)~PF6-/CH& I+(collidine)2 PF~-/CHZC~Z 24% (ref.10) 20% 59% R 76% 49% 79% 72% 5% 4% Scheme 23 n =l,R=H n =l,R=Me n = 1, R = Cyclohexyl n =I,R=BU' n =2,R=H n=5,R=H Scheme 24 rn =n =l rn =n =1 rn =n=1 rn =n =1 rn =O,n =2 rn =O,n =2 R', R2, R3 = H R', R2 = H; R3 = Me R' = Me; R2, R3 = H R', R2 = Me; R3 = H R1,R2,R3=H R',R2,=H;R3=Me 458 Chemical Society Reviews, 1997, volume 26 -1 rn =O,n =3 23% 23% rn =1,n =2 40% 5% rn =2,n =I 8% rn =n =2 24% 24% rn =O,n =5 I 8% 18% rn =O,n =6 7% 21% rn =2,n=5 16% 16% Scheme 26 G C O * H CHzCIz rn =3,n =O 23% rn =2,n =1 41yo m =l,n =2 26% 17% rn =3,n =1 19% rn =5,n =I 23% 4% Scheme 27 n=l ao yo n=2 ia yo 12Yo n=3 10Yo 10O/O n=4 6 Yo 9 Yo n =I 83% n =2 40% 22% n =3 30% 30% n =4 20% 36% Scheme 28 formation of an E-caprolactone.In this case the yield was almost quantitative. The presence of a gern-dimethyl group on the terminal double bond explains the large proportion of endo- cyclization (Scheme 29).35 3 Nitrogen as nucleophile Currently, there are very few results concerning the possibility of preparing nitrogen heterocycles larger than six-membered rings by electrophilic cyclization.A hexahydroazepine was obtained using Pdll salts as electrophiles. In this reaction, the presence of the allene unsaturation appeared necessary (Scheme 30).36 With these substrates no reaction was observed in the presence of silver salts. However, starting from an oxime, the cyclization did occur in the presence of silver tetrafluoroborate, giving an unstable nitrone which could be trapped with different dienophiles.3Ga Interestingly, the same authors found that these allenic substrates could lead to 7-1 1 membered azacycles in moderate yields using a two step reaction. Addition of iodine led to the diiodo products which were cyclized by slow addition to a NaH suspension (Scheme 3 1). The preference for the endo products was explained by the easier SN2 substitution of the primary iodide compare to the SN.2 substitution.36b Recently, we investigated the possibility of preparing %membered ring lactams by halolactamization.The cyclization Chemical Society Reviews, 1997, volume 26 459 Pd( PhCN)2C12 c\==N-H CO, MeOH I Ph/ Me’ 60% Scheme 30 0 25% 75% Scheme 29 did occur if a sulfone was fixed on the nitrogen atom to increase its nucleophilicity. We were able to obtain the azocanones in moderate yields (Scheme 32).34 These results open new possibilities in this area, in particular for the preparation of azepanones. 4 Conclusion This article shows that contrary to what was previously thought, it is now possible to obtain heterocycles larger than 6-mem- bered rings by electrophilic cyclization.As we have seen, several reagents are already available for the formation of 7-membered heterocycles. For larger membered heterocycles the situation is more critical. To have a successful cyclization one of the two criteria (or better both) must be fulfilled: (i) choice of the electrophile and (ii) choice of the substrate. For the moment few reagents are available and it appears necessary to 460 Chemical Society Reviews, 1997, volume 26 Ts TS n =1 03% 10% n =2 22% 35 Yo n=3 -55% n=4 -62% n=5 -35% Scheme 31 design new ones. Indeed, the nucleophile part of the reagent should be non-reactive to avoid competition with the nucleo- phile part of the substrate.Also, it is necessay to modify the substrates to decrease the activation entropy of these reactions. This can be done either by formation of an intermediate chelate or by introduction of a steric restraint. As we have seen in this review, depending on the R group present on the C-C double bond, and the nature of the electrophile, it is possible to orientate the cyclization to an exu-mode or an endo-mode. From our experimental results and following Baldwin’s rules37 we can add two complementary rules for ring sizes 2 7. Rule 1 :for these ring sizes both the ex0 and endo cyclizations are favoured. Rule 2: the exo to endo ratio decreases with increasing ring size. This phenomenon is more or less pronounced depending on the electrophile, and the substitution pattern. X= Br 33% X=I 44% Scheme 32 5 References 1 M J Bougault, C R Acad Sci , 1904,139, 864 2 For recent reviews see G Cardillo and M Orena, Tetrahedron, 1990, 46, 3321, K E Harding and T H Tmer, 111 Comprehensive Organic Synthesis, ed B M Trost, Pergamon Press, New York, 1991, vol 4, p 363 3 E E Van Tamelen and M Shamma, J Am Chem SOC, 1954, 76, 23 15 4 J Sicher, Progr Stereochem , 1962,3,202 5 G Illuminati and L Mandohni, Acc Chem Res , 1981, 14, 95 6 A R Chamberlin, R L Mulholland Jr, S C Kahn and W J Hehre, J Am Chem Soc , 1987,109,672 7 G Lassalle and R Grde, CR Acad Sci , 1990, 31011, 907 8 M J Kurth, M J Rodnguez and M M Olmstead, J Urg Chem ,1990, 55, 283 9 Y Brunel and G Rousseau, J Org Chem, 1996,61, 5793 10 R D Evans, J W Magee and J H Schauble, Synthesis, 1988, 863 11 K Takaku, H Shinokubo and K Oshima, Tetrahedron Lett , 1996,37, 6781 12 (a)T R Hoye,A J Caruso, J F DellanaandM J Kurth, J Am Chem Soc ,1982,104,6704,(b)J D White, T Nishiguchi and R W Skeean, J Am Chem Soc, 1982, 104, 3923, (c) H M Shieh and G D Prestwich, Tetrahedron Lett, 1982, 23, 4643 13 Y Ushigoe, Y Kano and M Nojima, J Chem SOC Perkin Trans I, 1997,5 14 P L Lopez-Tudance, K Jones and P Brownbridge, Tetrahedron Lett , 1991,32, 2261 15 H Inoue, S Murata and T Suzuki, Liebigs Ann Chem , 1994,901 16 W P Jackson, S V Ley and J A Morton, J Chem Soc Chem Commun , 1980, 1029, L Xiang and A P Kozikowski, Synfett, 1990, 279 17 M Petrzilka, Helv Chim Acta, 1978, 61, 3075 18 A Toshimitsu, T Aoai, H Owada, S Uemura and M Okano, Tetrahedron, 1985, 41, 5301 19 Z Cekovic, R Saicic and M L Mihailovic, Res Chem Int , 1989, 11, 257 20 M L Mihailovic, R Vukicevic, S Konstantinovic, G Milosauljevic and G Schroth, Liebigs Ann Chem , 1992, 305 21 N X Hu, Y Aso, T Otsubo and F Ogura, Tetrahedron Lett, 1987,28, 1281 22 B M Trost and A Tenaglia, Tetrahedron Lett, 1988,29, 2927 23 K Miura, S Okajima, T Hondo and A Hosomi, Tetrahedron Lett, 1995,36, 1483 24 G Melloni, G Modena and U Tonellato, Acc Chem Res , 1981, 14, 227 25 J Grimaldi and A Cormons, Tetrahedron Lett, 1985,26, 825 26 R M Williams and T D Cushing, Tetrahedron Lett ,1990, 31, 6325 27 Y Tamaru, M HOJO, H Higashimura and Z -I Yoshida, Angew Chenz Int Ed Engl , 1986,25,735 28 (a)S S Jew, S Terashima and K Koga, Tetrahedron 1979,35,2337, (b)J C Bottaro and G A Berchtold, J Org Chem , 1980,45, 1176 29 K C Nicolaou, S P Seitz, W J Sipio and J F Blount, J Am Chem Soc , 1979,101,3885,K C Nicolaou, D A Claremon, W E Barnette and S P Seitz, J Am Chem Soc , 1979,101, 3704 30 B Simonot and G Rousseau, J Org Chem , 1994, 59, 5912 31 L Mandolini, Adv Phys Org Chem 1986, 22, 1 32 A J Kirby, Adv Phys Org Chem , 1980, 17, 183 33 B Simonot and G Rousseau, Tetrahedron Lett, 1993, 34,4527 34 F Homsi and G Rousseau, submitted for publication 35 B Simonot and G Rousseau, Tetrahedron Lett, 1993, 34, 5723 36 (a)T Gallagher, I W Davies, S W Jones, D Lathbury, M F Mahon, K C Molloy, R W Shaw and P Vernon, J Chem Soc Perkin Trans I, 1992, 433, (b)R W Shaw and T Gallagher, J Chem Soc Perkin Trans I, 1994, 3549 37 J E Baldwin, J Chem SOC Chem Commun , 1976,734 Received, 19th May I997 Accepted, 14th July I997 Chemical Society Reviews, 1997, volume 26 461

ISSN:0306-0012    

DOI:10.1039/CS9972600453

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Glycosylation employing bio-systems: from enzymes to whole cells Vladimir KPen and Joachim Thiem Prague 4, Czech Republic; E-mail: kren@biomed.cas.cz Institute of Organic Chemistry, University of Hamburg, 0-201 46 Hamburg, Germany; E-mail: thiem@chemie.uni-hamhurg.de This contribution will highlight chemoenzymatic ap-proaches to the rather complex task of stereospecific and regiospecific glycosylation. Advantages and problems asso- ciated with the application of enzymes from carbohydrate metabolism as organic reagents in aqueous and non-aqueous solvent systems are discussed. One chapter will report on the use of glycosidases in reverse hydrolysis and tranglycosyla- tion reactions. Another section focuses on the syntheses with simple as well as more complex and delicate glycosyl- transferases including cofactor regeneration. Further, effec- tive combination of both degrading and synthesising enzyme systems, and finally glycosylations employing cells with direct access to their complete enzyme equipment are treated.1 Introduction Within recent years the application of enzymes in synthetic organic chemistry has become a rather well-established tech- nique. In fact, there are now textbooks and laboratory manuals available describing many useful and accessible techniques. Besides preparation of chiral synthons, the synthesis and modification of carbohydrates by enzymes is one of the most intensely exploited areas of enzyme applications. Whereas some applications of enzymes in sugar chemistry are indeed simple if not trivial several other sophisticated methods put high demands upon a broad knowledge and skills not only in organic chemistry but also in biochemistry, microbiology and im- munology.Vladimir KFen, born 1956 in Prague, Czech Republic, is a head of the Laboratory of Biotransformation of the Institute of Microbiology, Czech Academy of Sciences, Prague. A graduate of the University of Chemical Technology, Prague, he obtained his PhD from the Czechoslovak Academy of Sciences in 1986 and has worked since that time in the Institute of Microbiology. He spent his postdoctoral years with Professor David H. G. Crout at University of Warwick, UK and then as an A. von Humboldt fellow with Professor Joachim Thiem at the University of Hamburg, Ger- many.His research insterests are bioproduction of fungal secondary metabolites, their biotransformations, and re-cently preparation of enzymes and their use in glycoscience. He was awarded a Matsumae medal from the Matsumae In- ternational Foundation, Tokyo.He is a member of the RSC and author of over 60 papers in international journals. Vladimir KPen This review aims to point out some biological aspects and uses of rather complex, sophisticated systems for carbohydrate chemistry. Glycosylation is considered to be an important method for the structural modification of compounds with useful biological activities. It allows conversion of lipophilic compounds into hydrophilic ones, thus improving their pharmacokinetic proper- ties.Sometimes, by attaching a sugar pharmacodynamic , properties are also changed or novel and more effective drug delivery systems (prodrugs) obtained. Enzymatic glycosylation methods in comparison with chemical methods are especially useful in the glycosylation of complex biologically active substances, where generally harsh conditions or use of toxic (heavy metals) catalysts are undesirable. The enzymatic approach is also a good alternative in the chemistry of food additives where the use of synthetic chemistry is sometimes not acceptable. 2 Glycosidases Glycosidases are cheap enzymes, quite robust to handle, they use cheap donors and show absolute stereoselectivity .The main drawbacks for their use in glycoside synthesis are lower yields and generally low regioselectivity.This, however, can be overcome by the rational choice of appropriate glycosidases with more pronounced regioselectivity and by new sophisti- cated methods of transglycosylation as, e.g. controlled dosing of the reactants and use of co-solvents (Scheme 1). One of the main advantages of using glycosidases in glycosylation is their good stereoselectivity. These enzymes are considered to be ‘retaining enzymes’. There are, however, a few exceptions of inverting glycosidases which lead to products Joachim Thiem, born in Hamburg, received his Dr rer. nat. with Professor Hans Paulsen in 1972. Following his ‘habilitation’ in 1978 he became professor at the Westfalische Wilhelms- Universitat Miinster in 1983.He succeeded Hans Paulsen in the chair in 1989 at the Institute of Organic Chemistry in Ham- burg, spent sabbaticals at vari- ous universities and is asso-ciated with many foreign groups in carbohydrate chem- istry. His research interests, discussed in more than 250 lectures and reported in over 300 publications, are con-cerned with various aspects of classical as well as chemoenzy- matic carbohydrate chemistry, including some activities in more applied areas. Joachim Thiem Chemical Society Reviews, 1997, volume 26 463 Transglycosylation I Gly-OH I Hydrolysis Hydrolysis Scheme 1 Transfer reactions catalysed by glycosidases with an inverted anomeric configuration by using glycosyl fluorides or, e.g.glycosyl pyridinium ions. In contrast to glycosyltransferases, glycosidases are able to glycosylate many ‘xeno-substrates’ with primary or secondary hydroxy groups. Even though the activity of water as acceptor in the reaction mixture is usually one to two orders higher than the activity of the alcohol to be glycosylated, the yields of the glycosides are often much higher than would be expected from simple thermodynamic calculations. This is considered to be caused by the higher affinity of the alcoholic substrates for binding to the ‘aglycone site’ of the glycosidase compared to water. Speculation about the possible influence of the groups adjacent to the respective hydroxy group is common, and electron rich moieties such as double bonds in the allylic position, conjugated systems of double bonds, aromatic or heteroaromatic systems generally support glycosylation.Ali- phatic alcohols are also good acceptors for glycosyl moieties transferred by glycosidases. Amphiphatic properties and the presence of nitrogen in the acceptor molecule usually enhance transglycosylation yields. There exists a broad variety of substrates that are glycosy- lated by glycosidases (Scheme 2)-aliphatic and alicyclic alcohols,2 phenols, oximes, steroids and terpenes3 and amino acids-generally, only protected amino acids or peptides can be GalPo-OH Enzyme: P-galactosidase (E. cola donor: lactose yield: 42% ref. 2 GalNAcPO, CH2 I CH3 HN Enzyme: P-N-acetylhexosaminidase (A.oryzae) donor: p-nitrophenyl-P-GalNAc yield: 12% ref.5 NH Boc GalPO *CONpCOOCH3 Enzyme: P-galactosidase (E. coh)donor: lactose yield: 13% ref. 4 Scheme 2 Examples of various glycosides obtained by glycosylations with glycosidases 464 Chemical Society Reviews, 1997, volume 26 glycosylated by glycosidases,4 alkaloids5 and many other substances. The substrate to be glycosylated by a glycosidase should be at least partly soluble in water. The solubility can be enhanced by addition of water-miscible solvents (e.g. acetonitrile, dimethylformamide, dimethylsulfoxide, dioxane, tert-butyl al- cohol). Concentrations of co-solvents up to 30% are usually well tolerated by most glycosidases. Contrary to, e.g.lipases, lowering the water activity by co-solvents usually does not improve yields, the improved solubility of the acceptor and of the donor is the only decisive factor. There are, however, examples of lowering the water activity by, e.g. addition of various salts6 that improves the glycosylation yields. There are few examples of enzymatic glycosylation of simple alcohols in bi-phasic systems with organic solvents or by polyethyleneglycol-modified glycosidases in organic sol-vents.7 2.1 Enzymatic versus chemical glycosylation The strategy for choice of enzymatic glycosylation vs. chemical glycosylation is dictated by several factors. (1) There is no reliable chemical method for the glycosyla- tion-harsh methods decompose the aglycone, the reaction does not give reasonable yields or it is impracticable. Unwanted anomers are produced by chemical glycosylation, e.g.en route to syntheses of a-galactosides or P-mannosides. (2) Use of heavy metals as catalysts is not acceptable because of their toxicity or high price. This is especially true in the case of products with potential use in the pharmaceutical area or in nutrition. (3) Enzymatic glycosylation can help to overcome steric hindrance because an unprotected glycosyl residue is intro- duced that is less bulky than the protected glycosyl donor in the case of chemical reaction. (4)There exists a suitable enzymatic system for effective glycosylation and the substrate is at least partly water-soluble. (5) The substrate is rare or expensive-in the case of enzymatic glycosylation, even if the yields are lower, the unreacted substrate can be recovered near quantitatively. If there exists a suitable chemical method giving sufficient glycosylation yields of the desired product in the appropriate quality then the use of enzymatic glycosylation would be only l’art pour l’art.Nevertheless, even in these cases methodology orientated studies may be of interest with regard to future limits of the chemical processes. Chemical syntheses of sialyl-glycosides or fucosyl-glyco- sides often involve considerable problems. The use of glycosyl- transferases is limited due to their narrow selectivity and the high price of the enzymes and glycosyl donors-vide infra. Therefore, using sialidases and fucosidases represents an interesting alternative.Sialidases are usually isolated from pathogenic bacteria where they often act as toxins causing desialylation of cell surface glycoconjugates which then causes cell lysis. Sialidase from Vibrio cholerae was used for reversed glycosylation of methyl P-galactopyranoside using as a glycosyl donor-free 5-N-acetylneuraminic acid (Neu5Ac) affording both the a,2-3- and the a,2-6-linked disaccharides (ratio 1:10, 1H NMR) in a mere 1% overall yield.8 The tranglycosylation concept using as activated donor the a-p-nitrophenyl glycoside of NeuSAc and sialidase immobilized on ‘VA-epoxy’ as carrier gave con-siderably higher yields (16%) of a,2-6- and a,2-3-disaccharides in a ratio of 2: 1.In the case of methyl P-lactoside (a) or N-acetyllactosamine (b) (Scheme 3) as acceptors, a 1616% yield of trisaccharides was obtained. The a,2-6 isomer, and the a,2-3 isomer, representing the defucosylated sialyl LewisX were obtained in a ratio of 2.8 : 1 (Scheme 3).* These examples demonstrate well the complementarity of the two ‘glycosidase synthesis approaches’: reversed glycosylation and trans-glycosylation. HO HO b0H cop-AcNH HO R Sialidase from V. cholerae OH X* 4-t--HO RHO R AcN H OH X HO OH X X=OH, R=OCH3 (a) X = NHAc, R = OH (b) Scheme 3 Tranglycosylations catalysed by sialidase from Vibrio choleraex One of the few examples of the use of fucosidases for transglycosylation is a chemoenzymatic synthesis of galactosyl fucosides employing porcine a-~-fucosidase9 giving yields up to 16% (Scheme 4).Here, two types of activated donors were used, e.g.p-nitrophenyl a-L-fucoside and a-L-fucosyl fluoride with comparable yield and regioselectivity. Use of other glycosyl fluorides (a-galactosyl fluoride, a-glucosyl fluoride) as activated donors in the enzymatic transglycosylation is an attractive alternative to the nitrophenyl a-glycosides. P-Glycosyl fluorides cannot be used because of their instabil- ity. ? HO ,OH HO MOCH3 OH a-L-Fucosidasefrom porcine liver HO OH OHoH HO-0CH3 Ho@CH3 0 + HO ,O Scheme 4 Transglycosylation catalysed by a-L-fucosidase from porcine liver9 2.2 Regioselectivity of glycosidases The regioselectivity of glycosidases is rather poor.If more OH groups are available as acceptors, several products are usually encountered in various proportions. Often the regioselectivity of the glycosidases can be partly predicted-e.g. in hexopyranoses the affinity of OH groups for the glycosyl transfer generally decreases in the following order: 6-OH >> 4-OH 33-OH > 2-OH. This selectivity also correlates with chemical reactivity of these groups. The often cited papers of NilssonIo show a possible control for regioselectivity by changing the substituent configuration at the anomeric carbon. This case is, however, quite special and subtle changes in structure and conformation in this region could bring about alternative regio-effects.Until now, it has not been demonstrated whether the change of a methyl group in the above caselo into, e.g. an ethyl group would result in the conservation of regioselectivity . A systematic approach to the regioselectivity problem is, however, the use of glycosidases with their own structural preferences towards a specific position in the carbohydrate pyranose or furanose rings. Provided a broad variety of glycosidases having synthetic properties are available, it is possible to choose the particular glycosidases with appropriate regioselectivity. It seems that the regioselectivity of the transglycosylation reactions is analogous to the hydrolysis specificity towards the same type of linkages. This was rather convincingly demon- strated by Ajisaka et d.11 by a kinetic study on the isomers of N-acetyl lactosamine and on their sialylation by sialidases.The same authors11 and many others demonstrated that by rational selection of galactosidases from various sources it is possible to prepare various regioisomers of N-acetyllactosamine, e.g. the 1+6 isomer using (3-galactosidase from Aspergillus oryzae, the 1-4 isomer using that from Bacillus circulans or B. hifiidum and the 1+3 isomer using that from Streptococcus sp. or from bovine testes12 (Scheme 5). Glycosidases also display regioselectivity in a hydrolytic mode that can be well used for preparation of selectively glycosylated substances. 6-Glucuronidase from the limpet Patella vulgata can selectively cleave glucuronic acid from the aromatic hydroxy group of morphine-3,6-diglucuronide l3 (Scheme 6).This enables the production of morphine-6-glucur-onide that has about three times higher analgesic activity than the aglycone. There is, however, a constant need for novel biocatalysts. Glycosidases with pronounced regioselectivity should be searched mainly among endoglycosidases and screened for the cleavage of a particular glycosidic bond that is to be synthesised. Good substrates for such a screening can be found among natural heteropolysaccharides often bearing the desired glycosyl residue(s). 2.3 Stereoselectivity of glycosidases Besides regioselectivity of glycosidases one can also observe chiral discrimination displayed by these enzymes.This is a rather common fact observed in many enzymes, e.g. lipases, esterases, oxidoreductases and nitrilases-that are able to discriminate between two enantiomers in both the synthetic or the lytic mode of action. There are several examples demon- strating that glycosidases are able to discriminate enantiomers during glycoside synthesis. 14~~5Systematic studies of this problem were undertaken quite recently.2 Chemical Society Reviews, 1997, volume 26 465 D-Galactosidasefrom: bovine testes Streptococcus sp. OH bH NHAc HO p-nitrophenyl HO HO "*OH HO Bifidobacterium bifidum NHAc OH NHAc Kluyveromyces lactis HO Aspergillus oryzae Penicillium multicolor HO OH NHAc Scheme 5 Regioselectivity of glycosidases from various sources Boos et al.14 have studied P-D-galactosyl transfer to sn-glycerol using a P-D-galactosidase from E.coli and demon- strated that only the (2R)-glycerol- 1 -0-(3-D-galactopyranoside (in addition to smaller amounts of 2-0-(3-~-galactosyl-glycerol) are formed (Scheme 7). Racemic 2,3-epoxypropanol has also been used as an acceptor.lS A diastereomeric excess of (2R)-2,3-epoxypropyl (3-D-galactoside (R :S = 7 :3) was ob- tained with o-nitrophenyl P-D-galactoside as donor but no diastereomeric excess could be detected with lactose. This indicates that the stereochemical outcome of the formation- and probably then also of the hydrolysis-of the glycosides is kinetically controlled because both the (3-galactosyl donors used have different kinetic properties (I&, V,,,).Regio- and stereo-chemical studies have been carried out by Crout et a1.2 using lactose as a donor with racemic butan-2-01 and with a variety of diols acting as acceptors. If racemic propane- 1,2-diol was used as acceptor all four possible products were formed. The major product contained the galactose HO H \ HO OH P-Glucuronidasefrom Patella vulgata OH 24 % Scheme 6 Regioselective hydrolytic cleavage of morphine-3,6-diglucu-ronide by (3-glucuronidase from Patella vulgataI3 466 Chemical Society Reviews, 1997, volume 26 attached to the primary OH group of the diol (R :S = 1.0 :0.86). A similar selectivity (R:S = 1 :0.77) was observed in formation of the minor product by transfer to the secondary hydroxy group.Overall, transfer to the primary hydroxy group was favoured over the secondary by a factor of 1 :0.35. Somewhat different results were obtained when racemic butane- 1,3-diol was used as acceptor2 (Scheme 2). In this case there was only a very slight diastereoselectivity in transfer to the primary hydroxy group (R:S = 1.O :0.9), but in the transfer to the secondary hydroxy group there was a marked selectivity in favour of the (R)-enantiomer (R:S = 1:0.5). Overall transfer to the primary hydroxy group was favoured over the secondary by a factor of 1 :0.15. Various results for these diols were interpreted in terms of their possible conformational similarity with glucose and eventual interaction with the hydrophobic 'glucose' binding site of the (3-galactosidase (lactase). A number of other examples of glycosidase stereoselectivity have been reviewed recently.16 The preceding examples show that some glycosidases are able to distinguish between steric neighbourhoods of the hydroxy group to be glycosylated. It implies the possibility of analogous stereo-discrimination during hydrolysis.We recently demonstrated that 6-galactosidases display a significant ster- eoselectivity during hydrolysis. 16 Chemically synthesised ga- lactopyranosides of several racemic alcohols were subjected to enzymatic hydrolysis by (3-galactosidases from different micro- bial sources, e.g. E. coli, A. oryzae, Kluyveromyces lactis and Bacillus circulans (Scheme 9). From all P-galactosidases tested one from E.coli displayed the most pronounced stereoselectiv- y JoH P-Galactosidase(E.coh) GalPO-0 H Gal p(1-4)Glc + HO (4 OHGal p(1-4)Glc 1 +LOHP-Galactosidase(Ecol,) ~ GalPo GalPO 1 Scheme 7 Stereoselective galactosylation of racemic polyols by 0-galactosidase from E. c0li2.I~ R' R3 Glycosidase R2+, R3 3 Glycosyltransferases R' %O-Gly O-Gly + RIAOH 3.1 Leloir-type glycosyltransferases Scheme 8 Principle of chiral discrimination by glycosidases Glycosyltransferases are more expensive and rarer enzymes. They are also usually more sensitive to environmental condi- ity. The selectivity can be improved by higher reaction tions, often demanding special buffers or detergents for temperature and by a short reaction time.solubilization. They show, however, better selectivities and give nearly quantitative yields. Many glycosyltransferases have HO been cloned-for a survey of cloned sialyltransferases see e.g. ref. 17 Immobilization of the glycosyltransferases is another OH effective way to reduce the costs. HOeoJR P-GalactosidaseIHO ,OH > I.. OH Scheme 9 Chiral discrimination of racemic alcohols by 0-galactosidase from E. coliI6 Poor selectivities were observed for the (3-galactosides of pentan-2-01, the 1-P-O-galactoside of propane- 1,2-diol, and both galactosides of pentane- 1,4-diol. On the other hand, (3-galactosides of 1,2-0-isopropylidene-glycerol,butan-2-01, and both galactosides of butane- 1,3-diol were hydrolysed with good stereoselectivity (Table 1).In comparing these results to Table 1 Chiral discrimination of synthetic galactosides by P-galactosidase from E. coliu D.e. of E.e. of synthetic alcohol P-Galactoside of galactoside Conversion released aglycone (%) (%I Isopropylidene-glycerol 22 (S) 65 40 7 14 12 Pentan-2-01 Propane-1,2-diolh Butane-1,3-diolh Butane-1,3-diolh Pentane- 1,4-diol 11 (R> 23 (R) 9 (R) 23 (R)0 8 61 22 23 41 Pentane- 1,4-diol 8 (S) 16 24 a Concentration of substrates 4 pmol cm-3, reaction temperature 40 "C, enzyme activity 0.8 U cm-3. b Incubation time 15 min. c' Incubation time 30 min. those obtained by Crout et a1.2 it can be concluded that in the case of propane-1,2-diol the extent of chiral discrimination is about the same, however, the preferred absolute configuration appears to be opposite, e.g.(Rj in the transglycosylation and (S) in the hydrolytic process. For butan-2-01 and butane-l,3-diol higher selectivities were obtained in the hydrolytic process. Both transgalactosylation of butan-2-01 and the secondary OH group of butane-1,3-diol and hydrolysis of their galactosides showed the same preference for the (R)-enantiomer. In contrast to the observations in transgalactosylation2 it was observed that in the hydrolytic process16 the selectivity could not be associated with a similarity with the natural substrate of P-galactosidase, lactose, in particular, to the glucopyranose moiety. The distinguishing effect of P-galactosidase from E.coli could be based on hydrophobic interactions. In summary, galactosidase mediated hydrolysis of racemic (S-galactopyranosides led to significant enantiomeric enrich- ments in some of the alcohols released. Glycosyltransferases are often referred to as being rather stringent towards the distal one to two saccharidic moieties and also very specific to the glycosyl donor (nucleotidej. At present, there are, however, numerous examples demonstrating that also glycosyltranferases can be 'persuaded' to work with both unnatural donors and/or acceptors maintaining their main advantages, i.e. strict regio- and stereo-selectivity and high yields. Only a few glycosyltransferases are readily available and among those a considerable number of experiments have been performed with galactosyltransferase (GalT).The substrate specificity of this enzyme has been extensively studied and often reviewed. The problem of in situ UDP-Gal regeneration and GalT feedback inhibition by UDP was solved in 1982 by White- sides.l8 A recent extension of this approach with regard to donor modification opened up further possibilities for enzymatic approaches (Scheme 10). Hexokinase (viij allowed the forma- Pvr PEP / ATP ADP ,0P03H -?/ HO OP03H-PYr PEP R = NHAc R=OH Scheme 10 GalT glycosylation with UDP-2d-Gal as donor (enzymes: i phosphoglucomutase; ii UDP-glucose-pyrophosphorylase; iii, inorganic pyrophosphorylase; iv, UDP-galactose-4'-epimerase; v, galac-tosyltransferase (GalT); vi, pyruvate kinase; vii hexokinase) 19 tion of 2-deoxy-~-glucose-6-phosphatestarting from 2-deoxy- glucose (2-d-Glc) and ATP, which in turn could be regenerated from PEP with pyruvate kinase (vi).In analogy to ref. 18 the corresponding activated UDP-2-d-Glc and UDP-2-d-Gal could be obtained. Its reaction with GlcNAc or with glucose and GalT alone or in the presence of a-lactalbumin gave 2'-deoxy Chemical Society Reviews, 1997, volume 26 467 analogues of N-acetyllactosamine and of lactose in 40 or 25% yields, respectively. l9 A rather unexpected reaction with GalT (v) resulted when unnatural amino sugars were subjected to galactosylation (see Scheme 20 below). The reaction using GalT together with UDP- Gal 4'-epimerase with the 3-amino sugar N-acetylkanosamine (X = 0,R = CH20H) gave a P-galactosylation at the anomeric position in the P-configuration and led to GalP 1 -1PGlc3NAc in 22% yield.20 Analogously, N-acetylgentosamine (Xyl3NAc, X = 0, R = H) and N-acetylthiogentosamine (5SXyl3NAc, X = S, R = H) gave the corresponding trehalose-type linked disaccharide derivatives in good yields.20 Recently, xylose was shown to be the first ambident substrate for GalT-catalysed galactosylation. Both the P1-4 and the 1-P 1 transfer products (Galpl-4Xyl and GalP1-1PXyl) could be obtained in a 2: 1 ratio.2 UDP-Glc J.iv UDP, UDP-Gal / OH OH X=O, R=CH20H X=O, R=H X=S, R=H Scheme 11Frame shifted galactosylation with GalT (v)20,21 Galactosyltransferases can also be used in glycosylation of non-sugar substances, e.g.some natural products. Complex glycosides of ergot alkaloids, required for immunological studies were prepared by use of GalT: for preparation of p-D- by the use of bovine P, 1-4-galactosyltransferase was chosen (Scheme 12). For generation of UDP-Gal in situ, UDP-Glc and UDP-Gal 4'-epimerase were used. It was found, however, that P, 1-4-galactosyltransferase was able to transfer glucose form- ing in parallel also P-D-glucopyranosyl (1 +.4)-2-acetamido-2-deoxy-P-~-glucopyranosyl-(1-+0)-elymoclavine 3 (Scheme 12). This was later confirmed on a semipreparatory scale using UDP-Glc without the epimerase. The transfer of glucose was confirmed fully by spectral methods.5 This is the first example of concomitant transfer of glucose and galactose by galactos yltransferase.P-Lactosyl elymoclavine was prepared from the respective P-glucoside by use of bovine P, 1-4-galactosyltranferasein the presence of a-lactalbumin.5 Analogously, P-Lac and P-LacNAc derivatives of other ergot alkaloids, e.g. 9,lO-dihydrolysergene were prepared.5 The concept of the use of non-natural donors has already been extended to other glycosyltransferases. Recently, a, 1 -3/4 fucosyltransferase from human milk was shown to be able to use as donors GDP-L-Gal, GDp-3d-~-Gal and GDP-3 ,6d2-~- Gal the structure of which resembles that of the natural glycosyl donor GDP-L-FUC. The reactions were accomplished with high yields up to 93% (Scheme 13).22 Some other examples of glycosyltransferase use-mostly sialyltransferases-are given in Section 4.3.2 Non-Leloir glycosyl transferases A special group of non-Leloir glycosyltransferases are, e.g. cyclodextrin glucanotransferases (CGTase). These enzymes are produced by microorganisms and many of them are commer- cially available. They catalyse cyclodextrination of starch but also a transfer of one or more a-glucosyl units to various acceptors. They can be used for extending glycosides or for a-glucosylation of many compounds, e.g. monosaccharides, stevioside, rubuoside, hesperidin, (+) catechin and others. The acceptor specificity of CGTases is rather broad and the enzymes galactopyranosyl (1+4)-2-acetamido-2-deoxy-~-~-gluco-of various origin are also able to glycosylate phenolic OH pyranosyl-(1 +0)-elymoclavine 25 the extension of 2-ace- groups in very good yields.In sugars or glycosides with the tamido-2-deoxy-~-~-glucopyranosyl-( moiety, the transfer is rather regioselective with 1-+0)-elymoclavine l5 D-G~c~ 1 UDP-Galactose 4'-epimerase from yeast (E.C. 5.1.3.2) 3 O\1I CH2 p,1-4 Galactosyl transferase Ho% + (bovine colostrum E.C. 2.4.1.22)* HO OH 0-UDP HN HN'I 1 2 Scheme 12 Concomitant transfer of glucose and galactose by GalT5 468 Chemical Society Reviews, 1997, volume 26 OH6HO& OH0 0(CH2)8-CO2Me &OH HO 'OH NHAC R2w2GDpa,l-3/4Fuc T HO R' HO 'OH ID2 NHAc Scheme 13 Glycosylation of lactosamine derivative by a,1-3/4 fucosyl-transferase from human milk employing activated donors based on L-galactopyranose; (a) GDP-6-L-galactopyranose, (b) GDP-P-3-deoxy-L-galactose, (c) GDP-(3-~-3,6-dideoxy-~-galactose~~ preference for 4-OH so that the second Glc unit is always attached by an a( 1-4) bond.In other sugars the regioselectivity is, however, rather poor. CGTase from Bacillus stearothermophilus was used for transglucosylation of rutin. The glucosyl unit of rutin was extended by one or more a-glucosyl units and the population of oligoglucosides was trimmed to 4G-a-~-glucopyranosyl rutin by Rhizopus sp. glucoamylase (Scheme 14).23 The natural task of another group of non-Leloir glycosyl- transferases-phosphorylases-is the phosphorolytic cleavage of a glycosidic bond transferring the glycosyl unit released on the phosphate.Glycosyl transfer can, however, also be accomplished on another acceptor molecule, often with a HO Ho7 \ 0 HO PI 4G-a-D-GlucopyranosyI-rutin Scheme 14 Product of rutin glucosylation by CGTase from Bacillus stearothermophilus23 phenolic OH group. Sucrose phosphorylase is rather com- plementary-having different regioselectivity-to the CGTases in preparation of a-glucosides of various aglycones, e.g. catechins, polyphenols and cyclitols.24 Some phosphorylases are even able to accept non-natural sugars as glycosyl donors. Potato phosphorylase was able to transfer 2-deoxy-a-glucosyl residues from D-glucal to maltotetraose thus forming 2-deoxy- maltooligosaccharides (Scheme 15).25 4 Multienzyme glycosylation approaches The first basic approaches using the multienzyme system in glycosylation were mostly targeted to regeneration of the activated glycosyl donors (kinases, phosphorylation sequences) or the generation of glycosyl donors in situ.5-N-Acetylneuraminic acid is still a rather expensive starting material and therefore its generation in situ from N-acetyl- mannosamine with neuraminic acid aldolase was developed. This procedure can be integrated well with further enzymatic sequences generating the CMP-Neu5Ac required for enzymatic sialylation.26 Generation of dinucleoside glycosyl donors, e.g. UDP-Gal was developed by Whitesides et al. 18 for N-acetyllactosamine synthesis in 1982.This reaction sequence started from glucose- HO Chemical Society Reviews, 1997, volume 26 469 Hoeo*HO OH HO N HAc OH HO t%:oh UDP Pyruvate UTP kinase HoD-FructoseGlc-1-P ,(.OH p,1-4 Gal-v UDP-Glc v V Sucrose pyrophosphorylase phosphorylase UDP-Gal HO \ k-HO OH N HAc UDP-Gal e-epirner 'ase OH HO Sucrose Scheme 16 Galactosyltransferase reaction combined with sucrose phosphorylase reaction27 HO"' AcNH OR OSE OH 1 CTP a,2-3sialyltransferase Trans-sialidase CMPNeu5Ac Neu5Ac synthase CMPNeu5Ac HO OR AcNH OSE 2 p, OH OH SE: CH2CH2SiMe3 Scheme 17 Multienzyme system for synthesis of a,2-3 sialosaccharides using a,2-3-sialyltransferase from Trypanosoma cruzi28 1-phosphate, and recently this approach was improved by Ichikawa et al.27 Starting from sucrose, glucose- 1 -phosphate could be favourably generated with sucrose phosphorylase (Scheme 16).One of the main advantages of this method is the removal of part of the potential ultimate inhibitor-inorganic phosphate. All enzymes used in this multienzyme system are now commercially available. Recently, more sophisticated approaches have emerged, i.e. sequential use of glycosyltransferases including cofactor re- generation or even sequential use of glycosidases and glycosyl- transferases. These approaches sometimes simulate real enzy- matic sequences working in vivo. Multienzyme systems were applied mainly in preparation of complex sialooligosaccharides.Sialic acid is a common struc- ture in glycoconjugates and sialylated structures are involved in a variety of biological processes. Besides their role in secretion, immunogenicity, circulation half-life of glycoproteins, and cellular recognition phenomena they are identified as crucial structural elements of some antigenic determinants which have been identified as tumour markers. Terminally substituted sialyl glycosides can either be pre- pared by sialyltransferases or by sialidases-vide supra. In a quite prominent place among the sialidases used in sialoglycoside synthesis are the sialidases from the blood parasite Trypanosoma cruzi, which causes Chagas disease. This enzyme, named trans-sialidase (TS), catalyses the transfer of sialic acid from host glycoconjugates to acceptor molecules of the parasite plasma membrane.There is evidence that sialic acid receptors on the surface of T. cruzi mediate the initial stages of 470 Chemical Society Reviews, 1997, volume 26 trypanosome invasion of the host cells. TS from T. cruzi has the unique property of catalysing the reversible transfer of NeuSAc from a donor substrate of the sequence NeuSAca2- 3PGal-0-R to virtually any galactoside acceptor (3Gal-O-Rz to yield as new product NeuSAca2-3Gal-0-R2. Its use for synthetic purposes was, however, limited by the fact that the desired product is produced at the expense of another sialoside used as the donor substrate. This problem and also the shift of equilibrium in favour of the desired sialoside was solved by coupling this reaction to a complex system generating a donor of NeuSAc for TS. This multienzyme system consists of a catalytic CMP-NeuSAc regeneration cycle and a,2-3 sialyl- transferase28 (Scheme 17) and it allows for the enzymatic preparation of virtually any terminal NeuSAca2-3Gal sequence without limitation by acceptor specificity (as with sialyl-transferases) or low yield and poor regioselectivity (as with sialidases).A multienzyme system consisting of sequential action of glycosidase and glycosyltransferase coupled with in situ regeneration of sugar nucleosides was demonstrated in the production of sialyl Thomsen-Friedenreich (T-antigen) epi- topeI2 (Scheme 18). The concept is based on the pronounced selectivity of (3-galactosidase from bovine testes for the p, 1-3 glycosidic bond formation.Although small amounts of other regiomers are formed, only the p,1-3 isomer is accepted by a,2-3 sialyltransferase. The product is no longer a substrate for galactosidase and this shifts the equilibrium of the trans-galactosylation reaction towards the desired product and allows the synthesis of the trisaccharide in an irreversible manner. In HO OH NHAc GalP(1-4)Glc(pNPPGal) (PNP) OH HO AcHN OH 36 Oh Scheme 18 Multienzyme system this multienzyme reaction the problem of bringing rather distant pH optima of sialyltransferase, the cofactor regenerating system (pH 7.5-9.0) and P-galactosidase (pH 4.3) into harmony was faced.This was achieved by rational manipulation of reaction conditions and the use of activated substrates. Thus, it could be demonstrated that the enzymes with rather distant pH optima could be employed, when carefully rationalised and optimised. This concept opens new perspectives for the syntheses of glycosides having up to three or four glycosyl units in one-pot reactions. The systems described above are rather complicated, but they help to avoid multistep reactions and laborious purification procedures of the intermediates. They also considerably lower the overall costs mostly due to the integrated cofactor regeneration in situ. When this concept is combined with the use of immobilised enzymes or e.g. membrane reactor96 they can be used for effective production of some rare complex oligosaccarides. 5 Glycosylation by whole living cells The above described multienzyme systems tend to simulate the situation in the living cell where much more complicated multistep reactions take place.Use of living cells combines the advantages of the multienzyme approaches, e.g. high selectivity and cofactor regeneration and helps to avoid expensive and laborious isolation of the respective enzymes. For this purpose microorganisms, both prokaryotic and eukaryotic, and plant cell cultures are often used. Mostly, glycosyltransferases are responsible for the glycosyl transfer in these systems. There exist few examples, in which living microbial cells were used for glycosylations. An example of the glycosylation of a rather complicated molecule by resting cells of Bacillus suhtilis is a preparation of the 24-0-P-glucopyranoside of the immunosuppressive drugs FK 506 and of immunomycin29 (Scheme 19). A strict regioselectivity of the enzyme can be noted because glycosylation at only one of the three available OH groups occurs.The selectivity is also documented by the fact that the glycosylating strain was identified after screening of approximately 1000 strains. The above microbial glycosylation was catalysed by glyco- syltransferase. There are also whole microbial cell biosystems possessing glycosidases with transglycosylation activity. These enzymes are usually bound to cells quite often in the periplasmatic space. Well documented cases are, e.g.fructosy-lations. The growing culture of Claviceps purpurea is able to glycosylate ergot alkaloids, either introduced or produced per 2 PI l/phosphataseInorganic pp, CMP-Neu5Ac CMP-Neu5Ac CTP Pyruvate Pyruvatekinase PEP OH k /N HAc kinase Pyruvate PEP for synthesis of T-antigen epitopeI2 Me0 OR2 yMe0 OMe 1 R’ = CH2CH=CH2, R2 = H 2 R’= CH2CH3, R2=H 3 R’= CHzCH=CHz, R2=P-Glc 4 R’ = CH2CH3, R2 = P-Gk Scheme 19 Structure of FK 506 1 (Rl = CH2-CH=CH2, R2 = H), immunomycin 2 (Rl = CH2-CH3, R2 = H) and their 24-P-glucopyranosyl derivatives 3, 4 (R2 = OGlc) produced by Bacillus subtilis ATCC 5506029 HO HO HO HO I CH3 Scheme 20 Tetrafructoside of elymoclavine produced by trans-fructosylation activity of Claviceps purpureas se, forming alkaloid-oligofructofuranosides.~Sucrose served as a donor for this transfi-uctosylation reaction. Glucose was utilised by the fungus for its growth and the fructose released was transferred onto the aglycone.Various fructosides having up to four fructosyl units were isolated (Scheme 20). A special case of glycosylation by microorganisms is a preparation of N-2’-deoxy-(3-ribosides of heterocyclic bases. The chemical synthesis of 2’-deoxy-P-ribosides suffers, com- pared to the preparation of P-ribosides, often from poor enantioselectivity, giving anomeric mixtures of deoxy-ribo- sides. The enzyme responsible for these reactions-nucleoside Chemical Society Reviews, 1997, volume 26 471 E.COlI, PI E. coh, CI-Ade GH HoTo>P03H-“Toy -PI IHO HO HO G = Guanine Scheme 21 Synthesis of 2-chloro-2’-deoxyadenosineby E. coli BMT-lD/1A cells selected for high purine nucleoside phosphorylase30 HO HO OMe OMe HO HO OMe Scheme 22 Glycosylation of 3.5-dimethoxyphenol by Panax ginseng plant cell culture32 2’-deoxyribosyl transferase-ensures full stereoselectivity and regioselectivity. A preparation of the potent antitumour drug 2-chloro-2’-deoxyadenosinewas accomplished on a preparative scale by using glutaraldehyde-crosslinked whole cells of a selected strain of Escherichia ~0li30 with the glycosyl donor deoxy-guanosine and 10 mM inorganic phosphate necessary for a good yield (65%) (Scheme 21).Whereas glycosylations by growing microoorganisms are rather scarce, plant cell cultures have often been used mostly for (3-glucosylation. Plant glycosyl-transferases are often specific for various flavonoids. Some of them have somehow ‘wob- bling’ specificity and therefore they can be used for glycosyla- tion of, e.g. phenols, steroids, flavonoids, cardenolides and stevioles. These systems can be well used for glycosylation of phenolic hydroxy groups and also for the preparation of glycosyl esters in deprotected form. The main disadvantages are longer reaction times (days) and sometimes complicated handling of the biosystems, but the main advantages are the glycosylation of barely accessible phenolic hydroxy groups and carboxy groups.For glycosylations by plant cells, plant homogenates, tissue slices or even whole plants (e.g. seedlings) can be used. However, the most suitable are suspension plant cell cultures. In the last two decades their application has been developed so that they can be used without special prerequisites. Salicylic acid can be P-glucosylated by a cell suspension culture of Mallotus japonicus. The yield can be increased up to 1.1 g 1-1 by continuous feeding of the substrate and the process can be scaled up (5 1 fermenter). It was demonstrated that this glycoside exhibited quicker and longer lasting analgesic effects than the aglycone alone.31 The root culture of Panax ginseng has very good glycosylation activity towards aromatic phenols and carboxylic acid and was found to glucosylate, e.g.3,5-dimethoxyphenol, methyl salicylate, p-hydroxyacetophe- none and coniferyl alcohol, forming in addition to p-D-glucosides, 13-D-gentiobiosides and (3-D-primeverosides (Scheme 22).32 Flavonoid and flavolignan glycosides can often be found in plant tissues and respective glycosylation systems (usually UDP-Glc dependent glycosyltransferases) can be used for 472 Chemical Society Reviews, 1997, volume 26 glycosylation of various flavonoid compounds. Glycosylating systems with quite pronounced regioselectivity can often be found. Selective glycosylation of silybin (an effective hepato- protective compound with low water solubility) was accompli- shed in nearly quantitative yield by suspension cell culture of Papaver sornniferurn. The resulting compound was many times more water soluble than the parent compound and therefore its bioavailability greatly increased33 (Scheme 23). H o r nu0%OMe \ OH OH OH 0 Papaver sornniferurn suspension plant cell culture HO eOMe OH OH 0 Scheme 23 Glucosylation of silybin by Papaver sornniferum var.setigerurn plant cell culture33 The chemical preparation of glycosyl esters in their de- blocked form is sometimes a rather complicated task due to their lability during deblocking procedures. Plant cell cultures facilitate their preparation under very mild conditions in a single step. A good example is the preparation of the potential antitumour agent 6-O-butyryl-~-glucose by the cell suspension culture of Nicotiana plumbaginifolia.34 Using plant cultures anomeric glycosyl esters can also be prepared. Using plant tissue slices or homogenates is now rather obsolete but sometimes good results can be obtained with highly metabolically active whole young plants, seedlings or with germinating seeds.5’-O-(P-~-Glucopyranosyl)pyridoxine and 4’-O-(f3-~-glucopyranosyl)pyridoxine were formed by germinating seeds of wheat, barley and rice cultured on a pyridoxine containing medium; the ratio was 1 : 1 but germi- nating soy bean seeds formed only 5’-O-(~-~-glucopyranosyl) pyridoxine.35 Preparation of P-glucuronides of various xenobiotics is a good task to be accomplished by bio-systems. f3-Glucuronides are important for the catabolism and pharmacokinetic study of most drugs.Moreover, some glucuronides have different pharmacodynamic properties, compared to the original drugs. In man, morphine is glucuronidated at position 6 and the resulting 6-~-glucuronide of morphine has a three times higher analgesic effect than the aglycone36 (contrary to the morphine- 3-(3-glucuronide-see also Scheme 6). P-Glucuronides could be prepared by the use of a subcellular microsomal fraction from beef liver containing glucuronyl transferase. The extremely expensive UDP-glucuronic acid serves as a glucuronic acid donor in this case. Regeneration of this substrate can be accomplished well by the intact organ- ism-vide infra-or by, e.g. using whole perfused liver.This, rather sophisticated, technique recently became commonly used for liver physiology or drug metabolism studies37 and its further application in glycobiology is expected. There is probably a single example of glycosylation by using whole living organisms, e.g. preparative glucuronylation of various xenobiotics. Animals-usually rabbits-are fed or injected with the substance to be glucuronylated. The respective glucuronide is isolated from the urine. Fishman38 used this method for preparation of P-glucuronide of phenolphthalein used as a chromogenic substrate for p-glucuronidase determi- nation. Rabbits were injected with sodium phenolphthalein phosphate and the phenolphthalein glucuronide was isolated as the cinchonidine derivative. This method has certain draw- backs, e.g.possible degradation of substrate or unwanted conjugations, sulfonation, and also ethical problems. 6 Conclusions Synthetic applications of enzymes effectively supplement and/ or complement established carbohydrate synthetic methods. Individual applications should be carefully adopted and ration- alised. Enzymatic methods should always be used in a way complementary to synthetic steps. Glycosidases have become widely used in more subtle ways due to pronounced knowledge of their regio- and stereo-selectivity. This allows use of these cheaper and more robust enzymes in applications where previously only glycosyltrans- ferases or laborious synthetic sequences could be employed. Glycosyltransferases, formerly believed to be rather stringent towards acceptor and donor, can now be used in more flexible ways with structurally diverse analogues.Multienzyme ap- proaches combine both the advantages of glycosidases and glycosyltransferases and they enable one-pot syntheses of complex trisaccharides or even tetrasaccharides. Whole-cell systems can selectively glycosylate often very complex sub- strates that would be easily available neither by synthetic nor by enzymatic methods. These methods demand, however, broad screening of suitable systems but the eventual results are often rewarding. 7 Acknowledgements This study was supported in part by the Volkswagen Stiftung, the Deutscher Akademischer Austauschdienst, the Deutsche Forschungsgemeinschaft (SFB 470) and the Grant Agency of the Czech Republic (Grant No.203/96/1267). 8 References 1 C.-H. Wong and G. M. Whitesides, Enzymes in Synthetic Organic Chemistry, Tetrahedron Org. Chem. Ser., ed. J. E. Baldwin and P. D. Magnus, Pergamon, 1994,vol. 12, and references therein. 2 D. H. G. Crout, D. A. MacManus and P. Critchley, J. Chem.Soc.. Perkin Trans. I, 1990, 1865. 3 Y. Ooi, T. Hashimoto, N. Mitsuo and T. Satoh, Chem. Pharm. Bull., 1985,33, 1808. 4 S. Attal, S. Bay and D. Cantacuzene, Tetrahedron, 1993, 48, 925 1. 5 V. Kien, Top. Curr. Chem., 1997, 186, 45, and references therein. 6 E. Rajnochovi, J. Dvoiikovi, Z. HuiikovA and V. Kien, Bzotechnol. Lett., 1997, 19, 869. 7 Y. Okahata and T. Mori, J. Chem. Soc., Perkin Trans.I, 1996, 2861. 8 J. Thiem and B. Sauerbrei, Angew. Chem., Int. Ed. Engl., 1991, 30, 1503. 9 S. C. T. Svensson and J. Thiem, Carhohydr. Res., 1990, 200, 391. 10 K. G. I. Nilsson, Carhohydr. Res., 1987, 167, 95; 1988, 180, 53. 11 K. Ajisaka, H. Fujimoto and M. Isomura, Carhohydr.Res., 1994, 259, 103, and the references therein. 12 V. Kien and J. Thiem, Angew. Chem., Inr. Ed. Engl., 1995, 34, 893. 13 R. T. Brown, N. E. Carter, F. Scheinmann and N. J. Turner, Tetrahedron Lett., 1995, 36, 11 17. 14 W. Boos, J. Lehmann and K. Wallenfels, Carhohydr. Res., 1968, 7, 381. 15 F. B. Bjorkling and S. E. Gotfredsen, Tetrahedron, 1988, 44, 2957. 16 B. Werschkun, W. A. Konig, V. Kfen and J. Thiem, J. Chem. Soc., Perkin Trans.I, 1995, 2459. 17 S. Tsuji, J. Biochem., 1996, 120, 1. 18 C.-H. Wong, S. L. Haynie and G. M. Whitesides, J. Org. Chem., 1982, 47, 5416. 19 J. Thiem and T. Wiemann, Angew. Chem., Int. Ed. Engl., 1991, 30, 1163. 20 Y. Nishida, T. Wiemann, V. Sinnwell and J. Thiem, J. Am. Chem. Soc., 1993,115, 2536, 21 T. Wiemann, Y. Nishida, V. Sinnwell and J. Thiem, J Org. Chem., 1994,59, 6744. 22 K. Stangier, M. M. Palcic, D. R. Brundle, 0.Hindsgaul and J. Thiem, Carhohydr. Res., in press. 23 Y. Suzuki and K. Suzuki, Agric. Bid. Chem., 1991,55, 181. 24 S. Kitao and H. Sekine, Biosci. Biotech. Biochem., 1994, 58, 419. 25 B. Evers, P. Mischnick and J. Thiem, Carhohydr. ReJ., 1994, 262, 335. 26 U. Kragl, M. Kittelmann, 0.Gisalba and C. Wandrey, Ann. N. Y. Acad. Sci., 1995, 750, 300. 27 M. Ichikawa, R. L. Schnaar and Y. Ichikawa, Tetrahedron Lett., 1995, 36, 8731. 28 Y. Ito and J. C. Paulson, J. Am. Chem. Soc., 1993, 115, 7862. 29 B. R. Petuch, B. Arison, A. Hsu, R. Monaghan, F. J. Dumont and T. S. Chen, J. Industrial Microbiol., 1994, 13, 131. 30 I. A. Mikhailopulo, A. I. Zinchenko, Z. Kdzimierczuk, V. N. Barai, S. B. Bokut and E. N. Kalinichenko, Nucleosides Nucleotides, 1993,12, 417. 31 Y. Umetani, E. Kodakari, T. Yamamura, S. Tanaka and M. Tabata, Plant Cell Reports, 1990, 9, 325. 32 M. Ushiyama and T. Furuya, Phytochemistry, 1989, 28, 3009. 33 V. Kfen, A. Minghetti, P. Sedmera, V. HavliEek, V. Pfikrylovi and N. Crespi-Perellino, Phytochemistry, 1997, in press. 34 S. Kamel, M. Brazier, G. Desmet, M.-A. Fliniaux and A. Jacquin- Dubreuil, Phytochemistry, 1992,31, 1581. 35 Y. Suzuki, Y. Inada and K. Uchida, Phytochemistry, 1986, 25, 2049. 36 R. Osbome, P. Thompson, S. Joel, D. Trew, N. Patel and M. Slevin, BI.. J. Clin. Pharmucol., -1992, 34, 130. 37 T. A. Aasmundstad, A. Ripel, E. Bodd, A. Bjorneboe and J. Morland, Biochem. Pharmacoi., 1993, 46, 961. 38 W. H. Fishman, Method Enzymol., 1957, 3, 55. Received, 14th April 1997 Accepted, 4th July 1997 Chemical Society Reviews, 1997, volume 26 473

ISSN:0306-0012    

DOI:10.1039/CS9972600463

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Chemical Society Reviews 1997 Indexes Volume 26 Index of Authors Armstrong, Fraser A., 169 Hayashi, Takashi, 355 McGarvey, Glenn J., 407 Reid, Katharine L., 223 Aubk, Jeffrey, 269 Heering, Hendrik A., 169 Maier, John P., 21 Reynolds, Kevin A., 337 Behr, Jean-Paul 63 Higgins, Simon, 247 Mason, Stephen F., 29 Rousseau, Gkrard, 453 Bernasconi, Claude F., 299 Hill, Steve J., 291 Mason, Timothy, J., 443 Rzepa, Henry S., 1 Boyall, Dean, 223 Hirst, Judy, 169 Mathew, Jessy, 127 Sanders, Jeremy K. M., 327 Brady, Paul A., 327 Hodge, Philip, 417 Morris, Russell E., 309 Simonsson, Daniel, 181 Bruce, James E., 191 Holland, Koren, A., 337 Mortimer, Roger J., 147 Smith, Richard D., 191 Catallo, W. James, 401 Homsi, Fadi, 453 Murray-Rust, Peter, 1 Snook, Richard, 319 Chivers, Tristram, 345 Hynes, Michael J., 133 Nair, Vijay, 127 Takayama, Shuichi, 407 Cooper, D.L., 87 Iqbal, Abul, 203 Ng, Dennis K. P., 433 Thiem, Joachim, 463 Covington, Anthony D., 11 1 Jiang, Jianzhuang, 423 Ng, Sheila B. L., 425 Utley, James, 157 Davies, Malonne I., 215 Jones, Anthony C., 101 Nielsen, Peter E., 73 Wadso, Ingemar, 79 Delaloge, Francette, 377 Jonson, Bo, 133 Ogoshi, Hisanobu, 355 Ward, Michael D., 365 Doig, Andrew J., 425 Junk, Thomas, 401 Ohtaki, Hitoshi, 41 Weigel, Scott J., 309 Doronina, Svetlana O., 63 Karadakov, P. B., 87 Ojima, Iwao, 377 Whitaker, Benjamin J., 1 Evans, John, 11 Kim, Sanghee, 387 Owen, John R., 259 Winkler, Jeffrey, 387 Galema, Saskia A,, 233 men, Vladimir, 463 Piers, Warren E., 345 Wong, Chi-Huey, 407 Gerratt, J., 87 Lei, Q.Paula, 191 Prabhakaran, Jaya, 127 Wu, Qinyuan, 191 Hadima, Gerald, 73 Lewerenz, H. J., 239 Radnai, Tamas, 41 Yamaguchi, Toshio, 41 Hao, Zhimin, 203 Longridge, John J., 53 Raimondi, M., 87 Harris, Kenneth D. M., 279 Lunte, Craig E., 215 Rawson, Jeremy M., 53 Index of Titles The World-Wide Web as a chemical information tool Peter Murray-Rust, Henry S. Rzepa and Benjamin J. Whitaker 1-10 Shining light on metal catalysts John Evans 11-20 Electronic spectroscopy of carbon chains John P. Maier 2 1-28 The science and humanism of Linus Pauling (1901-1994) Stephen F. Mason 2940 Structure of water under subcritical and supercritical conditions studied by solution X-ray diffraction Hitoshi Ohtaki, Tamas Radnai and Toshio Yamaguchi 41-52 Sulfur-nitrogen chains: rational and irrational behaviour Jeremy M.Rawson and John J. Longridge 53-62 Towards a general triple helix mediated DNA recognition scheme Svetlana 0.Doronina and Jean-Paul Behr 63-72 Peptide nucleic acid. A DNA mimic with a pseudopeptide backbone Peter E. Nielsen and Gerald Haaima 73-78 Trends in isothermal microcalorimetry Ingemar Wadso 79-86 Modern valence bond theory J. Gerratt, D. L. Cooper, P. B. Karadakov and M. Raimondi 87-1 00 Developments in metalorganic precursors for semiconductor growth from the vapour phase Anthony C. Jones 101-1 10 Modern tanning chemistry Anthony D. Covington 11 1-126 Carbon+arbon bond-forming reactions mediated by cerium(1v) reagents Vijay Nair, Jessy Mathew and Jaya Prabhakaran 127-1 32 Lead, glass and the environment Michael J.Hynes and Bo Jonson 133-146 Electrochromic materials Roger J. Mortimer 147-156 Index of Titles continued on page 476 475 Index of Titles continued from page 475 Trends in organic electrosynthesis James Utley 157-168 Reactions of complex metalloproteins studied by protein film voltammetry Fraser A. Armstrong, Hendrik A. Heering and Judy Hirst 169-1 80 Electrochemistry for a cleaner environment Daniel Simonsson 181-190 New mass spectrometric methods for the study of noncovalent associations of biopolymers Richard D. Smith, James E. Bruce, Qinyuan Wu and Q. Paula Lei 191-202 Some aspects of organic pigments Zhimin Hao and Abul Iqbal 203-2 14 Microdialysis sampling coupled on-line to microseparation techniques Malonne I.Davies and Craig E. Lunte 2 15-222 Modern studies of intramolecular vibrational energy redistribution Dean Boyall and Katharine L. Reid 223-232 Microwave chemistry Saskia A. Galema 233-238 Surface scientific aspects in semiconductor electrochemistry H. J. Lewerenz 239-246 Conjugated polymers incorporating pendant functional groups-synthesis and characterisation Simon Higgins 247-258 Rechargeable lithium batteries John R. Owen 259-268 Oxaziridine rearrangements in asymmetric synthesis Jeffrey Aube 269-278 MELDOLA LECTURE: understanding the properties of urea and thiourea compounds Kenneth D. M. Harris 27 9-290 Speciation of trace metals in the environment Steve J. Hill 29 1-298 Developing the physical organic chemistry of Fischer carbene complexes Claude F.Bernasconi 2 99-3 0 8 The synthesis of molecular sieves from non-aqueous solvents Russell E. Morris and Scott J. Weigel 309-3 18 Laser techniques for chemical analysis Richard Snook 3 19-326 Selection approaches to catalytic systems Paul A. Brady and Jeremy K. M. Sanders 3 27-3 3 6 The mechanistic and evolutionary basis of stereospecificity for hydrogen transfers in enzyme-catalyzed processes Kevin A. Reynolds and Koren A. Holland 337-344 Pentafluorophenylboranes: from obscurity to applications Warren E. Piers and Tristram Chivers 345-354 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-covalently bonded supramolecular assemblies Michael D.Ward 365-376 Asymmetric synthesis of building-blocks for peptides and peptidomimetics by means of the (3-lactam synthon method Iwao Ojima and Francette Delaloge 3 77-3 86 Approaches to the synthesis of ingenol Sanghee Kim and Jeffrey Winkler 387400 Hydrogen isotope exchange reactions involving C-H (D,T) bonds Thomas Junk and W. James Catallo 40 1-406 Enzymes in organic synthesis: recent developments in aldol reactions and glycosylations Shuichi Takayama, Glenn J. McGarvey and Chi-Huey Wong 407-4 16 Polymer-supported organic reactions: what takes place in the beads? Philip Hodge 417-424 Molecular and chemical basis of prion-related diseases Sheila B. L. Ng and Andrew Doig 425-432 Sandwich-type heteroleptic phthalocyaninato and porphyrinato metal complexes Dennis K. P. Ng and Jianzhuang Jiang 433-442 Ultrasound in synthetic organic chemistry Timothy J. Mason 443452 Preparation of seven and larger membered heterocycles by electrophilic heteroatom cyclization Gerard Rousseau and Fadi Homsi 453-462 Glycosylation employing bio-systems: from enzymes to whole cells Vladimir KFen and Joachim Thiem 463474 476

ISSN:0306-0012    

DOI:10.1039/CS9972600475

出版商:RSC    

年代:1997

数据来源: RSC